BJR

The British Journal of Radiology

2006, Volume 79
The British Journal of Radiology
January 2006, Volume 79, Issue 937

February 2006, Volume 79, Issue 938

March 2006, Volume 79, Issue 939

April 2006, Volume 79, Issue 940

May 2006, Volume 79, Issue 941

June 2006, Volume 79, Issue 942

July 2006, Volume 79, Issue 943

August 2006, Volume 79, Issue 944

September 2006, Volume 79, 945

October 2006, Volume 79, 946

November 2006, Volume 79, 947

September 2006, Volume 79, Special Issue 1

Volume 79 (2006), Case reports

BJR
The British Journal
of Radiology

January
2006
Volume 79
Issue 937
January 2006, Volume 79, Issue 937

● The President’s Conference 2005: ‘‘Technology in Imaging

and Radiotherapy – towards improved workflow and
productivity’’
● CT scanning the early days

● Cardiac applications of multislice computed tomography

● Technology solutions for better outcomes: integrated

information management in key to productivity increases in

medicine
● The case for particle therapy

● The contribution of PET/CT to improved patient management

● Mesenteric panniculitis in oncologic patients: PET-CT findings

● Diagnostic efficacy of SonoVueH, a second generation

contrast agent, in the assessment of extracranial carotid or

peripheral arteries using colour and spectral Doppler
ultrasound: a multicentre study
● Lymphoepithelioma-like carcinoma of salivary glands:

treatment results and failure patterns

● Comparison of patient doses in 256-slice CT and 16-slice CT

scanners
● Assessment of tube current modulation in pelvic CT

● Radiosurgical palliation of aggressive murine SCCVII

squamous cell carcinomas using synchrotron-generated X-ray

microbeams
● Solitary pulmonary nodule with growth and contrast

enhancement at CT: inflammatory pseudotumour as an

unusual benign cause
● Non-haemorrhagic subdural collection complicating rupture of

a middle cranial fossa arachnoid cyst

● Correspondence

● A deformed skull with enlarging hand and feet in a young

female
● Acknowledgment to Referees
The British Journal of Radiology, 79 (2006), 1–4 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/19232533
Commentary
The President’s Conference 2005: ‘‘Technology in Imaging
and Radiotherapy – towards improved workflow and
productivity’’
P P DENDY
Günter Dombrowe, the President of the British Institute
of Radiology (BIR), introduced the theme of this year’s
Conference, and explained its dual purpose – to highlight
the contributions of medical and information technologies
towards improving clinical practice, patient outcome and
health economics; and to pay tribute to the pioneering
work of Sir Godfrey Hounsfield, the inventor of CT
scanning, perhaps the key technology of the digital
imaging age.
This Commentary provides an overview of some of
the important topics discussed at the Conference.
Some of the key presentations are also included in this
issue.
Elizabeth Beckmann reminded the audience of the early
days of CT – the excitement generated by the images of the
brain shown at the 32nd Congress of the BIR on 20 April
1972, the delightfully understated title of Sir Godfrey’s
lecture – ‘‘Computerised axial tomography, a new means
of demonstrating some of the soft tissue structures of the
brain without the use of contrast media’’, and the
subsequent publications in the BJR [1, 2]. The enduring
memory of this and other early developments is that so
much was achieved with so little money. Was Sir Godfrey
one of the last brilliant, intuitive, string and sealing wax
physics brigade?
The first of the two nominated Hounsfield lecturers,
Willi Kalender gave a comprehensive review of the past,
present and future of CT from a physics and technology
standpoint. He pointed out that there had been three
distinct phases of development: (1) the 1970s had been a
time of rapid development with second, third and fourth
generation scanners; (2) the 1980s had been a period of
stagnation with the competing development of MRI (the
late 1980s was the only time during a 30 year period when
there was no increase in the number of CT scanners in
Germany); (3) the 1990s were the renaissance years,
particularly with the introduction of spiral CT and
multidetector arrays.
Scan times are now typically 0.3 s to 0.5 s per full 360 ˚
scan and 10–30 s for the whole body. The first figure is
important for temporal resolution, especially in cardiac
applications, and one of the limitations on faster times is
the centrifugal force to which sensitive components such as
the X-ray tube are subjected [3]. To achieve better
temporal resolution increased electronic control of the
beam and possibly multiple tube designs are being
explored.
The British Journal of Radiology, January 2006
Improvements in total scan time will be achieved
through further development of wider detector arrays,
possibly towards flat panel detectors. This will in turn
require X-ray tubes with an even higher peak output, as
the total flux of photons required to image a given volume
remains roughly the same.
Like for like, patient doses have been reduced with tube
current modulation both on rotation from anteroposterior
(AP) to lateral projections and as the beam traverses the
body from high to low attenuating regions. Achieving
the same counting statistics on all data is a worthwhile
goal [4, 5].
Since 1990 the emphasis has been on scanning volumes
rather than slices and one of the landmarks has been to
achieve isotropically uniform spatial resolution, typically
in the range 0.4–0.6 mm [6]. It is important to recall that
for isotropic resolution, radiation dose to the patient
increases with the fourth power of the resolution element.
These improvements must also be seen in the context of
global use of radiology. CT is a relatively high dose
technique, now accounting for 25% of all radiation
exposure, and there must be strong clinical justification
for its use, and in particular serial, repeat whole body
scans.
The future for CT is hidden from view but there are
many possibilities and it is worthwhile to summarize
Kalender’s predictions – more detector rows; shorter
effective scan times; higher resolutions and more tissue
parameters (there is renewed interest in superimposing, e.g.
a calcium density map on a real density map obtained by
dual energy CT [7]); lower doses (of course!).
The second nominated Hounsfield lecturer, Adrian
Dixon, reviewed the clinical advances in CT. Two
important issues in particular were addressed:
(1) Do the ‘‘advances’’ in CT technology make any
difference to the patient?
(2) Many cutting-edge CT investigations are still chari-
tably funded and if the NHS is to become
responsible for their provision, they must be shown
to be cost-effective.
As a specific example of the clinical issues, he considered
the impact of multidetector CT on abdominal problems.
The improved anatomical resolution of modern helical CT
scanners enables the diagnosis of acute appendicitis or the
cause of small bowel obstruction to be made with a high
degree of accuracy [8]. Consequent on its multitasking
1

abilities, CT is increasingly being used as a means of
triaging patients and facilitating early discharge for those
without serious disease – with obvious benefits to the
patients and cost savings to the NHS [9].
CT has become so good that in many areas of radiology
the real questions are now (a) is there a role for plain film
radiography? (b) when should ultrasound be used? (c) is
there a role for MR other than to avoid the use of ionizing
radiation?
This success has come at a price: clinicians are tending
to request a CT scan without fully examining the patient;
surgeons are reluctant to operate without high quality
imaging; for outpatients in oncology the number of
requests for CT staging is starting to approximate the
number of visits to hospital. However, Dixon was able to
conclude on a positive note. For the patient CT has
replaced some very unpleasant investigations.
The tribute to Hounsfield concluded with a more
specialized lecture from Albert de Roos on cardiac CT.
Roos summarized the technical considerations for multi-
slice CT in cardiac scanning – low contrast detection,
spatial resolution at high contrast, temporal resolution,
scan time and patient dose. The choice of acquisition
variables and reconstruction characteristics is very depen-
dent on the clinical problem under investigation.
De Roos then reviewed a wide range of applications
including: the quantitative assessment of coronary artery
calcification [10, 11]; the assessment of coronary artery
morphology; stent and graft patency; the selection of
patients for invasive therapy; assessment of the anatomy of
pulmonary veins and the investigation of acute chest pain.
In the last of these applications there is now a one-stage
protocol, i.e. the nirvana of the ‘‘one stop shop’’ to
diagnose accurately both cardiac and non-cardiac causes
of chest pain [12].
The Mackenzie Davidson lecture, delivered by Nicola
Strickland, touched on many aspects of modern imaging
but concentrated on information technology, especially
PACS.
PACS has now become a mature technology, especially
as a result of the DICOM standard and network
protocols. It clearly has the potential to improve workflow
and productivity but does not, in itself, solve departmental
inefficiencies and may highlight them. It is not a ‘‘quick
fix’’ and must be an integral part of workflow engineering.
Looking to the future, speech recognition and web
browsers will be developed further. The electronic patient
record remains a major challenge, since the facilities
provided need to match the service being provided. A
good example is home reporting – a full work load
requires a full diagnostic service, emergency reporting
needs only more limited facilities.
Strickland concluded that technology provides the
means for improving workflow and productivity – the
challenge is to optimize the use of technology to maximize
productivity in a clinically efficient way.
Manufacturers’ views of the use and development of
technology were also presented. Hermanns Requardt from
Siemens Medical Solutions reminded us that, worldwide,
challenges to healthcare systems are dominated by two
main topics – demographic factors and progress in
medicine. In diagnostic radiology, as in some other
branches of medicine, for example molecular/genetic
medicine, the challenge now is not a lack of information
2 The British Journal of Radiology, January 2006
P P Dendy
but a flood of information. Drawing an analogy from
industry where knowledge management systems are
commonplace, Requardt predicted that information tech-
nology would bring about a paradigm shift in medicine if
it could facilitate the formation of a clinical knowledge
database and enable this to be used to complement the
data from the individual patient.
Jacques Souquet from Philips Medical Systems con-
sidered some other aspects of the impact of future
technology on medical imaging. Picking up a theme
from the previous speaker on progress in medicine, he
pointed out that knowledge doubling times have fallen
from about 8 years in 1970 to 1 year in 2001. Increased
use of computer-aided decisions is one way to improve
management of data, for example nodule identification in
a radiograph, using embedded medical knowledge to
reduce avoidable medical errors, genetic algorithms to
discover diagnostic patterns in huge data sets.
Souquet reminded us that much remains to be done.
There are still several diseases for which no diagnostic test
is available and the development of drugs to correct
specific genetic flaws that are biological causes of cancer
has a long way to go. In conclusion, he threw out two
challenges:
(1) How can the translation from cell to mouse to man be
speeded up?
(2) How can the multidisciplinary constituencies contri-
buting to progress (basic sciences, engineering, medi-
cine, industry) be coordinated? This is a challenge that
is close to one of the fundamental aims of the BIR.
Jane Guinn from Kodak Ltd concluded the session by
comparing the techniques of computed radiography (CR)
and digital radiography (DR) from the standpoint of
workflow patterns. She listed 16 distinct stages in the
production of a traditional analogue film, many
involving radiographer movement. CR removed only
two steps, DR removed nine. This had a big impact on
average examination time and in a busy general radio-
graphy room, on patient waiting time. Unfortunately DR
does not provide the flexibility of CR for several
examinations.
Peter Williams delivered the Silvanus Thomson
Memorial Lecture. With the somewhat enigmatic title
‘‘Things can only get better’’ he reviewed the development
of external beam radiotherapy treatment delivery, con-
centrating on current developments and future promises.
Early examples of ‘‘things getting better’’ included
megavoltage therapy with Co-60; isocentric mounting;
electrons as well as X-rays; anatomical data from the CT
scanner for treatment planning. For a few years the ability
to model tumours exceeded the ability to treat, which was
restricted to a cylinder.
In 1987 the multileaf collimator (MLC) became avail-
able for beam shaping and as with most really worthwhile
medical developments, there were no formal health quality
assessments or clinical trials.
MLCs led to intensity-modulated radiotherapy (IMRT),
essentially conformal therapy for difficult targets [13, 14],
and at the same time electronic portal imaging was being
developed to provide active control of beam direction
rather than a passive verification system.
Williams then discussed the current development of real
time tumour tracking to counteract patient movement by
Commentary: President’s conference 2005
mounting a diagnostic machine with fluoroscopic, radio-
graphic and CT capabilities onto the treatment linear
accelerator. Examples of improved set-up were shown for
lung and bladder treatments – image-guided radiotherapy
will certainly make things better!
For the future, although physicists and engineers are not
yet spent (vide the next topic of proton therapy), they will
need help from other disciplines, e.g. molecular biologists
and geneticists (biological targeting for anoxia and
metabolism, and selective targeting of tumour cells), and
from radiobiologists (for example to exploit the
information on bystander effects coming from microbeam
studies).
As a fitting sequel to the Silvanus Thomson Memorial
Lecture, Bleddyn Jones presented the case for particle
therapy, especially with protons. The theoretical advan-
tages of using the Bragg dose peak to improve the
therapeutic ratio have been known for many years.
Unfortunately, for a 60 MeV beam the peak is at only
3 cm depth and treatment is limited to quite superficial
tumours. Notwithstanding, over 1200 choroidal melano-
mas have been treated successfully at the Clatterbridge
Hospital.
Work by Lomax et al [15] has shown that for treatment
of the breast and regional nodes, a 9-field photon IMRT
approach can either produce similar dose homogeneity
across the planning treatment volumes to that of a proton
plan, or similar sparing of dose to both lungs and the
heart, but not both.
Jones estimated that 10–20% of patients might be better
treated by particle radiotherapy and believes that technical
improvements in physics, bioengineering and computing,
especially in robotics and particle delivery, now make
treatment with a 200 MeV beam, with Bragg peak depths
approaching 20 cm, a practical proposition. It is antici-
pated that this will lead to a big increase in demand for
particle therapy in the UK [16].
The Conference concluded with two further papers in
diagnostic imaging. Catherine Owens gave a wide-ranging
review of the changing practice of paediatric imaging. The
diagnostic capability and accuracy of multidetector CT
(MDCT) angiography was compared with echocardio-
graphy, cardiac catheterization and surgery in the assess-
ment of the great vessels in 40 consecutive patients (mean
age 5 years) with congenital heart disease. MDCT was
accurate, showing good agreement with interventional
catheter and surgery and provided additional information.
Effective doses of radiation were low – ranging from
0.97 mSv in neonates to 1.7 mSv in adolescents [17].
Magnetic resonance coronary angiography and late-
enhancement imaging have been shown to be feasible in
children who had undergone arterial switch for transposi-
tion of the great arteries. Diagnostic quality images were
acquired in 72% of the coronary arteries imaged and this
rose to 100% in subjects over 10 years old [18].
Finally, Peter Ell discussed the contribution of PET/CT
to improved patient management. Whilst acknowledging
the important contribution in neurology and cardiology, in
the limited time available and in the context of the
Conference, Ell concentrated on oncology. Four distinct
areas were covered, diagnosis, staging, radiotherapy
planning and treatment monitoring.
Two very different challenges for this wonderful
technique were highlighted. At the cutting edge of research
The British Journal of Radiology, January 2006
there are almost unlimited opportunities for PET/CT to be
used to assess the biology of individual response to
treatment [19]. Whilst recognizing the importance of F-18
fluorodeoxyglucose in oncology, Ell emphasised the need
to look at a wide range of other novel markers that are
being developed, aimed at imaging proliferation [20, 21],
hypoxia, angiogenesis, apoptosis, etc.
At the other extreme there is the huge problem of
diffusion of technology in a cost-effective way so that, on a
day-to-day basis, many more of the millions of cancer
sufferers can benefit from the power of multimodality
imaging.
Ell’s concluding remarks were:
N PET/CT has changed patient management;
N It is best at assessing extent and severity of cancer;
N It informs radiotherapy planning; and
N It combines the power of CT with the unique
metabolic mapping obtained with PET.
These remarks were, of course, addressed to PET/CT
but, in many respects, with suitable changes of wording,
could be applied to the impact of other technological
advances discussed during the 2005 President’s
Conference. We commend to you the full articles
contributed by the speakers in this issue of the Journal.
Acknowledgments
I am grateful to Fergus Gleeson and Günter Dombrowe
for helpful contributions to this Commentary.
References
1. Hounsfield GN. Computerised transverse axial scanning
(tomography). Part 1 description of system. Br J Radiol
1973;46:1016–22.
2. Ambrose J. Computerised transverse axial scanning (tomo-
graphy). Part 2 clinical application. Br J Radiol
1973;46:1023–47.
3. Shardt P, Deuringer J, Freudenberger J, Hall E, Knipfer W,
Mattern D, et al. New X-ray tube performance in computed
tomography by introducing the rotating envelope tube
technology. Med Phys 2004;31:2699–706.
4. Kalender WA, Wolf H, Seuss C. Dose reduction in CT by an
anatomically adapted tube current modulation. Med Phys
1999;26:2248–53.
5. Greess HR, Wolf H, Suess C, Lutze J, Kalender WA, Bautz
WA. Automatic exposure control to reduce dose in subsecond
multislice spiral CT – Phantom measurements and clinical
results. Radiology 2002;225 Suppl. RSNA programme p 593.
6. Kalender WA. Thin-section three dimensional spiral CT. Is
isotropic imaging possible? Radiology 1995;197:578–80.
7. Kalender WA, Klotz E, Suess C. Vertebral bone mineral
analysis: an integrated approach with CT. Radiology
1987;164:419–23.
8. See TC, Ng CS, Watson CJE, Dixon AK. Appendicitis:
spectrum of appearances in helical CT. Br J Radiol
2002;75:775–81.
9. Ng CS, Watson CJE, Palmer CR, See RC, Beharry NA,
Housden BA, et al. Evaluation of early abdominopelvic
computed tomography in patients with acute abdominal pain
of unknown cause – prospective randomised study. BMJ
2002;325:1387–9.
10. Girshman J, Wolff SD. Techniques for quantifying coronary
artery calcification. Semin Ultrasound CT MR 2003;24:33–8.
3

11. Thompson GR, Partridge J. Coronary calcification score:
the coronary-risk impact factor. Lancet 2004;363:
557–9.
12. White CS, Kuo D, Keleman M, Jain V, Musk A, Zaidi E,
et al. Chest pain evaluation in the emergency department; can
MDCT provide a comprehensive evaluation? AJR Am
J Roentgenol 2005;185:533–40.
13. Williams PC. IMRT: delivery techniques and quality
assurance. Br J Radiol 2003;76:766–76.
14. James HV, Scrase CD, Poynter AJ. Practical experience
with intensity modulated radiotherapy. Br J Radiol
2004;77:3–14.
15. Lomax AJ, Cella L, Weber D, Kurtz JM, Mirabell
R. Potential role of intensity-modulated photons and protons
in the treatment of the breast and regional nodes. Int J Radiat
Oncol Biol Phys 2003;55:785–92.
16. Jones B, Burnet NG, Price P, Roberts JT. Modelling the
expected increase in demand for particle therapy: implications
for the UK. Br J Radiol 2005;78:832–5.
4 The British Journal of Radiology, January 2006
P P Dendy
17. Benson C, Taylor A, Ross UG, et al. Three-dimensional anatomy
of the great vessels defined by 16-slice multi-detector CT
angiography in neonates, infants, children and adolescents with
congenital heart disease. Presented at the 42nd Congress of the
European Society for Paediatric Radiology, Dublin, June 2005.
18. Taylor AM, Dymarkowski S, Hamaerkers P, et al.
MR coronary angiography and late-enhancement myocardial
MR in children who underwent arterial switch surgery
for transposition of great arteries. Radiology 2005;234:542–7.
19. Bugarolas J, Clark JW, Chabner B. Using ‘‘rationally
designed drugs’’ rationally. Lancet 2003;361:1758–9.
20. Shields AF, Grierson JR, Dohmen BM, et al. Imaging in vivo
proliferation with 18FLT and positron emission tomography.
Nature Medicine 1998;11:1334–6.
21. Francis DL, Visvikis D, Costa DC, Croasdale I,
Arulampalam TH, Luthra SK, et al. Assessment of recurrent
colorectal cancer following 5-fluorouracil chemotherapy using
both 18FDG and 18FLT. Eur J Nucl Med Mol Imaging
2004;31:928.
The British Journal of Radiology, 79 (2006), 5–8 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/29444122

President’s conference paper

CT scanning the early days
E C BECKMANN, BSc(Eng)
Lanmark, Beaconsfield, Bucks, UK

Abstract. CT scanning has become an established diagnostic tool within the radiology department. This article
covers some of the history of the development and early days of CT scanning. It is based upon the lecture given
on the Memorial Day for Sir Godfrey Hounsfield during the British Institute of Radiology President’s
Conference 2005.

It is less than 34 years ago, on 20th April 1972, that an
unknown engineer from EMI Ltd, the company better
known at the time for publishing the Beatles records,
gave a presentation at the 32nd Congress of the British
Institute of Radiology. The Engineer, Godfrey Hounsfield,
was lecturing with Dr James Ambrose from Atkinson
Morley’s Hospital on ‘‘Computerised Axial Tomography
(A new means of demonstrating some of the soft tissue
structures of the brain without the use of contrast media)’’
[1, 2]. Many people attending that BIR congress will never
forget the experience of hearing a presentation on CT
scanning for the first time. In fact Hounsfield had
presented the results of some of his animal experiments
the previous year at the 2nd congress of the European
Association of Radiology, in Amsterdam, but they had not
excited much interest. The same might have happened in
the USA because at a Neuro Postgraduate Course at the
Albert Einstein College of Medicine, New York, on
Monday 15th May 1972, only about a dozen people
stayed to hear an extra lunchtime lecture by Hounsfield
and Dr Bull, where they showed the first clinical images.
However these people realised the significance of what they
had seen and the news spread rapidly.
The beginning
In the mid 1960s Hounsfield was working on the pattern
recognition of letters when he began to consider whether
he could reconstruct a three-dimensional representation of
the contents of a box from a set of readings taken through
the box at randomly selected directions. He found that by
considering the three-dimensional object within the box as
a series of slices, reconstruction was easier than treating
the content as a volume.
He tested the theoretical principal by working with a matrix
of numbers set to zero with a square in the middle where each
number was set at 1000. He entered these data into a com-
puter programme to get simulated absorption values and
then reconstructed the picture using another programme.
Hounsfield recalled his surprise at how accurate the result was.
Here he stated ‘‘The purpose of the study was to investigate
the employment of a computer to make better use of the
information obtained when an object is examined by gamma
rays or X-rays’’. In this proposal Hounsfield compared the
classic conventional X-ray technique producing a confused
and fuzzy picture to the clear outline produced by the
proposed system. Hounsfield proposed a system as shown in
Figure 1 based upon reconstructing pictures of slices through
an object and in detailing the expected benefits he indicated a
theoretical accuracy of detection better than 1%.
The lathe bed model
The initial test rig was built on the bed of an old lathe which
Hounsfield had been using in a previous project working on
computer stores. Hence the early test unit became referred to
as the ‘‘Lathe bed model’’. The initial rig utilized a gamma
source, Americium 95, with a photon counter as the detector.
On this rig, the source made 160 traverses of the object, which
was rotated 1 ˚ at the end of each traverse for a total of 180 ˚. It
took 9 days to collect sufficient information, and 2.5 h to
reconstruct the image on an ICL 1905 mainframe computer.
However, the resultant images proved the feasibility of the
technique and with the replacement of the gamma source by
an X-ray source as shown in Figure 2, the scanning time was
reduced to 9 h.
Initial images were of inert objects, then specimens from an
abattoir, including bullocks brains and pigs bodies as shown
in Figure 3. Due to the long scan times, particularly with the
gamma source, many of these specimens decayed while the

The project proposal

Once Hounsfield had proved the theoretical principle he
went on to generate the original project proposal in 1968.

Received 12 September 2005 and accepted 16 September 2005. Figure 1. Extract of the original 1968 project proposal.

The British Journal of Radiology, January 2006 5


Figure 2. The original lathe bed model (copyright EMI Ltd).
pictures were acquired producing gas bubbles which caused
artefacts in the images. This initial work was done by a very
small team comprising Hounsfield, Stephen Bates (program-
ming), Peter Langstone (electronics) and Mel King
(mechanics) working on a very low budget of £25 000.
Dr James Ambrose recalls that, in about 1969, he received a
call from an old acquaintance, Dr Evan Lennon then
principal medical officer in Radiology at the Department of
Health, asking if he would see a man called ‘‘Godfrey
Hounsfield’’ and listen to him. Lennon had found him
confusing but was reluctant to dismiss him as a crank
(Ambrose later learnt that other eminent radiologists had
already dismissed him as a crank!). Ambrose recalls that when
he and his senior physicist Dr John Perry met Hounsfield, the
conversation was difficult. Hounsfield would only say that the
method was fundamentally different from other methods of
X-ray imaging, more efficient in photon usage and likely to be
more sensitive to small density variations. In order to
demonstrate a clinically relevant image, Ambrose arranged
for a bottled specimen of a brain to be borrowed from a
museum and was amazed at the image Hounsfield showed
him 5 weeks later. An image of the first brain scanned is
shown in Figure 4.
Building the prototype
Having shown some clinically interesting images the
project was then ready to move to the next stage of
building a full prototype. However funding was an issue. It
was Gordon Higson at the Department of Health who had
Figure 3. Early scan of a pig.
6 The British Journal of Radiology, January 2006
E C Beckmann
Figure 4. First image of a brain specimen.
the foresight to place an order for a machine with a
theoretical specification which included a 4–5 min scan
time and an 0.5% pixel accuracy, and this enabled the
project to continue. This order was for a prototype and
three clinical machines that would generate sufficient
income to fund a fifth machine for Hounsfield and his
team to keep and work on. The Department of Health
order would also fund half the remaining research costs
and in exchange they would receive a small royalty on
sales. At the time it was calculated that it would cost
£69 000 to build a complete working system and so it was
agreed that the Department of Health would pay £150 000
for each of the four systems.
The first clinical patient
The prototype was installed at Atkinson Morley’s
Hospital in South London where the first patient, a
middle aged lady with a suspected frontal lobe tumour,
was scanned on 1st October 1971. The surgeon who
operated on her shortly afterwards reported that ‘‘it looks
exactly like the picture’’ shown in Figure 5.
Hounsfield remained cautious. He recalled ‘‘I’ve had this
before, first time is always lucky and then everything else
goes wrong after that. So I thought, the next ones are not
going to be any good, but they did another ten more
patients and every one of them came out as being obvious
diseases of the brain showing up in various forms. Dr
Ambrose found that, by injecting iodine-based contrast
agent that would localize the particular spot where the
tumour was and it showed up even better’’. Hounsfield
took some of the contrast enhanced images and subtracted
without contrast images to compare the blood flow on
either side of the brain.
In the original system the patient’s head was placed in a
rubber cap surrounded by water. This water bag was used
to reduce the dynamic range of the detected X-rays and
improve the absolute values of the attenuation figures.
Using one sodium iodide (NaI) crystal and photomul-
tiplier tube detector per slice, plus one as a reference
detector with a scan time of 4.5–20 min per 180 ˚ scan, the
system acquired two contiguous slices per scan each with a
80680 matrix of 3 mm63 mm613 mm voxels. Early
images showed the ability to meet the pixel density
accuracy of 0.5% in the absorption coefficient as defined in
the theoretical specification.
CT scanning the early days
Figure 5. First patient image scanned on the prototype EMI
scanner at Atkinson Morley’s Hospital on 1st October 1971.
The three systems ordered by the Department of Health
were installed at the National Hospital for Neurology and
Neurosurgery in London, Manchester and Glasgow. After
this, the first CT scanners were installed in the USA at the
Massachusetts General Hospital and the Mayo Clinic, where
the first scan in the USA was done on 19th June 1973.
Reconstructing the picture
Early scan data were actually taken back to EMI on tape
for processing overnight which took 20 min per image on an
ICL 1905 computer. In production this was done on a mini-
computer which fortuitously had emerged at the right time.
Images were taken back the next day on tape to Atkinson
Morley’s Hospital to be displayed. The early images were
displayed in three ways; paper printout, cathode ray tube
(CRT) display or as a Polaroid picture of the CRT display.
The early images were generated using iterative algebraic
reconstruction implemented by Steve Bates on the ICL 1905
mainframe. Subsequently reconstruction used the filtered
back projection or convolution method invented and
patented by Chris Lemay, one of the many patents filed
and held by Hounsfield and his team. On the original EMI
Mk1 scanner an 80680 image took 7 min to process, with
filtered back projection on the same computer a 1606160
image could be processed in 30 s after the end of the scan.
It had been thought that image reconstruction and
processing was so complicated that it would have to be
The British Journal of Radiology, January 2006
done at a central processing unit on a suitable large and
fast main frame machine.
But the introduction of the mini computer and the
implementation of the new improved reconstruction
algorithms were to change this.
CT1010 scanner
A challenge with the original EMI Mk1 scanner was the
water bag, both as regards the ease of use with patients
and also due to the occasional water leak! Replacement of
the water bag with shaped carbon fibre wedges and bean
bags was a significant improvement. This was further
enhanced by the increase to eight detectors per slice in the
CT1010 which was still a two contiguous slice scanner
offering 1606160 and 3206320 matrix sizes over a
210 mm scan diameter and with the minimum scan time
improved to 1 min. The prototype of this system was
installed in 1975 at Atkinson Morley’s Hospital and
showed significant improvement in clinical image quality.
Body scanning
The feasibility of body scanning was proved when a slim
member of the EMI team, Tony Williams, was scanned in
a head scanner.
The first body images taken in the body prototype
machine were of Hounsfield himself on 20th December
1974. The first body images were shown to a meeting at
the first International Conference on CT Scanning in
Bermuda on Friday 14th March 1975, one of these images
is shown in Figure 6.
All the research machines were named after stones:
Opal, Pearl, Garnet and the body prototype was Emerald.
This Emerald system was first installed clinically at
Northwick Park Hospital in March 1975. The first body
scan carried out in the USA was in October 1975 at the
Mallinkrodt Institute St Louis. Dr Ron Evans recalled
that this was a jaundiced patient, in whom it had been
difficult to differentiate between medical and surgical
jaundice. The CT scans showed that it was surgical
jaundice which was subsequently clinically confirmed.
Initially known as the CT5000, the body scanner was
developed into the commercial production machine, the
CT5005. These body scanners were single slice machines using
a gantry with 30 detectors plus a reference detector to reduce
scan time to 20 s. The matrix had been increased to 3206320
over a selectable 240 mm, 320 mm or 400 mm scan field.
The generation game
All these early scanners were the so called 1st or 2nd
generation utilizing the translate/rotate technology where
the gantry scanned across the patient before indexing by
one degree and scanning back.
An early problem in CT scanner design was detector
stabilization and the need for calibration. The EMI scanners
were using NaI crystal photon detectors and photo multiplier
tubes, and the translate/rotate technology enabled detector
calibration by taking air readings at the end of each translate
movement. This gave high accuracy but limited the speed of
the scan. By 1976 there were 17 companies offering CT
scanners with 3rd generation rotate/rotate scanners having
7

Figure 6. Body scan of Hounsfield taken on the prototype
scanner in the laboratories and shown at Bermuda conference
on 14th March 1975.
been introduced, to offer fast scan times, most based upon
xenon gas detectors arranged in an arc [3].
Hounsfield realised the need for a system that was faster
than translate/rotate and that could overcome the
calibration and artefact issues of rotate/rotate systems.
Topaz
The patent for a scanning focus system to produce a
true volume scanner was filed on 19th October 1976. The
Topaz research system, also named after a stone and
shown in Figure 7, was a 3rd generation system with a
flying X-ray spot. The X-ray flying spot scanned in a
direction opposite to the direction of rotation of the
machine which meant that the body could be scanned with
arcs of detector readings which overlapped in such a way
that they could be compared and continuously calibrated.
Built with 612 detectors including a central zoom region,
Topaz had a resolution in the x-y plane of 0.65 mm.
Volume scans taken in June 1980 were displayed in three
dimensions in real time as 1200612006270 pixels.
Recognition
Initially the scale for describing the attenuation
coefficients was referred to as EMI numbers. This was
then expanded by a factor of two and became known as
Hounsfield units (H) where
ktissue {kwater
H~ |1000
kwater
and m is the linear attenuation coefficient. Each Hounsfield
unit is equivalent to 0.1% of the attenuation of water [3].
In addition to giving his name to the unit of attenuation,
Hounsfield received many awards including the BJR
Barclay prize jointly with Ambrose in 1974, the Nobel Prize
for Physiology or Medicine in 1979 [4] and a Knighthood
in 1981.
8 The British Journal of Radiology, January 2006
E C Beckmann
Figure 7. Topaz 3rd generation flying focal spot scanner.
Hounsfield and his team created the CT scanner,
which has had an explosive impact on diagnostic
radiology, with little money and few resources. By the
end of the 1970s they already had plans for many of the
technologies which were to develop the CT scanner over
the next 30 years, including helical multislice scanners and
high power continuously rated scanned beam X-ray tubes.
They developed many of the techniques which formed
the foundation of modern imaging including image
subtraction. By 1976 the reconstruction techniques used
in CT were already being applied to other areas including
ultrasound and nuclear magnetic resonance.
Acknowledgments
The author is indebted to many people especially those
members of the original EMI team who worked with Sir
Godfrey Hounsfield for their input to the original lecture
and material used in this article.
References
1. Hounsfield GN. Computerised transverse axial scanning
(tomography): Part 1. Description of system. Br J Radiol
1973;46:1016–22.
2. Ambrose J. Computerised transverse axial scanning (tomo-
graphy): Part 2. Clinical application. Br J Radiol
1973;46:1023–47.
3. Brooks RA, Di Chiro G. Principles of computer assisted
tomography (CAT) in radiographic and radioisotropic ima-
ging. Phys Med Biol 1976;21:689–732.
4. Computed medical imaging. Nobel lectures in physiology or
medicine 1971–1980; 568–86.
The British Journal of Radiology, 79 (2006), 9–16
DOI: 10.1259/bjr/67045628
President’s conference paper
Cardiac applications of multislice computed tomography
1
A DE ROOS, MD, 1L J M KROFT, MD, 2J J BAX, MD, 1H J LAMB, MD and 1J GELEIJNS, PhD
Departments of 1Radiology and 2Cardiology, Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The
Netherlands
Multislice CT (MSCT) is gaining clinical acceptance for
cardiac imaging owing to improved temporal and spatial
resolution of the latest 16-slice and 64-slice technology.
Although the cardiac MSCT applications are promising,
there is still room for further technical improvements and
optimization of post-processing techniques for cardiac
evaluation.
Interestingly, the data acquired for CT angiography of
the coronary arteries can also be used to create volumetric
cine loops of cardiac function. The functional data are
available without the need for repeat scanning or for
administration of additional contrast material [1].
Furthermore, MSCT allows assessment of first-pass
perfusion and delayed enhancement imaging in patients
with subacute myocardial infarction. Recently, it has been
reported that MSCT reveals microvascular obstruction or
the so-called no-reflow phenomenon as a late perfusion
defect in patients with re-perfused acute infarctions,
similar to observations made by other techniques like
MRI [2]. With further development MSCT may allow
combined assessment of the presence and extent of
coronary atherosclerosis, the percent diameter stenosis,
plaque characterization and the effect of the lesion on
perfusion and myocardial function.
In this review, the technical requirements of cardiac
MSCT and some frequent clinical applications are
discussed.
MSCT imaging requirements
Requirements for cardiac MSCT image acquisition
depend strongly on the clinical problem. For example,
CT coronary angiography requires excellent spatial and
temporal resolution, whereas only modest spatial and
temporal resolution is sufficient for the assessment of the
anatomy of pulmonary veins and the left atrium. In
general, the higher the requirements for image quality
become, the more complex the acquisition, the longer scan
time and the higher patient dose. Main aspects with regard
to imaging performance are low-contrast and spatial
resolution, temporal resolution, and scan time. Patient
dose and radiation risk should always be considered as the
counterpart of image acquisition and image quality.
Low-contrast resolution and spatial resolution
Low-contrast resolution is the ability to visualize
structures that demonstrate only a small difference in
Received 22 September 2005 and accepted 5 October 2005.
The British Journal of Radiology, January 2006
E 2006 The British Institute of Radiology
Hounsfield units compared with their direct environment.
In cardiac applications of CT, native tissue contrasts are in
general not sufficient to differentiate between, for example,
the vessel wall and its unenhanced lumen, or the heart and
the inner chambers. Contrast enhancement is thus
mandatory for visualizing the lumen of coronary arteries,
the heart chambers, pathology of the myocardium or
anatomy of pulmonary veins. Low-contrast resolution
depends on tube current (mA), the reconstructed slice
thickness, tube voltage, beam filtration and the reconstruc-
tion algorithm, and is strongly correlated to radiation
exposure. In general, low-contrast resolution performance
of CT scanners is not a limitation for the application of
cardiac CT.
Spatial resolution, or high-contrast resolution, deter-
mines the ability to visualize contours of small structures
within the scanned volume. Small objects can only be
resolved when there is a rather large contrast with the
direct environment. Considerable improvement of spatial
resolution in clinical acquisitions was achieved with the
latest generations of multislice CT scanners. This is of
importance, particularly for the application of CT
coronary calcification scoring and CT coronary angio-
graphy. The actual diameters of the lumen of normal
coronary artery segments range from 5 mm in the
proximal segments to less than 1 mm in the distal
segments [3]. This means that spatial resolution of
1.0 mm in all three dimensions should be sufficient for
imaging of the coronary arteries, except for distal segments
that would require a spatial resolution of at least 0.5 mm.
Bypass graft diameter typically ranges from 4 mm to
6 mm. A spatial resolution of 2 mm3 (voxel size) might
thus be sufficient for imaging the lumen of bypass grafts.
For imaging of small structures within the coronary
arteries, such as atherosclerotic plaque and stents, excellent
spatial resolution, even better than 0.5 mm3, might be
required. Voxel size is often used as an indicator of
spatial resolution. However, voxel size should be inter-
preted with care since smaller voxel size does not
necessarily imply better spatial resolution. Spatial
resolution is preferably expressed as the response of a
delta-function; in CT, this response is either called a point-
spread-function (spatial resolution in the axial plane) or a
slice sensitivity profile (spatial resolution along the z-axis).
Spatial resolution is limited by the acquisition geometry of
the CT scanner, the reconstruction algorithm and the
reconstructed slice thickness. The performance of current
64-slice scanners with regard to spatial resolution,
expressed as the full-width half-maximum of the response
of a delta-function, is within the range 0.6–1.0 mm in all
three dimensions.
9
 A de Roos, L J M Kroft, J J Bax et al

Temporal resolution cycles. The more heart cycles that can be included in the
reconstruction, the better the temporal resolution. A low
Temporal resolution determines whether fast moving pitch factor, which is typical for cardiac CT acquisition, is
objects can be resolved in the CT image. Good temporal required to acquire data from more than one heart cycle.
resolution limits motion artefacts and blurring of the A pitch factor as low as 0.2 is required to record at least
image. Principally, good temporal resolution can be two heart cycles and to achieve a temporal resolution in
achieved by a short reconstruction window providing the order of magnitude of 100 ms for typical heart rates
snap shots of the beating heart and coronary arteries. between 60–80 BPM. Figure 1 shows, as an example, the
Good temporal resolution in cardiac CT is realised by fast temporal resolution that is achievable with a reconstruc-
data acquisition (fast rotation of the X-ray tube), but even tion algorithm that can merge transmission data from an
more importantly by a dedicated reconstruction algorithm. unlimited number of heart cycles. The figure illustrates the
A recent paper [4] provides information on the rest dependence of the reconstruction window on rotation time
period of the heart, which is a measure for the required and heart rate and was calculated for a pitch factor of 0.2.
reconstruction window. The rest period is defined as the From Figure 1 it can be concluded that, for achieving the
time during which the 3D motion of a coronary artery is shortest reconstruction window, rotation time should be
less than 1 mm. It was reported that, for patients with a adapted to the heart rate.
heart rate of 64¡9 beats per minute (BPM), the end-
systolic rest period duration was 76¡34 ms; and the mid-
diastolic rest period duration was 65¡42 ms for the
proximal to middle segments of the right coronary artery. Scan time
For the left coronary artery tree, the end-systolic rest Scan time is the time interval between the start and the
period duration was 80¡25 ms; the mid-diastolic rest end of one acquisition, sometimes referred to as a
period duration 112¡42 ms. From these data it is sequence. To avoid breathing artefacts and to limit the
concluded that the duration of a ‘‘snap shot’’ of the amount of contrast material in contrast enhanced acquisi-
coronary arteries, or in other words the reconstruction tions, scan time in cardiac CT should remain at least below
window, should be shorter than 65–110 ms. This is in good 30 s, but preferably below 20 s. The extent of the target
agreement with earlier papers; in one paper it is suggested volume, as well as acquisition parameters such as rotation
that the reconstruction window should be lower than time, pitch factor, section thickness and number of
100 ms for coronary angiography in mid-diastole at simultaneously acquired sections, define scan time. In
62¡10 BPM [5], and in another paper it is stated that a general a much shorter scan time than 20 s can now be
100 ms reconstruction window is relatively optimal for realised with the current generation of 64-slice scanners for
most patients at heart rates up to 90 BPM [6]. All of these typical cardiac CT examinations; for example, a typical
considerations assume image reconstruction at the cardiac
phase point that is associated with least motion, e.g. a
reconstruction window starting between 60% and 80% of
the interval between two consecutive R-waves. More strict
criteria for the reconstruction window apply if the heart
should be assessed at more than one cardiac phase point,
including those that are associated with rapid movement
of the heart wall, e.g. for studying the dynamics of the
myocardium. More strict criteria apply as well when a
1 mm displacement of a coronary artery within the
duration of the snap shot becomes unacceptable. This
may happen, for example, when imaging small distal parts
of the coronary arteries, quantifying coronary stenoses and
assessment of coronary atherosclerotic plaque.
General reconstruction algorithms that are used for
general CT applications provide, in principle, a temporal
resolution equal to the rotation time (360 ˚ rotation, full
reconstruction), the best achievable temporal resolution
with general reconstruction algorithms is slightly longer
than 50% of the rotation time (180 ˚ rotation, half
reconstruction). Current 64-slice scanners that are used
for cardiac applications provide a rotation time of 330–
400 ms. These typical rotation times are not short enough
for achieving a 100 ms or shorter snap-shot of the heart,
even if a 180 ˚ rotation half-reconstruction is applied. Figure 1. Temporal resolution of CT coronary angiography.
Therefore, dedicated reconstruction algorithms are used in The temporal resolution depends strongly on the rotation time
and the reconstruction algorithm. In segmented (multiphase)
cardiac CT that allow for reconstruction of synchronized
reconstructions, temporal resolution depends also strongly on
images from transmission data acquired during two or the pitch factor. The lower the pitch factor, the more cardiac
more successive heart cycles according to a method phases are captured during the acquisition and the better tem-
described already in 1977 [7]. These so-called segmented poral resolution. The graphs are calculated for a pitch factor
(multicycle) reconstruction algorithms allow for merging of 0.2. The graphs clearly show the dependence of temporal
synchronized transmission data from successive heart resolution on heart rate and rotation time.

10 The British Journal of Radiology, January 2006

Cardiac MSCT

Reconstruction
windowb (ms)
scan time for calcium scoring is 2.5 s, for coronary
angiography 10 s and for an ungated acquisition of the
pulmonary veins 3.0 s.

Half reconstruction 250

100

100

100

Half reconstruction 250

Patient dose in MSCT
Radiation protection of patients is based on justification

Multisegmental

Multisegmental

Multisegmental
Reconstruction
and optimization. Justification implies that the benefit for

time (s) algorithm

the patient outweighs the risk of radiation exposure.
Patient dose assessment is required for balancing harm
and benefit of the CT examination and to assess the effect
of measures for optimization of cardiac CT. Nowadays,
most CT scanners provide the user with an indication of

Scan range Scan

17.3
patient dose in the form of the CT dose index (CTDI) and

2.5

9.5

9.0

2.7
dose–length product (DLP). Effective dose can be derived
from these dose quantities. Effective dose from cardiac CT
coronary angiography is relatively high, mainly due to the

(mm)
need to catch more than one cardiac cycle and the

Not applicable 120

120

240

120

120
resulting low pitch factor. On the other hand, effective
dose from an ungated acquisition, such as in ungated
pulmonary vein CT angiography, is relatively low due to
the high pitch factor. Effective dose for calcium scoring,
assessment of ventricle function or pulmonary veins is in

Pitch

0.19

0.19

0.19

0.83
the range 1–3 mSv, effective dose for CT coronary
angiography is considerably higher, e.g. in the range 10–

configuration times (s) voltage (kV) current (mA)

15 mSv. Concern about radiation exposure stimulates
the development of methods for dose reduction in
cardiac CT coronary angiography. The field of view of
Tube

200

300
300

300

40
interest in cardiac CT is rather small and therefore
radiation exposure of tissue outside this field of view
can be limited by means of a special ‘‘small field’’ beam-
Table 1. Typical acquisition and reconstruction characteristics of some cardiac CT examinations

shaping filter. Another method for dose reduction is to

Acquisition Rotation Tube

reduce X-ray output during the systolic phases that are

120

120

120

120

100
expected to be of less interest for the evaluation of
the coronary arteries (ECG triggered modulation of
dose). Pitfalls of small field scanning are the occurrence
0.25a

0.4

0.4

0.4

of artefacts and reduced image quality. A pitfall of 0.4

tube modulation is reduced image quality at certain
(n6T mm)

relevant phases of cardiac cycle, e.g. due to an irregular

heart rate.
6460.5

6460.5

1662.0

6460.5
463

Clinical applications
No synchronization
Synchronization

MSCT provides special opportunities for cardiovascular

Retrospective

Retrospective

Retrospective

CT in addition to angiography of the coronary arteries

triggering
Prospective

and coronary bypass grafts. These options include

gating

gating

gating

assessment of left ventricular (LV) and right ventricular

(RV) function, coronary calcification score, myocardial
infarction imaging and assessment of the anatomy of
(ml s-1, ml)

pulmonary veins in patients with atrial fibrillation. Each of

Acquisition Contrast

these applications can be characterized by their specific

2.5/40
4/100

4/100

techniques for acquisition and reconstruction. Table 1

none

5/70

provides information about typical acquisition and

60 beats per minute assumed.

reconstruction parameters for some clinically established

Calcium scoring Sequential

cardiac CT applications.
Spiral

Spiral

Spiral

Pulmonary veins Spiral

Quantitative assessment of coronary artery calcification

Partial rotation.
CTA coronary

CTA coronary

Coronary artery calcification is a marker for athero-

Examination

RV function

sclerotic lesions in the coronary arteries. The amount of

arteries

bypass

coronary artery calcification is correlated to the risk of

coronary events. However absence of coronary artery
calcification does not rule out atherosclerosis. Applications
a
b

The British Journal of Radiology, January 2006 11


of quantitative assessment of coronary artery calcification
are screening of asymptomatic individuals with risk factors
for coronary artery disease and follow-up of patients who
received medication for the treatment of coronary artery
disease.
Coronary artery calcification is well visualized with
X-ray techniques such as radiography but only CT
provides a non-invasive method for detecting and
quantifying coronary artery calcification [8]. Coronary
calcification is best detected and measured in a plain CT
acquisition without contrast enhancement.
Quantification of coronary calcium was introduced in
1990 by Agatston et al [9]. They used electron beam
tomography and established the ‘‘Agatston score’’. The
Agatston score requires an acquisition with a special
protocol (3 mm contiguous slices, 130 kV). The Agatston
score is achieved by setting a threshold for the Hounsfield
unit (130 HU) and for the size of the lesion (1 mm2). Then
a pragmatic weighting of the calcified area is applied
depending on the maximum HU in the lesions for each
image. The total calcium score is calculated by summing
the weighted areas for all images (Figure 2).
With the introduction of MSCT, new acquisition
protocols came into use; prospective ECG triggering in
combination with a half (180 ˚) reconstruction at 120 kV is
now generally used for calcium scoring. In prospective
ECG triggered MSCT acquisitions, the patient is only
exposed within the 170–200 ms acquisition window at
diastole and radiation exposure is therefore significantly
less compared with retrospective gated MSCT cardiovas-
cular examinations. The application of MSCT for
quantification of coronary calcium made it mandatory
Figure 2. Coronary artery calcification imaging at 64-row
multidetector CT (MDCT). 64-row MDCT of a 52-year-old
male patient with risk factors for coronary artery disease.
Small calcifications in the left anterior descending artery. The
total calcium score according to Agatston was 21, and the
total volumetric score was 25, indicating mild atherosclerotic
plaque with mild or minimal coronary artery narrowings likely.
CT-angiography revealed no coronary artery stenoses.
12
A de Roos, L J M Kroft, J J Bax et al
to switch to new quantification methods that can be
compared for different scanners and that are robust with
respect to different scanners and acquisition protocols.
Alternatives for the Agatston score are the volume score
(the volume of all voxels exceeding a certain threshold)
and calcium mass (mg) [10]. The latter quantity holds the
promise of providing the best physical measure for
coronary artery calcification. Unfortunately, there is still
a lack of standardization of the MSCT techniques with
regard to image acquisition as well as to the methodologies
for quantitative coronary calcification scoring. The devel-
opment of standardized and reproducible protocols is a
technical prerequisite for coronary calcification scoring to
become a useful clinical tool. In addition, for screening
purposes, the coronary calcification score will have to be
established as an independent predictor of existing risk
factors for cardiovascular disease [11].
Coronary angiography
MSCT has rapidly evolved through different stages of
technological innovation, allowing high-quality non-inva-
sive 3D imaging of coronary artery morphology (Figures 3
and 4). Recently the diagnostic accuracy of 64-slice MSCT
for the identification and quantification of coronary artery
stenoses has been reported [12, 13]. The patient-based
analysis revealed that 94% of patients who required
revascularization were correctly diagnosed by CT.
Although excellent accuracy for stenosis detection was
noted, technical restrictions for exact quantification of the
degree of stenosis and reliable visualization of small vessel
segments remain [12]. In an accompanying editorial the
authors express the expectation that MSCT will be used in
the near future on a routine basis for the identification of
patients who do not need revascularization therapy despite
the presence of symptoms [14].
The potential value of MSCT for stenosis quantification
is currently under active investigation. Recently, a good
correlation between MSCT and quantitative coronary
X-ray angiography was shown for stenosis quantification
with the use of 16-slice technology, although MSCT
revealed a systematic overestimation as compared with the
reference standard [15]. Perfusion defects related to
previous myocardial infarction or ischaemia may be well
visualized with the use of MSCT (Figure 5).
In CT coronary angiography, beta-blockers may be used
to reduce the heart rate to a lower range, e.g. to 50–
60 BPM to increase the cardiac rest period and with this to
reduce motion artefacts. The resulting imaging perfor-
mance is more predictable and of more consistent quality
when using such medication. Special reconstruction
algorithms for the reconstruction pose an alternative to
the use of medication. The segmented reconstruction
algorithm yields good temporal resolution even at higher
heart rates. Also, when total scanning time is short, e.g.
below 10 s, the quality of the scan improves since, due to
the reduction of the total amount of heart beats in the
scan, less variation can be expected in the heart rate
during the acquisition. Hyperventilation and administra-
tion of oxygen may be used to stabilize heart rate
particularly at scan times of approximately 20 s scanning
time or longer.
The British Journal of Radiology, January 2006
Cardiac MSCT

Figure 3. Normal coronary artery anatomy at 64-row multidetector CT (MDCT). 64-row MDCT of a 62-year-old male patient with
risk factors for coronary artery stenosis. No stenoses were found at MDCT coronary angiography. Left anterior (a) oblique view
and (b) caudal view. LAD, left anterior descending coronary artery; D, diagonal branch of the LAD; IM, intermediate coronary
artery branch; Cx, circumflex coronary artery; MO, obtuse marginal branch (of the Cx); DP, descending posterior branch (of the
right coronary artery).

Figure 4. Bypass imaging at 64-row multidetector CT (MDCT). 64-row MDCT of a 78-year-old male patient after coronary artery
bypass graft operation (CABG). Occlusion of multiple venous bypass grafts (nr 1 in a). Left internal mammarian artery bypass graft
(nr 2 in a,b) with open anastomosis (nr 3 in a,b,c) on the left anterior descending coronary artery (nr 4 in a,c). Poor quality native
coronary artery system with multiple stenoses and poor contrast enhancement (nr 4 in a,c). b and c are displayed in two perpen-
dicular longitudinal directions.

The British Journal of Radiology, January 2006 13


Figure 5. Multiple perfusion defects imaged with 64-row multi-
detector CT (MDCT). Same patient (78-year-old male) as in
Figure 4 after coronary artery bypass graft operation and mul-
tiple venous bypass graft occlusions. Multiple perfusion defects
with regional wall thinning.
Assessment of ventricular function
With retrospective gated 180 ˚ segmented sinogram space
reconstruction the data can be reconstructed for evalua-
tion of ventricular function [16]. Diastolic and systolic
images can easily be extracted and reconstructed in any
orientation for functional evaluation (Figure 6).
Global ventricular function is generally measured as the
end-systolic and end-diastolic volume (ESV, EDV).
Subsequently, stroke volume (SV) and ejection fraction
(EF) can easily be derived from ESV and EDV.
Semiautomatic software may be used for ventricular
cavity contour detection and for the calculation of
global ventricular function. Regional LV wall motion
Figure 6. Ventricular function imaging at 64-row multidetector CT (MDCT). 26-year-old male patient after surgery for congenital
heart disease. Ventricular function can be assessed after drawing the endocardial ventricular contours in (a) end-diastolic and (b) end-
systolic phases at multiple cardiac levels, thereby including the ventricular volumes.
14
A de Roos, L J M Kroft, J J Bax et al
can be assessed by visual scoring of cinematic loops of well
described myocardial segments [17].
Integrated CT assessment of the coronary arteries and
regional myocardial function allows assessment of the
functional consequences of a coronary artery stenosis
leading to ischaemia and contraction abnormalities. The
usefulness of this combined approach has been reported in
patients with hypertension and diabetes mellitus [18, 19].
From the same data set global function and left ventricular
mass can also be determined, which have clinical relevance
in patients with hypertension for prognosis and guidance
of therapy.
Several studies have shown that right ventricular
function can also be accurately measured by gated
MSCT. The assessment of right ventricular function may
have special interest in patients with acute pulmonary
embolism. Right ventricular enlargement on chest CT has
been shown to be a predictor of early death in patients
with acute pulmonary embolism [20, 21]. Even the
dimensions of the right ventricle in non-gated CT
images may be predictive for mortality in this setting.
The potential value of gated MSCT for assessing right
ventricular function in patients with pulmonary embolism
is now under investigation.
Assessment of pulmonary veins
Atrial arrhythmias often originate in the pulmonary
veins and can be treated with percutaneous radiofrequency
catheter ablation. With this technique, the arrhythmic foci
are electrically disconnected from the left atrium by means
of catheters placed in the left atrium [22]. Pre-procedural
MSCT examination is helpful to depict the anatomy of the
pulmonary veins and left atrium and particularly to
demonstrate additional pulmonary veins (e.g. middle
lobe vein), which is important for planning the interven-
tional procedure. Variations in pulmonary venous anat-
omy are quite common and comprise variation in the
number of veins as well as the occurrence of common ostia
The British Journal of Radiology, January 2006
Cardiac MSCT
Figure 7. Pulmonary vein imaging at 64-row multidetector CT
(MDCT). 64-row MDCT, non-ECG-synchronized imaging. 59-
year-old male patient. Pre-interventional assessment of pulmon-
ary veins for radiofrequency ablation. Posterior view of the
patient’s heart. Common ostium for the left pulmonary veins,
i.e. the pulmonary veins join before entering the left atrium.
Separate ostia for the right pulmonary veins. LS, left superior
pulmonary vein; LI, left inferior pulmonary vein; RS, right
superior pulmonary vein; RI, right inferior pulmonary vein;
LA, left atrium; LPA, left pulmonary artery; RPA, right pul-
monary artery; VC, inferior vena cava.
and early branching [23]. Three-dimensional surface
rendering reconstructions provide a quick overview of
the pulmonary venous anatomy, but cross-sectional
reconstruction in coronal, sagittal and transverse orienta-
tions is necessary for full appreciation of the morphology
of the pulmonary veins (Figure 7) [24]. Post-procedural
MSCT also offers an opportunity for follow-up of the
pulmonary vein after ablation [25].
MSCT pulmonary venography requires a contrast
enhanced helical acquisition. To avoid motion artefacts
a half reconstruction is generally performed, yielding a
reconstruction window of about 165–200 ms. This is
sufficiently short for imaging the rather large pulmonary
veins with diameters well above 10 mm. Reliable images
can be acquired without the use of ECG gating. Breath-
hold acquisitions with a high pitch factor and resulting
rather low patient dose are routinely obtained. The
potential additional value of ECG synchronized MSCT
is under investigation.
Conclusion
MSCT is a highly accurate tool for the non-invasive
detection of coronary artery disease. Further technical
advances are expected in acquisition techniques as well as
post-processing of the CT data. Detector technology and
arrays may be further expanded, allowing shorter imaging
times. Improved temporal and spatial resolution will
contribute to better stenosis quantification and plaque
characterization. Integration of coronary artery imaging
and functional data are feasible with current MSCT.
Shorter scanning times may allow integration of coronary
imaging, first-pass perfusion imaging as well as wall
motion analysis from the same data set. Other cardiovas-
cular applications also benefit from the improvements in
CT technology. Recently, the value of MSCT for the
The British Journal of Radiology, January 2006
evaluation of patients with chest pain presenting to the
emergency department was reported [26]. It was shown
that MSCT is feasible to evaluate chest pain patients
comprehensively. During one comprehensive MSCT pro-
tocol cardiac and non-cardiac causes of chest pain can
accurately be diagnosed. It is expected that MSCT will
become a gatekeeper in patients presenting with chest pain
from various sources.
References
1. Schuijf JD, Bax JJ, Salm LP, Jukema JW, Lamb HJ, van der
Wall EE, et al. Noninvasive coronary imaging and assessment
of left ventricular function using 16-slice computed tomo-
graphy. Am J Cardiol 2005;95:571–4.
2. Paul JF, Wartski M, Caussin C, Sigal-Cinqualbre A, Lancelin
B, Angel C, et al. Late defect on delayed contrast-enhanced
multi-detector row CT scans in the prediction of SPECT
infarct size after reperfused acute myocardial infarction:
initial experience. Radiology 2005;236:485–9.
3. Dodge JT Jr, Brown BG, Bolson EL, Dodge HT. Lumen
diameter of normal human coronary arteries. Influence of
age, sex, anatomic variation, and left ventricular hypertrophy
or dilation. Circulation 1992;86:232–46.
4. Shechter G, Resar JR, McVeigh ER. Rest period duration of
the coronary arteries: implications for magnetic resonance
coronary angiography. Med Phys 2005;32:255–62.
5. Hofman MB, Wickline SA, Lorenz CH. Quantification of
in-plane motion of the coronary arteries during the cardiac
cycle: implications for acquisition window duration for MR
flow quantification. J Magn Reson Imaging 1998;8:568–76.
6. Lu B, Mao SS, Zhuang N, Bakhsheshi H, Yamamoto H,
Takasu J, et al. Coronary artery motion during the cardiac
cycle and optimal ECG triggering for coronary artery
imaging. Invest Radiol 2001;36:250–6.
7. Harell GS, Guthaner DF, Breiman RS, Morehouse CC, Seppi
EJ, Marshall WH Jr, et al. Stop-action cardiac computed
tomography. Radiology 1977;123:515–7.
8. Girshman J, Wolff SD. Techniques for quantifying
coronary artery calcification. Semin Ultrasound CT MR
2003;24:33–8.
9. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR,
Viamonte M Jr, Detrano R. Quantification of coronary
artery calcium using ultrafast computed tomography. J Am
Coll Cardiol 1990;15:827–32.
10. Ulzheimer S, Kalender WA. Assessment of calcium scoring
performance in cardiac computed tomography. Eur Radiol
2003;13:484–97.
11. Thompson GR, Partridge J. Coronary calcification score: the
coronary-risk impact factor. Lancet 2004;363:557–9.
12. Leber AW, Knez A, von Ziegler F, Becker A, Nikolaou K,
Paul S, et al. Quantification of obstructive and nonobstructive
coronary lesions by 64-slice computed tomography: a
comparative study with quantitative coronary angiography
and intravascular ultrasound. J Am Coll Cardiol
2005;46:147–54.
13. Leschka S, Alkadhi H, Plass A, Desbiolles L, Grunenfelder J,
Marincek B, et al. Accuracy of MSCT coronary angiography
with 64-slice technology: first experience. Eur Heart J
2005;26:1482–7.
14. Achenbach S, Daniel WG. Computed tomography of the
coronary arteries: more than meets the (angiographic) eye. J
Am Coll Cardiol 2005;46:155–7.
15. Cury RC, Pomerantsev EV, Ferencik M. Comparison of the
degree of coronary stenoses by multidetector computed
tomography versus by quantitative coronary angiography.
Am J Cardiol (In press).
15

16. Dirksen MS, Bax JJ, de Roos A, Jukema JW, van der Geest
RJ, Geleijns K, et al. Usefulness of dynamic multislice
computed tomography of left ventricular function in unstable
angina pectoris and comparison with echocardiography. Am J
Cardiol 2002;90:1157–60.
17. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S,
Laskey WK, et al. Standardized myocardial segmentation and
nomenclature for tomographic imaging of the heart: a statement
for healthcare professionals from the Cardiac Imaging
Committee of the Council on Clinical Cardiology of the
American Heart Association. Circulation 2002;105:539–42.
18. Schuijf JD, Bax JJ, Jukema JW, Lamb HJ, Vliegen HW, van
der Wall EE, et al. Noninvasive evaluation of the coronary
arteries with multislice computed tomography in hypertensive
patients. Hypertension 2005;45:227–32.
19. Schuijf JD, Bax JJ, Jukema JW, Lamb HJ, Vliegen HW, Salm
LP, et al. Noninvasive angiography and assessment of left
ventricular function using multislice computed tomography in
patients with type 2 diabetes. Diabetes Care 2004;27:2905–10.
20. Schoepf UJ, Kucher N, Kipfmueller F, Quiroz R, Costello P,
Goldhaber SZ. Right ventricular enlargement on chest
computed tomography: a predictor of early death in acute
pulmonary embolism. Circulation 2004;110:3276–80.
21. van der Meer RW, Pattynama PM, van Strijen MJ, van den
Berg-Huijsmans AA, Hartmann IJ, Putter H, et al. Right
ventricular dysfunction and pulmonary obstruction index at
helical CT: prediction of clinical outcome during 3-month
follow-up in patients with acute pulmonary embolism.
Radiology 2005;235:798–803.
16
A de Roos, L J M Kroft, J J Bax et al
22. Pappone C, Rosanio S, Oreto G, Tocchi M, Gugliotta F,
Vicedomini G, et al. Circumferential radiofrequency ablation
of pulmonary vein ostia: A new anatomic approach for curing
atrial fibrillation. Circulation 2000;102:2619–28.
23. Ghaye B, Szapiro D, Dacher JN, Rodriguez LM,
Timmermans C, Devillers D, et al. Percutaneous ablation
for atrial fibrillation: the role of cross-sectional imaging.
Radiographics 2003;23 Spec No:S19–S33.
24. Jongbloed MR, Dirksen MS, Bax JJ, Geleijns K, Lamb HJ,
Van der Wall EE, et al. Multislice computed tomography to
evaluate pulmonary vein anatomy prior to radiofrequency
catheter ablation of atrial fibrillation. Radiology 2005.
25. Maksimovic R, Cademartiri F, Scholten M, Jordaens LJ,
Pattynama PM. Sixteen-row multislice computed tomography
in the assessment of pulmonary veins prior to ablative
treatment: validation vs conventional pulmonary venography
and study of reproducibility. Eur Radiol 2004;14:369–74.
26. White CS, Kuo D, Kelemen M, Jain V, Musk A, Zaidi E, et
al. Chest pain evaluation in the emergency department: can
MDCT provide a comprehensive evaluation? AJR Am J
Roentgenol 2005;185:533–40.
The British Journal of Radiology, January 2006
The British Journal of Radiology, 79 (2006), 17–23 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/23726774

President’s conference paper

Technology solutions for better outcomes: integrated
information management in key to productivity increases in
medicine
H REQUARDT, PhD
Group Executive Management, Siemens Medical Solutions, Henkestrasse 127, 91052 Erlangen, Germany

Abstract. The challenges to healthcare systems around the world are primarily impacted by two topics:
demographic factors and progress in medicine. An ageing population inherently needs more medical services
which add financial burdens, in particular, to public healthcare. Since the level of medical education is growing
at the same time, we are observing an increased demand for sophisticated (in general expensive) medicine.
Drastic changes in financing seem unavoidable. Multiple diagnoses, repeated examinations, trial-and-error,
overcapacities and other signs of missing economical considerations are reinforced by reimbursement systems.
In a world where, in principle, all information is available everywhere, more than a patient’s history should be
accessible. Other industries have knowledge management systems in place that make state-of-the-art expertise
available everywhere. Intelligent patient databases could consist of learning cycles that (i) enable the individual
to benefit from structured knowledge, in addition to personal experience of the physician, and (ii) use the
knowledge generated from the individual to extend the database. The novel area of molecular medicine fits
perfectly well into these scenarios. Only attached to an IT backbone can the flood of information be managed in
a beneficial way. Efficiency improvements in healthcare address the needs of all parties in the system: patients,
providers, and payers. The opportunities, however, can only materialize if everyone is prepared to change. IT
will set the standards for the biggest challenge in healthcare: The paradigm shift in medicine.

Introduction this question translates to ‘‘What levers do we see to

Demographic developments are placing tremendous
pressure on healthcare systems around the world.
Although age distribution varies significantly in different
countries (e.g. China’s one-child policy versus India’s fir-
tree distribution), problems come down to one common
denominator: We are all living longer.
Figure 1 [1, 2] shows the age distribution in more
developed regions and the prognoses for 2025. It is
obvious that health is a major macroeconomic factor. If
we want to avoid the situation that fewer and fewer payers
have to support more and more users of healthcare
services, we will need to see more elderly people working.
The prerequisite for this development is that they stay
healthy. Healthcare systems thus would need to prove that
the investment in them pays off as a productivity factor.
A related challenge is reflected in the fact that a growing
population is increasingly demanding to actively partici-
pate in medical progress. Mass media and the Internet
depict what is possible today; with the majority of research
being funded by the public purse. Thus, it is a natural
desire that the same paying public also wants to enjoy the
benefits that are generated.
The basic question is: How can all of this remain
affordable? Cutting cost by cutting services is not helpful
for addressing both the need for higher quality care and
the necessity to reduce cost. Instead, all contributors to the
delivery of healthcare need to ask themselves ‘‘How can we
do more with less?’’ If we draw an analogy with industry,
Received 16 August 2005 and accepted 16 September 2005.
improve efficiency?’’
Innovations drive efficiency
Medical industries are developing not only more cost
effective and reliable systems, but are also generating more
and more relevant patient information in less examination
time.
Figure 2 shows a standard way of looking at CT
datasets. The approximately 2 GB of raw image data that
are typically acquired in a 5 s scan are stored in cache
memories, are post-processed with volume renderers and
can be displayed according to the interpreter’s comfort
view.
A different example is given in Figure 3: Not only has
the amount of data dramatically increased, but so has the
quality. In this case, a high-resolution three-dimensional
(3D) image of the moving heart displays the stent
structures with superb resolution.
The broadening of the application scope is typical for
each of the imaging modalities: Angio suites do excellent
3D imaging with cone beam reconstruction algorithms,
linear accelerators deliver kV and MV images, magnetic
resonance scanners have left the domain of pure
morphologic imaging, and now measure functions in
various ways. As an example, Figure 4 shows colour
coded diffusion spectral imaging that is highly correlated
with the directions of nerve bundles.
The international medical industry has developed many
technologies that can be utilized to improve efficiency in

The British Journal of Radiology, January 2006 17

 H Requardt

Figure 1. The change of age distribution in more developed regions. The qualitative cost curve reflects the current status. If nothing
changes, the real overall cost can be the integral over the age distribution multiplied by the cost curve.

Figure 3. High resolution CT image of the heart. The stent

structure is clearly delineated.

potential is not only a matter of technology; reimburse-

ment systems and workflow structures have to be adjusted
accordingly.

Figure 2. Volume-rendered abdominal CT image. The underly- Is more always better?

ing image dataset consists of approximately 800 images.
diagnostic and therapeutic processes. Figure 5 shows in a
schematic diagram how these developments can be locked
into the learning cycles of healthcare providers. The
potential for cost savings without sacrificing quality of
care is clear. It is, however, evident that leveraging this
The basic question ‘‘Do I get enough information about
my patient?’’ is no longer appropriate from a technology
perspective. The medical industry has established time-to-
market cycles that can rapidly turn a novel clinical
parameter into a product standard. Only 6 weeks elapsed
between the identification of the SARS virus and the
availability of a clinical test. The problem is no longer the

18 The British Journal of Radiology, January 2006

Technology solutions for better outcomes
Figure 4. Diffusion spectral MR image. Colours code for
spatial directions.
lack of data; the problem lies in filtering out the relevant
information.
There are various technological solutions for filtering. A
widely practiced method uses overlay of images with different
measurement parameters. Figure 6 shows an example in
which a positron emission tomography (PET) image shows us
where to focus in a set of hundreds of CT images. The overlaid
images help us to select the slices of interest.
A totally different approach with similar outcome
is represented by ‘‘computer-aided detection’’ (CAD)
Figure 5. Contiguous improvement cycles stimulated by technology (examples).
The British Journal of Radiology, January 2006
Figure 6. PET images overlaid to a volume-rendered CT data-
set. The primary breast cancer is clearly delineated. Metastasis
search is done within the same dataset.
algorithms. Figure 7 shows CAD-detected polyps within
a virtual colonoscopy dataset acquired with CT. These
algorithms have now reached a performance level that is
comparable with human readers. It is, however, still
applicable only for simple structures, but can help us to
focus our attention on the more complex features.
Progression of CAD into more complex structures will
be subject to the availability of standardized reference
cases.
It can be implied that innovation pressure for medical
devices will in future not only focus on the generation of
more data, but more and more on the generation of
19

Figure 7. Computer-aided detection (CAD) algorithms detect polyps in a virtual colonoscopy. The sensitivity for polyps ¢ 6 mm is
on average 90%; and the median false positive rate is a manageable 3 per volume [3].
‘‘smarter data’’. Yes, there will be CT scans that do 256
slices. But at the same time there will be an industry focus
on systems with two or three X-ray detector systems that
can generate not only increased temporal resolution, but
also open up new degrees of freedom with respect to
contrast by applying different anode voltages in the sub-
systems. Figure 8 shows a basic set-up for such a
system.
Figure 8. Multitube CT set-up. The system enables a new
degree of freedom allowing for double temporal resolution
and/or novel contrast opportunities.
20
H Requardt
Overall, the focus of industry will move from ‘‘genera-
tion of data’’ towards ‘‘exploitation of data’’. It is evident
that information technology is a key enabler for that shift.
IT enables process optimization
In a patient-centric system, the ultimate outcome of the
treatment is reflected by the status of the patient. The
typical patient process in a hospital usually starts with
diagnostic steps (radiology, ECG, lab, …), iterates with
various therapeutic procedures (medication, surgery,
radiation, …), and terminates with the recovery of the
patient (ICU, ward, rehab, …). The most competitive
healthcare provider will be the one that optimizes the
entire procedure chain rather than the individual steps
(this does not relieve the individual departments from
delivering the best quality; ‘‘best’’ according to cost
optimization criteria means ‘‘adequate and intelligent’’).
In industrial analogy this means analysis, mapping and
continuous improvement of workflow.
Workflow optimization comprises the moving of
patients, resources and information within the healthcare
continuum according to certain rules. Everything (includ-
ing the rules) is subject to best practice shared across all
relevant healthcare participants throughout the world.
Workflow can be referenced in ‘‘hospital information
systems’’ by so-called workflow engines. An example of
what a workflow engine can contribute is given in
Figure 9: The emergency treatment of an acute stroke
patient is managed by a computer network. The state-of-
the-art workflow engine would not only draft a work list,
it would also monitor all activities in feedback loops.
Cross-checks with rules engines ensure that the patient
experiences state-of-the-art stroke treatment procedures.
Figure 9 gives an impression how a workflow engine can
be programmed according to the local conditions. It is
obvious that workflow engines not only synchronize
The British Journal of Radiology, January 2006
Technology solutions for better outcomes

Figure 9. Workflow engine editor. The various decision steps reflect the time-critical diagnosis and treatment of an acute stroke. The
time window for initiating thrombolysis is computer controlled.

Figure 10. Steps for cancer development. Today’s procedures detect cancer at a very late stage associated with high treatment cost
and reduced prognosis. Early detection schemes lead to cellular and molecular levels; one of the exciting novel areas of ‘‘molecular
medicine’’.

The British Journal of Radiology, January 2006 21


Figure 11. Process chart of future treatment planning. Data access for both the patient’s individual electronic patient record and a
comprehensive knowledge data base are crucial to enable state-of-the-art medical treatment for everyone, everywhere.
clinical activities, but also other day-to-day operations,
e.g. discharge (paper work needs to be ready, transporta-
tion needs to be arranged, room needs to be made up, bed
needs to be cleaned, etc.).
Workflow engines will not only change the way care is
delivered, but will also define the requirements for newly
developed systems. Requirements and job descriptions in
both arenas, industry and healthcare services, will be
affected.
The patient is an individual
The process chain within healthcare environments
(prevention R diagnosis R therapy R care) is obviously
not limited to hospitals. If we look at a schematic
development of cancer in Figure 10, we realise that with
today’s diagnostic methods we detect cancer only at a very
late stage with higher cost and lower quality of life.
Patient-focused healthcare systems will bring the interven-
tion point forward to an earlier stage of the disease. With
early detection and prevention capabilities, healthcare will
increasingly be looked at just like every other service
industry. The patient will behave like any other customer,
but still with one fundamental difference: He/she is not
free in selecting the disease.
To shift the intervention point in an efficient way, much
basic research remains to be done: The complexity of the
‘‘omics’’ (genomics, proteomics, metabolomics) needs to be
understood and standardized with respect to the develop-
ment of individual diseases. The potential, however, is big
and every single day new cancer genes are being discovered
or proteins identified that originate in specific tissue
alterations. The diagnostic industry is asked not only to
22
H Requardt
deliver blood sample tests, but also software modules that
make the associated knowledge available.
The individualization, however, is not only subject to
the diagnosis of the individual patient. It also needs to give
clear recommendations for an optimized treatment. The
entire arena of pharmacogenomics will be closely asso-
ciated with ‘‘omics’’ analysis. Also, specific tumour
metabolisms can be clearly understood and thus indivi-
dually treated. It becomes evident that in scenarios like
these, the diagnostic process moves from primary diag-
nostic to optimized treatment planning and follow up.
The holistic scenario
The topics discussed so far lead to a few characteristics
of future healthcare systems:
(1) they will be patient-focused and workflow-driven;
(2) the patient’s entire history will be accessible through
an electronic patient record (EPR);
(3) the providers will be in a competitive situation and
thus will publish proven outcome statistics to differ-
entiate themselves;
(4) the capability of sharing best practices with best-in-
class providers will be a differentiating factor.
The patient of the future will no longer rely just on the
individual experience of his physician, but on the entire
medical knowledge that is available. Obviously, the
individual experience becomes part of that knowledge,
but there are also other contributors. Figure 11 shows a
scenario of how the individual patient information can be
matched with the available knowledge. The individual
The British Journal of Radiology, January 2006
Technology solutions for better outcomes
treatment plan for the patient is mainly impacted by two
elements: (1) the clinical knowledge database with rules for
utilization of equipment and drugs, contraindications,
standardizations, procedures and others; (2) the EPR
consisting of images, lab data, structured reports, ‘‘omics’’,
etc.
Those databases will be mined by software agents for
reference cases with proven outcome data to derive the
most promising treatment plans. This enables the primary
care physician (PCP) to match his individual experience
with all the information that is available in the data stores.
The databases will not only be filled with expert knowl-
edge from medicine, but will also include related
disciplines like pharmacology, radiation biology, biome-
chanics and others. In short, the PCP has a real, powerful
tool that leaves him with a high degree of confidence that
he has done all he can to help the patient.
It will certainly be a long way to reach this scenario, but
at the same time it is worth defining and working towards
a common vision. Enabling technologies are there to help
make this vision reality. Many new problems will come up
including topics like data protection, ethics, business
models or simply operational realization, and a social
consensus will be required to address them all.
The British Journal of Radiology, January 2006
Medicine will never become deductive, but managing its
complexity will become easier. Although basic work
remains to be done, the technological solutions are
available today. It is now a question of political desire
to launch the paradigm shift in medicine.
References
1. Population Division of the Department of Economic and
Social Affairs of the United Nations Secretariat. World
Population Prospects: The 2004 Revision Population
Database. [Online]. 2005 [cited 2005 March 15]. Available
from: URL: http://esa.un.org/unpp/
2. Economic Policy Committee (EPC). Budgetary challenges
posed by ageing populations: the impact on public spending
on pensions, health and long-term care for the elderly and
possible indicators of the long-term sustainability of public
finances. Brussels. 2001 October 24 (EPC/ECFIN/655/01-EN
final). p. 34.
3. Bogoni L, Cathier P, Dundar M, Jerebko A, Lakare S, Liang
J, et al. Computer-aided detection (CAD) for CT colonogra-
phy: a tool to address a growing need. Br J Radiol 2005;78:57–
62.
23
The British Journal of Radiology, 79 (2006), 24–31 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/81790390

President’s conference paper

The case for particle therapy
B JONES, MD, FRCR, MedFIPEM
Queen Elizabeth University Hospital, Birmingham B15 2TH, UK

Abstract. Among the most important decisions facing the British Government regarding the treatment of cancer
in the National Health Service (NHS) is the purchase of charged particle therapy (CPT) centres. CPT is
different from conventional radiotherapy: the dose is deposited far more selectively in Bragg Peaks by either
protons or ‘‘heavy’’ ions, such as carbon. In this way, it is possible to ‘‘dose paint’’ targets, voxel by voxel, with
far less dose to surrounding tissues than with X-ray techniques. At present the UK possesses a 62 MeV
cyclotron proton facility at Clatterbridge (Wirral), which provides therapy for intraocular cancers such as
melanoma; for deeper situated cancers in the pelvis, chest etc., much higher energies, over 200 MeV are required
from a synchrotron facility. There is an impressive expansion in particle beam therapy (PBT) centres worldwide,
since they offer good prospects of improved quality of life with enhanced cancer cures in situations where
conventional therapy is limited due to radioresistance or by the close proximity of critical normal tissues. There
is a threat to UK Oncology, since it is anticipated that several thousand British patients may require referral
abroad for therapy; this would severely disrupt their multidisciplinary management and require demanding
logistical support.

The benefits of an increase in charged particle therapy
(CPT) centres in the UK would be not only for children
and young adults with cancer, where a reduced risk of
radiation induced malignancy is predicted, but also in
older patients where it is necessary to avoid abnormal
tissues such as an enlarged heart/restricted lung irradiation
and where artificial (metallic) joints may cause difficulties
in the use of conventional radiotherapy techniques. The
results of phase I and II clinical studies are extremely
encouraging. The UK must obtain at least one CPT centre
with protons/ions in order to conduct research and
development; it is suggested that quality adjusted life
years should be used to assess outcomes. It is anticipated
that the UK might eventually require 7–8 such centres in
10–15 years from now. In the meantime, healthcare
purchasers and providers need to put in place mechanisms
and personnel for patient referrals abroad, as well as the
establishment of UK CPT facilities.
NHS became disadvantaged in terms of expensive tech-
nological acquisition.
Dr R D Errington related the history of cyclotron
radiotherapy at the BIR President’s Day conference in
2003. He detailed how the initial promising results
obtained with neutron therapy at The Hammersmith
Hospital were not subsequently confirmed in randomized
trials at Edinburgh and at the Clatterbridge facility [1, 2],
which produced neutrons that matched a 5 MeV X-ray
beam. The latter facility was converted to produce protons
on the recommendation of the late Prof. Arthur Jones of
St Bartholomew’s Hospital. This enabled patients with
choroidal melanoma of the eye to receive radical radio-
therapy using protons; this technique was the first example
of three-dimensional (3D) radiotherapy in the UK. Over
1400 patients have by now received this therapy with a
local control rate of 98% – an outstanding achievement
within British medicine [3].

Background Past attempts to obtain a higher energy facility in

The connection between subatomic particles and health
delivery improvements may seem rather tenuous, but the
narrative begins in 1879, when J J Thompson discovered
the negatively charged electron in Cambridge, and
Aneurin Bevan was born in Wales. The subsequent
discoveries of the positively charged proton (a term
coined by Ernest Rutherford in 1920) and the uncharged
neutron by James Chadwick in 1931, confirmed the pre-
eminence of our science. Bevan, with similar precision of
thought, digested the wide recommendations of the Beveridge
Report (1942) and transformed most of its principles to
practical achievements, including the National Health Service
Act of Parliament (1946) and the inception of the service in
1948. Subsequently, Britain was at the forefront of practical
applications of physics and engineering developments in
cancer therapy until the early 1990s, when the reorganized
the UK
Since 1992, Clatterbridge, Oxford and the National
Physical Laboratory at Daresbury (near Warrington)
have all unsuccessfully attempted to obtain a higher
energy CPT facility [4]. All these bids were rejected
because of perceived lack of clinical support,
intermittent beam availability, the lack of clinical trial
evidence, the recommendation that a facility should be
sited in a University Hospital campus and perhaps
mostly, the expected high initial costs incurred at a
time when NHS reforms discouraged large-scale
projects, even the provision of new (replacement) linear
accelerators.
More recently, there has emerged a more collective
response from clinical oncologists and medical physicists
who appreciate that obtaining a CPT facility is essential

24 The British Journal of Radiology, January 2006

The case for particle therapy
for the advancement of radiation oncology standards in
the UK. The Royal College of Radiologists (RCR), British
Institute of Radiology (BIR) and Institute of Physics and
Engineering in Medicine (IPEM) for example all support
the case for a CPT facility. Recent improvements in the
quality of cancer imaging and the availability of
industrially produced turnkey facilities, has allowed the
question to be carefully re-considered and better under-
stood, particularly in relation to the rapid expansion in
CPT facilities abroad.
Technical aspects
The velocity of heavy charged particles (electrons are
considered to be light) is reduced as they traverse deeper
through tissues. The interaction probability to cause
ionization increases as the velocity falls, so that a peak
of dose occurs at a depth proportional to the energy
imparted to each particle. William Bragg, a British
physicist, described this phenomenon over 100 years ago
[5]. The so called Bragg peak can be ‘‘spread out’’ to
achieve a plateau of uniform dose that covers a target by
use of rotating range-shifting modulators of variable
thickness. In the past, passively scattered beams were
used in this way to provide wide circular or rectangular
beams with spread out Bragg peaks (Figure 1). More
recently, the spot scanning method allows smaller beams
to deposit their peaks within individual voxel targets
defined by good imaging techniques: by the use of
‘‘wobbler’’ magnets and particle energy selection, the
raster scanning system allows cancer bearing voxels
(defined by x, y, z, co-ordinates), to be ‘‘dose painted’’.
The Bragg peak position will depend on the initial
energy imparted to the particles as well as their mass and
charge; the Bethe-Bloch equation contains all the neces-
sary parameters. It can be seen from Figure 2 that the
range for clinical use should be at least 200 MeV in the
case of protons; higher energies – up to 400 MeV – for
carbon ions.
Figure 1. Schematic depth dose diagram of a proton beam
Bragg peak, the spread out Bragg peak and a megavoltage
X-ray beam (modified from Suit et al [12]). The grey shaded
areas indicate the extent of dose reduction within normal tis-
sues situated proximal and distal to the tumour target.
The British Journal of Radiology, January 2006
Figure 2. Approximate depth dose positions of partially spread
out Bragg peaks for protons of different energies.
Gantries and robots
Within treatment rooms there are options for beam
arrangements. The simplest approach is to have either
fixed horizontal or vertical beams, or a combination of the
two for the simplest treatments. An isocentric rotating
gantry is required for more complex geometrical problems.
These consist of large cylindrical rotating structures that
contain the beam bending magnets: they weigh 100 tonnes
for protons and 200 tonnes for ions and require movement
with 1 mm precision of beam placement. Future engineer-
ing innovations may reduce the tonnage and costs.
Robotic treatment couches are desirable in order to
rapidly position the patient at predetermined angles
relative to the beams; they may also transport patients
in fixed positions from image guided or other localization
devices in the treatment rooms to the actual treatment
location. Radiographers may feel sensitive about robotics,
but it will always be the radiographer who commands the
robot and remotely monitors their performance.
Typical centre
The typical layout of a centre is illustrated in Figure 3.
The particles are injected from a small linear accelerator
and further accelerated to higher energies around the
synchrotron, then extracted and delivered selectively to
different rooms; the beam switching time between rooms is
Figure 3. A schematic diagram of a synchrotron treatment
centre.
25

as short as 10–20 s. A high throughput of patients can be
achieved by efficient placement and preparation of patient
position in advance of the beam availability in each room.
Larger synchrotrons can deliver carbon ions or protons.
Some rooms may be equipped with positron emission
tomography (PET) scanning facilities and other image
guided devices. The overall arrangement is quite different
from standard radiotherapy departments where there is a
linear accelerator in each treatment room. For more
detailed plans see various chapters in Supplement 2 of
Radiotherapy & Oncology (volume 73), 2004 [10].
The dose distribution advantages
Many authors have made important contributions by
means of comparative dose distributions using X-rays and
protons, which are summarized elsewhere [6, 7]. The
essential principles may be better realised by inspection of
relatively simple depth dose diagrams as seen in Figure 4.
In Figure 4A, the spread out Bragg peak (SOBP) is seen
from a single beam entering from the left hand side. In
contrast, the X-ray fall off of dose is pseudo-exponential
as shown in Figure 4D. When two opposed fields are used
there is approximately uniform dosage in the case of
X-rays (as in Figure 4E), whereas for particles there is a
preferential dose deposition where the SOBPs coincide, as
in Figure 4B. For three intersecting beams, there is now
some degree of selectivity for X-rays as seen in Figure 4F,
but the ratios of dose in the centre to that near the surface
is considerably better for the particles as shown in
Figure 4C.
Inspection of axial views of three intersecting beams, as
in Figure 5, shows the different dose distributions achie-
vable. These figures can be normalized to give the same
dose in the central region, with resulting lower peripheral
doses for particles. The absence of dose in one direction
beyond the target is striking – this arrangement may be
used to reduce exposure to critical structures such as
Figure 4. (A–C) Simplified schematic diagrams of protons and
(D–F) X-ray percentage depth dose distributions for three
simple field arrangements. In B, C, E, F depth is measured
along the direction of opposing fields. Relatively small changes
in dose are not included in these fields.
26
B Jones
Figure 5. (a,b). Axial views of simplified schematic dose distri-
butions for three field coplanar techniques using X-rays and
protons.
rectum, spinal cord, etc. Rotation of the beams may also
be used to avoid beam traversion through, or scattered
radiation from metal prostheses, which cause dose
uncertainties in treatment planning.
The reduction in the so called integral dose, which is an
assessment of dose to wider volumes within a patient, is
considerable – proton beams generally reduce this by 50%
and frequently by more in some cases [7]. This effect alone
should reduce the risk of second cancer formation [8],
which may be enhanced with the use of some modern
linear accelerator based techniques such as intensity-
modulated radiotherapy (IMRT), where there is a ‘‘dose
bath’’ effect due to increased integral dose. Not only is the
risk of second cancers reduced, but also substantial
reductions occur in dose commitment to organs that are
sensitive to radiation, e.g. kidneys, eyes, lung, heart, and
parts of the nervous system. Low doses to substantial
proportions of these organs can cause functional pro-
blems. For example, consider the treatment plans shown in
Figure 6, where multiple field IMRT is compared with
single field spot scanning protons. Whilst the target
volume is covered equally well with both techniques, the
dose bath effect is readily seen for IMRT, with significant
dose to spinal cord and kidneys; the proton plan
effectively spares these critical organs. Even a tissue
such as bone is highly relevant: bone marrow cell
production is not supported at doses above 30 Gy and
longer term effects include osteoporosis, micro-fractures
and fractures; in practice, low backache is not infrequent
following pelvic radiotherapy, and bone density changes,
revealed by MRI, are seen to exactly correspond to the
beam portals used.
For a wide variety of cancers the advantages of the
improved dose distributions should provide substantial
improvements in the quality of life where normal tissue
doses are reduced and improved cure potential when
tumour dose is increased. These are considered in further
detail in Table 1, although the generic reduction of second
malignancy is not included.
Meticulous studies in Japan, using carbon ions, with
respiratory movement gating compensation, have shown
two extremely important results. They are:
(1) Cure of small peripheral screen detected lung cancers in a
single exposure and without loss of lung function; similar
The British Journal of Radiology, January 2006
The case for particle therapy

(a) (b)

Figure 6. (a, b) Comparative dose distributions for IMRT and protons for a recurrent sarcoma in a young 12-year-old boy (repro-
duced by kind permission of Dr A Lomax, PSI, Switzerland and Prof. P Hoskins, Editor of Clinical Oncology).

Table 1. The advantages of charged particle therapy (CPT) in a range of anatomical situations

Cancer bearing region Advantage of CPT

Breast Avoid irradiation of heart, lung and brachial plexus
Head and neck Reduced dose to spinal cord, salivary glands, eyes, bone and brain
Pelvis (e.g. prostate, bladder, rectum) Reduced irradiation of bone, sparing of organs such as bladder, rectum; large sarcomas are
safely treated without sacral plexus damage
Gynaecological system As in pelvis, but also improved dose to lateral parametrium, better distribution for vulvar
cancers; can be used where brachytherapy not feasible; field extension to para-aortic region
with less toxicity
Limbs Reduced lymphoedema and deformities
Lung Better preservation of lung and heart function
Liver/pancreas Marked reduction in acute effects, can safely dose escalate for radio-resistant cancers, e.g.
hepatoma, cholangiocarcinoma
Paraspinal/para-aortic Sparing of small bowel, spine and kidneys
CNS Reduction of irradiation to sensitive structures such as hypothalamus, pituitary, reduced risk
of stroke
Reduction of collateral irradiation to tissues outside the CNS, e.g. all tissues anterior to spine
and reduced irradiation of appendages e.g. external auditory apparatus and eye, etc.

cure rates can be achieved by surgery, but with inevitable to treat a hepatoma or cholangiocarcinoma. The colour
loss of lung function [9]. wash dose distribution shows how restricted the dose is to
(2) Cure of patients with primary liver cancers treated in target; this spares the patient the acute side effects of
four exposures; again similar rates of cure can be nausea, vomiting and severe malaise which occur with
achieved following surgery but with considerable X-ray traversion of the stomach, duodenum and liver.
morbidity and some mortality [10].
These results suggest that radiotherapy might eventually
replace radical surgery in deeply situated anatomical
locations. The risks and costs of radical surgery are
likely to increase with time in an ageing population. In
addition, these results confirm previous theoretical predic-
tions based on radiobiological modelling that as dose is
better localized to the target and markedly reduced in a
wider range of surrounding tissues, the principles of
fractionation become less important [11]. Thus treatment
can be delivered in far fewer exposures; the economics of
CPT then become more favourable. In addition, the
treatment is more elegant, involves fewer beams and is
potentially less liable to errors made in treatment delivery. Figure 7. Comparisons of dose distributions for a 4 field
Owing to space constraints it is only possible to show a X-ray (photon) plan and a proton plan for treatment of hepa-
limited number of treatment plans. Figure 7 shows the tocellular cancer (courtesy of Dr J Munzenrider, Northwest
advantages of a four field proton plan which could be used Proton Therapy Centre, Boston, USA).

The British Journal of Radiology, January 2006 27


Figure 8. An example of a single field application of protons
to treat a posterior orbital cancer (courtesy of Dr
J Munzenrider, Northwest Proton Therapy Centre, Boston,
USA). The colours denote different dose levels with red being
the full prescribed dose, with fall off to the limits of the beam.
The next example (Figure 8) shows how the brain and
other bony structures in the head and neck can be spared
due to the sheer elegance of a single field proton approach
to treat cancers in the posterior orbit, such as lachrymal
gland cancer or rhabdomyosarcoma. To obtain equivalent
uniformity of dose across the target region, at least 2 or 3
X-ray fields would be required, with resultant exit doses
into the brain.
The existing evidence base
The clinical evidence base consists of phase I/II dose
escalation studies. There are no randomized control trials
that compare CPT with conventional radiotherapies [6],
although there are randomized phase II ‘‘dose searching’’
studies. One example is the randomization between 72 Gy
and 78 Gy cobalt Gray equivalent (CGE) for skull base
chordomas at Massachusetts General. Some international
authorities consider that randomized studies that compare
conventional X-ray therapy with protons are not justified
because of the advantageous dose distributions for the
latter [12]. Whereas this may be true for skull base
tumours and in hepatic cancers, there must be greater
justification elsewhere, e.g. the comparison of IMRT/
implants with protons in prostate cancer. Whether phase
III studies (comparisons with conventional radiotherapy)
will be performed remains to be seen: some authorities
consider that such research would be unethical [12]. It is
inevitable that randomized comparisons of CPT against
radical surgery will have to be done for small screen
detected cancers in deeply situated tissues (see below).
Misconceptions
It is not surprising that misconceptions abound
when referring to CPT. Comparisons are often made
with neutrons due to their production from similar
28
B Jones
sophisticated equipment. It must be remembered that
neutrons are neutral particles and consequently do not
have Bragg peak characteristics: the additional toxicity
seen with neutron therapy was due to the higher relative
biological effect (RBE) and high integral doses.
Precision is another issue: are protons and ions too
precise? Certainly, the dose can be painted onto any safe
volume, so that tumour margins can be fully respected.
There is no reason why, in certain tumours, one cannot do
wide initial volumes, shrinking down to smaller targets
with increasing dose; protons could be used with three
definite dose volume regions, e.g. 55 Gy, 65 Gy and 75 Gy
volumes defined around a target simultaneously.
Many oncologists assume that the advantages are only
seen in tumours such as skull base chordomas. It must be
realised that such tumours were treated because of poor
results with conventional therapy and with limited proton
beam time coupled with relatively low energy beams that
precluded treatment of deeper structures. Greater beam
availability has allowed testing of CPT in a wider variety
of tumours in different locations.
Added value for science research and teaching
A clinical facility could also be used for radionuclide
production: the particles can activate stable elements to
become radioactive, with applications in healthcare and
industry. Overnight production allows income generating
use of short-lived radionuclide on the following day.
Synchrotron radiation, essentially mono-energetic brem-
strahlung emitted when the particles are deviated by
magnets, can be used for X-ray crystallography studies.
Particle micro-beam analysis of solid state and biological
material can also be pursued, e.g. intracellular diagnostic
capacity at nanometre levels, testing of materials for their
resistance to cosmic rays prior to space flights. A detailed
case is presently being written by the Engineering and
Physical Sciences Research Council (EPSRC) Medical
Applications of Ion Beams Network.
Contributions from molecular biology
The vast expansion in knowledge gained by research in
molecular biology applied to oncology will inevitably
result in more reliable early diagnosis of cancer. Screening
of a population by ‘‘PCR (polymerase chain reaction)
amplification’’ techniques and proteomic techniques
should detect aberrant DNA and protein products from
quite small cancers in body fluids. Further gene specific or
target protein imaging using sophisticated forms of PET
scanning may be sufficient to confirm the presence of small
cancers in deeply situated organs. Image guided biopsies
may also be necessary in some cases. These approaches are
probably more practical than the more distant Holy Grail
of cancer cure following the application of such
approaches. This is not to say that such approaches will
not be useful, particularly in modifying cancer growth
patterns and metastatic potential; but when used alone,
molecular approaches may be doomed to failure because
of the capacity of a cancer to produce further mutations
and to bypass metabolic blockade even when multiple
approaches are used. However, the reliable earlier
diagnosis of cancer would create a high demand for
The British Journal of Radiology, January 2006
The case for particle therapy
surgery and radiotherapy, particularly highly focal forms
of radiotherapy that enable a high localized dose to be
delivered with good sparing of normal tissues, as in CPT.
The decisive clinical trials of the future may be those that
compare CPT with surgery, particularly in sites where the
latter has a high morbidity, mortality and cost, e.g.
hepatic, pancreatic and renal surgery.
Contributions from medical oncology
The reduction of exit dose radiation to skeletal regions
that contain active bone marrow will reduce the risk of
severe neutropenia and the morbidity and mortality that
follow septicaemia. Thus CPT radiotherapy may be
combined with more aggressive chemotherapy regimens.
In addition, the risk of subsequent organ failure on
exposure to certain classes of radiotherapy may be
reduced. For example, the cochlear sparing associated
with medulloblastoma proton-therapy is likely to reduce
the high tone deafness associated with the use of Cis-
platinum treatment [13]; the risk of renal failure may be
reduced when using protons instead of IMRT to treat the
para-aortic nodes in metastatic or advanced local cervix
cancers. Also, the risk of severe cardiomyopathy may be
reduced – even in the case of later exposure to
anthracycline drugs – if the heart has not been exposed
to significant radiation dose by use of CPT, e.g. in the case
of left sided breast cancer. There is clearly a wide
prospectus for research with a major input from medical
oncologists with an interest in radiotherapy in this
important area of oncology.
Contributions from surgery
The increasing future role of radiotherapy in small
volume deep-seated cancers has already been mentioned.
For larger cancers, volume reduction using surgery may
still be desirable, as might the concept of ‘‘improving
treatment geometry’’ by selective resection and restoring a
finite space between tumour and critical normal tissues.
Prolonged surgery will always reduce tissue tolerance
owing to accumulated vascular damage. Decisions regard-
ing operability, the extent of surgery and the necessary
dose of radiation will always need careful consideration
according to circumstances. The possibility of pre-
operative CPT in some situations would be useful: in
Massachusetts General Hospital there is already some
experience of pre-operative proton therapy to paraspinal
bone tumours in order to reduce the potential for
brachytherapy catheter implantation of tumour cells
when radio-iodine seed implants are made into the
adjacent bone situated distally to the tumour. There is
clearly considerable scope for research in the degree to
which surgery and CPT can be combined.
Research and development: quality adjusted survival
end points
There is increasing disquiet that very large trials are
required to detect small incremental changes in outcomes,
with a tendency to favour patient survival as the primary
end point, possibly with inclusion of some separate quality
of life study. This stance is not unreasonable for
The British Journal of Radiology, January 2006
comparisons of chemotherapy schedules, where severe
acute toxicity is life threatening and influences survival.
Such approaches are far from ideal for the assessment of
new radiation techniques where subtle long-term differ-
ences in a wide spectrum of tissues are more relevant.
Newer forms of trial assessment will probably be
necessary. One such approach is considered here. In a
computer generated survival curve with only 100 patients
in each treatment arm, with a survival advantage of ,10%
for CPT c.f. X-rays, the p-value exceeds 0.05 using the log-
rank test (p.0.05). The side effect profiles (graded in four
categories according to ascending severity) show subtle
improvements with CPT, although when tested using a
contingency table the Chi-squared statistic shows a non
significant trend (p.0.05) because of the low numbers in
each category. But when survival is adjusted by using the
toxicity grade factor F defined as (5-x)/5, where x is
the toxicity grade with five categories, the quality adjusted
survival (F times the actual survival) becomes highly
significant (p,0.0001). More work is required to justify
and encourage these approaches, but the potential
advantages in terms of cost and rapidity of obtaining
results with a greater number of trial arms containing
different doses/treatment combinations are readily appar-
ent from the example given. Such a novel approach could
be used within CPT studies.
The threat to British oncology
If the UK will not invest sufficiently rapidly in CPT
facilities, there is a real risk of there being between 5000
and 12 000 patients who will require or demand therapy
abroad in around 10 years from now [14]. These estimates
were arrived at using the logistic equation to simulate
supply and demand with best and worst case scenarios for
overall capacity to accept UK referrals abroad. Treatment
abroad would undoubtedly cause severe disruption of
multidisciplinary cancer care as well as anticipated social
and linguistic problems. In terms of staff retention, there is
a risk that many British physicists, radiographers and
oncologists might be attracted to work abroad. Also, the
UK clinical trial portfolio may not contain state of the art
radiotherapy and consequently our trials may become
irrelevant and ignored elsewhere in the world.
Costs
It has become politically incorrect to mention costs in
medical circles, although cost effectiveness is deemed
respectable and quotable. Such restrictive criteria are, for
example, accepted by The British Medical Journal for its
publications. One cannot escape the fact that the costs for
synchrotron commissioning are large, of the order of
£70–100 million depending upon the specifications for
protons and the more expensive ions and how many large
gantries are required. Some consideration has already been
given to cost benefit and patient demand in Switzerland,
Sweden, France and Austria [15–18]. Cost benefit will be
most accurately measured prospectively within clinical
trials. The costs charged will vary with the number of
exposures: presently around £12 000 for 4 exposures at
Clatterbridge; but with some economies of scale and
improved throughput one can envisage CPT for around
29

£8000–25 000 per patient, depending on the fractionation
used; this is less than the cost of renal dialysis necessary to
keep a patient alive for 1 year and compares favourably
with the cost of prolonged radical surgery.
A single UK centre should recoup its own initial and
running costs within 6 years providing it can treat 2500
patients by its third year of operation. However, the UK
would depend on a multitude of healthcare purchasing
agreements – a most unsatisfactory system for the
provision of complex healthcare. Definitive cancer treat-
ment using radiation should be separated from these
cumbersome procedures, with a clear assurance that all
British patients with a diagnosis of cancer will receive
equal access to more complex therapy where necessary.
Dr Neil Burnet has estimated from Swedish data
(Burnet N, personal communication) that the proportion
of total cancer care costs spent on radiotherapy would
increase from the present 5% to 6% if 15% of all
radiotherapy is given by protons [18]. This is likely to
be cost effective in the long term because of the reduced
side effects and compares well with the present expenditure
on cytotoxic chemotherapy, which accounts for around
12% of total cancer care.
It remains unclear as to how funding can be achieved
without a high level political decision. Even the new
Foundation NHS Trusts cannot borrow the necessary
monies to enable CPT. Our NHS needs better structures
that can arrange finance, whether public or private:
perhaps a return to regional and supraregional systems
for cancer care?
Logistics for a National Centre
The NHS has developed impressive Cancer Networks as
part of its Cancer Plan, and CPT will need to be
imaginatively superimposed on this framework. These
existing networks are essential to ensure equity of access
for CPT. Each local Network should form the basis of
referral to special multidisciplinary team (MDT) meetings
concerned with CPT. When a clinical indication is
identified, then appropriate dose planning assessments
are necessary: this might be achieved by electronic transfer
of data to a national reference centre which itself might be
virtual, i.e. it can be envisaged that all cases of tumour
type X might be independently assessed in City A, and for
tumour type Y in City B as for the physical appropriate-
ness of IMRT or CPT. The referring city could also plan
with the two modalities and confer with the national CPT
centre. Encouragement for physicists and oncologists to
attend a National Centre on a rotational or frequent basis,
e.g. for specific MDT and treatment planning meetings,
should also be encouraged. A national service will need to
have strong links with other centres abroad for the
treatment of rare conditions.
Logistics for referral abroad
The prospect of referring hundreds or thousands of
patients abroad is daunting. The time taken to assess and
counsel, and to send all diagnostic information away is
significant. There is an immediate need for full time staff
devoted to these logistics, with attention to transfer
funding for provision of appropriate care abroad. British
30
B Jones
staff should be put in place to support patients and
families whilst abroad and also to promote training in how
to deliver CPT. Eventually, the number of treatment
facilities in the UK should become appropriate to meet the
needs of the British people. However, UK healthcare
planners should urgently apply themselves to these
problems and produce appropriate plans that meet the
most likely short and long-term requirements.
Politics/Government/Research Councils and Charities
CPT needs to be fully researched, with major UK
participation. At least one high-energy UK CPT facility
should be established to conduct clinical research and
trials, with equitable patient referral via the Cancer
Networks. The immediate questions for the UK autho-
rities are ‘‘when’’ and ‘‘how many facilities’’ do we need?
These important decisions confront the UK Government
for future cancer care, and must be judged in the context
of the proposed increased investment in the scientific base
of this country [19]. The concept of joined up working
across the various Research Councils (EPSRC, MRC,
Accelerator Science, N-Tech), and linked to the major
cancer charities (Cancer Research UK) should allow the
UK to further develop the technology that underpins the
most sophisticated form of radiation therapy against
cancer. It would be tragic to wait until public awareness
forces the issue. Bevan, an astute politician and cancer
sufferer, would surely have sensed that the NHS should
possess the weapon of particle radiotherapy within its
arsenal against cancer, in the same way as he bravely
supported an independent nuclear deterrent. He wanted
only the best for the British people and so should we.
Acknowledgments
The author is indebted to the following for discussions
and their encouragement. Oncologists & Physicians: Prof.
Pat Price, Dr Neil Burnet, Dr Trevor Roberts, Dr R D
Errington, Dr P R Blake, Dr F Saran, Dr D Dearnaley,
Dr D A L Morgan, Dr R Taylor, Dr A Cassoni, Dr M
Gaze, Dr K I Hopkins, Dr D J Cole, Dr P N Plowman, Dr
R Beaney, Dr R Rampling, Dr D Spooner, Dr A Crellin,
Dr N G Glaser, Dr R H Phillips, Dr D V Ash, Prof. A
Price, Prof. W Duncan, Prof. A Munro, Dr J Staffurth,
Prof. C Coombes, Dr R K Coker, Prof. W Littler. RCR:
Dr A Barrett, Dr R Hunter, Dr M V Williams, Prof. P
Hoskins, Dr F Calman. Surgeons: Mr I McIndoe, Mr R E
Kingston, Prof. G Cruickshank. Medical Physicists: Dr A
Kacparek, Dr Roger Dale, Dr Ivan Rosenberg, Dr Stuart
Green, Prof. A Beddoe, Dr S Blake, Dr D Thwaites, Dr A
Nahum, Dr A Carabe. Academic Physicists: Dr K Kirkby,
Dr D Parker, Prof. J Nelson. National Physics
Laboratory: Dr H Palmens, Dr D Rayner.
Declarations: BJ is a Trustee of The Cyclotron Trust
(UK Charity) and a member of the EPSRC Medical
Applications of Ion Beams Network.
References
1. Errington RD, Ashby D, Gore SM, et al. High energy
neutron treatment for pelvic cancers: study stopped because
of increased mortality. Br Med J 1991;302:1045–51.
The British Journal of Radiology, January 2006
The case for particle therapy
2. Maor MH, Errington RD, Caplan RJ, et al. Fast neutron
therapy in advanced head & neck cancer: a collaborative
internal randomised trial. Int J Radiat Oncol Biol Phys
1995;32:99–604.
3. Damato B, Lecuona K. Conservation of eyes with choroidal
melanoma by a multimodality approach to treatment: an
audit of 1632 patients. Opthalmology 2004;111:977–83.
4. Price P, Errington RD, Jones B. The clinical and scientific
case for a high energy proton therapy facility in the UK. Clin
Oncol 2003;15:S1–S9.
5. Brown A, Suit H. The centenary of the discovery of the Bragg
Peak. Radiother Oncol 2004;73:265–8.
6. Jones B, Rosenberg I. Particle therapy Cooperative Oncology
Group (PTCOG40), Institut Curie 2004. Br J Radiol
2005;78:99–102.
7. Jones B, Burnet NG. Radiotherapy for the future. Br Med J
2005;330:979–80.
8. Miralbell R, Lomax A, Cella L, Schneider U. Potential
reduction of the incidence of radiation-induced second
cancers by using proton beams in the treatment of
pediatric tumors. Int J Radiat Oncol Biol Phys
2002;54:824–9.
9. Miyamoto T, Yanamoto N, Nishimura H, Koto M, Tsujii H,
et al. Carbon ion radiotherapy for stage 1 non-small cell lung
cancer. Radiother Oncol 2003;66:127–40.
10. Tsuji H. Overview of clinical experiences with carbon ions at
NIRS. Radiother Oncol 2004;73(Suppl. 2):S41–9.
The British Journal of Radiology, January 2006
11. Jones B, Dale RG. Radiobiologically based assessments of
the net costs of fractionated focal radiotherapy. Int J Radiat
Oncol Biol Phys 1998;41:1139–48.
12. Suit H, Goldberg S, Niemerko A, Trofimov A, Adams J, et al.
Proton beams to replace photon beams in radical dose
treatments. Acta Oncologica 2003;42:800–8.
13. Plowman PN, Usher C. Authors’ reply. Br J Radiol
2005;78:285–6.
14. Jones B, Price P, Burnet NG, Roberts JT. Modelling the
expected increase in demand for particle radiotherapy:
implications for the UK. Br J Radiol 2005;78:832–5.
15. Goitein M, Jermann M. The relative costs of proton and
X-ray radiation therapy. Clin Oncol 2003;15:S37–50.
16. Lundkvist J, Ekman M, Ericsson SR, Jonsson B, Glimelius B.
Cost-effectiveness of proton radiation in the treatment of
childhood medulloblastoma. Cancer 2005;103:793–801.
17. Mayer R, Mock U, Jager R, Potter R, et al. Epidemiological
aspects of hadron therapy. Radiother Oncol 2004;73(Suppl.
2):S24–8.
18. Ringborg U, Bergqvist D, Brorsson B, Cavallin-Stahl E,
Ceberg J, Einhorn N, et al. The Swedish Council on
Technology Assessment in health care (SBU) systematic
overview of radiotherapy for cancer including a prospective
survey of radiotherapy practice in Sweden 2001 – summary
and conclusions. Acta Oncologica 2003;42:357–65.
19. Jones B. Science and Innovation Summit 2004. Br J Radiol
2005;78:17–9.
31
The British Journal of Radiology, 79 (2006), 32–36 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/18454286

President’s conference paper

The contribution of PET/CT to improved patient management
P J ELL, FMedSci, FRCP, FRCR
Institute of Nuclear Medicine, UCL, London, UK

Abstract. With the introduction of both SPET/CT and PET/CT, multimodality imaging has truly entered
routine clinical practice. Multiple slice spiral CT scanners have been incorporated with multiple detector gamma
cameras or PET systems, such that the benefit of these modalities can be achieved in one patient sitting. The
subject of this manuscript is PET/CT and its impact on patient management. Applications of PET/CT span the
whole field of medical and surgical oncology since very few cancers do not take up the labelled glucose tracer,
18
F-FDG. Given the contrast achieved, high-quality data can be obtained with FDG PET/CT. This technology
has now spread worldwide and has been the subject of intense interest, as witnessed by the vast body of
published evidence. In this short overview, only a brief discussion of the main clinical applications is possible.
Novel applications of PET/CT outside the field of oncology are expected in the near future.

Introduction in tracer use. There are realistic expectations that a

The technologies of positron emission tomography
(PET) and spiral computed tomography (CT) have been
combined in a single multimodality detection instrument.
The PET/CT scanner provides, in a single patient sitting,
both the data to be expected from a high-end advanced
spiral CT scanner and information recorded by a top of
the range PET scanner, capable of depicting the distribu-
tion of positron-labelled tracers such as fluorodeoxyglu-
cose (FDG). Routine image fusion is obtained, CT data
being merged with PET data to aid in the exact
localization of the site of FDG uptake. CT information
is also used for the purpose of attenuation correction,
which is now almost instantaneous; as a consequence,
whole-body PET/CT studies can be obtained in less than
30 min. This has led to an increase in patient acceptance
and throughput (30% over that achieved with PET alone).
Scanning times are expected to improve further in the near
future. With PET/CT studies obtained from a flat bed, this
information can be used to improve radiotherapy plan-
ning, a novel and rapidly evolving application of this
technology. PET/CT leads to improved lesion detection
and localization and a faster learning curve for all
involved; it has achieved significant acceptance at multi-
disciplinary case conferences [1, 2].
Applications
Tables 1 and 2 summarize the present and predicted
areas of application of PET/CT, and anticipated changes
Table 1. Present and predicted areas of applications of
PET/CT
Present Future
Oncology 97% Oncology 70%
Infection 2% Infection + Musculoskeletal 5%
Cardiology 1% Cardiology 15%
Psychiatry 10%
Received 31 May 2005 and accepted 6 September 2005.
number of novel tracers, labelled with, for example, 18F or
even 68Ga (to mention just two radionuclides), will lead to
useful clinical studies on atherosclerosis [3], angiogenesis,
hypoxia and detection of amyloid plaque in Alzheimer’s
disease. Other tracers such as 18F-labelled thymidine
(FLT: a marker of TK1 activity and indirectly of cellular
proliferation) and 18F-labelled dopamine have already
been applied in the fields of oncology (FLT and dopamine)
[34, 35] and movement disorders (dopamine). The discus-
sion below will, however, be restricted to the use of FDG
in oncology.
Labelled FDG provides some of the highest signal-to-
noise ratios to have been observed in nuclear medicine.
This is the result of a number of factors which play a
role in the cellular uptake of FDG: over-expression of
membrane GLUT transporters, increased glucose trans-
port in malignancy, increased glycolysis, and increased
hexokinase activity coupled with a decrease in glucose-6-
phosphatase activity. It is also now well known that
maximal FDG uptake in the lesion is not reached within
the first hour of intravenous administration. Invariably, a
further increase in the signal-to-noise ratio can be
observed at 2 h, and a plateau is reached much later. It
must also be stressed that FDG is not a cancer-specific
ligand: macrophages actively take up FDG [4, 5], and
granulomas and inflammatory lesions can be falsely
interpreted as malignant.
From a practical point of view, the unit most often used
to quantitate FDG uptake is the standardized uptake
value (SUV). This normalizes the FDG taken up in a
region of interest to the total amount of tracer injected and
the patient’s body weight. The SUV is time dependent,
since FDG continues to accumulate during the period of
imaging. For each study, SUVs have to be measured at the
Table 2. Anticipated changes in tracer use
Present Future
FDG 95% FDG 85%
68
Ga + Others : 15%

32 The British Journal of Radiology, January 2006

PET/CT to improve patient management
same time after administration of FDG. From a region of
interest, average or maximum SUVs can be obtained, the
maximum SUV being the most reproducible value for
comparative purposes. SUVs greater than three are most
often associated with malignancy. Whilst this cut-off is
somewhat arbitrary, it is of value since it helps to
distinguish malignant from benign nodal disease: enlarged
nodes on CT with low SUVs are almost always benign.
PET/CT and FDG in oncology
As one might expect, the main areas of interest are in
diagnosis, staging, treatment monitoring and radiotherapy
planning [6].
Diagnosis
PET/CT is infrequently used to offer or aid in the
diagnosis of a patient’s primary condition, the principal
indications for this purpose being suspected paraneoplastic
syndrome, pyrexia of unknown origin and unresolved
suspicion of a CNS tumour (the more frequent application
is for differential diagnosis of post-treatment radiation
necrosis versus recurrence, rather than diagnosis at
presentation). Impressive data have been obtained in the
diagnosis of paraneoplastic syndromes and a variety of
vasculitides and arteritides [7]. Occasionally PET has
helped in the evaluation of patients with malignant
paragangliomas and carcinoid tumours [8], and PET/CT
holds promise for this indication.
Staging and re-staging
It is in the setting of cancer staging that PET/CT comes
into its own. Combined PET/CT has been shown to be
superior to other imaging modalities in most tumour types.
A gain of 20% was documented when the TNM tumour
classification was used as the comparator and PET/CT was
compared with whole-body MRI [9].
In lymphoma, PET/CT is better than CT in the
diagnosis of both nodal and extranodal disease, and can
detect disease in normal-sized lymph nodes that will be
overlooked by CT. As a consequence, PET/CT upstages
approximately 40% of all cases of lymphoma. PET/CT is
also better than CT for the purpose of post-therapy
evaluation owing to its greater predictive value: a positive
post-treatment PET study is associated with poorer
prognosis, whilst a scan performed after the first cycle
of treatment is often predictive of response, especially in
cases of aggressive Hodgkin’s disease and non-Hodgkin’s
lymphoma [10–12]. FDG PET is useful to guide adoptive
immunotherapy with donor lymphocyte infusions post
transplant [13].
With regard to non-small cell lung cancer (NSCLC),
three major studies have shown that PET/CT prevents
unnecessary surgery in one out of five patients deemed
operable by other criteria [14–16]. This is because PET/CT
upstages a large proportion of patients by demonstrating
both soft tissue and skeletal involvement. A further study
found that PET/CT resulted in a change in management in
30% of patients with NSCLC [17]. Recently, Goren et al
[18] discussed the relative roles of CT, PET and
The British Journal of Radiology, January 2006
endoscopic-guided ultrasound with needle aspiration in
the management of patients with lung cancer.
There is a clear clinical role for PET/CT in colorectal
cancer. It is of value for staging of recurrent disease,
detection of liver involvement, detection of local recur-
rence, differential diagnosis of recurrent disease from scar
and assessment of patients who present with rising tumour
markers [19–21]. A meta-analysis carried out over a 5-year
period showed that FDG PET changed the management in
approximately 35% of patients in the setting of colorectal
cancer. Often PET/CT demonstrates multiple liver deposits
not seen on other imaging modalities [22]. A case could
now be made that PET/CT should be the first imaging
modality to be employed in the staging and re-staging of
colorectal cancer.
PET/CT is also applied to the staging and re-staging of
patients with cancers in the head and neck, breast,
oesophagus, pancreas, cervix and testicle, as well as
patients with sarcomas and melanomas.
In the head and neck, PET/CT misses micrometastatic
disease (as do all imaging modalities) but it is useful in the
context of upstaging N0 disease [23, 24]. In patients who
present with cervical adenopathy and negative cross-
sectional imaging (CT/MRI), PET/CT is a useful investi-
gation [25]. Patients with advanced disease tend to be
upstaged with PET/CT. PET/CT is useful in disease
monitoring after therapy (surgery, chemotherapy or
radiation), but the optimal timing of this application
remains controversial. The possibility of a false positive
inflammatory response must be borne in mind. In thyroid
cancer, PET/CT should be restricted to the re-staging of
patients with raised serum markers (thyroxine-binding
globulin, calcitonin, carcinoembryonic antigen) who
present with negative cross-sectional imaging and
negative 131I scans [26, 27].
In breast cancer, PET/CT is not used to stage the axilla
owing to its failure to detect micrometastatic disease. PET/CT
is, however, useful in re-staging, in the detection of nodal
disease and in the visualization of distant disease in
unsuspected sites. PET/CT scanning uncovers deposits in
the skeleton and can be helpful in the evaluation of internal
mammary and mediastinal node involvement. It also appears
useful in the evaluation of response to treatment, absence of
response on PET/CT carrying a worse prognosis. Scarring
and fibrotic masses can be distinguished from active disease
on the basis of FDG uptake.
In the curative setting, PET/CT is used for the
investigation of the nodal spread of oesophageal cancer.
Here, PET/CT is better than CT alone. A growing body of
evidence shows the utility of PET/CT in the evaluation of
response to therapy. A study by Weber et al [28]
investigated 40 patients. A PET study was performed at
baseline and 2 weeks after initiation of chemotherapy. The
first scan had a sensitivity of 93% and a specificity of 95%.
Patients who responded to therapy had a reduction in
FDG uptake by 54%, whilst in non-responders the
reduction in FDG was of the order of 15% or less. In a
similar study by Brucher et al [29], 27 patients with
oesophageal carcinoma were given chemotherapy and
radiotherapy. Patients responding to the treatment had a
reduction in FDG uptake of 72%, whereas those who did
not respond had a reduction of only 22%. Most studies of
this type now point to the utility of FDG PET in the
assessment of early response to treatment.
33

In melanoma patients, PET/CT is not useful for initial
staging or in early disease, but it is of value for re-staging
of more advanced disease. Melanoma metastases are
intensely FDG avid. PET/CT is also used in the re-staging
of patients with carcinoma of the cervix. Recurrent disease
can be distinguished from non-viable necrotic or fibrotic
post-therapy tissue. PET/CT has been used in a variety of
other cancer types, such as GIST tumours, mesotheliomas,
multiple myelomas and sarcomas. Pancreatic cancer,
neuroendocrine tumours and germ cell tumours and
their deposits can all exhibit intense FDG uptake. In
contrast, prostate cancer and deposits from this tumour
often exhibit poor FDG avidity; hence PET/CT with FDG
is not useful in this context.
Treatment monitoring
In part, this application has already been alluded to. It
is evident that a metabolic response can precede a change
in tumour size, and a reduction in FDG uptake can be
seen within a matter of hours in patients with lymphoma
or germ cell cancer in whom treatment is effective [30, 31].
Eventually PET/CT will be used to assess the biology of
the individual tumour and its response to treatment [32],
with novel markers aimed at imaging proliferation [33–36],
hypoxia, angiogenesis, apoptosis, etc.
PET/CT is useful to assess the efficacy of novel
therapies. This has been demonstrated with Gleevec in
the treatment of germ cell cancers, but PET/CT will have
wide applicability in a number of new settings. It will be
used as a surrogate marker for drug response, and this
might imply yet another revision of the established but still
insufficiently used RECIST criteria for tumour response to
therapy.
Eary et al [37], studied the effect of tumour hetero-
geneity, reflected in heterogeneity in FDG uptake, in
patients with sarcomas. A 30% increase in risk of death
was observed for every increase of 1 standard deviation
(SD) in tumour heterogeneity, and there was a 12%
increase in risk of death for every increase of 1 SD in the
maximum SUV. However, the concept of a metabolic
response as assessed by FDG will need to be validated in
larger studies. In breast carcinoma patients treated with
Tamoxifen, a flare response, albeit transient, has been
described [38]. When such a response occurs it tends to do
so 8–10 days after the commencement of Tamoxifen, and
is usually an indicator of subsequent patient response to
the treatment [38]. It is also recognized that patients
studied soon after radiotherapy may exhibit an increase in
FDG activity owing to an inflammatory response [39].
MacManus et al [40] have nevertheless shown the utility of
evaluation of the metabolic response by PET in patients
with NSCLC.
Radiotherapy planning
A PET/CT scanner can be used to inform radiotherapy
planning. The CT component of the instrument is identical
to a conventional spiral CT and modern PET/CT scanners
are available with 4-, 8- or 16-slice spiral CT scans. The
CT component can be used for attenuation purposes only,
in order to aid in the localization of the abnormality seen
on the PET scanner, or it can be used at high power to
34
P J Ell
record data identical to those that would be obtained using
a conventional CT. In patients with cancer, radiation
exposure should often be considered of secondary
importance, given their age, survival rates and therapeutic
aspects. It can therefore be argued that PET/CT should
become the first imaging study in a significant proportion
of patients with cancer. From the above it can be seen that
the CT information obtained from the PET/CT instrument
can also be used for the purpose of volume planning and
that the available PET information can be similarly used
to better delineate tumour margins, whilst also distinguish-
ing viable from non-viable tumour and aggressive from
less aggressive disease. Ultimately, a more rational
approach to radiotherapy planning is an achievable goal.
Data are beginning to accrue that confirm this approach
and its utility [41].
Imaging the skeleton
With PET, it is possible to obtain data from skeletal
metastases via two tracers: 18F-labelled fluoride ion, which
is directly taken up by the skeleton, and 18F-FDG. More
data need to be obtained before final recommendations
can be made regarding the use of these two tracers for
skeletal imaging. It is already apparent, however, that in
many cancers, FDG can demonstrate both soft tissue and
skeletal involvement; indeed, it has been advocated that
conventional bone scanning is no longer required when
staging NSCLC patients with FDG. In multiple myeloma,
FDG is superior to conventional bone scanning in the
detection of bony deposits. If scanner availability and
tracer costs were not limiting factors, 18F-fluoride scanning
of the skeleton would come to replace the conventional
bone scan owing to the merits of PET/CT co-registration
in the context of both malignant and benign bone disease
[42–44].
Future developments
Multimodality imaging is here to stay and image fusion
will become routine. The first truly routine implementation
of image fusion involving a large number of patients has
been achieved with PET/CT. The design properties of
PET/MRI are under consideration, and progress has
already been made in this field with small animal scanners.
The next generation of PET/CT technology is likely to
make use of new radiation detectors and electronics.
Discussions are now focusing, for example, on the
reduction of whole-body imaging times to less than
15 min and the introduction of routine respiratory and
cardiac gating for improvement of lesion localization and
margin definition. Multiple slice spiral CT scans will open
the way for cardiac imaging, and interesting developments
are expected in this field, which, as with nuclear medicine
in general, is heavily dependent on the emergence of new,
clinically useful ligands. There is realistic hope that these
new ligands will lead to novel practical applications in
neurology, cardiology and oncology. As individually
tailored medicines begin to impact on healthcare, these
technologies will find special relevance in determining
patient response to these therapies. An early indicator of
lack of response may be not only beneficial but also
immensely important in economic terms. The future for
The British Journal of Radiology, January 2006
PET/CT to improve patient management
PET/CT imaging as a surrogate endpoint for novel
therapeutic interventions is bright. This will imply a
rethink of traditional criteria for lesion response –
conventional RECIST criteria will need to be re-assessed
in the light of the metabolic parameter made available by
PET [45].
References
1. Ell PJ and von Schulthess GK. PET/CT: A new road map.
Eur J Nucl Med 2002;29:719–20.
2. Schoder H, Erdi YE, Larson SM, Yeund, HWD. PET/CT: a
new imaging technology in nuclear medicine. Eur J Nucl Med
Mol Imaging 2003;30:1419–37.
3. Davies RJ, Rudd JH, Weissberg PL. Molecular and
metabolic imaging of atherosclerosis. J Nucl Med
2004;45:1898–907.
4. Bomanji JB, Syed R, Brock C, Jankowska P, Dogan A, Costa
DC, et al. Challenging cases and diagnostic dilemmas: case 2.
Pitfalls of positron emission tomography for assessing
residual mediastinal mass after chemotherapy for Hodgkin’s
disease. J Clin Oncol 2002;20:3347–9.
5. Kjaer A, Lebech AM, Eigtved A, et al. Fever of unknown
origin: prospective comparison of diagnostic value of 18F-
FDG PET and 111ln-granulocyte scintigraphy. Eur J Nucl
Med Mol Imaging 2004;31:622–6.
6. Bomanji JB, Costa DC, Ell PJ. The clinical role of positron
emission tomography. The Lancet Oncology 2001;3:157–64.
7. Rees J, Hain, Johnson N, Hughes, Costa DC, Ell PJ. The role
of fluorodeoxyglucose positron emission tomography scan-
ning in the diagnosis of paraneoplastic neurological disorders.
Brain 2001;124:2223–31.
8. Le Rest C, Bomanji JB, Costa DC, Townsend CE, Visvikis D,
Ell PJ. Functional imaging of malignant paragangliomas and
carcinoid tumours. Eur J Nucl Med 2001;38:478–82.
9. Antoch G, Vogt FM, Freudenberg LS, Nazaradeh F, Goehde
SC, Barkhausen J, et al. Whole-body dual-modality PET/CT
and whole-body MRI for tumor staging in oncology. JAMA
2003;290:3199–206.
10. Burton C, Ell PJ, Linch D. The role of PET imaging in
lymphoma. Br J Haematol 2004;126:772–84.
11. Hillner BE, Tunuguntla R, Fratkin M. Clinical decisions
associated with positron emission tomography in a prospec-
tive cohort of patients with suspected or known cancer at one
United States center. J Clin Oncol 2004;22:4147–56.
12. Hutchings M, Eigtved AI, Specht L. FDG-PET in the clinical
management of Hodgkin’s lymphoma. Crit Rev Oncol/
Hematol 2004;52:19–32.
13. Hart DP, Avivi I, Thomson KJ, Peggs KS, Morris EC,
Goldstone AH, et al. Use of 18F-FDG positron emission
tomography following allogeneic transplantation to guide
adoptive immunotherapy with donor lymphocyte infusions.
Br J Haematol 2005;128:824–9.
14. Lardinois D, Weder W, Hany TF, Kamel EM, Korom S,
Seifert B, et al. Staging of non-small-cell lung cancer with
integrated positron-emission tomography and computed
tomography. N Engl J Med 2003;348:2500–7.
15. van Tinteren H, Hoekstra OS, Smit EF, van den Bergh JH,
Schreurs AJ, Stallaert RA, et al. Effectiveness of positron
emission tomography in the preoperative assessment of
patients with suspected non-small-cell lung cancer: the
PLUS multicentre randomised trial. Lancet 2002;359:1388–
93.
16. Verboom P, van Tinteren H, Hoekstra OS, Smit EF, van den
Bergh HAM, Schreurs AJM, et al, and the PLUS study
group. Cost-effectiveness of FDG-PET in staging non-small
cell lung cancer: the PLUS study. Eur J Nucl Med Mol
Imaging 2003;30:1444–9.
The British Journal of Radiology, January 2006
17. Keidar Z, Haim N, Guralnik L, et al. PET/CT using 18-FDG
in suspected lung cancer recurrence: diagnostic value and
impact on patient management. J Nucl Med 2004;45:1640–6.
18. Goren HJM, Slijfer DTh, deVries EGE. Positron emission
tomography, computerized tomography, and endoscopic
ultrasound with needle aspiration of lung cancer. ASCO
Educational Book 2005:579–86.
19. Arulampalam THA, Costa DC, Loizidou M, Visvikis D, Ell
PJ, Taylor I. Positron emission tomography in colorectal
cancer. Br J Surg 2001;88:176–89.
20. Arulampalam THA, Francis DL, Visvikis D, Taylor I, Ell PJ.
FDG-PET for the pre-operative staging of colorectal liver
metastases. Eur J Surg Oncol 2004;30:286–91.
21. Huebner RH, Park KC, Shepherd JE, Schwimmer J, Czernin
J, Phelps ME, et al. A meta-analysis of the literature for
whole-body FDG PET detection of recurrent colorectal
cancer. J Nucl Med 2000;41:1177–89.
22. Dietlein M, Weber W, Schwaiger M, Schicha H. [18F-
Fluorodeoxyglucose positron emission tomography in resta-
ging of colorectal cancer. Nuklearmedizin 2003;42:145–56.
23. Hyde NC, Prvulovich E, Newman L, Waddington WA,
Visvikis D, Ell PJ. A new approach to pre-treatment
assessment of the NO neck in oral squamous cell carcinoma:
the role of sentinel node biopsy and positron emission
tomography. Oral Oncol 2003;29:350–60.
24. Myers LL, Wax MK, Nabi H, Simpson GT, Lamonica D.
Positron emission tomography in the evaluation of the N0
neck. Laryngoscope 1998;108:232–6.
25. Hughes SJ, Prvulovich EM, Witherow H, Kalavrezos N, Ell
PJ. A comparison of FDG PET/CT and MRI versus
histology for staging of primary head and neck cancers
and detection of recurrent disease. J Nucl Med
2004;45:80.
26. Chung JK, So Y, Lee JS, Choi CW, Lim SM, Lee DS, et al.
Value of FDG PET in papillary thyroid carcinoma
with negative 131I whole-body scan. J Nucl Med
1999;40:986–92.
27. Schluter B, Bohuslavizki KH, Beyer W, Plotkin M, Buchert
R, Clausen M. Impact of FDG PET on patients with
differentiated thyroid cancer who present with elevated
thyroglobulin and negative 131I scan. J Nucl Med
2001;42:71–6.
28. Weber WA, Ott K, Becker K, Dittler HJ, Helmberger H,
Avril NE, et al. Prediction of response to preoperative
chemotherapy in adenocarcinomas of the esophago-
gastric junction by metabolic imaging. J Clin Oncol
2001;19:3058–65.
29. Brucher BL, Weber W, Bauer M, Fink U, Avril N, Stein HJ,
et al. Neoadjuvant therapy of esophageal squamous cell
carcinoma: response evaluation by positron emission tomo-
graphy. Ann Surg 2001;233:300–9.
30. Hoekstra OS, Ossenkoppele GJ, Golding R, van Lingen A,
Visser GW, Teule GJ, et al. Early treatment response in
malignant lymphoma, as determined by planar fluorine-18-
fluorodeoxyglucose scintigraphy. J Nucl Med 1993;34:1706–10.
31. Oliver T, Shamash J, Powles T, Somassundram U, Ell PJ. 20
years phase K study of single agent carboplatin in metastatic
seminoma: could it have been accelerated by 72 hour PET
scan response? ASCO, New Orleans, June 2004.
32. Brugarolas J, Clark JW, Chabner B. Using ‘‘rationally
designed drugs’’ rationally. Lancet 2003;361:1758–9.
33. Francis DL, Visvikis D, Costa DC, Croasdale I,
Arulampalam TH, Luthra SK, et al. Assessment of recurrent
colorectal cancer following 5-fluorouracil chemotherapy using
both 18FDG and 18FLT PET. Eur J Nucl Med Mol Imaging
2004;31:928.
34. Francis DL, Freeman A, Visvikis D, Costa DC, Luthra SK,
Novelli M, et al. In vivo imaging of cellular proferation in
colorectal cancer using Positron Emission Tomography. Gut
2003;52:1602–6.
35

35. Francis DL, Visvikis D, Costa DC, Arulampalam THA,
Townsend C, Luthra I, et al. Potential impact of [18F]39-
deoxy-39-fluorothymidine versus [18F] fluoro-2-deoxy-D-glu-
cose in Postron Emission Tomography for colorectal cancer.
Eur J Nucl Med 2003;30:988–94.
36. Shields AF, Grierson JR, Dohmen BM, et al. Imaging in vivo
proliferation with 18FLT and positron emission tomography.
Nature Med 1998;11:1334–6.
37. Eary JF, Brenner W, Vernon C, and O’Sullivan F. Tumor
heterogeneity in sarcoma patients is a significant predictor of
survival. Eur J Nucl Med Mol Imaging 2004;31 Suppl.
2:S232.
38. Dehdashti F, Flanagan FL, Mortimer JE, Katzenellenbogen
JA, Welch MJ, Siegel BA. Positron emission tomographic
assessment of ‘‘metabolic flare’’ to predict response of
metastatic breast cancer to antiestrogen therapy. Eur J Nucl
Med 1999;26:51–6.
39. Strauss LG. Fluorine-18-deoxyglucose and false-positive
results: a major problem in the diagnostics of oncological
patients. Eur J Nucl Med 1996;23:1409–15.
40. MacManus MP, Hicks RJ, Matthew JP, et al. Positron
emission tomography is superior to computed tomography
scanning for response-assessment after radical radiotherapy
or chemotherapy in patients with non-small-call lung cancer.
J Clin Oncol 2003;21:1285–92.
36
P J Ell
41. Scarfone C, Lavely WC, Cmelak AJ, Delbeke D, Martin WH,
Billheimer D, et al. Prospective feasibility trial of radiotherapy
target definition for head and neck cancer using 3-dimentional
PET and CT imaging. J Nucl Med 2004;45:543–52.
42. Schirrmeister H, Guhlmann A, Kotzerke J, Santjohanser C,
Kuhn T, Kreienberg R, et al. Early detection and accurate
description of extent of metastatic bone disease in breast
cancer with fluoride ion and positron emission tomography.
J Clin Oncol 1999;17:2381–9.
43. Cook GJ, Houston S, Rubens R, Maisey MN, Fogelman I.
Detection of bone metastases in breast cancer by 18FDG
PET: differing metabolic activity in osteoblastic and osteolytic
lesions. J Clin Oncol 1998;16:3375–9.
44. Gayed I, Vu T, Johnson M, Macapinlac H, Podoloff D.
Comparison of bone and 2-deoxy-2-[18F]fluoro-D-glucose
positron emission tomography in the evaluation of bony
metastases in lung cancer. Mol Imaging Biol 2003;5:26–31.
45. Schuetze SM, Eary JF, Griffith KA, Rubin BP, Hawkins DS,
Vernon CB, et al. FDG PET but not RECIST agrees with
histological response of soft tissue sarcoma to neoadjuvant
chemotherapy. ASCO 2005.
The British Journal of Radiology, January 2006
The British Journal of Radiology, 79 (2006), 37–43 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/29320216

Mesenteric panniculitis in oncologic patients: PET-CT findings

1,2,3 1,3
R ZISSIN, MD, U METSER, MD, 4D HAIN, MD and 1,3
E EVEN-SAPIR, MD, PhD
1
Department of Nuclear Medicine, Tel-Aviv Sourasky Medical Center, and the 2Department of Diagnostic Imaging, Sapir
Medical Center, Kfar Saba, both affiliated to the 3Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, and
4
Nuclear Medicine Institute, Shaare Zedek Medical Center, Jerusalem, Israel

Abstract. The aim of this study is to assess the role of PET/CT in differentiating between mesenteric panniculitis
(MP) and co-existing MP and mesenteric tumoural involvement. A total of 33 PET/CT examinations, of 19
oncologic patients (16 men and three women with ages ranging from 48 years to 83 years) with findings of MP
on the CT part of the study were retrospectively reviewed. The FDG uptake in mesenteric nodules was
recorded. The final diagnosis of malignant mesenteric involvement was based on clinical and imaging follow-up.
Based on the FDG uptake in mesenteric nodules, patients were categorized as group A: increased mesenteric
uptake (n58) and group B: no mesenteric uptake (n511). In seven of the eight patients in group A, a co-existing
MP and mesenteric tumour involvement was found: one patient had a recurrent cervical carcinoma and the
other six patients had lymphoma. In four of these six patients, the positive PET findings disappeared on follow-
up PET/CT with complete remission while the CT findings of the MP remained unchanged. In the other two,
the PET findings progressed along with clinical deterioration. In the last patient of group A, with rectal
carcinoma without evidence of recurrence, the mesenteric FDG uptake was a false positive uptake. In all 11
patients with CT findings of MP and negative PET, no malignant involvement of the mesentery was diagnosed.
To conclude, a negative PET has a high diagnostic accuracy in excluding tumoural mesenteric involvement
while increased uptake suggests the co-existing of mesenteric deposits, particularly in patients with lymphoma.

Mesenteric panniculitis (MP), also entitled liposclerotic
mesenteritis, mesenteric lipodystrophy, mesenteric lipoma-
tosis and lipogranuloma of the mesentery, is a benign
condition characterized by non-specific inflammation
involving the adipose tissue of the mesentery, with acute
inflammatory changes and fat necrosis being the pre-
dominant histological findings. In its chronic phase when
fibrosis is dominant, the disease is known as retractile
mesenteritis [1–4]. Sclerosing mesenteritis seems the most
appropriate diagnostic term of this entity, characterized by
a spectrum of histological findings [4]. The specific aetiology
of the disease is unknown, although various causes have been
suggested, including infection, trauma or ischaemia of the
mesentery. The disease has been related to other pathological
processes such as vasculitis, granulomatous disease, pancrea-
titis and malignancy [2]. Its prevalence in abdominal CT
examinations is approximately 0.6%, commonly appearing
as an incidental finding, mostly in middle or late adulthood
[5]. An association between MP and pre-existing malignancy
has been reported [5, 6].
The CT features of MP are well recognized and may
suggest the diagnosis, but they are non-specific and can
appear in other conditions such as mesenteric oedema,
granulomatous diseases, primary or secondary abdominal
neoplasms and lymphoma [1]. In cases of MP and known
intra-abdominal malignancies, differentiating MP from
tumoural involvement of mesenteric lymph nodes (LNs) is
of crucial importance.
18
F-Fluorodeoxyglucose (FDG)/PET imaging has been
introduced in addition to conventional cross-sectional
Received 17 March 2005 and in final form 3 May 2005, accepted 1 June
2005.
Address correspondence to: Einat Even-Sapir, Department of Nuclear
Medicine, Tel-Aviv Sourasky Medical Center, 6 Weizman Street, Tel-
Aviv, 64239 Israel.
imaging methods in the routine practice of oncologic
patients. Recently, hybrid systems composed of PET and
CT have been introduced and its use is increasing steadily
[7]. PET and CT are performed at the same clinical setting
resulting with generation of fused PET/CT images, which
provides both functional and anatomical data. The
potential role of PET/CT in differentiating benign MP
from tumoural mesenteric involvement is the topic of the
current study. We have reviewed 19 patients with a history
of known malignancy, who had incidental MP on the CT
component of the PET/CT study, and report the PET/CT
features of MP in this oncologic population.
Material and methods
The clinical data and PET/CT findings of 19 consecutive
patients with MP incidentally diagnosed on the CT part of
the study were retrospectively reviewed. The patient group
consisted of 16 men and 3 women with ages ranging from
48 years to 83 years (mean age 62¡11 years). Five of the
19 patients underwent a PET/CT study for staging and 14
for suspected recurrence or for monitoring response to
treatment. Known malignancies included lymphoma
(n510), colorectal cancer (n55), melanoma (n52) and
lung and cervical carcinomas, one patient each. A total of
33 PET/CT examinations were performed and reviewed in
these 19 patients as 11 of them had one to three follow-up
(F/U) studies.
The patients fasted at least 4 h prior to the intravenous
(IV) injection of 370–666 MBq (10–18 mCi) FDG.
Iodinated oral contrast material was administered prior
to FDG injection. Glucose levels had been checked prior
to the injection of FDG. A PET/CT study was performed
only when blood glucose levels were bellow 8.32 mmo l21.
Scanning from the base of the skull through the mid-thigh

The British Journal of Radiology, January 2006 37


was performed using the Discovery LS PET/CT system
(GE Medical Systems, Milwaukee, WI). Low-dose CT
acquisition was performed first with 140 kV, 80 mA, 0.8 s
per CT rotation, a pitch of 6 and a table speed of
22.5 mm s21, without any specific breath-holding instruc-
tions. A PET emission scan was carried out immediately
following acquisition of the CT, without changing the
patient’s position. From 5 to 8 bed positions were
performed with an acquisition time of 5 min for each
one. CT data were used for attenuation correction. Images
were generated and interpreted on work station (Xeleris
Elgems, Haifa, Israel) equipped with fusion software that
enables the display of PET, CT and fused PET/CT images.
The CT criteria for the diagnosis of MP included a well-
defined, inhomogeneous fatty mass with higher attenua-
tion than the normal retroperitoneal fat, occasionally with
preserved perivascular fat, that contained small nodules
and surrounded by a tumoural pseudocapsule (Figure 1)
[5]. When interpreting the PET/CT, the uptake of FDG in
the mesenteric nodules was reported. Standardized uptake
value (SUV) was measured for any focal increased uptake
within the CT mesenteric abnormalities. The latter
semiquantitative parameter was automatically obtained
on the patient’s final report and was calculated as the ratio
of activity in tissue per millilitre to the activity in the
injected dose per patient body weight. The final diagnosis
of the mesenteric pathology was based on clinical and
imaging (PET/CT and/or diagnostic CT) F/U: co-existing
MP and malignancy was concluded when improvement or
disappearance of the mesenteric abnormalities were seen
on F/U imaging together with clinical evidence of a
favourable response to therapy, or when mesenteric
abnormalities progression was seen on F/U imaging
along with clinical and imaging evidence of disease
progression. The MP was regarded as benign if the
mesenteric findings remained stable in a patient who was
clinically disease-free or if the findings remained stable on
Figure 1. CT findings of mesenteric panniculitis (MP). Non-
enhanced abdominal CT at the mid-abdomen shows a well-
defined, inhomogeneous fatty lesion, with higher attenuation
than the normal retroperitoneal fat, confined by a highly-
attenuated stripe representing a tumoural pseudocapsule (thick
arrows), with an engorged mesenteric vessel and scattered dis-
crete nodules of soft-tissue density, some of which are engulfed
by a hypodense fatty halo (thin arrow).
38
R Zissin, U Metser, D Hain and E Even-Sapir
F/U imaging while other sites of disease resolved in
response to therapy and the patient was clinically
considered in complete remission.
Results
Based on the FDG uptake within the MP, the study
patients were divided into two groups: group A consisting
of eight patients with MP and focal increased FDG uptake
within mesenteric nodules and group B consisting of 11
patients with MP without increased FDG uptake.
Group A: FDG uptake within MP
FDG uptake was demonstrated within CT mesenteric
abnormalities, indistinguishable from a benign MP, in
eight patients. The clinical and imaging findings of these
patients are summarized in Table 1. A conclusion of
malignant mesenteric involvement was made in seven of
these patients, one with a metastatic cervical cancer and
six with non-Hodgkin’s lymphoma (NHL): In four
lymphoma patients both the increased FDG uptake and
the nodules themselves resolved following chemotherapy,
while other CT findings of MP remained unchanged on
F/U PET/CT (Figure 2). In the other two, clinical F/U and
repeat PET/CT were consistent with tumour progression
(Figure 3). In the case of metastatic cervical carcinoma
new mesenteric PET findings appeared within known MP,
seen previously on two PET/CT studies, along with clinical
evidence of tumour recurrence. In coexisting MP and
mesenteric tumoural involvement, the MP changes, seen
on the CT part of the examination, which were not
associated with increased FDG uptake, remained
unchanged on F/U PET/CT studies.
In the remaining patient with rectal cancer and focal
mesenteric FDG uptake (SUV – 3.5) the positive PET was
regarded as a false positive study as there was no clinical
or imaging evidence of active tumour, together with
stability of the MP findings on F/U diagnostic CT studies
during a long disease-free period of 28 months.
Group B: MP without increased FDG uptake
In 11 patients no FDG uptake was seen within typical
features of MP. The mesenteric soft-tissue nodes ranged
between immeasurable, numerous small nodules, to
discrete nodes measuring up to 0.9 cm in the short axis
and 1.9 cm in the long axis (Figure 4). In all these patients
the mesenteric abnormalities seen on the CT part of the
study were stable on imaging F/U of a mean of 10.5
months (range: 5–30 months) and we therefore believe that
the mesenteric findings were benign.
Discussion
MP is a non-neoplastic inflammatory process of
unknown aetiology, affecting the small bowel mesentery.
It was rarely diagnosed before the era of ultrasound and
CT, but currently it is not uncommonly encountered, often
as an incidental imaging finding. Male predominance, as
was found in our group, has been previously reported [1–3]
though a slight female predominance has been reported
in ?a single publication [5]. Most cases of MP are
The British Journal of Radiology, January 2006
PET-CT finding of mesenteric panniculitis
Table 1. Clinical, PET/CT and F/U of eight patients with CT findings of mesenteric panniculitis (MP) and increased FDG uptake
within mesenteric nodules
Patient no., sex, Primary Indication for PET/CT findings
age (years) tumour the 1st PET/CT (on the 1st study)
1. M, 76 NHL Staging at diagnosis CT: increased-attenuation
fat, pseudocapsule and
numerous, slightly enlarged
soft-tissue nodules
PET: uptake (SUV – 13.2)
in a 1.3 cm mesenteric
nodule
2. M, 61 NHL Monitoring response CT: markedly increased
to treatment. attenuation of the
- No baseline study mesenteric fat, enlarged
mesenteric LNs (up to
2.5 cm62 cm), pseudocapsule,
‘‘fat ring’’ sign.
PET: diffuse uptake (SUV – 2.7)
in enlarged mesenteric nodules,
and in mesenteric fat
3. M, 50 NHL Restaging for suspected CT: increased-attenuation fat,
recurrence pseudocapsule and numerous,
slightly enlarged soft-tissue
nodules
PET: uptake (SUV – 3.6) in a
0.8 cm mesenteric nodule
4. M, 55 NHL Staging at diagnosis CT: increased-attenuation
fat, pseudocapsule and
numerous, slightly enlarged
soft-tissue nodules
PET: uptake (SUV – 4.5)
in a 1 cm61.3 cm
mesenteric nodule
5. F, 57 NHL Staging at diagnosis CT: increased-attenuation
fat, pseudocapsule and
numerous, slightly enlarged
soft-tissue nodules
PET: uptake (SUV – 8.2) within
markedly enlarged mesenteric
nodules, up to 2.3 cm61.7 cm
6. M, 58 NHL Monitoring response to CT: increased-attenuation fat,
treatment. pseudocapsule and numerous,
- No baseline study slightly enlarged soft-tissue
nodules
PET: uptake (SUV – 3.8) in
several mesenteric nodules,
up to 0.8 cm.
The British Journal of Radiology, January 2006
Final diagnosis and
imaging F/U
Co-existing MP and
mesenteric lymphoma
F/U PET/CT (4 M later):
the hypermetabolic
mesenteric nodule enlarged
to 3.8 cm64 cm (SUV – 18.8)
with the appearance of two
new hypermetabolic nodules
(1.1 cm61.5 cm, SUV – 7.5),
indicating disease progression
No change in the other MP
findings
Co-existing MP and mesenteric
lymphoma
F/U PET/CT (5 M later-without
treatment): no change in
the mesenteric abnormalities
and their uptake. The patient
was clinically considered with
active disease
Co-existing MP and mesenteric
lymphoma
F/U PET/CT (17 M
later-following
chemotherapy):
no focal increased uptake.
No change in the other
MP findings.
Co-existing MP and mesenteric
lymphoma
F/U PET/CT (6 M later – following
chemotherapy):
No focal increased uptake
No change in other MP findings
Co-existing MP and mesenteric
lymphoma
F/U PET/CT (4 M later-following
chemotherapy): the mesenteric
lymphadenopathy decreased in
size to 1 cm, FDG uptake
disappeared
No change in other MP findings
Co-existing MP and mesenteric
lymphoma
F/U PET/CT (2 M later- following
chemotherapy): the hypermetabolic
mesenteric nodules and FDG
uptake disappeared
No change in other MP findings
(Continued)
39
 R Zissin, U Metser, D Hain and E Even-Sapir

Table 1. (Cont.) Clinical, PET/CT and F/U of eight patients with CT findings of mesenteric panniculitis (MP) and increased FDG
uptake within mesenteric nodules

Patient no., sex, Primary Indication for PET/CT findings Final diagnosis and
age (years) tumour the 1st PET/CT (on the 1st study) imaging F/U
7. F, 56 Metastatic Monitoring response CT: increased-attenuation fat, Co-existing MP and metastases
cervical to treatment. pseudocapsule and numerous,
carcinoma - No baseline study slightly enlarged soft-tissue
nodules
PET: negative F/U PET/CT (11 M later): no
change
F/U PET/CT (after 5 M): disease
progression with a new
1.7 cm60.9 cm mesenteric
nodule with FDG uptake
(SUV – 5.7)
No change in other MP findings
8. F, 55 Rectal Misty mesentery CT: increased-attenuation fat, No clinical evidence for active
carcinoma on a diagnostic CT, pseudocapsule and numerous, disease: A false positive PET
performed for F/U, enlarged LNs up to
negative markers 1.4 cm62.5 cm
PET: uptake (SUV – 3.5) in 3 - No change in the MP findings
small mesenteric nodules. in a previous CT study, 2 years
No uptake in the enlarged earlier and in a F/U diagnostic
nodes CT after 4 M, negative markers

F/U, follow-up; M, months; LN, lymph node.

asymptomatic and are incidentally detected on abdominal
CT performed for unrelated conditions [5]. On CT, MP
appears as a mass of increased-attenuation mesenteric fat
containing small soft-tissue nodes, with a maximal
transverse diameter directed toward the left abdomen
consistent with the orientation of the jejunal mesentery.
The infiltrated fat typically engulfs the mesenteric vessels
and displaces adjacent bowel loops without invading them
[1, 2, 5]. Hypodense, cystic-like areas and calcifications due
to fat necrosis are infrequently seen within this mass [1].
Increased fatty attenuation and small mesenteric nodules,
also termed ‘‘misty mesentery’’ may, however, be seen in
any pathological process infiltrating the mesentery, such as
inflammation, oedema, haemorrhage or metastases [8].
Two CT findings are considered more specific for the
diagnosis of MP as they have not been reported in other
mesenteric diseases: the presence of tumoural pseudocap-
sule (found in up to 60% of MP cases) and the ‘‘fat ring’’
sign of hypodense fatty halo surrounding mesenteric
nodules and vessels (seen in up to 75% of cases) [1, 2,
4]. Daskalogianki et al have reported the co-existence of
MP and various neoplastic diseases, especially lymphoma
and gastrointestinal and urogenital adenocarcinomas, in
up to 69% of patients with MP [5]. 10 of the 19 study
patients with CT findings of MP had lymphoma as the
underlying malignancy. Co-existing MP with malignant
mesenteric involvement was found in six of the lymphoma
patient (60%), representing 85.7% of the 7 study patients
with malignant mesenteric involvement. In oncologic
patients, therefore, the small soft-tissue mesenteric nodules
typically seen within the infiltrated mesenteric fat of MP
may be misdiagnosed as metastatic implants. On the other
hand, metastatic deposits within a pre-existing MP can
also be present, as was described in a single case report on
a patient with uterine papillary serous adenocarcinoma in
whom multiple nodular metastases were detected on CT
within typical MP findings [9].
The results of our study emphasise the potential role of
PET/CT in differentiating benign MP and MP with
mesenteric tumoural involvement. Fused PET/CT images
provide both metabolic and anatomic information with a
high accuracy. On CT, lymph node pathology is based on
size criteria alone. Enlarged lymph nodes may be reactive
while normal-sized nodes may contain early metastatic
deposits, which can be reliably detected by the functional
(PET) part of the study.
The majority of our patients, including those with a
malignant mesenteric involvement, had only subtle CT
findings and the differentiation between benign and
malignant causes could not be made with confidence
based on the CT alone. Our results suggest a potential role
for integrated PET/CT in the assessment of MP detected
on CT in oncologic patients. PET/CT study can be used to
correctly exclude mesenteric tumoural involvement when
no FDG uptake is seen within typical CT features of MP.
Alternatively, in a patient with an oncologic history, the
demonstration of FDG uptake, even in small-sized nodules
within characteristic CT findings of MP, is highly
suggestive of neoplastic involvement of the mesentery. In
PET/CTs of co-existing MP and mesenteric metastatic
deposits, the increased FDG uptake was detected in
nodules smaller than the benign nodules of the MP that
had no increased uptake. The increased FDG uptake of
these malignant mesenteric deposits resolved on a F/U
study following a favourable response to treatment while
the findings of the benign MP remained unchanged.
Increased FDG uptake is, however, not tumour-specific as
FDG uptake may be seen in benign inflammatory
conditions [10], as was the case in one of our patients in
whom a slightly increased FDG uptake was detected
within MP findings without evidence of malignancy on a
long-term F/U of 28 months. As the most consistent
histological finding of MP is the presence of an
inflammatory infiltrate, it may explain the uptake in that

40 The British Journal of Radiology, January 2006

PET-CT finding of mesenteric panniculitis

(a)

(b)

Figure 2. A 50-year-old man with follicular lymphoma: mesenteric panniculitis (MP) with meseneteric tumoural involvement before
and after a favourable response to chemotherapy. (a) A fused PET/CT image shows increased 18F-FDG uptake in an 8 mm nodule
(dashed arrow) in the background of MP (arrows). The latter appears as a mesenteric mass of inhomogeneous fatty tissue containing
scattered soft-tissue nodules which are not 18FDG-avid. (b) PET/CT images at diagnosis (top images) and following chemotherapy
(lower images) show regression in size of the nodule and disappearance of 18FDG uptake (arrows). No change is seen in the other
findings of the MP.

The British Journal of Radiology, January 2006 41

 R Zissin, U Metser, D Hain and E Even-Sapir

(a)

(b)

Figure 3. A 76-year-old man with non-Hodgkin’s lymphoma (NHL): mesenteric panniculitis (MP) with meseneteric tumoural involve-
ment with disease progression. PET/CT images at diagnosis (top images) and 4 months later (lower images): (a) At the mid-abdomen
typical findings of MP with no 18F-FDG uptake are seen, stable on F/U. (b) More caudally, increased 18F-FDG uptake is detected
at diagnosis within a 1.2 cm61.5 cm mesenteric nodule (SUV – 13.2) (arrows). On F/U the hypermetabolic node, most likely
involved with lymphoma, enlarged to 3.8 cm64 cm with increasing 18F-FDG uptake (SUV – 18.8) (arrows).

42 The British Journal of Radiology, January 2006

PET-CT finding of mesenteric panniculitis
Figure 4. A 78-year-old man with diffuse B cell lymphoma
and co-existing mesenteric panniculitis (MP). A fused PET/CT
image shows typical findings of MP, including inhomogeneous
fatty density mass with well-defined nodules of soft-tissue den-
sity, confined by a pseudocapsule (arrows), without 18F-FDG
uptake. The findings remain stable on a F/U PET/CT, 5 months
later, with no clinical evidence of active lymphoma.
case [4]. We have found a single case report in the English
literature regarding the FDG/PET in a patient with sclerosing
mesenteritis; a large, speculated, soft-tissue mesenteric mass
showed peripheral increased FDG uptake, probably repre-
senting the peripheral high metabolic inflammation and
inactive central area of fibrosis [11].
The limitations of our study are the relatively small
number of patients and the lack of a pathological proof
for all lesions. However, as often happens in tumour
imaging, not all detected lesions have histological diag-
nosis and their nature is sometimes based on clinical and
imaging F/U. Validation of our findings in larger patient
groups is warranted.
To conclude, if MP is suspected on the CT part of
the PET/CT study, special attention should be paid to the
18
F-FDG-avidity of the findings. A negative PET has high
diagnostic accuracy in excluding tumoural mesenteric
involvement while increased uptake may suggest the co-
existing of metastatic deposits, particularly in patients with
lymphoma.
The British Journal of Radiology, January 2006
Acknowledgments
The authors wish to thank Mrs Limor Zuriel, MSc, for
her assistance in the preparation of the manuscript.
References
1. Horton KM, Lawler LP, Fishman EK. CT findings in
sclerosing mesenteritis (panniculitis): spectrum of disease.
Radiographics 2003;23:1561–7.
2. Sabate JM, Torrubia S, Maideu J, Franquet T, Monill JM,
Perez C. Sclerosing mesenteritis: imaging findings in 17
patients. AJR Am J Roentgenol 1999;172:625–9.
3. Parra-Davila E, McKenney MG, Sleeman D, Hartmann R,
Rao RK, McKenney K, Compton RP. Mesenteric pannicu-
litis: case report and literature review. Am J Surg
1998;64:768–71.
4. Emory TS, Monihan JM, Carr NJ, Sobin LH. Sclerosing
mesenteritis, mesenteric panniculitis and mesenteric lipody-
strophy: a single entity? Am J Surg Pathol 1997;21:392–8.
5. Daskalogiannaki M, Voloudaki A, Prassopoulos P,
Magkanas E, Stefanaki K, Apostolaki E, et al. CT evaluation
of mesenteric panniculitis: prevalence and associated diseases.
AJR Am J Roentgenol 2000;174:427–31.
6. Kipfer RE, Moertel CG, Dahlin DC. Mesenteric lypodystro-
phy. Ann Intern Med 1974;80:582–8.
7. Schoder H, Larson SM, Yeung HW. PET/CT in oncology:
integration into clinical management of lymphoma, mela-
noma, and gastrointestinal malignancies. J Nucl Med. 2004;45
Suppl. 1:72S–81S.
8. Mindelzun RE, Jeffrey RB Jr, Lane MJ, Silverman PM. The
misty mesentery on CT: differential diagnosis. AJR Am J
Roentgenol 1996;167:61–5.
9. La Fianza A, Alberici E, Di Maggio EM, Preda L, Campani
R. Multiple nodular metastases in mesenteric panniculitis by
uterine papillary serous adenocarcinoma [UPSC]: CT appear-
ance of a case. Clin Imaging 1999;23:90–3.
10. Kazama T, Faria SC, Varavithya V, Phongkitkarun S, Ito H,
Macapinlac HA. FDG PET in the evaluation of treatment for
lymphoma: clinical usefulness and pitfalls. Radiographics
2005;25:191–207.
11. Nguyen BD. F-18 FDG PET demonstration of sclerosing
mesenteritis. Clin Nucl Med 2003;28:670–1.
43
The British Journal of Radiology, 79 (2006), 44–51 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/23954854

Diagnostic efficacy of SonoVueH, a second generation

contrast agent, in the assessment of extracranial carotid or
peripheral arteries using colour and spectral Doppler
ultrasound: a multicentre study
1
P S SIDHU, FRCR, 2P L ALLAN, FRCR, 3F CATTIN, MD, 4D O COSGROVE, FRCR, 5A H DAVIES, MD,
6
D D DO, MD, 7S KARAKAGIL, MD, 8J LANGHOLZ, MD, 9D A LEGEMATE, MD, PhD,
10
A MARTEGANI, MD, 11J-B LLULL, MD, 12C PEZZOLI, PhD and 11A SPINAZZI, MD
1
Department of Radiology, King’s College Hospital, Denmark Hill, London SE5 9RS, UK, 2Department of Radiology,
Edinburgh Royal Infirmary, Edinburgh, Lothian, UK, 3CHU de Besancon, Service de Radiologie B, Boulevard Fleming,
F-25030 Besancon, Belgium, 4Department of Imaging, Royal Postgraduate Medical School, Hammersmith Hospitals
Trust, Du Cane Road, London W12 OHS, UK, 5Department of Surgery, Imperial College School of Science, Technology
& Medicine, Charing Cross Hospital, Fulham Palace Road, London W6 8RP, UK, 6Inselspital-Division of Angiology,
Department of Internal Medicine, University of Berne, Freiburgstarsse 10, 3010 Berne, Switzerland, 7Department of
Surgery, Uppsala University Hospital, Akademiska sjukhuset, 75185 Uppsala, Sweden, 8Schwerpunktpraxis für
Angiologie, Wilsnacker Strabe 14, 10559, Berlin, Germany, 9Department of Vascular Surgery, Academic Medical Centre,
Meibergdreef 9, NL-1105 AC DE Amsterdam, The Netherlands, 10Servizio di Radiologia, Ospedale Valduce, Via Dante
Alighieri, 11, 22100 Como, Italy, 11Bracco Diagnostics Inc., 107 College Road East Princeton, Princeton, NJ 08540,
USA and 12Bracco Imaging SpA, Via E. Folli 50, 20134 Milan, Italy

Abstract. The purpose of this study was to demonstrate the improvement in diagnostic quality and diagnostic
accuracy of SonoVueH microbubble contrast-enhanced ultrasound (CE-US) versus unenhanced ultrasound
imaging during the investigation of extracranial carotid or peripheral arteries. 82 patients with suspected
extracranial carotid or peripheral arterial disease received four SonoVue doses (0.3 ml, 0.6 ml, 1.2 ml and
2.4 ml) with Doppler ultrasound performed before and following each dose. Diagnostic quality of the CE-US
examinations was evaluated off-site for duration of clinically useful contrast enhancement, artefact effects and
percentage of examinations converted from non-diagnostic to diagnostic. Accuracy, sensitivity and specificity
were assessed as agreement of CE-US diagnosis evaluated by an independent panel of experts with reference
standard modality. The median duration of clinically useful signal enhancement significantly increased with
increasing SonoVue doses (p¡0.002). At the dose of 2.4 ml of SonoVue, diagnostic quality evaluated as number
of inconclusive examinations significantly improved, falling from 40.7% at baseline down to 5.1%. Furthermore,
SonoVue significantly (p,0.01) increased the accuracy, sensitivity and specificity of assessment of disease
compared with baseline ultrasound. SonoVue increases the diagnostic quality of Doppler images and improves
the accuracy of both spectral and colour Doppler examinations of extracranial carotid or peripheral arterial
disease.

Colour and spectral Doppler ultrasound examination of
the peripheral [1, 2] and extracranial carotid [3, 4] arterial
systems is a well established non-invasive method of
assessment of arterial disease. Frequently, Doppler
ultrasound replaces conventional angiography with the
associated cost savings and reduction in patient morbidity
[5–8]. However, Doppler ultrasound does not always
provide a full diagnostic assessment and there is no
alternative but to seek confirmatory evidence of arterial
disease with conventional angiography and increasingly
with helical CT angiography and MR angiography [9–12].
A number of factors preclude a full Doppler ultrasound
examination of a vessel: heavily calcified plaque causes
acoustic shadowing, a deep-seated artery returns a poor
echo-signal and vessel tortuosity precludes a satisfactory
Doppler angle for accurate velocity measurements. In
Received 6 May 2005 and accepted 9 June 2005.
order to improve the diagnostic capability of a Doppler
ultrasound examination of the peripheral and carotid
arteries, introducing an echo-enhancing agent would be
expected to facilitate visualization of difficult arteries, thus
overcoming inherent problems associated with ultrasound,
and ultimately reducing unnecessary invasive and expen-
sive diagnostic procedures.
SonoVueH is the trademark name of a new ultrasound
contrast agent (BR1, Bracco, Italy) [13]. SonoVue is a
suspension of phospholipid stabilized sulphur hexafluoride
(SF6) microbubbles. When reconstituted with normal
saline the product is stable at room temperature for
several days, but should be used after reconstitution within
6 h as the product contains no preservative [14].
Reconstitution produces a high microbubble concentration
(up to 5 6 108 microbubbles ml21), a favourable size (90%
of microbubbles smaller than 8.0 mm, mean diameter
2.5 mm) and strong echogenicity over the range of

44 The British Journal of Radiology, January 2006

Diagnostic efficacy of SonoVue
frequencies used in medical ultrasound examinations [15].
The microbubbles produced are not trapped in the
capillary vasculature, and the use of SF6 (an innocuous
gas) renders the microbubbles more resistant to pressure
increases from the left ventricle of the heart, increas-
ing microbubble survival. SonoVue demonstrates a
maximum backscatter coefficient at about 3 MHz and
an elimination half-life of approximately 6 min. More
than 80% of the compound is exhaled via the lungs
in 11 min [16].
The efficacy of SonoVue in extracranial carotid or
peripheral arterial disease was evaluated in a multicentre
study where the quality end-points were as follows: to
ascertain the optimal dosage with regard to global quality
of images, to assess the duration of microbubble contrast
effect, to aid the interpretation of diagnostically difficult
colour and spectral Doppler ultrasound examinations and
to evaluate the potential of contrast-enhanced ultrasound
(CE-US) to change a non-diagnostic ultrasound examina-
tion into a diagnostic examination. In a subset study
population, the diagnostic accuracy, sensitivity and
specificity of SonoVue enhanced Doppler investigations
were evaluated in terms of agreement of CE-US examina-
tions in comparison with other recognized diagnostic
imaging modalities.
Material and methods
The overall study consisting of two (study A and B)
parallel multicentre studies was aimed at investigating,
with Doppler ultrasound and SonoVue, different vascular
territories: renal, abdominal, cerebral, extracranial carotid
or peripheral arteries, and the portal circulation.
We refer here to the results pertaining to extracranial
carotid or peripheral arteries. Local Medical Ethics
Committees granted approval for the study at each
hospital site according to local legal requirements and
the study was conducted in accordance with the
Declaration of Helsinki and European Good Clinical
Practice. All patients recruited gave written informed
consent.
Study population
The study population comprised 82 male and female
patients, over 18 years of age, with a suspected vascular
pathology, referred for Doppler ultrasound investigations
of carotid, iliac, femoral, popliteal or tibial arteries and for
whom the observers could not make an interpretation with
confidence at baseline unenhanced colour and spectral
Doppler ultrasound examination. The main criteria for
patient exclusion from the study were: severe congestive
heart failure (New York Heart Association Class IV);
unstable angina; severe cardiac arrhythmia; recent myo-
cardial infarction; recent organ transplant or unstable
neurological disease. Lactating women or women known
or suspected to be pregnant were excluded. Patients were
also excluded if they were critically ill, medically unstable
or were in an intensive care setting. Patients receiving
another investigational drug within 30 days prior to the
study were not recruited.
The British Journal of Radiology, January 2006
Study design
The study was a multicentre, open-label (on-site),
blinded (off-site), randomized, dose-ranging, cross-over
study to compare four different doses of SonoVue (0.3 ml,
0.6 ml, 1.2 ml, and 2.4 ml) in Doppler ultrasound
investigations of extracranial carotid (22 patients) or
peripheral arteries (60 patients). Patients were randomized
to one of four dose sequences, according to a randomiza-
tion schedule with block size 4. The SonoVue doses were
administered as an intravenous bolus injection over 20 s via
a 20 gauge cannula (Introcan-W; Braun Melsungen AG,
Germany) placed in the forearm. All centres employed the
standard ultrasound machine used for routine colour
Doppler ultrasound vascular examinations within the
department. A variety of ultrasound machines were
used, depending on the centre. Once the optimum colour
and spectral Doppler ultrasound parameters were set for
each patient at the baseline examination to extract
maximum information (with the gain turned down to
the lowest informative level), the parameters were
unaltered for the remainder of the CE-US examination.
For each patient, a vessel of interest was designated for
further investigation, based on the vessel that would most
likely drive the patient’s diagnosis. Doppler ultrasound
investigations of the designated vessel were performed at
baseline and after each injection of SonoVue, with Super-
Video Home System (S-VHS) videotape recording of
images beginning 30 s prior to injection and continuing
until the end of the microbubble contrast effect. At each
time point, the designated vessel was studied first with
either colour Doppler ultrasound or power Doppler
ultrasound (only one mode was used for each patient
depending on investigator choice), and then with spectral
Doppler imaging. All SonoVue administration and ima-
ging procedures were completed on the same day. The
interval between administrations of the different doses was
at least 10 min or until disappearance of the microbubble
contrast effect from the previous administration.
Assessments
Four independent experienced readers, paired for each
of the studies (A and B) and unaffiliated with the study
sites performed an off-site assessment of the recorded
ultrasound images. These readers were blinded to study
agent dose (whether baseline or post-dose), and patient
information, including results of other imaging procedures.
The S-VHS videotape recordings were divided into sets
consisting of four post-injection images, one for each of
the four SonoVue injection doses, plus the baseline images.
Within each image set, an assigned random code number
determined the order of presentation of patient images to
the off-site readers. The off-site readers were provided with
the identification of the vessel under investigation for each
video sequence. Following completion of these unpaired
assessments, the baseline and corresponding post-injection
images for each dose of SonoVue in each patient were
then assessed in matched pairs. For patients with an
available reference diagnostic modality (conventional
angiography, MR angiography or CT angiography)
from which a diagnosis could be ascertained (on-site), a
committee of three experienced physicians (Accuracy
Review Committee) unaffiliated with the study sites
45

compared the diagnosis obtained by the off-site assess-
ments of Doppler ultrasound images with the diagnosis
obtained with the reference modality.
Diagnostic quality
Duration of clinically useful signal enhancement, defined
as the time from appearance until disappearance of a
microbubble contrast effect of sufficient intensity to be
diagnostically or clinically useful, was assessed and
documented by each off-site observer subjectively
during review of the video recordings of the individual
examinations.
Incidence and duration of artefactual microbubble
contrast effects (shadowing, blooming and saturation
effects) were assessed following each dose of SonoVue.
Artefacts were defined as follows: a shadowing effect
appeared as an obscured image and/or Doppler spectrum,
blooming appeared as the presence of colour in an area
without flow, while a saturation effect appeared as a
noisy Doppler spectrum with artificially high velocities
[17]. Duration was evaluated from the actual time of
appearance to the disappearance of shadowing and/or
blooming and/or saturation effects. Each of the artefacts
was evaluated at their maximal effect in accordance
with the following three-point scale: 05no artefactual
effect; 15artefactual effect not compromising the image
analysis; 25artefactual effect compromising the image
analysis.
Assessment of inconclusive Doppler examinations was
performed on each baseline or post-injection video clip
where off-site readers had to assess if a diagnosis was
possible or not and, in patients where it was possible,
make a diagnosis based on a pre-defined check list.
Diagnostic accuracy
Diagnostic accuracy was assessed for baseline and for
the clinically recommended dose of SonoVue only (2.4 ml).
Assessment of agreement was carried out by an
Accuracy Review Committee, based on a comparison of
the diagnosis recorded by each of the off-site blinded
readers from the Doppler ultrasound investigations with
the diagnosis from the reference imaging modality. The
following four-point scale was used: 15full agreement;
25basic agreement (differences in details but leading to
the same diagnostic conclusion); 35partial agreement
(differences in details possibly leading to a different
diagnostic conclusion); 45disagreement.
Sensitivity and specificity
For the evaluation of the diagnostic performance of
SonoVue CE-US in terms of sensitivity and specificity,
it was necessary to further define agreement with the
reference modality in terms of detection/exclusion
(presence/absence) of particular lesions in the investigated
vessels for each study patient. An independent experienced
radiologist, not previously involved in these studies, was
asked to classify the Doppler ultrasound off-site diagnoses
and the reference modality diagnoses according to the
following predetermined list of possible diagnoses for the
designated vessel of interest: (1) no abnormality;
46
P S Sidhu, P Allan, F Cattin et al
(2) abnormality present: (a) stenosis . 50% or occlusion;
(b) atheromatous plaque; (c) arteriovenous malformation;
(d) aneurysm; (e) vessel displacement/compression due to
extrinsic space-occupying mass; (f) collaterals or collatera-
lization of normal vessels; (g) arterial wall dissection;
(h) other. Sensitivity was defined as the proportion of
patients with a matching abnormality in the vessel of
interest using Doppler ultrasound and patients with an
abnormality in the vessel of interest using the reference
standard. Specificity was defined as the proportion of
patients with no abnormality in the vessel of interest using
Doppler ultrasound and patients with no abnormality in
the vessel of interest using the reference standard.
Statistical methods
Demography
Demographics and other baseline characteristics were
summarized using descriptive statistics.
Efficacy analysis
An analysis of variance (ANOVA) using ranked
durations of clinically useful signal enhancement was
performed to investigate overall differences between doses.
Summary statistics, frequency distributions and cross-
tabulations were elaborated for efficacy parameters, but no
formal statistical analyses were performed. For the
purposes of statistical summaries from the assessment by
the Accuracy Review Committee, these data were further
categorized as follows: agreement denoted full agreement
or basic agreement, and disagreement denoted partial
agreement or disagreement. Individual study results for
diagnostic accuracy were analysed using McNemar’s test
of association between baseline and post-dose in the
proportion of patients for whom agreement was recorded.
For all analyses, a two-sided p-value was used to test for
significance.
Results
Eighty-two patients (study A, n543; study B, n539),
49 male and 33 female subjects, median age of 71 years
(range 41–87 years), with suspected extracranial carotid
artery or peripheral vascular disease were recruited and
received SonoVue for the assessment of diagnostic quality
parameters. The diagnostic accuracy assessment was
performed on 59 patients where final diagnosis made
from a reference imaging modality (conventional angio-
graphy/CT angiography n558 and MR angiography n51)
was available for assessment by the Accuracy Review
Committee.
At Doppler examination 32 of 59 were found positive
and 27 of 59 were negative for the presence of pathology.
Sensitivity and specificity were calculated in the subset
of patients (n546) where the reference standard with the
pathology or no pathology in the vessel of interest was
available for assessment (n522 and 24 patients, respec-
tively, for the two studies).
The British Journal of Radiology, January 2006
Diagnostic efficacy of SonoVue

Diagnostic quality Assessment of inconclusive Doppler ultrasound

Duration of signal enhancement
A statistically significant dose response was observed in
the duration of clinically useful signal enhancement with a
significant increase in the median duration across the doses
(p,0.001 for 3 readers and p50.002 for 1 reader). At the
highest SonoVue dose of 2.4 ml, the average median
duration of clinically useful signal enhancement was of
3.9 min, range 0.0–14.3 (Table 1).
The pair of off-site reader assessments for each patient
was combined by calculating the average duration of
clinically useful signal enhancement.
examinations
Despite the limitations of the methodology used in this
study, the results of the statistical analysis performed
showed that, at the dose 2.4 ml, which is recommended for
Doppler ultrasound of macrovasculature, SonoVue mark-
edly decreased the number of baseline inconclusive
Doppler ultrasound examinations (rated as ‘‘no diagnosis
possible’’). Considering the entire population, the percen-
tage decreased from 40.7% to 7.4% (decrease533.3%)
while, in the patient population with a reference gold
standard control, the percentage decreased from 45.8% to
5.1% (decrease540.7%) (Figure 2).

Assessment of diagnostic accuracy

Artefactual effects
The percentage of agreement between diagnosis from
Due to methodology which did not permit gain Doppler ultrasound investigations and diagnosis from the
adjustment, a dose response was observed in the incidence reference imaging modality in the entire patient population
and duration of artefactual contrast effects, with median increased from 30.7% at baseline to 68.9% post-contrast
values increasing up to a maximum of 3.6 min at the (Table 2).
2.4 ml dose. The most common artefactual effects were
blooming in colour or power Doppler ultrasound (up to
92.9% with the 2.4 ml dose) followed by a saturation effect
on spectral Doppler ultrasound. Both of these artefacts are
related to the increase in Doppler signal intensity caused
by the microbubble contrast agent (Figure 1). Shadowing
was not reported to be a significant microbubble contrast
artefactual effect by any of the four off-site readers.

Table 1. Duration of clinically useful signal enhancement in the

two studies combined (A and B) from the off-site evaluation

SonoVue dose
n582 Baseline 0.3 ml 0.6 ml 1.2 ml 2.4 ml Figure 2. Bar chart diagram demonstrating the alteration in
the number of inconclusive off-site Doppler ultrasound assess-
Median (min) 0.00 2.5 2.9 3.4 3.9 ments at the dose of 2.4 ml of SonoVue (entire population: all
Range 0.0– 2.0 0.0–9.2 0.0–10.1 0.0–20.5 0.0–14.3 patients in the study, n582; population with ref. std: patients
with the reference gold standard, n546).

(a) (b)

Figure 1. Illustration of the ‘‘blooming’’ artefact. (a) Following the administration of SonoVue 1.2 ml, extensive blooming at 36 s
obscures the arterial anatomy precluding diagnostic interpretation. (b) Without adjustment of the ultrasound machine imaging para-
meters, at 63 s blooming has subsided and there is better delineation of the arterial anatomy.

The British Journal of Radiology, January 2006 47


The change in agreement rates from baseline was
statistically significant for three of the four off-site readers
(range p,0.05–0.001). Furthermore, considering the subset
of patients whose investigation was diagnostic and the
reference standard imaging modality available, agreement
between the Doppler ultrasound diagnosis and the
diagnosis from the reference imaging modality further
increased to 72.3% after microbubble contrast adminis-
tration (Figure 3).
Diagnostic performance, in terms of sensitivity and
specificity, was assessed in the subset of patients (n546)
with an abnormality or no abnormality in the vessel of
interest on the available reference gold standard examina-
tions. In study A, in the eight patients with an abnormality
in the carotid/peripheral vessel of interest on the reference
standard modality, the sensitivity increased from 13% pre-
contrast to 75% at the 2.4 ml dose for reader 1 and from
50% to 75% for reader 2 (Table 3).
In the 16 patients with no abnormality on the reference
modality, specificity increased from a pre-contrast value of
0% to 85 % at 2.4 ml for reader 1 and from 31% to 75%
for reader 2. In study B, in the nine patients with an
abnormality on the reference modality, the sensitivity
increased from 89% pre-contrast to 100% at the 2.4 ml
dose for reader 3 and increased from 22% to 67% for
reader 4. In the 13 patients with no abnormality on the
reference modality, specificity increased from a pre-
contrast value of 23% to 46% at 2.4 ml for reader 3 and
from 8% to 85% for reader 4.
Discussion
Failure to obtain a diagnostic colour Doppler ultra-
sound examination of the extracranial carotid and
peripheral arteries is typically a consequence of patient
factors. Rather than abandon the colour Doppler ultra-
sound examination instituting another examination, with
the implications of higher cost, the introduction of a
microbubble contrast agent would enable the examiner to
attempt to establish a conclusive diagnosis and reduce the
amount of time necessary to perform a peripheral arterial
examination. In the present study, a significant dose effect
was observed for the duration of clinically useful signal
enhancement for all four off-site readers; the average
median duration of useful enhancement was of 3.91 min
for the 2.4 ml dose. Due to a conservative approach in the
study design, the incidence and duration of artefactual
microbubble contrast effects also tended to increase with
increasing dose. These artefacts can normally be limited by
reducing the effective sensitivity of the system, by
decreasing the colour or power and spectral Doppler
ultrasound gains. Indeed, in order to maximize Doppler
quality, gain settings for both colour or power and spectral
Doppler ultrasound should be continuously adjusted as
enhancement returns to baseline. However, in this study
Table 2. Diagnostic accuracy. Percentage agreement between
diagnosis from Doppler ultrasound investigations (both unen-
hanced and SonoVue microbubble contrast-enhanced) and diag-
nosis from the reference standard in the entire population
Unenhanced SonoVue 2.4 ml
Agreement with gold standard 30.7% 68.9%
48
P S Sidhu, P Allan, F Cattin et al
the protocol required that the level of the gain for both
colour or power and spectral Doppler be set before the
first injection of microbubble contrast and could not be
modified after that. With gain adjustment, the duration of
artefacts would have been greatly reduced if not
completely eliminated and, as a result, the duration of
clinically useful signal enhancement would have been
increased. Moreover, now there is a tendency to use
infusions rather than bolus injection of microbubble
contrast during the investigation of vascular disease.
This has been demonstrated to further improve the
duration of useful enhancement and reduce artefactual
effects in the extracranial carotid and peripheral arteries
[18], in transcranial Doppler ultrasound [19] and in the
portal vein [20, 21].
Administration of microbubble contrast resulted in an
increase in agreement between colour Doppler ultrasound
diagnosis and diagnosis from a reference imaging mod-
ality. This is of importance where full reliance can be
placed on the results of a colour Doppler ultrasound
examination in order to bypass angiography prior to any
surgical procedure, particularly in carotid end-arterectomy
surgery [5] where there is a small but significant morbidity
attached to diagnostic angiography [22]. The effect of
introducing a microbubble contrast agent, the ‘‘Doppler
rescue’’ effect, has been successful in the imaging of the
renal arteries [23], the hepatic artery in the liver transplant
patient [24, 25] and the portal vein [20, 26, 27]. In the
assessment of renal artery disease, using a galactose based
microbubble contrast agent (LevovistTM; Schering AG,
Berlin, Germany), visualization of the renal arteries
improved from 65.7% to 78.3% (p,0.01) following the
administration of microbubble contrast [23]. The use of
SonoVue in the present study improved the ability of all
the off-site readers, presented with a minimum amount of
information, to make a confident interpretation of the
underlying vascular disorder on the CE-US examination.
On the baseline ultrasound examinations, a correct
diagnosis confirmed by the standard of reference was
achieved in 29.7% of studies, improving to 67.6% with the
2.4 ml dose of SonoVue. Moreover, if accuracy is
evaluated in the subset of patient population with
diagnostic examinations, a further increase in the percen-
tage of agreement with reference gold standard is observed
after SonoVue (72.3%) compared with unenhanced
examinations. There was an overall improvement in
sensitivity and specificity for all the off-site readers. The
ability shown by SonoVue to improve the diagnostic
information from a recorded ultrasound examination,
having knowledge of the vessel of interest only, is
remarkable since in the clinical practice the nature of
any ultrasound examination is one of examiner-patient
interaction, where the physician is allowed to develop an
overall concept of the diagnosis. This would suggest that
the use of a microbubble contrast agent as part of an on-
site ultrasound assessment would improve the diagnostic
ability to an even greater degree than what appeared under
the investigational conditions of the present study.
One limitation of the current study is the level of
sophistication of the ultrasound machines used. When this
multicentre study was commenced, each centre was
equipped with a ‘‘top-of-the-range’’ ultrasound machine,
but during the course of the study introduction on the
market of newer machines with digital capability, more
The British Journal of Radiology, January 2006
Diagnostic efficacy of SonoVue

(a)

(b)

(d)

(c)

Figure 3. (a) Baseline unenhanced colour and spectral Doppler ultrasound examination of a patient right lower limb. Insufficient
information for a firm conclusion about patency of the anterior tibial artery. (b) Following the administration of 2.4 ml of SonoVue,
clear depiction of a patent anterior tibial artery is seen (long arrow) with a large collateral artery seen in a superior position (short
arrow). (c) Spectral Doppler contrast ultrasound confirms a monophasic abnormal arterial trace. (d) Corresponding arteriogram con-
firms the patent anterior tibial artery (long arrow) and the collateral artery (short arrow). Collateral arteries have formed around an
occluded popliteal artery. (Courtesy of Dr J Langholz).

sensitive to blood flow, was seen to improve vascular microbubble contrast will still be advantageous to reduce
ultrasound diagnosis. Nevertheless, even with the improved the need for further imaging. The quality and standard of
capabilities of these newer ultrasound machines, problem the on-site colour Doppler ultrasound examinations
patients will still exist and the need for ‘‘Doppler-rescue’’ with were dependent on the experience of the examining

The British Journal of Radiology, January 2006 49

 P S Sidhu, P Allan, F Cattin et al

Table 3. Sensitivity and specificity of Doppler ultrasound Acknowledgments

investigations based on diagnosis from the reference gold
standard We wish to thank Dr Franca Heiman for her statistical
assistance.
Study A (n524) Baseline SonoVue 2.4 ml
Reader 1
Sensitivity (n58) (1/8) 0.13 (6/8) 0.75
CI (20.10, 0.36) (0.45, 1.05)
Specificity (n516) 0.00 (14/16) 0.88 References
CI (0.00, 0.00) (0.72, 1.04)
1. Kohler TR, Nance DR, Cramer MM, Vandenburghe N,
Reader 2
Strandness DEJ. Duplex scanning for diagnosis of aortailiac
Sensitivity (n58) (4/8) 0.50 (6/8) 0.75
and femoropopliteal disease: a prospective study. Circulation
CI (0.15, 0.85) (0.45, 1.05)
1987;76:1074–80.
Specificity (n516) (5/16) 0.31 (12/16) 0.75
2. Whelan JF, Barry MH, Moir JD. Color flow Doppler
CI (0.083, 0.54) (0.54, 0.96)
Study B (n522) Baseline SonoVue 2.4 ml ultrasonography: comparison with peripheral arteriography
Reader 3 for the investigation of peripheral arterial disease. J Clin
Sensitivity (n59) (8/9) 0.89 (9/9) 1.00 Ultrasound 1992;20:369–74.
CI (0.68, 1.09) (1.00, 1.00) 3. Sidhu PS, Allan PL. The extended role of carotid artery
Specificity (n513) (3/13) 0.23 (6/13) 0.46 ultrasound. Clin Radiol 1997;52:643–53.
CI (0.0012, 0.46) (0.19, 0.73) 4. Sidhu PS, Allan PL. Ultrasound assessment of internal
Reader 4 carotid artery stenosis. Clin Radiol 1997;52:654–8.
Sensitivity (n59) (2/9) 0.22 (6/9) 0.67 5. Khaw KT. Does carotid duplex imaging render angiography
CI (20.051, 0.49) (0.36, 0.98) redundant before carotid endarterectomy? Br J Radiol
Specificity (n513) (1/13) 0.08 (11/13) 0.85 1997;70:235–8.
CI (20.067, 0.23) (0.65, 1.04) 6. Garrard CL. Cost savings associated with the non-routine use
of carotid angiography. Am J Surg 2000;174:650–3.
7. Koelemay MJ, Legemate DA, de Vos H, van Gurp AJ, Balm
CI, 95% confidence interval.
R, Reekers JA, Jacobs MJ. Duplex scanning allows selective
use of arteriography in the management of patients with
sonographer, with likely variation between the centres severe lower leg arterial disease. J Vasc Surg 2001;34:661–7.
involved in the study. Accepting centres with an estab- 8. van der Zaag ES, Legemate DA, Nguyen T, Balm R, Jacobs
lished reputation for vascular ultrasound and ensuring that MJ. Aortoiliac reconstructive surgery based upon the results
of duplex scanning. Eur J Vasc Endovasc Surg 1998;16:383–9.
only the most experienced sonographers performed the
9. Rankin SC. CT angiography. Eur Radiol 1999;9:297–310.
examination minimized this variation. Not all of the 10. Dupuy DE, Boland GW. Non-invasive angiography with
patients had an acceptable standard of reference imaging magnetic resonance imaging and computed tomography.
examination, but even in the smaller number where this Imaging 1995;7:134–47.
was available, addition of microbubble contrast improved 11. Koelemay MJ, Lijmer JG, Stoker J, Legemate DA, Bossuyt
the diagnostic capability of the colour Doppler ultrasound PM. Magnetic resonance angiography for the evaluation of
examination. The variation in the assessment of the lower extremity arterial disease: a meta-analysis. JAMA
baseline colour Doppler ultrasound examinations by the 2001;285:1338–45.
off-site investigators highlights the difficulties of ultra- 12. Nelemans PJ, Leiner T, de Vet HC, van Engelshoven JM.
Peripheral arterial disease: meta-analysis of the diagnostic
sound interpretation and the subjective nature of conclu-
performance of MR angiography. Radiology 2000;217:105–14.
sions reached. However, addition of microbubble contrast, 13. Schneider M, Arditi M, Barrau MB, et al. BR 1: a new
although not completely eliminating this subjectivity, ultrasonographic contrast agent based on sulfur hexafluoride-
dramatically improved the confidence in interpretation filled microbubbles. Invest Radiol 1995;30:451–7.
allowing the off-site investigator to establish the correct 14. Schneider M. Characteristics of SonoVue. Echocardiography
diagnosis more consistently. 1999;16:743–6.
In conclusion, at the dose of 2.4 ml of SonoVue, the 15. Schneider M. SonoVue, a new ultrasound contrast agent. Eur
duration of useful enhancement achieved, which may be Radiol 1999;9:S347–8.
further extended by adjustment of the ultrasound machine 16. Morel DR, Schwieger I, Hohn L, et al. Human pharmaco-
kinetics and safety evaluation of SonoVue, a new contrast
settings, allowed a sufficiently prolonged period to
agent for ultrasound imaging. Invest Radiol 2000;35:80–5.
establish a definitive diagnosis avoiding further imaging. 17. Forsberg F, Liu JB, Burns PN, Merton DA, Goldberg BB.
Further studies to evaluate the potential use of infusion Artifacts in ultrasonic contrast agents studies. J Ultrasound
methods of SonoVue administration are needed; these may Med 1994;13:357–65.
have advantages over bolus methods of administration. 18. Albrecht T, Urbank A, Mahler M, et al. Prolongation and
The use of SonoVue allowed an improvement in diagnostic optimization of Doppler enhancement with a microbubble US
accuracy to be achieved in comparison with an contrast agent by using continuous infusion: preliminary
accepted reference examination. The administration of experience. Radiology 1998;207:339–47.
2.4 ml dose significantly produces an overall improvement 19. Goertler M, Kross R, Baeumer M, et al. Diagnostic impact
and prognostic relevance of early contrast-enhanced tran-
in terms of diagnostic performance. Microbubble ultra-
scranial color-coded duplex sonography in acute stroke.
sound contrast represents the next stage of development, Stroke 1998;29:955–62.
following on from the introduction of duplex Doppler 20. Schiedermaier P, Layer G, Sauerbruch T. Impact of the
and colour Doppler ultrasound, in the improving the continuous infusion of Levovist on color Doppler
overall diagnostic capability of ultrasound in the vascular sonography in portal hypertension. AJR Am J Roentgenol
system. 2002;178:61–5.

50 The British Journal of Radiology, January 2006

Diagnostic efficacy of SonoVue
21. Sidhu PS, Sellars ME, Heneghan M, Blomley MJK, Bauer A.
Visualization of the portal vein in normal subjects and
patients with cirrhosis: comparison of a bolus versus different
infusion injections of ultrasound contrast. Eur Radiol
2000;10:119.
22. Hankey GJ, Warlow CP, Sellar RJ. Cerebral angiographic
risk in mild cerebrovascular disease. Stroke 1990;21:209–22.
23. Claudon M, Plouin PF, Baxter GM, Devos DM. Renal
arteries in patients at risk of renal arterial stenosis: multi-
centre evaluation of the Echo-enhancer SH U 508A at color
and spectral Doppler US. Radiology 2000;214:737–46.
24. Sidhu PS, Shaw AS, Ellis SM, Karani JB, Ryan SM.
Microbubble ultrasound contrast in the assessment of hepatic
artery patency following liver transplantation: role in
reducing frequency of hepatic artery arteriography. Eur
Radiol 2004;14:21–30.
The British Journal of Radiology, January 2006
25. Sidhu PS, Ellis SM, Karani JB, Ryan SM. Hepatic artery
stenosis following transplantation: significance of the tardus
parvus waveform and the role of microbubble contrast media
in the detection of a focal stenosis. Clin Radiol 2002;57:789–
99.
26. Gebel M, Caselitz M, Bowen-Davies PE, Weber S. A
multicenter, prospective, open label, randomized, controlled
phase IIIb study of SH U 508A (Levovist) for Doppler signal
enhancement in the portal vascular system. Ultraschall in der
Medizin 1998;19:148–56.
27. Marshall MM, Beese RC, Muiesan P, Sarma DI, O’Grady J,
Sidhu PS. Assessment of portal venous system patency in the
liver transplant candidate: a prospective study comparing
ultrasound, microbubble contrast enhanced colour Doppler
ultrasound, with arteriography and surgery. Clin Radiol
2002;57:377–83.
51
The British Journal of Radiology, 79 (2006), 52–55 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/17905092

Lymphoepithelioma-like carcinoma of salivary glands:

treatment results and failure patterns
1
C-Y HSIUNG, MD, 2C-C HUANG, MD, 1C-J WANG, MD, 1E-Y HUANG, MD and 2H-Y HUANG, MD
Departments of 1Radiation Oncology and 2Pathology, Chang Gung Memorial Hospital-Kaohsiung, Taiwan, R.O.C.

Abstract. The purpose of this study was to evaluate the treatment results and failure patterns of
lymphoepithelioma-like carcinoma (LELC) of salivary glands. From June 1987 to May 2001, nine patients
with LELC of salivary glands were treated at our hospital. One patient was excluded due to the loss of clinical
follow-up after surgery. For the remaining eight patients, the primary tumour sites were parotid glands (4
patients), submandibular glands (3), and the minor salivary glands in right cheek (1), respectively. Seven
patients underwent surgical treatment and post-operative radiotherapy, while the other one patient was treated
with surgery only. The total radiation dose to the salivary tumour bed ranged from 39.6 Gy to 67.6 Gy (mean
dose: 58.3 Gy and median dose: 59 Gy). The treatment results and failure patterns were analysed. The survival
time ranged from 21.4 months to 145.2 months (mean: 69.1 months, median: 54.5 months). At the end of
follow-up, six patients were still alive and two died. One patient died of distant metastases 21.5 months after
the surgical treatment of LELC. The other case died of intercurrent disease (pontine haemorrhage) 53 months
after surgery. No patient had local or regional failure after the treatments. Distant failure was noted in two
patients. The patients with LELC of salivary glands were shown to have favourable prognoses. No local or
regional failure was noted. However, distant failure developed in two patients. The risk of distant metastasis
should be carefully monitored, especially for those patients with more advanced neck node involvement.

Lymphoepithelioma [1] consisted of poorly differen-
tiated cells with large nuclei and nucleoli within the
lymphoid stroma. Lymphoepithelioma occurs mainly in
the nasopharynx [2, 3]. Also, lymphoepithelioma-like
carcinoma (LELC) has been found in salivary glands [4–
6]. Because LELC is a rare histological type of cancer of
salivary glands [7, 8], the clinical data concerning LELC of
salivary glands is inadequate compared with other
common histological types. Also, the clinical course and
prognosis of this disease after the treatments have not been
thoroughly studied in the medical literature. As a result, a
retrospective study based on our patient database was
undertaken to analyse the treatment results and failure
patterns of LELC of salivary glands.
Patients and methods
From June 1987 to May 2001, nine patients with LELC of
salivary glands were treated at our hospital. One patient was
excluded due to the loss of clinical follow-up. The remaining
eight patients are followed up regularly after the treatments
and included in the current study. The general characteristics
of these patients were shown in Table 1. Three out of eight
patients were male and five were female. The primary tumour
sites were parotid glands (4 patients), submandibular glands
(3), and the minor salivary glands in right cheek (1),
respectively. These patients with LELC were staged according
to TNM classification of the American Joint Committee on
Cancer [9] (Table 1).
The treatment data of these patients are presented in
Table 2. All these eight patients underwent the excision of
Received 5 January 2005 and accepted 7 June 2005.
Address correspondence to: Hsuan-Ying Huang, Department of
Pathology, Chang Gung Memorial Hospital-Kaohsiung, 123, Ta-Pei
Road, Niao Sung Hsian, Kaohsiung Hsien, Taiwan, R.O.C.
primary salivary gland tumours. The dissection of enlarged
neck lymph nodes was also performed for the five patients
(patients 3, 4, 5, 6, and 8 in Table 2) with neck node
metastases noted by physical examination or CT scans. After
surgery, seven cases received post-operative radiotherapy with
a 60Co machine or 6–10 MV linear accelerator. Six (patients 1,
3, 4, 5, 6, and 8 in Table 2) out of these seven patients were
irradiated with two bilateral portals covering the salivary
tumour bed and upper neck and an anterior–posterior portal
covering the bilateral lower neck. The remaining one patient
(patient 2 in Table 2) received small-field radiotherapy
covering only salivary tumour bed without elective nodal
irradiation to bilateral low neck. In the seven patients treated
with post-operative radiotherapy, the total radiation dose to
the salivary tumour bed ranged from 39.6 Gy to 67.6 Gy
(mean dose: 58.3 Gy and median dose: 59 Gy). The dose to
spinal cord was no more than 45 Gy. For the six patients
receiving elective nodal irradiation to bilateral low neck, the
low-neck dose ranged from 34.2 Gy to 45 Gy (Table 2).
After the treatments, all the patients were followed
regularly at the clinics. The treatment results and failure
patterns were retrospectively reviewed. The survival time
was measured from the date of the first surgical treatment
to the date of last follow-up or death. The survival curves
were calculated by the Kaplan-Meier product-limit method
[10]. Local failure was defined as tumour recurrence in the
salivary tumour bed. Regional failure was defined as
tumour recurrence in the head and neck outside the
salivary tumour bed.
Results
The histology of LELC of one patient is shown in
Figure 1. The treatment results and failure patterns are
summarized in Table 3. The survival time ranged from

52 The British Journal of Radiology, January 2006

Lymphoepithelioma-like carcinoma of salivary glands

Table 1. The general characteristics of the eight patients with lymphoepithelioma-like carcinoma (LELC) of salivary glands

Age (years) Sex Primary site Stage [9]

Patient 1 42 Male Right submandibular gland T3 N0 M0
Patient 2 50 Female Minor salivary gland in right buccal area T1 N0 M0
Patient 3 40 Male Left submandibular gland T2 N2b M0
Patient 4 39 Female Right parotid gland T4 N1 M0
Patient 5 43 Female Right parotid gland T3 N2b M0
Patient 6 40 Male Left submandibular gland T3 N2b M0
Patient 7 42 Female Right parotid gland T2 N0 M0
Patient 8 46 Female Left parotid gland T3 N2b M0

21.4 months to 145.2 months (mean: 69.1 months, median:

54.5 months). The survival curve of these patients is shown
in Figure 2. At the last follow-up, six patients were still
alive and two had died. One patient died of distant
metastases 21.5 months after the surgical treatment of
LELC (patient 6 in Table 3). The other case died of
intercurrent disease (pontine haemorrhage) 53 months
after surgery (patient 1 in Table 3). No patient had local
or regional failure after the treatments. However, distant
metastases were noted in two patients (patients 6 and 8 in
Table 3). The interval between surgery and distant failure
was 6.3 months and 6.5 months for patient 6 and 8,
respectively. After the occurrence of distant metastases,
these two patients received chemotherapy with CDDP and
5-FU. At last follow-up, five patients were alive without
cancer, one was alive with distant metastases, another one
had died of distant metastases, and the remaining one had
died of intercurrent disease (Table 3).
During radiotherapy, oral mucositis and skin reaction
over radiation field were experienced in all the seven Figure 1. The histology of lymphoepithelioma-like carcinoma
patients irradiated. The major long-term complications (LELC) of one patient is shown here. Microscopically, diffuse
lymphoid infiltration is noted around the atrophic acini of sali-
after the treatments were xerostomia (8 patients), neck
vary gland and occasionally forms lymphoid follicles. Islands
fibrosis (6 patients), and facial palsy (3 patients). The of neoplastic epithelial cells bearing pleomorphic, vesicular
complication of facial palsy was due to tumour encase- nuclei and indistinct cell border are present within the lym-
ment of facial nerve and the surgical treatment. phoid tissue.

outside the post-operative radiation field 6 years after

Discussion
Lymphoepithelioma in nasopharynx is known as a
radiosensitive tumour and radiotherapy is the standard
treatment for nasopharyngeal lymphoepithelioma [2, 3].
Non-nasopharyngeal lymphoepithelioma of the head and
neck is also reported to be radiosensitive with high rates of
locoregional tumour control [5]. In the study of salivary
gland carcinoma by Teo et al [6], seven patients had LELC
from the parotid glands and only two of them experienced
locoregional relapses; one had isolated regional relapse
treatments and the other had in-field failure in the parotid
tumour bed 3.5 years after total parotidectomy and post-
operative radiation (50 Gy). In the current study, seven of
these eight patients with LELC of salivary glands received
surgery and post-operative radiotherapy and the other one
was treated with surgery only. No local or regional failure
was noted. From the results of this study and the above
literature [5, 6], surgery and post-operative radiotherapy
may be the appropriate treatment combination with
satisfactory locoregional control for patients with LELC

Table 2. The treatment data of the eight patients with lymphoepithelioma-like carcinoma (LELC) of salivary glands

Treatments Radiation dose (Gy)

Salivary tumour bed Bilateral low neck
Patient 1 Operation & radiotherapy 67.6 45
Patient 2 Operation & radiotherapy 65.2 0
Patient 3 Operation & radiotherapy 57.6 45
Patient 4 Operation & radiotherapy 39.6 39.6
Patient 5 Operation & radiotherapy 64.8 45
Patient 6 Operation & radiotherapy 59 45
Patient 7 Operation alone
Patient 8 Operation & radiotherapy 54.2 34.2

The British Journal of Radiology, January 2006 53

 C-Y Hsuing, C-C Huang, C-J Wang et al

Table 3. The treatment results and failure patterns of the eight patients with lymphoepithelioma-like carcinoma (LELC) of salivary
glands

Survival months Distant metastases Status at last follow-up

Patient 1 53 No Died of intercurrent disease
Patient 2 116.5 No Alive without cancer
Patient 3 56 No Alive without cancer
Patient 4 21.4 No Alive without cancer
Patient 5 103.6 No Alive without cancer
Patient 6 21.5 Lung, bone & liver Died of distant metastases
Patient 7 145.2 No Alive without cancer
Patient 8 35.4 Lung, bone & liver Alive with distant metastases

Figure 2. The survival curve of the

eight patients with lymphoepithe-
lioma-like carcinoma (LELC) of sali-
vary glands.

of salivary glands. In the future, further study with more
patients is needed to find the appropriate radiation field
and radiation dose for LELC of salivary glands.
Distant metastases to lung, bone, and liver were noted
in two patients (patient 6 and 8, Table 3). The duration
from the date of operation to distant metastases was
6.4 months and 6.6 months for patients 6 and 8,
respectively. Among these eight patients, there were four
patients with N0 or N1 stage (Table 1), and none of them
experienced distant metastases. The other four patients
were all staged as N2b and two of them had distant
metastases after the treatments. From this finding, the
neck node status might be associated with the risk of
distant metastases. In the study of non-nasopharyngeal
lymphoepithelioma of the head and neck [5], the main
cause of treatment failure was distant metastasis, which
occurred more frequently in patients with lymph node
involvement. As a result, the risk of distant metastasis
should not be overlooked for those patients with more
advanced neck node involvement.
Conclusion
In the current study, patients with LELC of salivary
glands were shown to have favourable prognoses. No local
or regional failure occurred in these patients. However,
distant failure developed in two patients. The risk of
distant metastasis should be carefully monitored, especially
for those patients with more advanced neck node
involvement.
Acknowledgments
The authors thank Yu-Ling Wu, M.S. for the kind
assistance with manuscript preparation.
References
1. Schmincke A. Uber lymphoepitheliale Geschevulste. Beitr
Pathol Anat 1921;68:161.
2. Perez CA. Nasopharynx. In: Perez CA, Brady LW, editors.
Principles and practice of radiation oncology. 2nd edn.
Philadelphia, PA: JB Lippincott; 1992:617–43.
3. Moss WT. The nasopharynx. In: Cox JD, editor. Moss’
radiation oncology: rationale, technique, results. 7th edn. St.
Louis, MO: Mosby, 1994:149–68.
4. Cleary KR, Batsakis JG. Undifferentiated carcinoma with
lymphoid stroma of the major salivary glands. Ann Otol
Rhinol Laryngol 1990;99:236–8.
5. Dubey P, Ha CS, Ang KK, El-Naggar AK, Knapp C, Byers
RM, et al. Nonnasopharyngeal lymphoepithelioma of the
head and neck. Cancer 1998;82:1556–62.

54 The British Journal of Radiology, January 2006

Lymphoepithelioma-like carcinoma of salivary glands

6. Teo PM, Chan AT, Lee WY, Leung SF, Chan ES, Mok CO.
Failure patterns and factors affecting prognosis of salivary
gland carcinoma: retrospective study. Hong Kong Med J
2000;6:29–36.
7. Simpson JR. Salivary glands. In: Perez CA, Brady LW,
editors. Principles and practice of radiation oncology. 2nd
edn. Philadelphia, PA: JB Lippincott, 1992:657–71.
8. Moss WT. The salivary glands. In: Cox JD, editor. Moss’
radiation oncology: rationale, technique, results. 7th edn. St.
Louis, MO: Mosby, 1994:121–31.
9. Major salivary glands (parotid, submandibular, and sub-
lingual). In: American Joint Committee on Cancer: AJCC
Cancer Staging Manual. Philadelphia, PA: Lippincott-Raven
Publishers, 5th edn, 1997:53–8.
10. Kaplan EL, Meier P. Nonparametric estimation from
incomplete observations. J Am Stat Assoc 1958;53:457–81.

The British Journal of Radiology, January 2006 55

The British Journal of Radiology, 79 (2006), 56–61 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/39775216

Comparison of patient doses in 256-slice CT and 16-slice CT

scanners
1,2
S MORI, MS, RT, MPR, 1M ENDO, PhD, MPH, 1K NISHIZAWA, PhD, MPH, 2K MURASE, PhD, MPH,
2
H FUJIWARA, PhD and 3S TANADA, MD
1
Department of Medical Physics, National Institute of Radiological Sciences, Chiba 263-8555, Japan, 2School of Allied
Health Sciences, Faculty of Medicine, Osaka University, Osaka 565-0871, Japan and 3Department of Medical Imaging,
National Institute of Radiological Sciences, Chiba 263-8555, Japan

Abstract. The 256-slice CT-scanner has been developed at the National Institute of Radiological Sciences.
Nominal beam width was 128 mm in the longitudinal direction. When scanning continuously at the same
position to obtain four-dimensional (4D) images, the effective dose is increased in proportion to the scan time.
Our purpose in this work was to measure the dose for the 256-slice CT, to compare it with that of the 16-slice
CT-scanner, and to make a preliminary assessment of dose for dynamic 3D imaging (volumetric cine imaging).
Our group reported previously that the phantom length and integration range for dosimetry needed to be at
least 300 mm to represent more than 90% of the line integral dose with the beam width between 20 mm and
138 mm. In order to obtain good estimates of the dose, we measured the line-integral dose over a 300 mm range
in PMMA (polymethylmethacrylate) phantoms of 160 mm or 320 mm diameter and 300 mm length. Doses for
both CT systems were compared for a clinical protocol. The results showed that the 256-slice CT generates a
smaller dose than the 16-slice CT in all examinations. For volumetric cine imaging, we found an acceptable scan
time would be 6 s to 11 s, depending on examinations, if dose must be limited to the same values as routine
examinations with a conventional multidetector CT. Finally, we discussed the studies necessary to make full use
of volumetric cine imaging.

In 2001 the introduction of a 16-slice CT-scanner raised
some new topics in CT technology development. 16-slice
CT allows applications of three-dimensional (3D) images
in clinical fields such as diagnosis, surgical simulation,
planning of radiation therapy and monitoring of inter-
ventional therapy. However, it is still difficult to take
dynamic 3D images of moving organs such as the heart or
lung to enlarge the application fields. In order to take
these images, we have developed a prototype 256-slice CT
at NIRS (National Institute of Radiological Sciences)
which employs continuous rotations of a cone-beam [1].
Clinical applications of CT techniques have continued
to increase the dose to patients during recent decades, as
CT examinations have come to provide higher quality
X-ray imaging with substantial benefits in clinical
diagnosis [2]. Notwithstanding the potential benefits to
the healthcare of patients using CT, the fundamental
concern in radiological protection is the optimization of
radiation exposure.
The maximum nominal beam width of the 256-slice CT
is 128 mm and is four times larger than the third-
generation 16-slice CT-scanner (Toshiba Aquilion;
Toshiba Medical Systems, Japan). A wider beam width
is more efficient for imaging in a wider coverage. However,
doses to patients with 256-slice CT are of considerable
concern if it is to be used for obtaining dynamic 3D
images (volumetric cine images). When scanning continu-
ously at the same position, the effective dose is increased in
Received 6 August 2004 and in revised form 8 April 2005, accepted 13
June 2005.
Address correspondence to: Shinichiro Mori, 4-9-1 Anagawa, Inage-
ku, Chiba-shi, Chiba, 263-8555, Japan.
proportion to the scan time and a wider coverage brings
larger doses to patients. Therefore, it is very important to
assess the dose of the 256-slice CT before volumetric cine
imaging for patients.
This work was carried out to compare doses, including
scattered radiation, of the 256-slice CT and 16-slice CT
and to make a preliminary assessment of dose for
volumetric cine imaging.
Materials and methods
Acquisition systems of 256-slice CT and 16-slice CT
scanners
The prototype 256-slice CT-scanner uses a wide-area 2D
detector designed on the basis of the present CT
technology and is mounted on the gantry frame of a
state-of-the-art CT-scanner (Figure 1) [3]. The number of
elements is 912 channels6256 segments; element size is
approximately 1 mm61 mm, corresponding to a 0.5 mm
(transverse)60.5 mm (longitudinal) beam width at the
centre of rotation. Gantry rotation time is 1.0 s. Data
sampling rate is 900 views/s, and the dynamic range of the
A/D converter is 16 bits. As shown in Appendix 1, the
reconstructed regions are cylinders of 240 mm diameter
and 102.4 mm length for the head scan and 320 mm
diameter and 93.9 mm length for the body scan. The
detector element consists of a scintillator and photodiode,
which are the same as for the scintillator of multidetector
CT (MDCT) (Toshiba Aquilion). Three wedge designs
(large, small, and flat) on the 256-slice CT are intended to
extend the conventional wedge designs of the third-
generation 16-slice CT-scanner (Toshiba Aquilion) in the

56 The British Journal of Radiology, January 2006

Comparison of patient dose in multislice CT
(a)
Figure 1. (a) Front view of 256-slice CT-scanner. (b) A wide-area 2D detector is designed on the basis of the present CT technology
and mounted on the gantry frame of the state-of-the-art CT-scanner.
longitudinal direction. The large and small wedges are
shaped to compensate for the variable path length of the
patient across the scan field of view (FOV). The small
wedge is used for an object under 240 mm FOV, and the
large wedge is used for over 240 mm FOV (e.g. chest and
abdomen). The flat wedge is thicker at the centre than the
other wedges.
A Feldkamp-Davis-Kress (FDK) algorithm [4] is used
for reconstruction. All further data processing and
interpretation are done with a high-speed image processor
with field programmable gate-array based-(FPGA) archi-
tecture. It takes less than 1 s to reconstruct volume data of
a 51265126256 matrix.
The 16-slice CT detector consists of 40 segments, which
can be electronically grouped to provide different image
slice configurations. The longitudinal FOV is 32 mm at the
maximum. Other major components are the same as those
of the 256-slice CT. In addition to the axial scan, the
helical scan mode can be selected to cover volumes beyond
the detector width.
Phantoms
The length of the IEC-recommended dosimetry phan-
tom [5] is at least 140 mm. This conventional phantom
contains holes just large enough to accept the pencil-
shaped ionization chamber. For dose measurement in
cone-beam CT, the length of the phantom should be
longer, because of the wider scatter distribution.
According to our previous results [6], the phantom
length and integration range for dosimetry needed to be
at least 300 mm to represent more than 90% of line
integral dose with the beam width between 20 mm and
138 mm. Therefore, in the present study we used 300 mm
long phantoms of PMMA (polymethylmethacrylate).
The diameters of the phantoms are 160 mm for head
and 320 mm for body examination. These phantoms
were provided by joining unit cylinders 150 mm long.
The details of the phantoms were described by Mori
et al [6].
The British Journal of Radiology, January 2006
(b)
Detectors
A pencil-shaped ionization chamber (CT-30; Oyogiken,
Japan) of active length 300 mm was connected to a
dosemeter (AE-132; Oyogiken, Japan) and used to
measure dose. The dosemeter was calibrated
(National Institute of Advanced Industrial Science and
Technology, Japan) for the appropriate radiation qualities.
Clinical scan conditions
We compared the doses of the 256-slice CT and the
16-slice CT for clinical scan conditions. These conditions
were mainly derived from those recommended by the
manufacture for the 16-slice CT. The X-ray tube current
was set such that the effective mAs should be the same
for both CTs, as given by (current)6(rotation time)/
(helical pitch) for the 16-slice CT and by
(current)6(rotation time) for the 256-slice CT. For the
256-slice CT, slice collimation was 224 mm60.5 mm for
the head, 128 mm61.0 mm for the pelvis, and
256 mm60.5 mm for other sites. For the 16-slice CT,
the slice collimation was set to 16 mm61.0 mm for pelvis
and 16 mm60.5 mm for other sites, helical pitch was 0.69
for the head, and 0.94 for other sites, because the scan
conditions were chosen to obtain the same spatial
resolution as for the 256-slice CT.
The whole scan ranges were 93.9 mm for chest,
187.8 mm for abdomen, and 281.7 mm for pelvis.
These scan ranges, except chest examination, were
beyond the detector width of the 256-slice CT in
the longitudinal direction, therefore they were set as
multiples of 93.9 mm, the maximum longitudinal FOV
of the 256-slice CT (Appendix 1). For the head
examination, because the recommended value for the
16-slice CT was shorter than the maximum FOV of
256-slice CT, the FOV was adjusted to narrow the
collimator width for the 256-slice CT. The clinical
scan conditions thus obtained are summarized in
Table 1.
57
 S Mori, M Endo, K Nishizawa et al

Table 1. Scan conditions for 256-slice CT and 16-slice CT-scanners

Examination Scanner Voltage Current Rotation Scan Beam collimation FOV Scan range Scan mode Helical pitch
(kV) (mA) time (s) time (s) (mm6mm) (mm) (mm)
Head 256-slice CT 120 326 1.0 1.0 22460.5 240 90.0 Axial N/A
16-slice CT 300 0.75 17.0 1660.5 Helical 0.69
Chest 256-slice CT 120 160 1.0 1.0 25660.5 320 93.9 Axial N/A
16-slice CT 300 0.5 8.3 1660.5 Helical 0.94
Abdomen 256-slice CT 120 213 1.0 1s62 25660.5 320 187.8 Axial N/A
16-slice CT 400 0.5 14.5 1660.5 Helical 0.94
Pelvis 256-slice CT 120 213 1.0 1s63 12861.0 320 281.7 Axial N/A
16-slice CT 400 0.5 11.4 1661.0 Helical 0.94

FOV, field of view.

Dose measurements
The dose for both CT systems was measured with the
300 mm long pencil-shaped ionization chamber and
300 mm long phantoms (160 mm and 320 mm diameter)
in one rotation scan. The measurement range in the
longitudinal direction was 300 mm (z5¡150 mm). The
phantom was placed on the patient table and its centre was
aligned at the isocentre. The ionization chamber was
inserted into either the central or one of the peripheral
cavities of the phantom (other cavities were filled with
PMMA rods). The exposure (expressed as Roentgens) was
obtained with the ionization chamber dosemeter and
converted to the values of absorbed dose to air measured
in PMMA with the f-factor 0.898 cGy R21.
Dose assessment
The dose was assessed using the dose profile integral
(DPI) over 300 mm (z5¡150 mm) (Appendix 2), which
was given by the output of the pencil ionization chamber
of 300 mm length [6].
The weighted average of DPI at the centre and
peripheries of the phantoms is given by
1 2
DPIw ~ DPIc z DPIp
3 3
if we assume a linear decrease (or increase) of DPI in the
radial direction, where DPIc is the DPI at the centre and
DPIp the average DPI on the peripheries.
was carried out as follows: (i) head, (ii) chest,
(iii) abdomen, and (iv) pelvis for one subject at each
anatomical site. The subjects held their breath at end-
inhale for the chest examination and end-exhale for the
abdomen and pelvis examinations during scanning. Scan
conditions were the same as the clinical conditions
(Table 1) except the scan ranges, which were 102.4 mm
for head (one scan), 375.6 mm for chest (four contiguous
scans), 93.9 mm for abdomen and pelvis (one scan). The
matrix size was 51265126111251265126205, and the
convolution kernel was the standard head kernel (FC43)
for the head examination and the standard body kernel
(FC10) for the others.
Image quality was evaluated by three board-certified
radiologists who had more than 10 years experience in
clinical diagnosis. They compared quality of the images
taken with the prototype scanner to their quality standard
formed by experience. It took about 1.5 h to read the
images obtained in multiple planes in all four cases.
Results
For both CTs, DPIc, DPIp, and DPIw in an axial scan
are summarized in Table 2. These values are
normalized to 100 mAs. For the 256-slice CT, DPIw is
ð1Þ 1966 mGy?mm/100 mAs for the head phantom and
1109 mGy?mm/100 mAs for the body phantom. For the
16-slice CT, DPIw is 181.6 mGy?mm/100 mAs with 8 mm
Table 3. Dose profile integral weighted average (DPIw) for
clinical protocols for 256-slice CT and 16-slice CT

Clinical image quality Examination DPIw (mGy?mm) DPIw percentage

(%)
We imaged four healthy male volunteers (mean age 30.0 256-slice CT 16-slice CT
years¡7.6 (standard deviation) (SD); age range 23–53 Head 6410 12127 52.9
years) using the 256-slice CT. The study was approved by Chest 1775 2462 72.1
the Institutional Review Board, and written informed Abdomen 4725 5773 81.9
consent was obtained from all subjects before starting. A Pelvis 7088 7981 88.8
non-enhanced examination with a step-and-shoot approval

Table 2. Dose profile integral (DPI) for the 256-slice CT and 16-slice CT

CT scanner Phantom Beam width (mm) DPIc (mGy mm/100 mAs) DPIp (mGy mm/100 mAs) DPIw (mGy mm/100 mAs)
256-slice CT Head 112 1829 2034 1966
Body 128 781 1273 1109
16-slice CT Head 8 174.2 185.3 181.6
Body 8 67.7 98.7 88.4
Body 16 117.6 175.0 155.9

58 The British Journal of Radiology, January 2006

Comparison of patient dose in multislice CT
beam width for the head phantom, 88.4 mGy?mm/
100 mAs with 8 mm beam width and 155.9 mGy?mm/
100 mAs with 16 mm beam width for the body phantom.
In Table 3 DPIw values are calculated for the clinical
protocols. Values for the 256-slice CT are smaller than
those for the 16-slice CT in all examinations. We note that
especially in the head examination, the DPIw for the
256-slice CT is approximately 47% smaller than that for
the 16-slice CT.
With regard to the clinical image quality, Figure 2
shows normal anatomical images from the 256-slice CT.
Auditory ossicles are observed clearly in the sagittal
section with the same image quality as the state-of-the-art
CT-scanner (Figure 2a). For the chest examination, 3D
visualization of the lung from four contiguous axial scans
is shown in Figure 2b. For the abdomen examination, the
coronal image has an image quality as good as that of
conventional CT (Figure 2c). For the pelvis examination,
three contiguous coronal images are shown in Figure 2d.
These images also show the same image quality as
conventional CT.
Discussion
In the present study, we compared doses in the 256-slice
CT and the 16-slice CT for clinical conditions. The results
showed that the dose for the 256-slice CT was smaller
than that of the 16-slice CT in all examinations (Table 3).
(a)
(b)
Figure 2. Clinical images. (a) The 0.5 mm isotropic normal anatomy images of auditory ossicles in sagittal section. (b) 3D visualiza-
tion of the chest with four contiguous scans. (c) Normal anatomy images of abdomen (0.63 mm reconstruction increment).
(d) Coronal image (0.63 mm reconstruction increment) of pelvis with three contiguous scans.
The British Journal of Radiology, January 2006
The percentages of DPIw for the 256-slice CT to that for
the 16-slice CT were 52.9%, 72.1%, 81.9% and 88.8% in the
examinations of head, chest, abdomen and pelvis,
respectively.
The dose for the 256-slice CT was less than that of the
16-slice CT in all examinations for the following reason. In
a MDCT-scanner the actual beam width is set as the
nominal beam width (slice thickness6slice number) plus a
certain margin, where the margin is added to cover
penumbra and mechanical errors. X-ray photons incident
on a marginal portion do not contribute to image
formation, but they do contribute to increased dose. If
the nominal beam width becomes large, the contribution
of this portion becomes smaller. Thus, the 256-slice CT
with larger beam width provides smaller DPIw values than
the 16-slice CT. For the 16-slice CT the pelvis examination
with 16 mm nominal beam width is more effective than the
others with 8 mm beam width. In general, helical scans
with pitch less than one caused overlap regions. Therefore
in the present study, we set the effective mAs value to be
the same to obtain the same signal-to-noise ratio in both
CT systems.
Notwithstanding the dose for the 256-slice CT being
smaller than that of the 16-slice CT, the 256-slice CT
provides sufficient image quality for diagnosis (Figure 2)
[7]. In these clinical conditions, the 256-slice CT achieved a
0.5–0.8 mm isotropic resolution and large volumes of data
were taken in a one-rotation scan [8]. Therefore coronal
(c)
(d)
59
 S Mori, M Endo, K Nishizawa et al

through during scanning by the tetra-angular pyramid

Figure A1. Reconstruction geometry of cone-beam CT. An
X-ray source and a 2D detector rotate around the z-axis. The
volume that can be reconstructed with the Feldkamp algorithm
is shown by the shaded region and is a double conical region
within a cylinder of radius Rmax, which is determined by the
detector size in the x-direction and shows the maximum field
of view in the transverse plane. R and H show diameter and
height, respectively, of a cylindrical reconstructed volume as it
varied with an object. N6T show the nominal beam width
where N is the number of slice and T is the slice collimation.
Table A1. Calculated weighted CT dose index (CTDIw), dose–
length product (DLP) and effective dose E for the 256-slice CT
whose apex and base are the X-ray source and the 2D
detector, respectively (Figure A1). The reconstructed
region is a double conical shape within a maximum
FOV (Rmax) in the transverse plane that is determined by
the detector size in the transverse direction.
Reconstruction is not made in the entire Rmax except at
the midplane and depends on a reconstructed FOV (R). In
the case of the 256 mm60.5 mm (5 N6T) beam
collimation, the length of the reconstruction region (H)
is 102.4 mm for R5 240 mm and 93.9 mm for R5
320 mm. As seen in this example, the reconstructed
region is generally smaller than the nominal beam width
in cone beam CT.
Appendix 2. Effective dose estimation
CT dose index (CTDI), dose–length product (DLP), and
effective dose (E) are usually used for CT dosimetry [2],
and they are derived from DPI described in the present
report.
CTDI is given as follows.
ð
1 l=2
CTDI~ d(z)dz ½mGy
(A1)
NT {l=2

DPIw CTDIw (mGy) DLP E (mSv) where N is the number of slices, T (mm) is the nominal
(mGy?mm) (mGy?cm) slice thickness, and d(z) is the dose profile for an axial
scan, l indicates the integration range. The International
Chest 1775 13.87 130.2 2.21 Electrotechnical Comission (IEC) recommended an inte-
Abdomen 2363 18.46 173.3 2.60
gration range of 100 mm. However we used the integration
Pelvis 2363 18.46 173.3 3.29
range of 300 mm for the reason described.
DPI is given with these notations as follows.
and sagittal images were obtained at sufficient spatial ð l=2
resolution without secondary reconstruction. DPI~ d(z)dz ½mGy mm
(A2)
{l=2
Regarding the diagnostic reference level, the effective
dose [9] for the MDCT was approximately 15 mSv for From Equations (A1) and (A2),
routine chest examinations and 30 mSv for routine abdo-
men or pelvis examinations [10]. If these values are taken 1
CTDI~ DPI (A3)
as upper limits and X-ray conditions are the same as those NT
in Table 1, the acceptable scan time in volumetric cine Weighted CTDI (CTDIw) is defined with CTDIs measured
imaging might be estimated in the following way. From at the centre and peripheries of the phantoms as follows.
Appendix 2, the estimated effective dose for a 1 s scan was
2.21 mSv, 2.60 mSv and 3.29 mSv for chest, abdomen and 1 2
CTDIw ~ CTDIC z CTDIP ½mGy
(A4)
pelvis, respectively. Therefore, the acceptable scan time 3 3
should be 6 s (5 15 [mSv]/2.21 [mSv]), 11 s (530 [mSv]/ CTDIc and CTDIp represent the CTDI measured at the
2.60 [mSv]) and 9 s (5 30 [mSv]/3.20 [mSv]) for chest, centre and the average CTDIs measured on the periphery
abdomen, and pelvis, respectively. As these scan times may
of the phantom, respectively. CTDIw is given by DPIw as
not be sufficient for a dynamic study in some cases, further
follows.
efforts are necessary to develop dose reduction methods
such as automatic dose control [11–13], as well as to justify 1
increasing the dose in dynamic studies consistent with risk- CTDIw ~ DPIw ½mGy
(A5)
NT
benefit. Resolution of these issues will allow full use of
Dose–length product (DLP) for a complete examination
volumetric cine images which will significantly increase the
is given as:
amount of diagnostic information available to radiologists.
In particular, we expect new applications such as DLP~CTDIw |L ½mGy cm
(A6)
computed tomographic angiography (CTA) of coronary
arteries or perfusion studies of the whole brain. where L (cm) is the scan range in the longitudinal
direction.
Estimation of effective dose (E) may be derived from
values of DLP for an examination using appropriately
Appendix 1. Field of view for the 256-slice CT
normalized coefficients:
In the 256-slice CT, the reconstructed images with
the Feldkamp algorithm is the region that is passed E~EDLP . DLP ½mSv
(A7)

60 The British Journal of Radiology, January 2006

Comparison of patient dose in multislice CT
EDLP is the region-specific normalized effective dose
(mSv mGy21 mm21) [9].
From these equations CTDIw, DLP and E can be
calculated from measured DPIw. Table A1 gives calculated
DPIw, CTDIw, DLP and E with one second scan of the
256-slice CT in the clinical conditions for chest, abdomen
and pelvis examinations, respectively.
References
1. Endo M, Mori S, Tsunoo T, Kandatsu S, Tanada S, Aradate H,
et al. Development and performance evaluation of the first model
of 256-slice CT-scanner. IEEE Trans Nucl Sci 2003;50:1667–71.
2. International Commission on Radiological Protection.
Managing patient dose in computed tomography. ICRP
Publication 87. Annals of the ICRP 30 (4), Pergamon Press,
Oxford, 2000.
3. Saito Y, Aradate H, Igarashi K, Ide H. ‘‘Large area 2-
dimensional detector for real-time 3-dimensional CT (256-
slice CT),’’ Proc SPIE 2001;4320:775–82.
4. Feldkamp LA, Davis LC, Kress JW. Practical cone-beam
algorithm. J Opt Soc Am 1984;A 1:612–9.
5. International Electrotechnical Commission. Evaluation and
routine testing in medical imaging departments constancy
tests – X-ray equipment for computed tomography.
Publication IEC 1994:1223-2-6.
6. Mori S, Endo M, Nishizawa K, Tsunoo T, Aoyama T, Fujiwara
H, et al. Enlarged longitudinal dose profiles in cone-beam CT
and the need for modified dosimetry. Med Phys 2005;32:1061–9.
The British Journal of Radiology, January 2006
7. Mori S, Endo M, Obata T, Murase K, Fujiwara H, Kandatsu
S, et al. Clinical potentials of the prototype 256-detector roe
CT-scanner. Acad Radiol 2005;22:149–55.
8. Mori S, Endo M, Tsunoo T, Kandatsu S, Tanada S, Aradate
H, et al. Physical performance evaluation of a 256-slice CT-
scanner for 4-dimensional imaging. Med Phys 2004;31:1348–
56.
9. European Guidelines. Quality Criteria for Computed
Tomography. EUR 16262, CEC Luxembourg, 1997.
10. Aoki C, Nishizawa K, Tonari A, Hachiya J. Effective dosing
for multi-detector CT scanning. Jpn J Med Imaging
2001;20:101–9.
11. Nagel HS, Galanski M, Hidajat N, Maier W, Schmidt T.
Radiation exposure in computed tomography-fundamentals,
influencing parameters, dose assessment, optimization,
scanner data, terminology. Hamburg: CTB Publications,
2001.
12. Thomas LT, Neil BB, Tin-Su P, Jerry R, Steven JW, Jianying Li,
et al. A dose reduction X-ray beam positioning system for
high-speed multislice CT scanners. Med Phys 2000;27:
2659–68.
13. Kachelrieß M, Kalender WA. Dose reduction by generalized
3D adaptive filtering for conventional and spiral single-,
multirow and conebeam CT. Radiology 1999;213:283–4.
61
The British Journal of Radiology, 79 (2006), 62–70 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/50019934

Assessment of tube current modulation in pelvic CT

G R IBALL, MSc, DipIPEM, D S BRETTLE, PhD and A C MOORE, MSc, DipIPEM
Department of Medical Physics & Engineering, Leeds Teaching Hospitals NHS Trust, Leeds General Infirmary, Leeds
LS1 3EX, UK

Abstract. An anatomically shaped polymethylmethacrylate (PMMA) phantom was used to assess the effect of
the Siemens CARE Dose mA modulation system on pelvic CT scans. The effect of the system on absorbed dose
to air, image percentage noise and the signal to noise ratio of clinically relevant details was assessed. The signal
to noise ratio was calculated using Polytetrafluoroethylene (PTFE) and distilled water inserts; PTFE was used
to represent bony structure and distilled water was used to represent soft tissue abscess. Pelvis protocols
identified from local hospitals and the UK CT Dose Survey (2002), were assessed and compared with those
provided by Siemens Medical (UK). These protocols were tested on a Siemens Sensation 4 CT scanner, both
with and without CARE Dose. Results were obtained which showed that dose savings were possible with no
significant increase in image noise. Dose reductions were 8% in the lateral positions in the phantom and 42% in
the centre, top and bottom. The calculated ‘‘CTDIvol’’ was 32% lower with CARE Dose than without CARE
Dose. This is slightly greater than the 25% change in the effective mAs values that was found. This implies that
the reduction in the effective mAs values is a reasonable predictor of the total reduction in absorbed dose to air,
whilst slightly underestimating the actual change. The results also showed a non-significant trend towards
decreased signal to noise ratios for clinically relevant CT numbers when CARE Dose was activated. This
suggests that tube current modulation may detrimentally affect signal detection due to changes in image noise.

CT examinations account for a large proportion of the
collective dose from medical X-ray examinations in the
UK. In 2000 this was reported as being 40% [1] but may
now be even higher due to an increase in the range and
volume of routine examinations and the uptake of CT
fluoroscopy and cardiac CT scanning. There is a require-
ment for all X-ray examinations to be optimized such that
the patient dose is ‘‘As Low As Reasonably Practicable’’
(ALARP) [2]. However, it is often difficult to implement
procedures which significantly lower the radiation dose
without decreasing the image quality to a non-diagnostic
level. One recent technological advance from CT manu-
facturers in terms of dose reduction has been to introduce
tube current modulation systems for CT scanning. The
approach taken by Siemens Medical Systems (Erlangen,
Germany) is a system called CARE Dose which claims to
reduce patient doses whilst having no significant adverse
effects on the image quality. This system has been
described extensively in the literature [3–5].
The human body varies in composition both along its
length and in the transverse plane at any given point along
the body. This produces variations in X-ray attenuation
due to both the external dimensions of the body and its
internal composition. In CT scanning, as the X-ray tube
and detectors rotate around the body, the attenuation can
change by two orders of magnitude [4]. These differences
in attenuation are most significant in the regions of the
shoulder and pelvis, where large thicknesses of bone are
found in the lateral projections, but a much smaller
thickness of bone is present in the anterior–posterior
projections. It is these examinations which provide the
greatest challenges, in terms of the dose–image quality
balance. As a result, using a constant tube current (mA)
for each scan angle within a given rotation may result in
Received 18 October 2005 and in final form 24 May 2005, accepted 31
May 2005.
either photon starvation artefacts on the high attenuation
projections or overdosing in the lower attenuation
projections.
In the CARE Dose system, during each rotation of the
tube and detector assembly around the patient, a small
number of the central detector channels provide attenua-
tion information, which is dependent upon the patient
cross section and scan angle, to the X-ray generating
system [3]. The information provided by these detector
channels is used to determine to what extent the mA can be
modulated, with respect to an initial tube current setting,
without adversely affecting the image quality. As a result
the tube current is modulated dynamically with a delay of
one rotation relative to the attenuation measurement.
The first patient based assessment by Greess et al [6]
showed that, when CARE Dose is used, a dose reduction
of approximately 25% (in terms of total mAs reduction) is
possible in pelvic scanning ‘‘with no significant decrease’’
in subjective assessments of image quality. Similar
percentage dose reductions have been demonstrated in
other clinical work [7] and these showed good agreement
with phantom based data [2, 5]. Most of the published
work has used image noise and/or subjective image
assessment to quantify image quality. A small number
of papers [8, 9] have used standard deviations from regions
of interest (ROIs) to yield a more objective assessment of
image noise.
Claims that the image quality was not affected by the
CARE Dose system were queried by local users. Having
used CARE Dose for a period of time, they perceived that
the quality of the images for pelvis scans was subjectively
worse when CARE Dose was used and this raised
concerns that it may have a detrimental effect on the
accuracy of diagnosis. This is despite the manufacturer’s
recommendation that CARE Dose is used for all
clinical situations other than for extremely large patients.
This discrepancy between the reported claims and

62 The British Journal of Radiology, January 2006

Assessment of tube current modulation
local experience prompted this investigation into the
relationships between patient dose, image percentage
noise and the signal to noise ratio (SNR) as an indicator
of diagnostic detectability. The objective was to clarify
whether the CARE Dose system can yield significant dose
reduction for no loss of image quality in pelvic scanning.
Materials and methods
A series of measurements were made using an
anatomically shaped polymethylmethacrylate (PMMA)
phantom which has been described in the literature [3]
as a ‘‘hip’’ phantom. A schematic diagram of the phantom
is shown in Figure 1. The thickness of the phantom is
14.5 cm in the z-axis. The hip phantom does not contain
any bony structure and therefore the similarity of the
phantom to the pelvic region is geometric only. As such it
may be expected that the magnitude of the tube current
modulation in clinical practice may be different from that
found for this phantom.
Protocol selection
Routine protocols for soft tissue assessment of the pelvis
on Siemens 4 slice CT scanners (Somatom Volume Zoom
and Somatom Sensation 4) were obtained from three local
hospitals, the UK CT Dose survey 2002 and from Siemens
Medical (UK) (Bracknell, UK). These protocols fell into
two main groups, those that used a pitch of 1.00 and those
that used a pitch of 1.25. There was some variation in the
mA/mAs setting that was used, but all of the protocols
used 120 kV and a rotation time of 0.5 s. In light of these
findings all measurements were performed at the standard
exposure factors given in Table 1. Reconstruction kernel
B40s was used.
Effective mAs is defined as the tube mAs per rotation
divided by the helical pitch, where the pitch is the ratio of
Figure 1. A schematic diagram of the hip phantom. Solid
cylinders within the phantom (1) represent the five ion chamber
positions, dashed circles within the phantom (2) represent the
five polymethylmethacrylate (PMMA) CT number and noise
measurement positions; the dashed circle outside the phantom
(3) represents the air CT number measurement position.
Table 1. Standard exposure factors for all scans
Tube voltage (kV) Effective mAs Rotation time (s) Beam collimation (mm)
120 165 0.5 462.5
SFOV, field of view.
The British Journal of Radiology, January 2006
the table feed per rotation and the total X-ray beam width
[10]. The effective mAs value of 165 was chosen as this was
representative of most of the protocols that were obtained.
Pitch settings of 1.00, 1.13 and 1.25 were used, both with
and without CARE Dose. All the measurements were
performed using the same PMMA hip phantom on the
same scanner, a Siemens Sensation 4.
Measurements were made which investigated how four
different parameters changed with the application of
CARE Dose. The four parameters that were investigated
were (i) absorbed dose to air, measured in the phantom,
(ii) image percentage noise, (iii) CT number for water and
polytetrafluoroethylene (PTFE), and (iv) the SNR for both
of these materials. These parameters were investigated for
each pitch setting, both with and without CARE Dose.
Water was chosen to represent low density abscess and
PTFE to represent bony structure.
Consistency tests
Prior to the testing, all of the test equipment was placed
in the scanning room for at least 4 h in order for the
temperature of the phantom and test equipment to
stabilize with the room temperature. At the start of each
visit the scanner was air-calibrated using the software on
the scanner. A short series of consistency tests were
performed immediately after the air-calibration which, on
all subsequent visits, enabled us to verify that the
performance of the scanner had not changed from the
previous visit.
On each occasion the hip phantom was positioned
15 cm from the end of the couch on top of the mattress in
order to maintain consistent scattering conditions. The
phantom was aligned using the laser lights on the scanner
and with a spirit level. The set up is shown in Figure 2.
A Scan Projection Radiograph (Topogram) of the
phantom was acquired and a helical acquisition was
planned from this image. The scan length for the helical
acquisition was the whole length of the phantom
Figure 2. The phantom as positioned for the dose and noise
measurements.
Image slice thickness (mm) SFOV (mm)
5 380
63

(145 mm), which gave a total scan time of 8.2 s, using a
pitch of 1.25. All of the consistency tests were performed
with CARE Dose on.
The first test was a measurement of the absorbed dose to
air. A scan was performed at the standard exposure factors
at a pitch of 1.25 with a calibrated 3 cm3 pencil ionization
chamber (Capintec Inc., Ramsey, NJ), having an active
length of 100 mm, in the central position. The chamber
was connected to a Keithley 35050A Dosimeter (Keithley
Instruments Inc., Cleveland, OH). The absorbed dose to
air was recorded and the mean PMMA CT number and
standard deviation (p) were measured adjacent to each of
the five possible chamber positions (see Figure 1), on the
CT slice closest to the centre of the phantom, using the
region of interest (ROI) tool on the scanner. The size of
the ROI that was used was kept constant throughout all of
the measurements. The mean CT number of air was also
measured at a standard position outside the phantom
using a ROI of the same size.
This scan and measurement procedure was then
repeated with the ion chamber in the right lateral
measurement position. On each occasion the ambient air
temperature and pressure were measured, in addition to
the phantom temperature, so that an air density correction
could be applied to the dose measurements.
The ion chamber was then removed from the phantom
and a PTFE rod was inserted into the central measurement
position. The scan was repeated and the mean CT number
and p of the PTFE rod were recorded in addition to the
measurements described above. Again this was repeated
with the PTFE rod in the right lateral position.
Absorbed dose to air and noise measurements
Absorbed dose to air measurements were made for each
of the five chamber positions both with and without
CARE Dose at each of the three pitch settings. For each
scan the measured dose and total mAs were recorded. On
the central slice the mean PMMA CT number and p were
recorded at each measurement position and the mean air
CT number was also recorded. For each measurement
position the image percentage noise was calculated using
Equation (1) [11]:
pPMMA
100
Image Percentage Noise~ ð1Þ
CTPMMA {CTAir
where: p is the standard deviation and CT is the mean CT
number (Hounsfield Unit) of the indicated material.
The absorbed dose to air was corrected for ambient
temperature and pressure and the ion chamber calibration
factor was applied. The volume averaged CT dose index
(CTDIvol) was then calculated for the scans with and
without CARE Dose, using Equation (2) [11]. This was
performed for each pitch setting:


CTDIvol ~ 1=3
CTDIcentre z2=3
CTDIperiphery =pitch ð2Þ
where: CTDIcentre is the CTDI measured in the centre of
the phantom and CTDIperiphery is the average of the four
CTDI values which were measured in the periphery of the
phantom. Pitch is the ratio of the table feed per rotation
and the total X-ray beam width.
CTDIvol is actually defined for a cylindrical phantom
and as such it is not strictly applicable to the hip phantom
64
G R Iball, D S Brettle and A C Moore
that was used in this study. However, the CTDIvol method
is an accepted way of accounting for the distribution of
dose within a phantom. Since, in this case, it is the
comparison between the CTDIvol values for two different
scanning situations, rather than the absolute value that
was of most importance, the CTDIvol was used simply as
an indicator of the relative change in absorbed dose to air.
As such, the term ‘‘CTDIvol’’ is used for all calculations
that relate to the hip phantom.
The effect of the CARE Dose system on the percentage
dose reduction was also evaluated over a range of initial
effective mAs settings (50–200 mAs).
Signal measurements
Two sets of signal to noise measurements were made,
for the water and PTFE inserts.
For the water measurements thin rubber sheaths were
inserted into each of the five holes in the phantom and
distilled water was inserted into each of the sheaths and
the ends were secured with plastic clips. The sheaths were
similar in diameter to the holes in the phantom which
made it possible to almost completely fill the holes with
water. The set up of the phantom for the water
measurements is shown in Figure 3.
For the PTFE measurements each individual rod was
manufactured in house from a single PTFE rod
(Barkston Plastics Ltd, Leeds, UK). All five rods were
manufactured from the same batch of PTFE to ensure that
there was no difference in composition between the
individual rods.
For each set of measurements the phantom was scanned
five times at each pitch setting with and without CARE
Dose. For each scan the mean CT number and p of the
water/PTFE and PMMA were recorded at each measure-
ment position on the central slice in addition to the mean
CT number of air at the standard position.
Measurements were also repeated 10 times on one scan
of the PTFE rods in order to establish the repeatability of
the measurements.
Figure 3. The phantom as set up for the water signal to noise
ratio (SNR) measurements.
The British Journal of Radiology, January 2006
Assessment of tube current modulation

The SNR for the inserts was calculated using Dose measurements
Equation (3).
  For the scans without CARE Dose (i.e. constant mA)
CTsignal {CTPMMA  the absorbed doses to air were significantly higher in the
SNR~ 
 ð3Þ
pPMMA
Where: CTsignal is the mean CT number of water or PTFE
and p is the standard deviation.
The modulus was used as the mean CT number for
water was sometimes below zero.
SNR calculations were performed for each measurement
point for each pitch setting. Error propagation was
performed for all of the parameters of interest and the
calculated values are shown with the results.
The pooled standard deviation of the SNRs was
calculated for each pitch setting for the water measure-
ments and this result was used to power the study. The
powering process showed that for a result to be
statistically significant at the 95% level 25 measurements
were required (both with and without CARE Dose).
As a result, a further set of SNR measurements were
made for both water and PTFE. The phantom was set
up as described earlier and 25 scans were performed
both with and without CARE Dose. For each scan the
mean CT number and p of the insert (PTFE/water) was
measured in the central position in addition to the mean
CT number and p of the PMMA adjacent to the
central insert. SNRs were calculated from these measure-
ments and errors were calculated as for the previous
measurements.
Statistical analysis was performed on these results
(Kruskal–Wallis non parametric test) to determine
whether the SNRs of water and PTFE changed signifi-
cantly for the scans with CARE Dose.
Results
The results of the consistency tests that were performed
showed that on each occasion the performance of the
scanner had not changed since the first visit.
For clarity all the results for the 1.25 pitch setting are
shown with summary results for the other pitches.
The British Journal of Radiology, January 2006
top and bottom positions than in the lateral positions. For
the scans with CARE Dose there was a significant decrease
in the absorbed dose to air in each position. The
reductions were approximately 42% in the central position,
42% in the top and bottom positions and 8% in the lateral
positions. These results were as expected and are shown in
Figure 4.
The calculated value of ‘‘CTDIvol’’ with CARE Dose
was 32% lower than the value for the scans without CARE
Dose. The scanner indicated reduction in effective mAs for
the scans with CARE Dose was 25% (relative to the
constant tube current case).
The percentage reduction in ‘‘CTDIvol’’ was indepen-
dent of pitch to within 0.5% over the pitch range of 1–1.25,
as shown in Figure 5. The error bars that are shown in
Figure 5 represent one standard deviation about the mean.
The percentage reduction in absorbed dose to air in the
central position for varying initial effective mAs settings is
shown in Figure 6. The reduction in absorbed dose to air
is approximately 40% for mAs settings between 50 mAs
and 165 mAs. However, this reduction in absorbed dose to
Figure 5. Percentage reduction in ‘‘CTDIvol’’ against pitch
setting.
Figure 4. Variation of absorbed dose
to air with position in phantom.
65

Figure 6. Reduction in absorbed dose to air in the central
position against initial effective mAs setting.
air rises to 50% at 200 mAs. It was not possible to obtain
results for scans with mAs settings above 200 mAs as this
would have exceeded the maximum tube loading at this
pitch setting.
66
G R Iball, D S Brettle and A C Moore
Noise measurements
The initial set of image percentage noise values,
calculated using Equation (1), across the five positions
showed that, in general, there was not a large difference
between the measured values with and without CARE
Dose, other than for the top and bottom positions, as
shown in Figure 7. These discrepancies between the noise
values in the top and bottom positions were not found for
the other pitch settings and are thought to be anomalous
results, relative to the other pitch settings. As expected the
noise values in the top and bottom were slightly different
from those found in the lateral positions.
For all pitches the highest noise values were found in the
centre of the phantom.
There was a general reduction in the image percentage
noise as the pitch setting was increased (Figure 8), for both
CARE Dose on and off, although this was not greater
than the experimental uncertainties.
For the 25 additional scans the image percentage noise
was assessed in the centre of the phantom as for the initial
tests. The difference between the noise values for CARE
Dose on and off, which was approximately 10%, was
tested for significance using the Kruskal–Wallis test. The
mean and standard deviations of the noise values for the
Figure 7. Variation of image percen-
tage noise with position in the
phantom.
Figure 8. Image percentage noise var-
iation with pitch setting.
The British Journal of Radiology, January 2006
Assessment of tube current modulation

25 additional scans are shown in Table 2 along with the of the phantom both with and without CARE Dose. This
calculated p-values. is as expected as the noise values were highest in the centre
These results show that there was no significant of the phantom. The differences seen between the PTFE
difference, at the 95% level, in the noise levels for the SNR values with and without CARE Dose, were generally
scans with and without CARE Dose. within the experimental uncertainties. There appears to be
a general decrease in the water SNR at each position for
the scans with CARE Dose, which is an undesirable trend.
However, the differences in water SNR were also within
SNR measurements the experimental uncertainties.
The measurements for the water and PTFE inserts were Figures 9 and 10 show that there are positional
used to calculate the SNR for each material (Equation (3)) variations in the SNR within the phantom. As such it is
and these results, for a pitch of 1.25, are shown in Figures not valid to average the SNR for the five different
9 and 10. These results from the initial tests show that the positions as this will mask the positional variations and
SNRs, for both water and PTFE, are lowest in the centre will result in large uncertainties in the results.
For a pitch of 1.25, the water SNR values measured
with CARE Dose were lower than those without CARE
Table 2. Calculated mean, standard deviation and p-values for Dose. This trend was observed for the other pitch values
the image percentage noise tests for the water scans but was not observed for the PTFE
scans.
Mean noise (SD) for Mean noise (SD) for p-value For the 25 additional scans the signal and noise values
CARE Dose off CARE Dose on
were measured for PTFE and water in the central position
PTFE 19.4 (3.7) 19.0 (3.6) 0.727 in the phantom. From these results the SNR for both
Water 17.3 (3.1) 19.1 (3.8) 0.099 inserts were calculated as for the original scans. The
differences between the values of CT number and SNR for
SD, standard deviation; PTFE, polytetrafluoroethylene. the scans with CARE Dose on and off were tested for

Figure 9. Signal to noise ratios

(SNR) for polytetrafluoroethylene
(PTFE) for each phantom position.

Figure 10. Signal to noise ratios

(SNR) for water for each phantom
position.

The British Journal of Radiology, January 2006 67


significance using the Kruskal–Wallis test. The mean and
standard deviations of the CT numbers and SNRs and the
resulting p-values are shown on Table 3. These results
show that there were no significant differences, at the 95%
level, in CT number or SNR between the scans with and
without CARE Dose despite the SNRs generally being
decreased when CARE Dose was used.
Discussion
Significant reductions in absorbed dose to air were
found in all five positions in the phantom for the scans
with CARE Dose relative to the constant tube current
situation (Figure 4). The largest reductions, up to 42%,
were found in the top, bottom and central positions as
these positions lie on the lowest attenuation paths through
the phantom and therefore experience the largest tube
current modulation and reduction in absorbed dose to air.
The dose reductions in the lateral positions, around 8%,
are much smaller in magnitude as the attenuation is at its
highest in these positions which means the tube current
will be at its maximum value. These dose reductions were
smaller than those shown by Kalender et al [5] who found
a 45% ‘‘average’’ dose reduction via direct dose measure-
ment. Kalender used a scan time of 1 s (compared with a
0.5 s scan time in our work) which allowed for a larger
modulation amplitude and therefore a greater dose
reduction than in this study. We also found that the
reduction in the effective mAs values were lower than
those found by Kalender et al, at approximately 25%
compared with 40%. However, Kalender’s work showed
an associated increase in image noise of approximately
10%, which was not found in the first part of this study.
Kalender’s work was performed with a prototype version
of the CARE Dose system which may also explain some of
the differences between those initial results and the results
of this study. Kalender et al [5] measured an average dose
reduction of 45% in the hip phantom with a 3 cm3
ionization chamber similar to that used in this study. This
45% reduction in dose, however, was a straightforward
average of the five measurement points rather than a
volume average (‘‘CTDIvol’’) which was calculated here. A
straightforward average of our results yields a dose
reduction of 30%.
The reduction in ‘‘CTDIvol’’ of approximately 32% was
in good agreement with the relative dose reduction found
by Gies et al [4], who found dose reductions of
approximately 38%, for computer simulations using the
hip phantom. The large reduction in absorbed dose to air
in the central position is of importance as most of the
more radiosensitive organs lie centrally. These results
imply that the reduction in an individual organ dose (with
an associated change in the effective dose) may be larger
than the reduction in the values of ‘‘CTDIvol’’ shown
Table 3. Calculated mean, standard deviation and p-values for CT number and SNR for water and PTFE
Mean CT number (SD) Mean CT number (SD)
for CARE Dose off for CARE Dose on
PTFE 962.4 (4.6) 961.2 (4.5)
Water 7.0 (3.6) 6.5 (3.3)
SD, standard deviation, SNR, signal to noise ratio, PTFE, polytetrafluoroethylene.
68
G R Iball, D S Brettle and A C Moore
here. These results have implications for calculating
effective doses in CT as the current Monte Carlo data
sets that are used do not reflect the distribution of dose
within the patient when a tube current modulation system
is used. The large dose reduction in the centre of the
phantom also has significant implications for pelvic scans
of pregnant patients. If CARE Dose was used for these
patients the risk to the fetus may be significantly reduced
relative to scans performed with a constant tube current.
Tack et al [10] showed that when using CARE Dose, the
percentage dose reduction was independent of the initial
effective mAs setting. They used six different mAs settings
between 20 mAs and 100 mAs for chest and abdomen CT
scans. Our results (Figure 6) show that the percentage dose
reduction is approximately constant at a value of around
40% for initial effective mAs values up to 165 mAs. Above
this value the percentage dose reduction increases, to
approximately 50% at 200 mAs. This occurred as the mAs
setting approached the maximum tube current rating for
the tube. The Manufacturers recommend that for extre-
mely large patients, where the mAs setting may be close to
the tube limit, CARE Dose is not used. No measurements
were made to determine whether or not the tube output
varied linearly with mAs so we cannot exclude poor
output linearity with mAs as a possible cause of the results
shown in Figure 6.
The image percentage noise level was not significantly
affected by the application of CARE Dose, as shown in
Figure 7, for the initial set of noise measurements. The
reduction in dose of approximately 8% in the left and right
positions occurs as a result of the integration of the
reduction in tube current over all scan angles as there is no
reduction in the tube current setting in the lateral
projections. Given that there has been a general reduction
in dose across the phantom there should have been an
associated increase in the image percentage noise. No such
increase in image percentage noise was found. Combining
these results and those for the dose measurements shows
that the reductions in absorbed dose to air that were
calculated are net dose savings, i.e. they come with no
significant noise penalty. Previous work [3, 5–7] showed
that dose reductions of 23–45% were possible in the pelvis
region with no significant difference in subjective assess-
ments of image quality.
The slight decrease in the image percentage noise with
pitch setting, for both CARE Dose on and off is thought
to be due to the combined effect of setting a constant
effective mAs value and the magnitude of the over-scan
which is necessary in helical scanning.
For the additional scans with the water inserts there was
an increase in the noise level of approximately 10% for the
scans with CARE Dose on relative to the scans with
CARE Dose off. This was not found to be significant at
the 95% level (p50.099). This 10% increase in noise agrees
Mean SNR (SD) for Mean SNR (SD) for p-value p-value
CARE Dose off CARE Dose on (CT number) (SNR)
44.3 (8.8) 44.9 (8.8) 0.393 0.764
7.6 (1.7) 7.0 (1.6) 0.421 0.197
The British Journal of Radiology, January 2006
Assessment of tube current modulation
well Kalender’s work [5]. A similar change in noise was
not found for the scans of the PTFE inserts (p50.727). As
the PTFE provides much greater X-ray attenuation than
water there is less scope for modulation of the tube current
when the PTFE inserts are scanned. As a result the slightly
larger reduction in the reported tube current that was
found when the water inserts were scanned results in a
larger percentage change in noise relative to the scans with
CARE Dose off.
Figures 9 and 10 show that there were differences
between the SNRs calculated for the scans with CARE
Dose on and off. These figures also show that the SNR
varied with position within the phantom. The highest
values of image percentage noise and the lowest values of
SNR were found in the central position which is as
expected from photon path length and reconstruction
theories. For PTFE the SNRs for the scans with CARE
Dose on showed no distinct trend relative to the SNRs for
the scans without CARE Dose. This is in contrast to the
situation for water where the SNRs for the scans with
CARE Dose on were lower than those for the scans with
CARE Dose off for 80% of the total number of scans. This
shows that there is a trend towards decreased SNR for
water when CARE Dose is used.
The larger set of SNR measurements showed a
difference in the SNRs of approximately 10% for water
whilst there was no difference for the PTFE measure-
ments. This is attributable to the similar percentage change
in the noise which was found (Table 2). Statistical analysis
showed that there was no statistically significant difference
in the SNRs for PTFE and water between the situations
with and without CARE Dose (p50.197 for water,
p50.764 for PTFE). Table 3 shows that, at the 95%
level, there was also no significant change in the CT
numbers for water and PTFE for the scans with and
without CARE Dose. Since the SNR depends on both the
signal and noise, neither of which showed a significant
change at the 95% level, there was no associated significant
change in the calculated SNRs for both water and PTFE.
This does not provide an explanation for the users’
subjective opinions that the images acquired with CARE
Dose, for imaging pelvic abscess, were unsatisfactory.
When the SNR values for water are error corrected
(mean value minus uncertainty), the average SNR for the
scans with CARE Dose is only just above the detectability
threshold of 5 as defined by Rose [12]. Water has an
inherently low SNR relative to the PMMA background,
but this is further reduced by 10% when CARE Dose is
activated. The worse case SNR (i.e. the lowest value of
SNR taking into account the calculated uncertainties) was
below the threshold value of 5 for 25% of the measure-
ments with CARE Dose off and for 40% of the
measurements with CARE Dose on. Although these
differences may not be statistically significant they may
be detectable by the person viewing the image and are
therefore important differences.
The X-ray attenuation path in clinical scanning is non-
homogeneous and the human pelvis may have an even
more asymmetric attenuation pattern than this phantom.
This may introduce a larger modulation in the tube
current which would affect the noise and serve to further
worsen the SNR situation. This may therefore reduce the
confidence with which the viewer of the image can detect
tissues which have subtle differences in SNRs. This
The British Journal of Radiology, January 2006
combination of the decrease in the water SNR and the
non-homogeneous attenuation path may therefore explain
why subjectively the images that were acquired with
CARE Dose had been reported as unsatisfactory for pelvic
abscess imaging.
There were large uncertainties in the results of this
study. However, the reductions in the SNRs that were
found were repeatable over a large number of scans and
are therefore considered to be a true representation of the
performance of the system. The main explanation for the
large uncertainties was that the ROIs that were used for
the water and PTFE measurements were small – these
were limited by the size of the inserts which were, in turn,
limited by the construction of the phantom. If measure-
ments were made too close to the edge of the insert then
the mean CT number would have been skewed by the
presence of any air around the insert or by the background
material itself. It was not possible to make any changes to
the phantom design. If it had been possible to use larger
inserts (and therefore larger ROIs) it may have been
possible to obtain results which were less error dominated.
We would recommend that any future studies should
consider using larger inserts and ROIs to improve the
noise statistics and to ensure homogeneity in the
measurements taken within the signal areas. However, it
should be noted that at 12 mm in diameter the size of the
water inserts were representative of abscesses which are
found in the pelvis.
Some differences were found between the results for the
left and right lateral positions in the phantom, in terms of
absorbed dose to air, noise and SNR for both inserts.
Further tests showed that the central alignment laser was
inaccurate by approximately 3 mm which resulted in a
relative difference between the left and right
measurement positions of around 6 mm and that the
differences were not due to the performance of the CARE
Dose system.
Conclusions
The CARE Dose system on Siemens 4 slice CT scanners
results in significant dose savings for scans of the pelvic
region. This yielded a reduction of approximately 32% in
the value of ‘‘CTDIvol’’ which agreed well with the 25%
reduction in the displayed effective mAs. This implies that
the reduction in the effective mAs value can be used as an
approximate indicator of the true dose reduction. This
reduction is a real, net dose saving as there was no
statistically significant increase in the noise.
There appears to be a trend towards decreased SNRs
for both water and PTFE when CARE Dose was used
although no significant differences were found at the 95%
level. These changes in SNR were mostly due to changes
in the image percentage noise values. The largest
decreases in SNR were found for water and were as
large as 14%. Since the water inserts were representative of
low-density abscess this suggests that the use of CARE
Dose may decrease the visibility of low-density
structures relative to the background. Therefore using
CARE Dose in situations where subtle differences in low
CT number tissue pathology are of interest may not be
advisable.
69

Acknowledgments
The authors wish to thank the following for their
invaluable assistance in the work: Leeds Nuffield Hospital,
especially Joanna Hartley for use of their scanner and for
involvement in the measurement procedures; ImPACT, St
George’s Hospital, London for loan of the phantom and
general advice; Harrogate District Hospital, York District
Hospital, UK CT Dose survey (2002) and Dr Paul
Shrimpton of the National Radiological Protection
Board (NRPB) for provision of protocol data; Siemens
Medical (UK), Bracknell for protocol data.
References
1. Hart D, Wall BF. Radiation exposure of the UK population
from medical and dental x-ray examinations (NRPB-W14).
Chilton: National Radiological Protection Board, 2002.
2. Annals of the ICRP 26 (2). Radiological protection and safety
in medicine. ICRP Publication 73, ICRP, 1996.
3. Kalender WA, Wolf H, Seuss C, Gies M, Greess H, Bautz
WA. Dose reduction in CT by online tube current control:
principles and validation on phantoms and cadavers. Eur
Radiol 1999;9:323–328.
4. Gies M, Kalender WA, Wolf H, Suess C. Dose reduction in
CT by anatomically adapted tube current modulation. I.
Simulation studies. Med Phys 1999;26:2235–47.
70
G R Iball, D S Brettle and A C Moore
5. Kalender WA, Wolf H, Seuss C. Dose reduction in CT by
anatomically adapted tube current modulation. II. Phantom
measurements. Med Phys 1999;26:2248–53.
6. Greess H, Wolf H, Baum U, Lell M, Pirkl M, Kalender WA,
et al. Dose reduction in computed tomography by attenuation
based online modulation of tube current: evaluation of six
anatomical regions. Eur Radiol 2000;10:391–4.
7. Greess H, Nomayr A, Wolf H, Baum U, Lell M, Bowing B,
et al. Dose reduction in CT examination of children by an
attenuation based online modulation of tube current (CARE
Dose). Eur Radiol 2002;12:1571–6.
8. Mastora I, Remy-Jardin M, Seuss C, Scherf C, Guillot JP,
Remy J. Dose reduction in spiral CT angiography of thoracic
outlet syndrome by anatomically adapted tube current
modulation. Eur Radiol 2001;11:590–6.
9. Jakobs TF, Becker CR, Ohnesorge B, Flohr T, Seuss C,
Schoepf UJ, et al. Multislice helical CT of the heart with
retrospective ECG gating: reduction of radiation exposure by
ECG-controlled tube current modulation. Eur Radiol
2002;12:1081–6.
10. Tack D, De Maertelaer V, Gevenois PA. Dose reduction in
multidetector CT using attenuation based online tube current
modulation. AJR Am J Roentgenol 2003;181:331–4.
11. Institute of Physics and Engineering in Medicine Report 32,
Measurement of the Performance Characteristics of
Diagnostic X-ray Systems used in Medicine Part III
Computed Tomography X-ray Scanners, 2nd Edition,
IPEM, 2003.
12. Rose A. Vision: human and electronic. New York: Plenum,
1974:21–3.
The British Journal of Radiology, January 2006
The British Journal of Radiology, 79 (2006), 71–75 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/50464795

Short communication
Radiosurgical palliation of aggressive murine SCCVII
squamous cell carcinomas using synchrotron-generated X-ray
microbeams
1
M MIURA, PhD, 2H BLATTMANN, PhD, 3E BRÄUER-KRISCH, BEng, 3A BRAVIN, PhD,
1
A L HANSON, PhD, 1M M NAWROCKY, BA, 1P L MICCA, BS, 1,4D N SLATKIN, MD and
4
J A LAISSUE, MD
1
Medical Department, Brookhaven National Laboratory, Upton, NY 11973-5000, USA, 2Niederwiesstrasse 13C,
Untersiggenthal, Switzerland, 3European Synchrotron Radiation Facility, 6 Rue Jules Horowitz, BP 220, Grenoble,
France and 4Pathologisches Institut der Universität Bern, Murtenstrasse 31, Bern, Switzerland

Abstract. Microbeam radiosurgery (MBRS), also referred to as microbeam radiation therapy (MRT), was tested
at the European Synchrotron Radiation Facility (ESRF). The left tibiofibular thigh of a mouse bearing a
subcutaneously (sc) implanted mouse model (SCCVII) of aggressive human squamous-cell carcinoma was
irradiated in two orthogonal exposures with or without a 16 mm aluminium filter through a multislit collimator
(MSC) by arrays of nearly parallel microbeams spaced 200 mm on centre (oc). The peak skin-entrance dose
from each exposure was 442 Gy, 625 Gy, or 884 Gy from 35 mm wide beams or 442 Gy from 70 mm wide
beams. The 442/35, 625/35, 884/35 and 442/70 MBRSs yielded 25 day, 29 day, 37 day and 35 day median
survival times (MST) (post-irradiation), respectively, exceeding the 20 day MST from 35 Gy-irradiation of
SCCVIIs with a seamless 100 kVp X-ray beam.

A century ago, radiotoxic doses of X-rays delivered
through a flexible grid of 1 mm thick strands of iron
woven 3.5 mm on centre and a thin, continuous underlay
of leather (a low-Z filter), pressed hard against the skin to
blanch it, were able to palliate deep malignancies safely;
iron-shielded epidermal cells healed the resultant punctate
skin burns within 2 weeks [1]. After half a century, such
millimetre-scale grid therapy (GT) was generally super-
seded by skin-sparing megavoltage radiotherapy, although
at least one centre is currently pursuing a version of GT
clinically [2].
It was, however, the radiobiological studies in mice,
which used a deuteron microbeam to simulate cosmic
radiation in space [3] that led to microbeam radiosurgery
(MBRS) investigations, GT’s micrometre-scale analogue.
The MBRS studies have continued since ,1990 using
,200–800 Gy doses of ,30–200 keV X-rays delivered
almost instantaneously through an array of multiple nearly
parallel microslices of tissues [4–13]. Putatively, MBRS
irreparably damages microsegments of neoplastic but not
of normal endothelium; surviving clonogenic tumour cells
may be insufficiently perfused and too sparse to re-grow.
Received 7 January 2005 and in revised form 2 June 2005, accepted 16
June 2005.
This manuscript has been sponsored by Brookhaven Science
Associates, LLC under Contract No. DE-AC02-98CH10886 with
the United States Department of Energy. The US Government
retains, and the publisher, by accepting the article for publication,
acknowledges, a world-wide license to publish or reproduce the
published form of this manuscript, or allow others to do so, for the
US Government purposes. Funding was provided by the DOE Office
of Biological and Environmental Research, the Institute of Pathology
of the University of Bern, and the European Synchrotron Radiation
Facility.
Imminently lethal intracerebral rat 9L gliosarcomas
have been palliated with 25 mm wide microbeams, 100 mm
on centre (oc). About 4 months later when untreated
controls had long been euthanized for tumour overgrowth,
50%, 18%, or 36% of rats remained alive after crossfired
625 Gy, crossfired 312 Gy, or unidirectional 625 Gy skin-
entrance doses, respectively [7].
Despite its weak immunogenicity [14] and robust
radioresistance [15], the deadly aggressive squamous-cell
carcinoma (SCCVII) can be ablated either by immu-
notherapy [16] or by X-irradiation using a radiosensitizer
[17]. However, the outcome of an experimental therapy for
the murine SCCVII carcinoma is generally informative in
terms of growth delay rather than ablation [18, 19].
Accordingly, we compared SCCVII growth delays and
their normal-tissue radiotoxicities following different
MBRS strategies to enable future ranking of various
proposed clinical MBRS treatment plans.
Material and methods
Radiation source
MBRS was performed at the ID17 beamline of the
European Synchrotron Radiation Facility (ESRF), a
6 GeV electron storage ring with an operating current of
180–200 mA. Beamline ID17 is equipped with a 1.6 T
wiggler, which produces a beam of X-rays [20, 21] with a
median energy of 38.1 keV. The beam is filtered with
1.5 m each of C and Al followed by 1.0 mm Cu. This
filtration hardened the spectrum to 93 keV at maximum
intensity, suitable for MBRS. The beam emerged from the
beam pipe through a beryllium window to air in the

The British Journal of Radiology, January 2006 71


radiation-shielded ID17 irradiation hutch, where it was
collimated to 18 mm60.5 mm.
Collimator and irradiations
The microbeams were created with a variable width
tungsten multislit collimator (MSC) (Tecomet, Woburn,
MA) before impinging on the animal [13]. For MBRS, the
anaesthetized mouse was placed prone, lengthwise, on the
15 cm61.5 cm surface of a 15 cm66.5 cm61.5 cm
PlexiglasH block, each foreleg and the left, tumour-bearing
hind leg gently taped to the sides of the block (Figure 1).
The first exposure (of the entire tumour-bearing left
tibiofibular thigh) was nearly anteroposterior, with the
mouse saddle rotated 5 ˚ clockwise (from the horizontal 0 ˚
reference direction of the oncoming beam) about a vertical
axis (as seen by an observer looking downward toward the
mouse, Figure 2) to avoid irradiating the left foreleg; the
second (orthogonal) exposure was implemented after the
block was rotated 95 ˚ clockwise from the 0 ˚ reference
direction about the same vertical axis. Although each of
the two 16 mm broad, 15 mm high anatomical (skin-
entrance) targets had its estimated vertical and horizontal
midplane at the estimated level of the centre of the tumour,
the actual upper horizontal limit of each target was parallel to
and 1 mm below the long edge of the block’s upper surface.
The right hind leg had been taped slightly backward to avoid
exposure to microbeams during the second exposure. For
each irradiation, a computer-guided platform moved the
mouse directly upward (at several cm s21) past the
microbeam array emerging from the MSC. The shutter-
activated exposure time was selected to conform to the slowly
decaying ring current and the pre-programmed upward
acceleration and speed of the platform.
Animal tumour model
SCCVII murine squamous cell carcinoma cells (Prof.
J Martin Brown, Stanford University) were cultured in
D-MEM enriched with 10% fetal bovine serum, 1%
Figure 1. Photograph of anaesthetized female C3H mouse
bearing a leg squamous-cell carcinoma (SCCVII) carcinoma
taped to PlexiglasH block, readied for microbeam radiosurgery
(MBRS) at the European Synchrotron Radiation Facility
(ESRF).
72
M Miura, H Blattmann, E Bräuer-Krisch et al
Figure 2. A PlexiglasH polymethylmethacrylate block (thick
black outline) served as a ‘‘saddle’’ for the mouse, viewed as it
would be by an observer directly above it. The mouse was
anaesthetized and placed prone on the block for its first
tumour irradiation. In this figure, the outline of the mouse is
represented by an ellipse. Two black dots represent its eyes. To
avoid irradiating its left foreleg, the 150 mm long axis of the
block was rotated 5 ˚ (about a vertical axis through the centre
of the block) clockwise from the reference 0 ˚ microbeam direc-
tion. The microbeam array, symbolized by thin arrows, was
propagated in a thin, wide, slightly divergent fan-beam, sub-
stantially in a horizontal plane, represented here as the plane
of this page. The second irradiation was implemented after the
block was rotated 95 ˚ clockwise from the 0 ˚ reference direction
about the same vertical axis.
penicillin/streptomycin, and 1% L-glutamine. Only pas-
sages 1–3 were used to initiate tumours. Cells (26105 in
0.05 ml of medium) were then implanted subcutaneously
(sc) into the left thighs of 20–25 g female C3H mice
(Taconic Farms, Germantown, NY or Charles River
Laboratories, Wilmington, MA). Alternatively, freshly
removed ,1 mm3 fragments of mouse tumours that had
been initiated sc on the dorsal thorax with 56105 cells in
0.1 ml of medium [22] were minced in saline, then
implanted sc in the left thighs through a 16-gauge
trocar. All MBRS-irradiated tumours and 16 of the 40
untreated control tumours grew from cell suspensions. Our
preliminary studies had shown that growth rates using cell
suspensions were the same as those using tumour
fragments; the former are preferred because the suspended
cells do not seem to form satellite tumours along the trocar
track when implanted. Mice bearing ,80–100 mg tumours
(as estimated from volume < x2y/2, where x , y) were
anaesthetized (0.01 ml per gram of body weight (gbw) of
an aqueous 6 mg ml21 sodium pentobarbital solution) by
intraperitoneal injection, (,60 mg gbw21) for MBRS.
Mice were irradiated 10 days after tumours were implanted.
Median survival times (MST) are defined as the time interval
between the day the treated groups were irradiated, which is
equivalent to 10 days after tumour implantation, and the day
they were euthanized unless otherwise stated.
Therapy studies
Single-exposure irradiations were used throughout.
Tumour dimensions were measured 2–3 times per week
and mice were euthanized either when estimated tumour
volumes exceeded 500 mm3 or when skin ulceration or
severe oedema (foot diameter . 5 mm) was observed.
Mice were weighed whenever the tumours were measured,
except during the first week after irradiation, when they
were weighed daily.
100 kVp seamless X-rays
In three groups of anaesthetized mice placed prone on
a horizontal surface, tumours were X-irradiated at
2.10 Gy min21 vertically downward, delivering 25 Gy or
The British Journal of Radiology, January 2006
Short communication: MBRS for murine SCCVII carcinomas

35 Gy. A Philips RT-100 generator was operated at

100 kVp and 8 mA with a 0.4 mm thick Cu filter, a 10 cm
focus-to-skin distance, and a 2.5 cm collimator aperture in
contact with the thigh. Radiation dosimetry was carried
out using an air-equivalent thimble ionization chamber,
adhering to the 1996 IPEB code of practice for 10–
300 kVp, Cu-filtered X-rays [23].

Irradiation groups
The rapidly growing SCCVII cancers were treated in a
clinically analogous way, i.e. after the tumours became
palpable, which took 7 days after implantation (volumes
¢50 mm3). They were then sorted into groups bearing
tumours of comparable size and were irradiated 3 days
later, 1 day after they arrived at the ESRF.
Microbeam widths were either 35 mm or 70 mm and the
on centre (oc) distances for each of the treatment groups
were 200 mm. Groups 1, 2 and 3 were irradiated at skin-
entrance doses of 442 Gy, 625 Gy, and 884 Gy, respec-
tively, using 35 mm microbeam widths in each direction.
Group 4 was similarly irradiated to Group 3 (884 Gy) but
with a 16 mm aluminium filter upstream from the
collimator. Group 5 was irradiated at a skin-entrance
dose of 442 Gy with 70 mm microbeam widths in each
direction and Group 6 was similarly irradiated but with
the 16 mm aluminium filter. The control group comprised
40 untreated SCCVII tumour-bearing mice from five
separate experiments.
Results
100 kVp seamless X-irradiation at 25 Gy and 35 Gy
yielded MSTs of 14 days and 20 days, but long-term
survivals were only 0/10 and 1/9, respectively (Figure 3a).
Untreated controls had a MST of only 6 days or a median
post-implantation survival time of 16 days.
MBRS survival data are shown in Figure 3b. Figure 4
shows average growth rates of various irradiated and
control SCCVIIs. In Groups 1 and 2, euthanasia was
Figure 4. Average relative tumour volumes of the various
microbeam radiosurgery (MBRS)-irradiated and control mice.
The lower tumour volumes noted in groups 3 to 6 relate to the
fact that those tumours had regressed to relatively small or
undetectable volumes when most of the mice had to be eutha-
nized due to severe radiodermatitis of the inner thigh.
usually for tumour overgrowth (volume ¢500 mm3); in
Groups 3–6, it was mainly for foot/leg damage (severe
oedema; diameter of the foot .5 mm) (Table 1). Figures 3
and 4 do not distinguish those reasons for euthanasia.
Euthanasia necessitated by skin radiotoxicity probably
prevented much longer survivals of the third of 884/35 and
442/70 MBRS mice that showed no residual tumour at
necropsy.
MBRS yielded long-term survival rates (up to 153 days)
of 0/12 in Groups 1 and 2, 1/10 in Group 3, and 0/10 in
Groups 4, 5, and 6. Group 3 (884/35 without aluminium)
showed the highest median survival time, and only 1/10
was euthanized for tumour overgrowth; but 8/10 were
euthanized for foot/leg damage and only 1 of those 8

Figure 3. Kaplan-Meier graphs of C3H mice bearing aggressive squamous-cell carcinoma (SCCVII) leg carcinomas irradiated with
various radiation modalities. The on-centre distances for microbeam radiosurgery (MBRS)-irradiations were 200 mm. Mice euthanized
due to foot/leg damage were not distinguished from those euthanized due to tumour overgrowth: (a) Survival graphs of mice bearing
SCCVII carcinomas treated with seamless 25 Gy or 35 Gy skin-entrance doses of X-rays in comparison with unirradiated controls.
(b) Survival graphs of similar mice in MBRS groups (1–6) with skin entrance doses of 442 Gy, 625 Gy, and 884 Gy at 35 mm and
442 at 70 mm beam width. ‘‘Al’’ designates a 16 mm-thick aluminium filter placed upstream from the collimator.

The British Journal of Radiology, January 2006 73

 M Miura, H Blattmann, E Bräuer-Krisch et al

Table 1. Number of ablated tumours, median survival times and explanations for euthanasia in mice treated with microbeam radio-
surgery (MBRS) tracked up to 153 days after irradiation compared with those treated with seamless X-rays and with
untreated controls

Group Dose/beam width Number 153-d tumour Post-irradiation Euthanized (tumour Euthanized Euthanized
(MBRS) of mice control median survival overgrowth) (foot/leg damage) (foot/leg damage)
time (days) mice with tumours
1 442 Gy/35 mm 12 0 25 11 1 1
2 625 Gy/35 mm 12 0 29 9 3 2
3 884 Gy/35 mm 10 1 41 1 8 7
4 884 Gy/35 mm + 10 0 33 3 7 5
aluminium filter
5 442 Gy/70 mm 10 0 38 1 9 5
6 442 Gy/70 mm + 10 0 31 1 9 6
aluminium filter
Unirradiated control 40 0 6 40 0 –
25 Gy seamless 10 0 14 10 0 0
35 Gy seamless 9 1 20 8 0 0

showed no tumour at necropsy (Table 1). In contrast, in
Group 5 (442/70 without aluminium), 9/10 mice were
euthanized due to severe foot/leg damage, of which 4/9
euthanized mice (two each on days 27 and 31) showed no
residual tumour at necropsy. On those same days (27 and
31) in Group 3, seven and four mice, respectively, had no
tumours, indicating that the lower incidence of tumours in
Group 5 compared with Group 3 was due to the earlier
time of euthanasia.
Some non-parametric Wilcoxon Two-Sample analyses
to rank palliation rates, using a morbidity/mortality index
technique [24] on days 9 and 23, before foot/leg damage
became apparent, are shown in Table 2. Without regard to
radiodermatitis, Group 3 (884 Gy) followed by Group 5
(442/70, showed the most effective tumour palliation,
which was expected from the survival graphs (Figure 3b)
and because they received the highest tumour ionization
energies.
Discussion
Figures 3 and 4, and Table 1, demonstrate that MBRS
delayed tumour growth more than did seamless 100 kVp
X-rays and that the former effect is dose-dependent.
However, in the higher dose groups, the plotted growth
rates after day 40 are based on only a few animals, as
many mice had to be euthanized due to radiodermatitis.
The MSTs of each MBRS-treated group were longer than
were the 14 day and 20 day MSTs observed for the
seamless 25 Gy or 35 Gy groups, respectively.
Normal tissue damage occurred more quickly in mice
irradiated with the broader microbeams imparting less
energy per beam (442/70) than in the group with the
narrower microbeam imparting greater energy per beam
(884/35). The radiation field of the SCCVII carcinoma
on the mouse leg encompassed the entire thigh, but not
the foot (Figure 1). Radiodermatitis was most marked in
the inner thigh and oedema was most severe in the left
hind foot below the irradiation field. We attribute the
latter to ablation of overirradiated lymphatics
proximal to the foot. At the higher radiation doses, such
damage limited survival time more than did tumour
overgrowth.
The radiodermatitis of the inner thigh was explained
with microdosimetry simulations using the MCNPX code
[25]. The simulations were performed assuming a water
phantom of the left mouse thigh, shaped as an
inverted, truncated cone (16 mm high with a 13 mm
diameter top and a 3 mm diameter bottom) in which a
0.4 mm diameter sphere of water, the phantom tumour,
was embedded. Computations showed that doses between
the microbeams (‘‘valley doses’’) in the epidermis adjoining
the PlexiglasH would have been ,25% less without
contributions from back-scattered X-rays. Even at
1.5 mm from the PlexiglasH, the dose would have been
reduced ,15% if the PlexiglasH was not present.
Conclusions
Palliation of the exceptionally radioresistant murine
SCCVII carcinoma was better from MBRS than from
seamless 35 Gy irradiation with no more risk to normal
tissue in the radiation field. Normal-tissue damage in the
higher-dose MBRS groups, especially to the left foot

Table 2. p-values from the non-parametric Wilcoxon Two-Sample Test on tumour volumes using morbidity/mortality indices [24] on
days 9 and 23 after irradiationa

Group Days after irradiation 625/35 (2) 884/35 (3) 884/35+Al (4) 442/70 (5) 442/70+Al (6)
442/35 (1) 9 0.011 0.001 0.003 0.036
442/35 (1) 23 0.001 0.001 0.017 0.001 0.004
625/35 (2) 23 0.017

a
Differences between one group (numbered in parentheses) in the top row and another numbered in the left column were deemed
significant if p ¡ 0.05, in which case that column shows the group with the better tumour palliation. No other pairs of groups showed
an advantage in palliation that was significant at the p ¡ 0.05 level.

74 The British Journal of Radiology, January 2006

Short communication: MBRS for murine SCCVII carcinomas
below the radiation field, could be deemed clinically
irrelevant as most of that damage was anatomically remote
from the cancer in structures that would have been spared
high doses under clinical circumstances. Left foot oedema
probably resulted from radiation-induced strictures of
proximal blood vessels and lymphatics. Thus our compu-
tations suggest that MBRS of such SCCVII tumours using
similar skin-entrance doses without the irradiated skin in
contact with the PlexiglasH may enable a greater propor-
tion of mice to survive long-term.
Acknowledgments
The authors thank Mr Seymour Brittman of Brittman &
Son, East Northport, New York, for constructing the
ventilated hardwood cases to enclose mouse cages for
intercontinental air transportation. We also thank Mr
Larry McMillan of Swiss International Airlines and his
coworkers for facilitating our air travel with mice at the
JFK Airport.
References
1. Köhler A. Zur Röntgentiefentherapie mit Massendosen.
Münchener medizinische Wochenschrift 1909;56:2314–6.
2. Zwicker RD, Meigooni A, Mohiuddin M. Therapeutic
advantage of grid irradiation for large single fractions. Int
J Radiat Oncol Biol Phys 2004;58:1309–15.
3. Curtis HJ. The microbeam as a tool in radiobiology. Adv Biol
Med Phys 1963;175:207–24.
4. Slatkin DN, Spanne P, Dilmanian FA, Sandborg M.
Microbeam radiation therapy. Med Phys 1992;19:1395–400.
5. Slatkin DN, Dilmanian FA, Nawrocky MM, Spanne P,
Gebbers J-O, Archer DW, et al. Design of a multislit,
variable-width collimator for microplanar beam radiotherapy.
Rev Sci Instr 1995;66:1346–7.
6. Slatkin DN, Spanne P, Dilmanian FA, Gebbers J-O, Laissue
JA. Subacute neuropathological effects of microplanar beams
of x-rays from a synchrotron wiggler. Proc Natl Acad Sci
USA 1995;92:8783–7.
7. Laissue JA, Geiser G, Spanne PO, Dilmanian FA, Gebbers
J-O, Geiser M, et al. Neuropathology of ablation of rat
gliosarcomas and contiguous brain tissues using a micro-
planar beam of synchrotron-wiggler-generated X rays. Int J
Cancer 1998;78:654–60.
8. Dilmanian FA, Morris GM, Zhong N, Bacarian T, Hainfeld
JF, Kalef-Ezra J, et al. Murine EMT-6 carcinoma: high
therapeutic efficacy of microbeam radiation therapy. Radiat
Res 2003;159:632–41.
9. Schweizer PM, Spanne P, Di Michiel M, Jauch U,
Blattmann H, Laissue JA. Tissue lesions caused by micro-
planar beams of synchrotron-generated X-rays in Drosophila
melanogaster. Int J Radiat Biol 2000;76:567–74.
10. Laissue JA, Lyubimova N, Wagner H-P, Archer DW, Slatkin
DN, Di Michiel M, et al. Microbeam radiation therapy, in
medical applications of penetrating radiation. In: Bradford
Barber H, Roehrig H, editors. Proceedings of SPIE
1999;3770:38–45.
The British Journal of Radiology, January 2006
11. Laissue JA, Blattmann H, Di Michiel M, Slatkin DN,
Lyubimova N, Guzman R, et al. The weanling piglet
cerebellum: a surrogate for tolerance to MRT (microbeam
radiation therapy) in pediatric neuro-oncology. In: Bradford
Barber H, Roehrig H, Doty FP, Schirato RC, Morton EJ,
editors. Penetrating radiation systems and applications III.
Proceedings of SPIE 2001;4508:65–73.
12. Blattmann H, Burkard W, Djonov V, Di Michiel M, Brauer
E, Stepanek J, et al. Microbeam irradiation of the chorio-
allantoic membrane (CAM) of chicken embryo. Strahlenther
Onkol 2002;178:118.
13. Bräuer-Krisch E, Bravin A, Zhang L, Siegbahn E, Stepanek
J, Blattmann H, et al. Characterization of a tungsten/gas
multislit collimator [TMSC] for microbeam radiation therapy
at the European Synchrotron Radiation Facility. Rev Sci
Instr 2005;76:0643031–7.
14. Kanazawa H, Rapacchietta D, Kallman RF. Schedule-
dependent therapeutic gain from the combination of fractio-
nated irradiation and cis-diamminedichloroplatinum(II) in
C3H/Km mouse model systems. Cancer Res 1988;48:3158–64.
15. Suit H, Allam A, Allalunis-Turner J, Brock W, Girinsky T,
Hill S, et al. Is tumor cell radiation resistance correlated with
metastatic ability? Cancer Res 1994;54:1736–41.
16. Mandpe AH, Tsung K, Norton JA. Cure of an established
nonimmunogenic tumor, SCC VII, with a novel interleukin
12-based immunotherapy regimen in C3H mice. Arch
Otolaryngol Head Neck Surg 2003;129:786–92.
17. Nakajima S, Fujii T, Murakami N, Aburano T, Sakata I, Nakae
Y, et al. Therapeutic and imaging capacity of tumor-localizing
radiosensitive Mn-porphyrin KADT-F10 for SCCVII tumors
in C3H/He mice. Cancer Lett 2002;181:173–8.
18. Ning S, Yu N, Brown DM, Kanekal S, Knox SJ.
Radiosensitization by intratumoral administration of cisplatin
in a sustained-release drug delivery system. Radiother Oncol
1999;50:215–23.
19. Katori K, Baba Y, Imagawa Y, Nishimura G, Kagesato,
Takagi E, et al. Reduction of in vivo tumor growth by MMI-
166, a selective matrix metalloproteinase inhibitor, through
inhibition of tumor angiogenesis in squamous cell carcinoma
cell lines of head and neck. Cancer Lett 2002;178:151–9.
20. Thomlinson W, Berkvens P, Berruyer G, Bertrand B,
Blattmann H, Bräuer-Krisch E, et al. Research at the
European Synchrotron Radiation Facility medical beamline.
Cell Mol Biol 2000;46:1053–63.
21. Bräuer-Krisch E, Bravin A, Lerch M, Rosenfeld A,
Stepanek J, Di Michiel M, et al. MOSFET dosimetry for
microbeam radiation therapy at the European Synchrotron
Radiation Facility. Med Phys 2003;30:583–9.
22. Fu KK, Rayner PA, Lam KN. Modification of the effects of
continuous low dose rate irradiation by concurrent
chemotherapy infusion. Int J Radiat Oncol Biol Phys
1984;10:1473–8.
23. Klevenhagen SC, Aukett RJ, Harrison RM, Moretti C,
Nahum AE, Rosser KE. The IPEMB code of practice for the
determination of absorbed dose for x-rays below 300 kV
generating potentials (0.035 mm Al-4 mm Cu HVL;
10-300 kV generating potential). Phys Med Biol
1996;41:2605–25.
24. Coderre JA, Slatkin DN, Micca PL, Ciallella JR. Boron
neutron capture therapy of a murine melanoma with p-
boronphenylalanine: dose-response analysis using a morbidity
index. Radiat Res 1991;128:177–85.
25. Los Alamos National Laboratory Report LA-UR-03-5916;
August, 2003.
75
The British Journal of Radiology, 79 (2006), 76–78 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/97645635

Case report
Solitary pulmonary nodule with growth and contrast
enhancement at CT: inflammatory pseudotumour as an
unusual benign cause
1
S DIEDERICH, Prof. Dr. med., 2D THEEGARTEN, Priv. Doz. Dr. med., 3G STAMATIS, Prof. Dr. med. and
4
R LÜTHEN, Priv. Doz. Dr. med.
1
Department of Diagnostic and Interventional Radiology, Marien Hospital, Academic Teaching Hospital, Rochusstr. 2,
D-40479 Düsseldorf, 2Institute of Pathology, BG-Kliniken Bergmannsheil, Ruhr University Bochum, 3Department of Thoracic
Surgery and Endoscopy, Ruhrland Hospital Essen and 4Department of Medicine, Marien Hospital Düsseldorf, Germany

Abstract. Small (¡10 mm) pulmonary nodules are frequently detected at modern chest CT. As most of these
nodules are benign, non-invasive classification is required – usually based on assessment of growth and
perfusion. Absence of growth and no evidence of perfusion, as demonstrated by lack of enhancement at
contrast-enhanced CT or MRI, strongly suggest a benign nodule. On the other hand, growth with a doubling of
the nodule’s volume between 20 days and 400 days or enhancement suggest a malignant nature of the lesion. We
present an example of a nodule with strong contrast enhancement and a doubling time of approximately 260
days, which histologically represented a benign inflammatory pseudotumour.

Case report
A 56-year-old asymptomatic male underwent chest
radiography in two views as part of a general health
survey. This revealed a small non-calcified nodule
projected over his right mid lung field not demonstrated
on a chest radiograph obtained 3 years previously. CT of
the chest (collimation 5 mm) confirmed a non-calcified
nodule in the lateral segment of the right middle lobe
adjacent to a subsegmental artery and bronchus with a
diameter of approximately 9 mm. No other abnormality
was demonstrated, in particular no hilar or mediastinal
lymphadenopathy was observed.
The patient presented to our hospital for a second
opinion 3 months later. Evaluation of size and contrast-
enhancement was performed obtaining limited spiral CT
data sets with a collimation of 1 mm (Somatom Plus 4;
Siemens, Erlangen, Germany) before and 1 min, 2 min,
3 min and 4 min after administration of 1.4 cm3 kg21
body weight iomeprerol (Imeron 300H; Altana Pharma,
Konstanz, Germany) with an injection rate of 2 cm3 s21.
Images were displayed at lung and mediastinal windows
(Figure 1a, b, c). Nodule density was measured in regions
of interest representing 70% of the nodule’s cross section
at anatomically identical levels. Density was 27 Hounsfield
Units (HU) before contrast injection and increased to
80 HU, 95 HU, 63 HU and 62 HU after 1 min, 2 min,
3 min, and 4 min. Thus, maximum enhancement after
2 min was 68 HU. The diameter of the nodule was again
measured to be 9 mm (Figure 1a).
It has been shown that lack of contrast-enhancement
almost excludes malignancy with a negative predictive
value of 96%, whereas demonstration of contrast-enhance-
ment allows no differentiation between benign and
Received 24 February 2005 and in revised form 18 April 2005, accepted
29 April 2005.
malignant nodules [1]. Thus, the patient was informed
that malignancy could not be excluded and biopsy was
recommended. Due to the central location of the relatively
small nodule adjacent to a subsegmental artery, it was felt
that percutanous biopsy was not appropriate and surgical
biopsy was suggested. The patient, however, did not agree
to immediate biopsy. Therefore, follow-up thin section CT
with 1 mm slice thickness was performed at 6 months and
10 months. There was questionable growth at 6 months
and definite growth at 10 months (Figure 2). The nodule’s
volume was calculated from measurements digitally on the
monitor of a workstation in the axial plane and also by
counting the number of contiguous 1 mm slices for
estimation of the diameter in craniocaudad direction. As
the nodule appeared almost ideally spherical at the
baseline measurement (9 mm) and at 10 month follow-
up (12 mm) its volume was calculated (V54/3 p r3) as
381 mm3 (baseline measurement) and 904 mm3 (10
months) resulting in a doubling time of 8.6 months.
Again surgical biopsy was recommended and now the
patient agreed.
As the nature of the nodule could not be established
prior to surgery it was decided to proceed to minimally
invasive thoracotomy. During surgery a tumour measuring
18 mm624 mm adjacent to the medial segmental
bronchus of the middle lobe was palpated rendering
wedge resection impossible. Due to the small volume of
the middle lobe, primary middle lobectomy was performed
including resection of regional lymph nodes. Final
histological assessment of the well-circumscribed lesion
(Figure 3) including immunostaining for CD-1a, CD-3,
CD-20, CD-68, and EMA showed a mixed inflammatory
infiltrate and connective tissue typically for a benign
inflammatory pseudotumour (Figure 4a). Diagnosis of a
malignant tumour could not be confirmed. Several vessels
were demonstrated within the nodule (Figure 4b). The

76 The British Journal of Radiology, January 2006

Case report: Pulmonary nodule with contrast enhancement and growth at CT

(a) (b) (c)

Figure 1. Dynamic thin-section CT scan before ((a) lung window, (b) mediastinal window) and 2 min after ((c) mediastinal window)
contrast enhancement: The nodule shows an enhancement of 68 Hounsfield units.

post-operative course was unremarkable and the patient
was discharged from hospital after 8 days.
Discussion
Pulmonary nodules are common findings at chest radio-
graphy and even more at chest CT. With the introduction
of spiral CT, and particularly multirow-detector spiral CT,
an increasing number of small nodules (¡10 mm) is
detected. The ratio of benign and malignant nodules
strongly depends on nodule size. In nodules .10 mm the
proportion of malignant nodules is high requiring biopsy
in many cases [2]. In nodules ¡10 mm more than 90% of
nodules are benign [3]. Therefore, biopsy is not routinely
performed in these lesions and non-invasive diagnostic
tests are required to differentiate between benign and
malignant nodules.
The two techniques used routinely are analysis of nodule
growth and perfusion. It has been shown that most
malignant tumours exhibit doubling times between 20 days
and 400 days, whereas faster or slower doubling times
suggest benign lesions [4, 5].
Also, assessment of tumour perfusion is helpful in
predicting a nodule’s nature. As malignant tumours
.1 mm require neoangiogenesis for further growth, all
malignant tumours visible at chest CT should exhibit
enhancement at contrast-enhanced CT or MRI [1].
It has been shown that absence of contrast enhancement
strongly predicts the benign nature of a nodule (e.g.
granuloma); on the other hand, not all enhancing nodules
are malignant due to enhancing benign lesions such as
inflammatory nodules or intrapulmonary lymph nodes
[1, 6, 7].

Figure 2. Follow-up thin-section CT scan after 10 months Figure 3. Histological slide showing the nodule with focal
revealing growth of the nodule to 12 mm. sclerosis (Haematoxylin and eosin, magnification62).

The British Journal of Radiology, January 2006 77


(a)
Figure 4. Histological slides demonstrating a mixed inflammatory infiltrate with lymphocytes and plasma cells ((a) Haematoxylin and
eosin, magnification6200) and involvement of medium-sized vessels with occlusion ((b) CD34 staining, ABC method, magnifica-
tion6100).
Our case is another example of a benign nodule with
strong enhancement that also exhibited growth with a
volume doubling time suspicious for malignancy.
Inflammatory pseudotumour (other diagnostic terms:
fibroxanthoma, xanthogranuloma, xanthofibroma, histio-
cytoma) is a rare entity histologically composed of a
mixture of inflammatory cells, including plasma cells,
lymphocytes, macrophages, a few eosinophils, fibroblasts
and connective tissue. In cases with dominance of plasma
cells, the term plasma cell granuloma is used. The lesions
are typically solitary, round and well circumscribed. The
diameter varies from 0.8 cm to 36 cm. Pulmonary
inflammatory pseudotumours clinically present in 60% of
patients with symptoms such as cough, dyspnoea and
haemoptysis; 40% are asymptomatic. The lesion presents
in patients ranging from 1 year to 77 years, but
approximately 60% are under the age of 40 years [8, 9].
Radiologically, most inflammatory pseudotumours pre-
sent as well-defined nodules or masses measuring between
1 cm and 10 cm. The large difference in the maximum size
reported in the literature probably depends on the mode of
detection as well as the presence or absence of symptoms.
Inflammatory pseudotumours are slightly more common
in the lower lobes. If followed radiographically, growth
has been documented. Contrast studies usually demon-
strate significant enhancement of the lesions. Cavitation or
calcification is rare. Infiltration of adjacent organs may be
observed and misinterpreted as evidence of malignancy
[9, 10]. In symptomatic patients surgical resection is the
therapy of choice.
In conclusion, inflammatory pseudotumour has to be
included in the differential diagnosis of enhancing
pulmonary nodules with growth particularly in children
78
S Diederich, D Theegarten, G Stamatis and R Lüthen
(b)
and young adults. As there is no specific imaging feature,
biopsy is required for the diagnosis.
References
1. Swensen SJ, Viggiano RW, Midthun DE, et al. Lung nodule
enhancement at CT: multicenter study. Radiology
2000;214:73–80.
2. Tan BB, Flaherty KR, Kazerooni EA, Iannettoni MD. The
solitary pulmonary nodule. Chest 2003;123:89S–96S.
3. Kim YH, Lee KS, Primack SL, et al. Small pulmonary
nodules on CT accompanying surgically resectable lung cancer:
likelihood of malignancy. J Thorac Imaging 2002;17:40–6.
4. Yankelevitz DF, Henschke CI. Does 2-year stability imply
that pulmonary nodules are benign? AJR Am J Roentgenol
1997;168:325–8.
5. Yankelevitz DF, Gupta R, Zhao B, Henschke CI. Small
pulmonary nodules: evaluation with repeat CT – preliminary
experience. Radiology 1999;212:561–6.
6. Bankoff MS, McEniff NJ, Bhadelia RA, Garcia-Moliner M,
Daly BDT. Prevalence of pathologically proven intrapulmon-
ary lymph nodes and their appearance on CT. AJR Am J
Roentgenol 1996;167:629–30.
7. Matsuki M, Noma S, Kuroda Y, Oida K, Shindo T, Kobashi
Y. Thin-section CT features of intrapulmonary lymph nodes.
J Comput Assist Tomogr 2001;25:753–6.
8. Colby TV, Koss MN, Travis WD. Inflammatory pseudotu-
mor. In: Tumors of the lower respiratory tract. Atlas of
tumor pathology (3rd edn), Fascicle 13. Washington, DC:
Armed Forces Institute of Pathology, 1995:327–38.
9. Agrons GA, Rosado de Christensen ML, Kirejczyk WM,
Conran RM, Stocker JT. Pulmonary inflammatory pseudo-
tumor: radiologic features. Radiology 1998;206:511–8.
10. McCall IW, Woo-ming M. The radiological appearances
of plasma cell granuloma of the lung. Clin Radiol
1978;29:145–50.
The British Journal of Radiology, January 2006
The British Journal of Radiology, 79 (2006), 79–82 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/94682952

Case report
Non-haemorrhagic subdural collection complicating rupture of
a middle cranial fossa arachnoid cyst
C OFFIAH, BSc, FRCS, FRCR, W ST CLAIR FORBES, MA, DMRD, FRCR and J THORNE, FRCS
Departments of Neuroradiology, Hope Hospital, Salford Royal Hospitals NHS Trust, Stott Lane, Salford, Manchester M6
8HD and Royal Manchester Children’s Hospital, Central Manchester and Manchester Children’s University Hospitals
NHS Trust, Manchester, UK

Abstract. Arachnoid cysts are a common incidental finding on routine brain imaging and, for the most part,
their presence is uneventful. Occasionally they may be associated with haemorrhage into the subdural
compartment. Rarer still is simple rupture of the contents of the arachnoid cyst into the extra-axial space. MRI
can help distinguish between these two rare occurrences – an important distinction to make as this may assist in
directing the treating clinician toward the most appropriate management plan.

Arachnoid cysts are a well-recognized benign intracra-
nial lesion occuring most commonly in the middle cranial
fossa. Although most are small and asymptomatic,
they may be associated with a complicated course most
typically causing mass effect or hydrocephalus.
Spontaneous and post-traumatic intracystic and subdural
haemorrhage has also been reported. We describe a case of
the very rare complication of symptomatic rupture of a
middle cranial fossa cyst into the subdural compartment
without haemorrhage. Despite extensive literature review,
there has been no previous description of this.
Case report
An 8-year-old boy presented with a history of
intermittent headaches, vomiting and double vision over
a period of several weeks, the onset of which was related
to a fall playing football when he struck his head on
concrete. No loss of consciousness occurred at the time
of the injury. On examination, his Glasgow coma scale
(GCS) was 15 and there was no focal neurology, cranial
nerve deficit or papilloedema. CT performed on
admission demonstrated a low-attenuation right-sided
subdural collection – isodense to cerebrospinal fluid
(CSF) – causing moderate mass effect, compression of
the ipsilateral ventricle and effacement of the cortical sulci.
No focal intraparenchymal abnormality was present. A
left middle cranial fossa arachnoid cyst was also noted
(Figure 1). The patient underwent burr-hole drainage of
the right subdural collection and at surgery fluid consistent
with CSF was seen to escape under pressure.
He was discharged a few days later following improve-
ment but re-admitted shortly after that with recurrence of
his symptoms and a notable right-sided scalp swelling. A
repeat CT scan (not shown) demonstrated re-accumulation
with a slight increase in size of the right subdural
collection, which remained low-attenuation. The left
arachnoid cyst remained unchanged and evidence of a
Received 3 August 2004 and in revised form 9 January 2005, accepted 6
May 2005.
small right Sylvian fissure arachnoid cyst persisted.
Right-sided extracranial soft-tissue swelling was also
demonstrated. Ultimately, a subdural-peritoneal shunt
was placed on the right and his subsequent recovery was
unremarkable. Prior to his discharge, a CT scan was
performed which showed only a thin residual right-sided
subdural collection with some associated subdural air, but
no residual mass effect. The left-sided arachnoid cyst was
noted as previously. The small right-sided arachnoid cyst
has become more readily appreciated with resolution of
the ipsilateral subdural fluid (Figure 2).
Since his discharge, a MR scan of the brain has been
performed (7 months since the index admission). The
subdural collection has completely resolved. The right-
sided middle cranial fossa arachnoid cyst has increased
significantly in size since the preceding CT examination
obtained during the admission some 6 months previously.
The left-sided arachnoid cyst has remained unchanged in
size (Figure 3).
The patient remains clinically well and the subdural-
peritoneal shunt in situ.
Discussion
Arachnoid cysts derive from the meninx primitive,
embryologically, which is the primitive membrane ensheath-
ing the developing central nervous system (CNS). As
subarachnoid CSF accumulates, this meninx cavitates and
resorbs under normal circumstances leaving only the
subarachnoid space and the arachnoid membrane. During
this process, the arachnoid membrane may split with
secretion of fluid by the arachnoid cells into the resulting
cleft ultimately yielding a cyst – the so called arachnoid
cyst – which is truly intra-arachnoid anatomically [1–3].
Both intracystic haemorrhage and rupture of middle
cranial fossa arachnoid cysts into the subdural space
resulting in acute or chronic subdural haematoma – either
spontaneous or post-traumatic – have been well docu-
mented in the medical literature [1–14]. Bleeding occurs
due to tearing of an unsupported bridging vein or veins
that are stretched by the cyst and susceptible to rupture by

The British Journal of Radiology, January 2006 79


(a)
Figure 1. (a) Unenhanced CT brain demonstrating a right subdural effusion causing mass effect and (b) a left middle cranial fossa
arachnoid cyst. The right Sylvian fissure demonstrates notable prominence of low (cerebrospinal fluid) density consistent with an
underlying right-sided middle cranial fossa arachnoid cyst.
a rise in intracystic pressure [5, 7, 15]. What has not been
highlighted in the radiological literature is the occurrence
of arachnoid cyst rupture into the subdural compartment
resulting in progressive symptoms of raised intracranial
pressure, despite the lack of haemorrhage. Appreciation of
this complication does appear to have implications in
relation to the management of these patients and is a
valuable differential to highlight to the referring clinician.
This is because there potentially remains a communication
between the arachnoid cyst and the subdural compartment
following rupture so that, despite burr hole drainage of the
collection, there remains predisposition to re-accumulation
of cyst fluid in the subdural compartment and therefore
the increased probability of drain insertion being required
as an immediate definitive treatment. An important
imaging manoeuvre to assist in this differentiation would
be early MRI as the signal characteristics of the subdural
collection would aid distinction between acute or subacute
haemorrhage, as opposed to rupture of arachnoid cyst
contents into the subdural compartment. In the case of the
latter, the signal characteristics of the subdural fluid would
present as isointense to cyst contents (and to CSF).
It would appear that very minor trauma, if any, is
required for arachnoid cyst rupture to occur [6]. In our
case study the head injury that preceded the onset of
symptoms was not associated with any loss of conscious-
ness at the time suggesting that the insult was indeed a
minor one. Rupture has been reported to occur in cases
following the Valsalva manoeuvre during various activities
80
C Offiah, W St Clair Forbes and J Thorne
(b)
such as swimming [4, 5]. There have been sporadic reports
in the medical literature regarding spontaneous disappear-
ance of middle cranial fossa arachnoid cysts following
rupture or haemorrhage into the subdural space with
eventual resorption [4, 5]. Various mechanisms have been
proposed for such resolution [4–6]. However, in our case
report the cyst was seen to increase in size consistent with
re-accumulation after the subdural collection had been
treated and had begun to resolve. Presumably, diversion of
the subdural accumulation with shunt placement reduced
the intracranial pressure enough for the arachnoid cyst to
re-accumulate. The re-accumulation of the right-sided
arachnoid cyst in this case may also have been aided by
the widely conjectured ‘‘flap-valve’’ effect that may result
after a tear in the inner cyst wall following rupture that
allows passage of CSF from subarachnoid space into the
cyst, but closure of the tear in the outer membrane that
allowed cyst contents to egress from the cyst into the
subdural compartment [4, 5, 9].
Forty-eight percent of arachnoid cysts occur in the
middle cranial fossa. Only 20% occur in the posterior fossa
[10–15]. It is reported that only middle cranial fossa cysts
rupture [15] and this is supported by a review by Rogers
et al that demonstrated six cases of subdural haematomas,
which were all associated with middle cranial fossa
arachnoid cysts [9]. In our experience arachnoid cysts
are frequently bilateral; the presence of a middle cranial
fossa arachnoid cyst and a contralateral subdural fluid
collection should therefore raise the possibility of rupture
The British Journal of Radiology, January 2006
Case report: Subdural collection complicating arachnoid cyst
Figure 2. Unenhanced CT brain following shunt drainage of
the right subdural collection. Only a small residual effusion
remains (with some air). The left middle cranial fossa cyst
appears unchanged. The presence of the right middle cranial
fossa cyst is more readily appreciated.
of a contralateral arachnoid cyst as a consideration,
particularly if early MRI fails to confirm the presence of
haemorrhagic subdural fluid contents.
Conclusion
We have demonstrated the rare complication of rupture
of a middle cranial fossa arachnoid cyst into the subdural
space without haemorrhage following minimal trauma.
Although it is well recognized that arachnoid cysts may be
associated with acute and eventually chronic subdural
blood following rupture due to tearing of the vessels that
bridge the cyst wall, non-haemorrhagic rupture into the
subdural compartment is an important radiological
differential diagnosis to consider in order to direct the
clinical/surgical management of such patients optimally as
the imaging appearances of these two entities on CT
examination can be identical. In such cases, early MRI
would be a valuable adjunct.
References
1. Robertson R, Caruso PA, Truwit CL, Barkovich AJ. Disorders
of brain development. In: Atlas SW, editor. Magnetic
resonance imaging of the brain and spine. Philadelphia, PA:
Lipincott Williams and Wilkins, 2002:316–9.
The British Journal of Radiology, January 2006
Figure 3. Axial T2 weighted MR brain performed 6 months
later, confirming the re-accumulation of the right middle cra-
nial fossa arachnoid cyst as indicated by the interval increase
in size as well as the presence of the unaltered left middle cra-
nial fossa arachnoid cyst. No subdural collection was present
this time.
2. Barkovich AJ. Hydrocephalus. In: Barkovich AJ, editor.
Pediatric neuroimaging. Philadelphia, PA: Lippincott
Williams and Wilkins, 2000:592–5.
3. Intracranial arachnoid and ependymal cysts. In: Wilkins RH,
Rengachary SS, editors. Neurosurgery. New York, NY:
McGraw-Hill, 1985:2160–72.
4. Cullis PA, Gilroy J. Arachnoid cyst with rupture into
the subdural space. J Neurol Neurosurg Psychiatry
1983;46:454–6.
5. Inoue T, Matsushima T, Tashima S, Fukui M, Hasuo K.
Spontaneous disappearance of a middle cranial fossa
arachnoid cyst associated with subdural hematoma. Surg
Neurol 1987;28:447–50.
6. Rakier A, Feinsod M. Gradual resolution of an arachnoid
cyst after spontaneous rupture into the subdural space.
J Neurosurg 1995;83:1085–6.
7. Eustace S, Toland J, Stack J. CT and MRI of arachnoid cysts
with complicating intracystic and subdural haemorrhage.
J Comput Assist Tomogr 1992;16:995–7.
8. Ochi M, Morikawa M, Ogino A, Nagoki K, Hayashi K.
Supratentorial arachnoid cyst and associated subdural
hematoma: neuroradiological studies. Eur Radiol
1996;6:640–4.
9. Rogers M, Klug GL, Siu KH. Middle fossa arachnoid cysts
in association with subdural haematomas. A review and
recommendations for management. Br J Neurosurg
1990;4:497–501.
10. Gupta R, Vaishya S, Mehta VS. Arachnoid cyst presenting as
subdural hygroma. J Clin Neurosci 2004;11:317–8.
11. Donaldson JW, Edwards-Brown M, Luerssen TG. Arachnoid
cyst rupture with concurrent subdural hygroma. Pediatr
Neurosurg 2000;32:137–9.
81

12. Gelabert-Gonzalez M, Fernandez-Villa J, Cutrin-Prieto J,
Allut AG, Martinez-Rumbo R. Arachnoid cyst rupture with
subdural hygroma: report of three cases and literature review.
Childs Nerv Syst 2002;18:609–13.
13. Poirrier AL, Ngosso-Tetanye I, Mouchamps M, Misson JP.
Spontaneous arachnoid cyst rupture in a previously
asymptomatic child: a case report. Eur J Paediatr Neurol
2004;8:247–51.
82
C Offiah, W St Clair Forbes and J Thorne
14. Cayli SR. Arachnoid cyst with spontaneous rupture into the
subdural space. Br J Neurosurg 2000;14:568–70.
15. Shapiro KN, Swift DM. Intracranial arachnoid cyst. In:
Tindall GT, Cooper PR, Barrow DL, editors. The practice of
neurosurgery. Baltimore, MD: Williams and Wilkins,
1996;2667–79.
The British Journal of Radiology, January 2006
The British Journal of Radiology, 79 (2006), 83 E 2006 The British Institute of Radiology
Correspondence
(The Editors do not hold themselves responsible for opinions expressed by correspondents)
Social factors in improving radiological
perception
The Editor—Sir,
Manning, Gale and Krupinski are absolutely correct when
they state ‘‘good displays and tools are clearly necessary
……but what we need to understand is how the
Radiologist interacts with the displayed information
during the reading process in order to determine how
we can further improve decision making’’ [1]. They identify
many of the perceptual, cognitive and ergonomic factors.
Social factors also need to be addressed. There is
considerable opportunity for improved knowledge sharing
among radiologists. Performance can be improved by
knowing where in particular to look and what exactly to
look for in different clinical scenarios. While subspeciali-
zation is important in improving perception, targeted
instruction and top up training can improve the perfor-
mance of all [2]. We need to have much more extensive
prompt feedback of our discrepancies and errors. The
prevalence of eye strain among radiologists has been reported
[3] as has the medicolegal implications of reporting at a
significantly faster rate than average [4]. There is no magic
solution that will produce a perfect imaging perceptual
process. However, the social dimension of reporting needs to
be included to optimize our performance.
Yours etc.,
R FITZGERALD
Consultant Radiologist
Radiology Department
New Cross
Wolverhampton
WV10 0QP
UK
(Received 12 August 2005 and accepted 23 August 2005)
DOI: 10.1259/bjr/18395574
References
1. Manning DJ, Gale A, Krupinski EA. Commentary: Perception
research in medical imaging. Br J Radiol 2005;78:683–5.
2. FitzGerald R. Radiological error: analysis, standard setting,
targeted instruction and teamworking. Eur Radiol 2005;15:1760–7.
3. Vertinsky T, Forster B. Prevalence of eye strain among
radiologists. Influence of viewing variables on symptoms. AJR
Am J Roentgenol 2005;184:681–6.
4. Berlin L. Liability of interpreting too many radiographs. AJR
Am J Roentgenol 2000;175:17–22.
The Grandfather of volume scanning
The Editor—Sir,
I would like to express my concern after some comments
made at the memorial lectures held at the Royal Society on
25 May 2005.
The British Journal of Radiology, January 2006
Sir Godfrey Hounsfield was well aware of the possibility
of what at EMI was called volume scanning. He
appreciated the difference between single slice and multiple
slice data acquisition. At EMI, the former idea was based
on movement of the patient through the scanner while the
continuous (slip-ring) gantry rotated using a continuous
power X-ray source. The latter technique was to have
involved the collection of data from set of contiguous
slices at the same time, a technique originally described in
Sir Godfrey’s first CT patent.
It was to achieve volume scanning that the TOPAZ
geometry was invented. The patent makes clear the
continuous rotation nature of the scanner. This system
was conceived in the mid 1970s with discussion for
implementation with the commercial Division later the
same decade. This system used solid-state detectors and an
X-ray tube with a directly oil cooled anode. A photograph
of the prototype, built by the Research team in the
Central Research Laboratories of EMI, was shown on
the 25 May 2005. It is clear that this geometry was the
first to make possible the matching of continuous
rotation with a continuous power X-ray source. The
apparent falter in the development of CT in the 1980s can
be traced to other causes and not to a lack of technical
innovation.
The nature of the TOPAZ configuration also uniquely
enabled focused layers to be obtained from the scanno-
gram or pilot scan data (Zonogram).
In the mid 1980s a 1 s, 1 mm slice thickness, version of
the system, based on Sir Godfrey’s ideas, was successfully
built and fully tested. From the scanner volume clinical
scans were obtained from which 3D images were
reconstructed.
The implication that there were any constraints placed
on the future of CT by Sir Godfrey is therefore wholly
inaccurate. If not the father of volume scanning, Sir
Godfrey Hounsfield must indeed be considered to be the
grandfather.
Yours etc.,
A BASKERVYLE STRONG
Engineering Manager EMI Medical Ltd (retired)
Broombank
267 Penistone Road
Kirkburton
Huddersfield
HD8 0PF
(Received 16 August 2005 and accepted 22 August 2005)
DOI: 10.1259/bjr/15532036
We hope the article in this issue by E Beckmann (p. 5)
and this 16
(Received letter will2005
August rectify any omission—Editor.
and accepted 22 August 2005)
83
The British Journal of Radiology, 79 (2006), 84–86 E 2006 The British Institute of Radiology
DOI: 10.1259/bjr/23776068

Case of the month

A deformed skull with enlarging hand and feet in a young
female
1
B GUGLANI, MD, 1C J DAS, MD, DNB, 1A SEITH, MD, 2N TANDON, DM and 2B A LOWAY, MD
1
Radiodiagnosis and 2Department of Endocrinology, All India Institute of Medical Sciences, New Dehli, India

A 25-year-old woman presented with a 6 year history of

gradually enlarging swelling at the back of her head. She
had also noticed enlarging hands and feet with increased
prominence of eyes for the last 3–4 years. She had been
amenorrhoeic for the past 2 years. On physical examina-
tion, her height was 163 cm. There was facial deformity
with a prominent right side of the face and bony swelling
in the region of the external occipital protuberance. Her
hands and feet were enlarged with a doughy consistency.
In addition, mild scoliosis was found in the mid-thoracic
region with convexity towards the right side. The left
humerus was short and bowed and the left rib cage was
deformed with multiple swellings. There was no evidence
of abnormal skin pigmentation. Her thyroid was mildly
enlarged. Galactorrhoea was also observed. Visual field
examination revealed bitemporal hemianopsia on perime-
try. Endocrine evaluation showed a non-suppressible
growth hormone level (GH) of 60 ng ml21 and an
increased prolactin level of 43 mg l21. Plain radiography
of the skull was obtained (Figure 1). Subsequently, Figure 1. Plain radiograph skull (lateral view).
contrast enhanced MRI (CEMRI) of the sella was
performed (Figure 2). What is the diagnosis in this case?

Received 25 February 2005 and in revised form 22 April 2005, accepted

31 May 2005.

(a) (b)

Figure 2. (a) Pre- and (b) post-contrast enhanced MR coronal images through the sella turcica.

84 The British Journal of Radiology, January 2006

Case of the month: A deformed skull

Discussion
The diagnosis of acromegaly was clinically and bio-
chemically unequivocal in this case. Subsequently, we
performed contrast enhanced MRI (CEMRI) of the sella
for evaluation of the patient’s acromegaly. A T1 weighted
image of the sella revealed a large sellar, suprasellar mass
compressing the optic chiasm and causing expansion of the
sella. The mass showed enhancement in post-contrast
imaging suggestive of a pituitary macroadenoma, thus
confirming the clinical diagnosis of acromegaly. Marked,
diffuse expansion of the skull base was also noted
enhancing on contrast administration (Figure 2).
The patient sought medical advice primarily for her
marked skull expansion seen in the occipital region, which
had increased gradually over the last 6 years. Plain
radiography of the skull showed gross expansion of the
skull base and occiput with areas of sclerosis and ground
glass density. Enlarged maxillary and frontal sinuses were
also seen. A CT performed for detailing of the foraminal
compression in the skull base showed the typical ground-
glass appearance of fibrous dysplasia (Figure 3).
Narrowing of the bilateral optic canal and orbital apices
were also seen. A subsequent skeletal survey revealed that
the involvement was indeed multifocal with expansile
lesions seen in the ribs, left humerus and radius (Figure 4).
All of these features pointed to a pathology in addition to
acromegaly due to a GH secreting pituitary adenoma.
Based on the radiological appearance of the ground glass
density and their characteristic distribution, a differential
diagnosis of coexisting fibrous dysplasia was made. As
both of these conditions are associated in only one genetic
abnormality, a final diagnosis of McCune-Albright
syndrome – polyostotic fibrous dysplasia with acromegaly
due to pituitary macroadenoma – was made.
The McCune-Albright syndrome (MCAS) is a sporadic
disorder characterized by polyostotic fibrous dysplasia, Figure 4. Radiographs of the left humerus and radius show
cutaneous pigmentation and endocrine hyperfunction. The classical lesions of fibrous dysplasia.
presence of any two of the three lesions (skin, bone and
endocrine glands) is sufficient for the diagnosis of MCAS
[1]. The genetic basis of MCAS is now reasonably
understood and is due to the post-zygotic activating
mutations of arginine 201 in the guanine-nucleotide-
binding protein (G protein) alpha-subunit (Gsalpha),
leading to a mosaic distribution of cells bearing
constitutively active adenylate cyclase [2]. The resultant
disorder depends on when the mutation occurs; during
embryonic development or post-natal life. The earlier it
takes place, the more cells are affected. Somatic mutations
in a small cell mass result in MCAS; whereas in a larger
cell mass, mutation results in polyostotic fibrous dysplasia
[3]. The distribution of affected cells follows embryological
lines of ectodermal migration, which explains the uni-
lateral and focal expression of MCAS in bones as well as
in endocrine tissue.
Various endocrinopathies reported in MCAS include
precocious puberty, thyrotoxicosis, Cushing’s syndrome,
acromegaly, hyperprolactinaemia and hypophosphataemic
rickets [4, 5]. The association of polyostotic fibrous
dysplasia and acromegaly, although rare, is a well
described entity [5]. In MCAS, gigantism/acromegaly
Figure 3. Axial CT image (bone-window) showing ground- usually present at an earlier age (less than 30 years)
glass expansion of skull base with narrowed basal foramina, than in classical acromegaly [6–8]. A pituitary adenoma
bilateral optic canal and orbital apices. may be found less often than in classical acromegaly [5, 8].

The British Journal of Radiology, January 2006 85


Moreover, the macroadenomas in MCAS are smaller than
those in classical acromegaly [5, 8].
The association of acromegaly and fibrous dysplasia
may pose a diagnostic challenge to the clinician and the
radiologist. The majority of patients with MCAS are short
in stature because of precocious puberty, recurrent
fractures and hypophosphataemic rickets, whereas those
with associated GH excess/acromegaly usually reach a
normal height [5, 9]. Also, since fibrous dysplasia has a
predilection for skull base involvement, the facial dys-
morphism may mask the usual features of acromegaly
causing delay in diagnosis.
CT and MRI play a pivotal role in the evaluation of
these patients. MRI is better than CT in assessment of
sella in the presence of bony skull base thickening due to
fibrous dysplasia. The distinction between pituitary gland
and abnormal fibrous bone tissue at skull base is better
made on MRI. The combination of pre- and post-contrast
images is useful in this regard. However, CT of skull
base plays a useful role in some cases for detailing
neural foraminal compression, especially if surgery is being
contemplated [10].
Craniofacial fibrous dysplasia may mimic hyperostotic
meningioma (meningioma en plaque) or even osteoma,
especially in a monostotic lesion [9]. Association of
acromegaly and meningioma has also been described [12, 13].
MCAS with acromegaly and skull base fibrous dysplasia
is also a therapeutic challenge as transpituitary surgery is
often not possible in the presence of fibrous dysplasia of
skull base whereas radiation therapy can induce bone
sarcomatous transformation [14]. Some authors have
suggested a transfrontal route to approach the pituitary
adenoma [15]. In a series by Akintoye et al, the authors
described a distinct clinical phenotype of MCAS due to
GH excess which is characterized by inappropriately
normal stature, thyroid releasing hormone (TRH) respon-
siveness, prolactin cosecretion, small or absent pituitary
tumours, a consistent but inadequate response to treat-
ment with cabergoline and an intermediate response to
long acting octreotide [5]. In our patient too, the medical
treatment was chosen because surgery was not possible
due to the fibrous dysplasia of the skull base and
radiotherapy would increase the risk of sarcomatous
transformation. The patient received octreotide LAR
40 mg (intramuscular) monthly. The plasma growth
hormone levels (post-oral glucose) decreased from an
initial 60 ng ml21 to 23 ng ml21 1 month after first
injection.
In conclusion, the association of acromegaly with
MCAS may pose a diagnostic and therapeutic challenge.
MRI is vital in the evaluation of such patients for the
delineation of the pituitary adenoma separate from the
86
B Guglani, C J Das, A Seith et al
skull base abnormality. CT is a useful adjunct pre-
operatively to delineate the foraminal compression.
References
1. Schwindinger WF, Levine MA. McCune-Albright Syndrome.
Trend Endocrinol Metab 1993;7:238–42.
2. Lumbroso S, Paris F, Sultan C; European Collaborative
Study. Activating Gsalpha mutations: analysis of 113 patients
with signs of McCune-Albright syndrome--a European
Collaborative Study. J Clin Endocrinol Metab
2004;89:2107–13.
3. Feldman F. Tuberous sclerosis, neurofibromatosis, and
fibrous dysplasia. In: Resnick D, editor. Diagnosis of bone
and joint disease. 4th edn. Philadelphia, PA: WB Saunder’s
Company 2002:4792–843.
4. Weinstein LS, Shenker A, Gejman PV, Merino MJ, Friedman
E, Spiegel AM. Activating mutations of the stimulatory G
protein in the McCune Albright Syndrome. N Engl J Med
1991;325:1688–95.
5. Akintoye SO, Chebli C, Booher S, Feuillan P, Kushner H,
Leroith D, et al. Characterization of gsp-mediated growth
hormone excess in the context of McCune-Albright syn-
drome. J Clin Endocrinol Metab 2002;87:5104–12.
6. Chanson P, Dib A, Visot A, Derome PJ. McCune-Albright
syndrome and acromegaly: clinical studies and responses to
treatment in five cases. Eur J Endocrinol 1994;131:229–34.
7. Lipson A, Hsu T. McCune Albright syndrome associated
with acromegaly: report of a case and review of literature.
Johns Hopkins Med J 1981;149:10–4.
8. Premawardhana LDKE, Vora JP, Mills R, Scanlon MF.
Acromegaly and its treatment in the McCune-Albright
syndrome. Clin Endocrinol (Oxf) 1992;36:605–8.
9. Leet AI, Chebli C, Kushner H, Chen CC, Kelly MH,
Brillante BA, et al. Fracture incidence in polyostotic fibrous
dysplasia and the McCune-Albright syndrome. J Bone Miner
Res 2004;19:571–7.
10. Kim KS, Rogers LF, Goldblatt D. CT features of
hyperostosing meningioma en plaque. AJR Am
J Roentgenol 1987;149:1017–23.
11. Daly BD, Chow CC, Cockram CS. Unusual manifestations of
craniofacial fibrous dysplasia: clinical, endocrinological and
computed tomographic features. Postgrad Med J 1994;70:10–6.
12. Cannavo S, Curto L, Fazio R, Paterniti S, Blandino A, Marafioti
T, et al. Coexistence of growth hormone-secreting pituitary
adenoma and intracranial meningioma: a case report and review
of the literature. J Endocrinol Invest 1993;16:703–8.
13. Bunick EM, Mills LC, Rose LI. Association of acromegaly
and meningiomas. JAMA 1975;240:1267–8.
14. Immenkamp M. Malignant change in fibrous dysplasia
(author’s transl). Z Orthop IhreGrenzgeb 1975;113:331–43.
15. Bhansali A, Sharma BS, Sreenivasulu P, Singh P, Vashisth
RK, Dash RJ. Acromegaly with fibrous dysplasia: McCune-
Albright Syndrome -- clinical studies in 3 cases and brief
review of literature. Endocr J 2003;50:793–9.
The British Journal of Radiology, January 2006
Acknowledgment to Referees
The The Editors
Editors would
would like like to thank
to thank all their
all their colleagues
colleagues whowho have
have contributed
contributed their
their valuable
valuable time
time and
and effortin inreviewing
effort
manuscripts submitted to The British Journal of Radiology. Listed below are the names of referees of papersreferees
reviewing manuscripts submitted to The British Journal of Radiology. Listed below are the names of submitted to
BJRof papers 1submitted
between Decemberto2004
BJRand
between 1 December
1 December 2005.2004 and 1 December 2005.

A Campbell, S Dowling, A Grier, D

Abernethy, L Carr, R Downes, M Griffiths, S
Adam, E Carrington, B Doyle, P Grubnic, S
Adams, J Carroll, N Driver, D M Grundy, A
A’Hern, R Carruthers, D M Drury, A Guest, P
Aird, E Casey, M Duck, F Guthrie, J A
Albrecht, T Castellano, I Durante, M
Allan, R Chalmers, N H
Allen, G Chambers, R E Hale, M
Allisy-Roberts, P Chandy, J Elabassy, M Hall, C
Al-Qaisieh, B Chapman, A Elford, J Hall, E
Amin, Z Chapple, C-L Elias, D Hall-Craggs, M
Anbarasu, A Chawla, T England, R Halligan, S
Andreyev, J Chinn, R Evans, D Halpin, S
Ansorge, O Clarke, S Evans, J Hanlon, R
Armstrong, P Cleveland, T Evans, S Hanson, M
Ashleigh, R Cochlin, D Harbinson, M
Ashton-Key, M Cole, D F Harden, S
Atchley, J Colligan, S Fairbairn, K Hardman, J
Collins, C Faithfull, S Hare, C
B Collins, M A Farrugia, M Harrison, R
Balogun, M Collins, M C Faulkner, K Hart, D
Barber, P Connolly, D Flynn, T Hart, G
Barker, C Conway, J Fogelman, I Hartley, A
Barker, D Cook, G Forbes, K Harvey, C
Barrington, S Cook, P Forsyth, L Haslam, P
Barron, D Cooper, P Fowler, J Hatton, M
Bearcroft, P Copley, S Fox, B Healy, J
Beavis, A Coral, A Francis, I Heenan, S
Beggs, I Corr, C Freeman, A Heinz-Peer, G
Bell, J Cosgrove, D Freeman, S Heneghan, M
Belli, A-M Cosgrove, V Fukuda, S Henson, J
Benham, J Cousins, C Heron, C
Birchall, D Cowan, N G Hide, G
Blake, P Cowling, M Gawne-Cain, M Hillier, J
Blease, S Crawley, C Gaze, M Hiorns, M
Bownes, P Crawley, T Geh, I Hirst, D
Boyle, G Crellin, A Geleijns, J Hoggard, N
Brada, M Crowe, P Gentle, D Holemans, J
Bradley, A Curtis, J George, C Hollaway, P
Bradley, D Czajka, J George, J Hoole, A
Brettle, D Gibson, M Hopewell, J
Broderick, N D Gillams, A Hopkins, K
Brown, P Dance, D Gillespie, J Hounsell, A
Bryant, P Darby, M Gilligan, P Hubscher, S
Budgell, G Darroudi, F Gishen, P Huda, W
Buffa, F Davies, M Given-Wilson, R Hufton, A
Burch, A Davies, N Glynne-Jones, R Hughes, D
Burling, D Davison, P Goddard, T Hutchinson, C
Burn, P Dawson, P Goh, V IIrving, H
Burrell, H Deane, C Goldstone, K
Butteriss, D Deehan, C Gordon, A J
Byrne, J Derchi, L Goss, D Jackson, A
Dhingsa, R Gould, D Jackson, J
C Ditchfield, A Greaves, S Jackson, S
Callaway, M Dixon, A Green, R Jackson, S A
Campbell, R Domjan, J Greener, A Jaspan, T

The British Journal of Radiology, January 2006 87

Jayakrishnan, V Matthews, S Olliff, S Rockall, A
Jobling, J Maxwell, A O’Neill, P Rodgers, P
Johnson, K Mayles, H O’Reilly, G Rogers, A
Julian, W Mayles, W Ormerod, O Ross, P
Julyan, P McCafferty, I Ostlere, S Rottenberg, G
McCall, J Owens, S Rowbottom, C
K McCallum, H Rowell, N
Kaanders, J McCavana, J P Rowland Hill, C
Karani, J McGee, S Padhani, A Rust, A
Kaufmann, P McHugh, K Padley, S Rutherford, E
Kay, C McHugo, J Page, A Ryan, P
Keane, A McKenzie, A Paisley, E Ryan, S
Kearney, S McLean, A Paley, M
Keat, N McNee, S Parker, G S
Kelly, C Meeson, S Parry, J Saada, J
Kerslake, R Mikhaeel, G Patankar, T Saidlear, C
Kessar, P Miles, K Patel, U Saifuddin, A
Kessel, D Miller, S Patsios, D Salim, F
Keston, P Mitchell, F Payne, G Saran, F
Khaw, K-T Mitra, D Pelling, M Saunders, M
Khoo, L Mohammed, S Pereira, P Saunders, T
Kinsella, D Mohan, H K Phillips, A Schofield, K
Kirby, M Mooney, R Phillips-Hughes, J Scholz, M
Koh, D-M Moore, C Pickles, T Senior, R
Koller, C Moores, M Pilcher, J Shah, P
Kuker, W Morcos, S Pilling, D Sharma, B
Kuntzsch, M Morgan, A Plant, G Shaw, A
Morgan-Fletcher, S Plowman, P N Shaw, A S
L Moriarty, M Pretorius, P Sheridan, M
Laitt, R Morrison, R Price, P Shorvon, P
Langton, C M Moss, H Prise, K Shrimpton, P
Larkin, E Moss, J Pullinger, R Sidhu, P
Lee, C Mothersill, C Sikdar, T
Lee, S Mott, J Q Silver, D
Lenthall, R Mountford, P Quaghebeur, G Slater, A
Leslie, M Moussa, S Slevin, N
Lewis, M Mueller-Klieser, W R Smart, J
Lewis, M A Muirhead, C Raby, N Sminia, P
Lewis-Jones, H Munro, A Ralleigh, G Smith, S
Lim, A Murphy, D Rampling, R Sohaib, A
Litherland, J Murphy, P Ramsdale, M Sprigg, A
Livsey, J Murray, D Rankine, J Staffurth, J
Locks, S Rawlings, D Stanley, S
Logan, M N Reading, P Steeds, R
Lomas, D Nahum, A Reddy, M Stevenson, G
Lopez, C Nakielny, R Reek, C Stratford, M
Lowdell, C Nanda Kumar, E Rees, M Strickland, N
Lowe, S Negus, S Reliene, R Sulkin, T
Lucraft, H Ng, C Reynolds, J Sutton, D
Ng, K K-C Rezvani, M Sykes, J
M Niven, S Richardson, D
Mackenzie, A Nolte-Ernsting, C Richenberg, J T
Maliakal, P Nunan, T Ridley, N Tait, D
Malone, J Nutting, C Riley, P Tait, N
Malone, L Rimoldi, O Tan, L T
Manning, D O Ritchie, D Tattersall, D
Marples, B O’Connor, R Roach, H Tawn, E
Marshall, M O’Donnell, P Robbins, M Taylor, A
Marshall, N O’Donovan, N Roberts, D Taylor, B
Marshall, T Odurny, A Robertson, G Taylor, D
Martin, C Offiah, C Robertson, I Taylor, P
Martin, D Ogunremi, T Robinson, P J A Taylor, R
Maskell, G Old, S Robinson, P S Taylor, S
Matson, M Olliff, J Robson, K Temperton, D

88 The British Journal of Radiology, January 2006

Thomas, S V Warren, R Worthy, S
Thompson, P van der Molen, A Warrington, A Wu, H
Thomsen, H van Zeeland, B Watson, J
Thurston, J Vaó, E West, R Y
Tibballs, J Varghese, A Weston, M Yates, S
Tins, B Vennart, W Whitby, E Ying, M
Traill, Z Vijayanathan, S Whitehouse, R
Travis, S Vinjamuri, S Wilkins, J Z
Troughton, A Vinnicombe, S Wilkinson, J Zoetelief, J
Tsalafoutas, I Wilkinson, L
Tuck, J W Williams, H
Tung, K Wade, P Williams, J
Turnbull, I Wakeley, C Winder, J
Twyman, N Waldman, A Wittkop, B
Walker, A Wood, A
U Waller, M Wood, C
Uberoi, R Ward, S Woolhouse, I
Uthappa, V Wardman, P Workman, A

The British Journal of Radiology, January 2006 89

BJR
The British Journal
of Radiology

February
2006
Volume 79
Issue 938
February 2006, Volume 79, Issue 938

● DNA repair: therapeutic implications

● Patients’ perception of tests in the assessment of faecal

incontinence
● Enhanced visualization and quantification of magnetic resonance

diffusion tensor imaging using the p:q tensor decomposition

● PET-CT findings in surgically transposed ovaries

● An audit of imaging test utilization for the management of

lymphoma in an oncology hospital: implications for resource

planning?
● Image quality and breast dose of 24 screen–film combinations for

mammography
● The effect of phantom type, beam quality, field size and field

position on X-ray scattering simulated using Monte Carlo

techniques
● Techniques for measurement of dose width product in panoramic

dental radiography
● A comparison of three-field and four-field techniques in different

clinical target volumes in prostate cancer irradiation using dose

volume histograms: a prospective three-dimensional analysis
● A comparative evaluation of two head and neck immobilization

devices using electronic portal imaging

● Excessive leakage radiation measured on two mobile X-ray units

due to the methodology used by the manufacturer to calculate

and specify the required tube shielding
● Improvements in dose homogeneity for tangential breast fields

from a selection of combinations of library compensators

● Ruptured spinal dermoid cyst with disseminated intracranial fat

droplets
● Colobronchial fistula: a late complication of childhood

radiotherapy
● Ingested foreign body mimicking an appendicolith in a child

● Misleading positioning of a Foley catheter balloon

● An unusual cause and presentation of a pelvic mass

The British Journal of Radiology, 79 (2006), 91–93
COMMENTARY
DNA repair: therapeutic implications
1
S R MCKEOWN, MA, PhD and 2B JONES, MSc, MD
1
School of Biomedical Sciences, University of Ulster at Coleraine BT52 1SA, Northern Ireland and
2
Queen Elizabeth University Hospital, Birmingham B15 2TH, UK
The goal of all anti-cancer treatments is to design
strategies that are specific for tumours and non-toxic to
the patient. Molecular targeting is now becoming a
reality with new treatments designed to target processes
that are thought to be tumour specific, or where there are
quantitative differences in target expression between
cancer and normal cells. On 3 March 2005, the Radiation
and Cancer Biology committee held a meeting to discuss
the targeting of DNA repair pathways, which are often
defective in tumours. Prof. Steve Jackson (Cambridge
University) started the programme with a discussion of
some of the main DNA damage response (DDR) path-
ways. Repair in normal cells is a hugely efficient process,
with individuals requiring repair of an estimated 1018
DNA lesions per day caused by reactive oxygen species
alone. Most of the inherited cancer predisposition
syndromes involve DDR dysfunction and similar muta-
tions are often found in sporadic cancers. The specificity
of DDR targeting agents comes from the need for the
faster dividing tumour cells to repair DNA damage more
quickly and efficiently than the mostly quiescent, or
more slowly cycling normal cells. This may also be
compromised by an already defective DDR pathway,
which further reduces the ability of the tumour cells to
repair efficiently, while being less critical in normal cells.
He pointed out the clearly integrated nature of stress
response in cells since DNA-PK, ATM and ATR have
overlapping roles in DDR, transcriptional regulation, cell
cycle control and cell death pathways. These processes
are relevant not only to cancer therapy, but also in
immune deficiency syndromes, neurodegenerative dis-
orders, infertility, premature ageing and impaired
telomere function.
Many laboratory approaches were discussed during
the day. Of particular relevance was the talk by Dr Kai
Rothkamm (Gray Cancer Institute, Northwood) who
discussed the potential uses of the cH2AX assay to
measure double strand breaks (DSB). This relatively new
assay allows quantitation of DSBs with more accuracy
and at much lower doses than was possible previously.
The British Journal of Radiology, February 2006
Received 24 March 2005
Revised 8 June 2005
Accepted 13 July 2005
DOI: 10.1259/bjr/22946335
’ 2006 The British Institute of
Radiology
However, as with many assays there are inherent pitfalls
as well as advantages of this method, and further work is
required to characterize the assay completely. Dr
Rothkamm identified possibilities for its use as a low
dose exposure assay, using blood lymphocytes, since it is
sufficiently sensitive to quantify exposure after diagnos-
tic CT scans. It can also be used in studies of DDR
inhibitors to quantitate responses and several speakers
during the day reported uses for this assay.
Prof. Jackson proposed that if certain types of cancer
possess inherent DNA repair disorders then, in principle,
inhibition of the remaining DDR mechanisms should
lead to cell death more efficiently that can be achieved in
normal cells where the full complement of repair
enzymes is available. This theme was exemplified by
several of the symposium speakers.
Dr Niall Martin (KuDOS Pharmaceuticals, Cambridge)
described two approaches to this strategy. In colorectal
tumour cell lines with mis-match repair (MMR) defects,
the response to standard cytotoxic agents such as
temozolamide is enhanced when combined with inhibi-
tors of PARP-1 (Poly (ADP-ribose) polymerase-1) – an
enzyme critical to the early response to single strand
breaks (SSB). This combination increases the yield of
both SSB and DSB; the latter have been shown using the
DSB specific cH2AX assay. MMR defects are not found in
normal bone marrow cells, so that enhanced acute
marrow toxicity is not to be expected. BRCA1/2 are also
known to be key proteins involved in the cellular
response to DDR. A significant number of breast
tumours contain defects in BRAC1/2, including almost
all inherited breast tumours. Inhibition of repair with
PARP-1 caused a profound sensitization of BRCA1/2
deficient cells affecting G2/M checkpoint arrest,
increased chromosome aberrations and tumour regres-
sion in the absence of other cytotoxic agents. This offers
an exciting opportunity to control this relatively large
subset of breast tumours. A similar approach, targeting
DNA-PK inhibitors, cause preferential cell kill in ATM
-/- cells, again with dramatic effects in vitro.
91

Prof. Penny Jeggo (Sussex University) described the
role in DNA repair of ATM, and a small ATM interacting
nuclease, called Artemis. Using quiescent fibroblasts, so
that cell cycle differences did not confound the inter-
pretation, she showed that for full restoration of DNA
damage caused by some agents, times in the region of
72 h are needed. This is significantly longer than most
reported repair studies and interestingly is longer than
the time allowed between fractions in conventional
radiotherapy; in part this may offer an explanation for
the poorer DNA repair capacity in tumour vs normal
cells. Although most (,90%) of the DNA repair occurs
rapidly, the residual damage is significantly more
difficult to deal with. The slower repair process appears
to be dependent on the integrity of cell cycle checkpoint
control. In addition, more severe lesions, using alpha
particles, show a longer time to complete resolution of
the damage. She showed evidence that Artemis is
required for this process suggesting another potential
drug target.
Prof. Hilary Calvert (University of Newcastle) gave a
keynote lecture on the current clinical trials involving
DNA repair inhibitors. Resistance to methylating agents
in many cells is caused by the repair enzyme alkylgua-
nine alkyltransferase (O6AT). This enzyme can be
inhibited using 6-benzylguanine (6BG) and 4 bromothe-
nylguanine (Patrin). Unfortunately the clinically toler-
ated dose of Carmustine must be reduced threefold in
combination with 6BG, whereas temozolamide is less
affected by combination with Patrin. This suggests that
O6AT plays an important role in normal tissue recovery.
Phase 2/3 trials are currently determining whether there
is an overall therapeutic benefit with this combination. A
Phase 1 trial combining a PARP-1 inhibitor with
temozolamide is just about to report and further trials
are in the planning stages. Prof. Calvert discussed briefly
the difficulties of setting up clinical trials in the
molecular targeting era, where precise control of sample
collection, storage and evaluation must be in place to
identify the molecular profile of the tumour and its likely
susceptibility to the treatment under investigation.
Dr Stephany Veuger (Newcastle University) presented
further work on PARP-1 inhibition. NF-kB is a stress
inducible transcription complex that induces genes that
control proliferation responses and suppress apoptotic
cascades. Aberrant activation of NF-kB is common in
tumours and recently it has been noticed that its
activation in PARP-1 deficient cells is reduced. The
involvement of these two proteins in the presence or
absence of a potent PARP-1 inhibitor (AG14361) was
investigated when cells were also exposed to 20 Gy
ionizing radiation (IR). The data provided evidence that
PARP-1 function is required for NF-kB activity following
exposure to IR. The results suggested that potentiation of
IR-induced radiosensitivity may be through inhibition of
NF-kB rather than as a direct consequence of PARP-1
mediated inhibition of DNA repair. This result clearly
has implications for rationale design of PARP-1 inhibi-
tors in the treatment of cancer.
Dr Paul Mullan (Queen’s University, Belfast) showed
the power of an initial microarray screen to identify
differences in BRCA1 competent and deficient cells. Dr
Mullan and colleagues have identified a family of
calcium binding proteins that are novel BRCA-1
92
S R McKeown and B Jones
repressed targets. S100A7 (psoriasin) is dependent on
functional c-Myc and is also inducible by DNA damage
in a BRCA-1 dependent manner. They linked this to a
novel pathway of p27kip1 down-regulation that has
previously been seen to be consistently down-regulated
in BRCA1 mutated cells. The data have allowed
identification of a novel pathway that could provide a
target for molecular targeting agents.
Targeting of DNA base excision repair was discussed
by Dr Srinivasan Madhusudan (CRUK, Weatherall
Institute of Molecular Medicine, Oxford). The multi-
functional protein endonuclease HAP-1/APE-1/Ref-1 is
involved in base excision repair and is implicated in the
pathogenesis of several human tumours. Its over-
expression is linked to both chemoresistance and radio-
resistance. Using a high throughput chemical screen, the
Oxford group has identified KM09181 as a lead inhibitor
of HAP-1 with an IC50 value of 3.5 mM. At non-toxic
concentrations it causes significant potentiation of the
cytotoxicity of a number of agents. This report is the first
biological evidence for the direct targeting of this DNA
repair enzyme.
A series of novel PARP-1 inhibitors were described by
Dr Esther Woon (University of Bath). Previously, they
had identified 5-aminoisoquinolin-1-one (5-AIQ), which
shows a wide range of therapeutic activity in vivo. Using
the PARP-1 crystal structure, they designed a series of
compounds similar to 5-AIQ, with the aim of identifying
novel compounds with more potent PARP-1 inhibitory
activity while retaining the excellent biopharmaceutical
properties of 5-AIQ. A compound, 5-amino-3-methyliso-
quinolin-1-one (3-Me-5AIQ) was identified, which was 7
times more potent than 5-AIQ.
A rather surprising result was reported by Dr S C Sak
(CRUK, Leeds) who used immunohistochemistry to
assess expression of two DDR proteins, APE-1 and
XRCC1, in biopsy samples from 90 muscle invasive
bladder tumours. High levels of these proteins correlated
with survival after radical radiotherapy. On first reflec-
tion, high expression should protect tumours from IR.
However, others factors may be invoked to explain this.
Since the median patient age was 75 years, it is possible
that depletion of natural radioprotectors, e.g. glutathione
and other sulphydryl compounds, might allow more
DSB damage to occur per unit dose with enhanced repair
responses in patients who are cured. Another potential
explanation could be the occurrence of enhanced mis-
repair in these patients. This enigmatic finding needs
further investigation.
The response of DDR pathways following exposure to
low dose radiation (0–2 Gy) was discussed by Dr Susan
Short (Gray Cancer Institute, Northwood). She reported
the response of a number of genes in two cell lines, +/-
for low dose hypersensitivity (HRS). ATM signalling to
downstream targets such as P53, CHK1 and CHK2 is
functional at doses as low as 0.2 Gy. The induction of
DSB, measured using cH2AX, appear to be linear with
dose, but inhibition of DNA repair produces an
exaggerated effect when using ATM inhibitors. DNA-
PK inhibitors have a lesser effect on low dose responses,
but Rad51/BRCA2 mediated repair events may increase
at doses below 1 Gy, which may be applicable to normal
tissue responses during radiotherapy.
The British Journal of Radiology, February 2006
Commentary: DNA repair
Overall, the workshop provided an excellent update
on the progress of molecular targeting of DNA repair as
a strategy for enhancing anti-cancer treatments.
Significant progress has been made in recent years. The
processes are better understood and the development of
the cH2AX assay has allowed the interrogation of effects
in the low clinically relevant range. New and better
drugs are currently being tested and there is an
expectation that these strategies will be successful in
controlling at least a subset of solid tumour treatment
responses where DDR pathways are already significantly
compromised. There are some important caveats and
implications to radiotherapy, including a theoretical risk
of enhanced carcinogenesis in normal tissues; malignant
transformation assays should be performed to investi-
gate the potential magnitude of this risk and whether
there is a synergy with concomitant radiation and/or
chemotherapy. The use of proton beam radiotherapy
might allow these agents to be used more safely due to
The British Journal of Radiology, February 2006
the reduced collateral radiation of normal tissues;
intensity-modulated radiotherapy (IMRT), associated
with a dose bath effect of low to medium dose in
surrounding normal tissues would need very careful
assessment, although dose escalation may not be so
necessary in the presence of DNA repair inhibitors. It is
also self-evident that the extant mathematical models of
repair used in radiotherapy might require specific
changes to accommodate the mechanisms described in
this paper. Changes in radiotherapy fractionation (dose
per fraction and interfraction interval) might follow the
determination of precise repair capacity in tumours
relative to normal tissues. Robust laboratory, clinical
and analytical methodology is necessary in order to
determine whether enhanced cure rates and an
improved therapeutic index can be achieved by exploita-
tion of altered repair systems in some types of cancer.
Based on the content of this meeting, the prospects seem
good.
93
 The British Journal of Radiology, 79 (2006), 94–100

Patients’ perception of tests in the assessment of faecal

incontinence
1
M DEUTEKOM, PhD, 2M P TERRA, MD, 1M G W DIJKGRAAF, PhD, 2A C DOBBEN, MSc,
2
J STOKER, MD,
3 1
PhD, G E BOECKXSTAENS, MD, PhD and P M M BOSSUYT, PhD

1
Department of Clinical Epidemiology and Biostatistics, 2Department of Radiology and
3
Department of Gastroenterology from the Academic Medical Center, Amsterdam, The
Netherlands

ABSTRACT. The objective of this study was to evaluate patient perception of endoanal
MRI compared with defecography and anorectal functional testing in the workup of
patients with faecal incontinence. Consenting consecutive patients underwent a
standard testing protocol consisting of endoanal MRI, defecography and anorectal
function combination. Patient experience was evaluated with a self-administered
questionnaire, addressing anxiety, embarrassment, pain and discomfort, each
measured on a 1 (none) to 5 (extreme) point-scale. Patients were also asked to rank the
three tests from least to most inconvenient. Statistical analysis was performed with
parametric tests. Data from 211 patients (23 men; mean age 59 years (SD¡12)) were
available. MRI had the lowest average score for embarrassment and discomfort (1.6)
and defecography the highest (1.9 and 2.0, respectively) (p,0.0001, tested with general
linear model for related samples). The average pain score was lowest for MRI (1.4) and Received 26 May 2005
highest for the anorectal function combination (1.7) (p,0.0001). Level of anxiety was Revised 15 June 2005
highest for MRI (1.6 versus 1.4; p50.03). MRI was scored as least inconvenient by 69% of Accepted 22 June 2005
patients. Endoanal MRI was scored as least inconvenient. However, the differences in
DOI: 10.1259/bjr/63269033
patient burden between the three diagnostic tests were small and absolute values were
low for all tests. Patient perception will not be a key feature in determining an optimal ’ 2006 The British Institute of
diagnostic strategy in faecal incontinence. Radiology

Faecal incontinence is defined as recurrent uncon-
trolled passage of faecal material at an inappropriate
time or in an inappropriate place more than twice a
month [1]. The reported prevalence values range from
1.4% in the general population [2] to 46% in institutio-
nalized elderly [3, 4]. It is possible that the real
prevalence is even higher than reported as faecal
incontinence is associated with high social stigma and
people do not easily seek help for this disorder out of
embarrassment [5, 6]. Childbearing injuries (sphincter
and/or pudendal nerve damage) and prior anorectal
surgery (sphincter trauma) are the main causes of faecal
incontinence [7, 8].
Diagnostic tools are used to determine the exact cause of
the faecal incontinence complaints and aim to guide future
therapy. In the evaluation of faecal incontinence clinicians
can use a large variety of diagnostic tools, including
anorectal function tests and anorectal imaging techniques
after medical history and physical examination.
Treatment guidance appears to be problematic as there
exists debate in the value of diagnostic tests with respect
to treatment outcomes [9]. It has been shown that the
existence of sphincter atrophy has a negative predictive
value on the success of sphincter repair [10]. About
10 years ago, endoluminal MRI of the rectum and the
This research was supported by grant 945-01-013 of the
Netherlands Organization for Health Research and Development.
anus was introduced [11]. It appeared that, although
endoanal MRI was comparable with endoanal ultra-
sound for identifying defects of the sphincter, only
endoanal MRI could reveal thinning of the external
sphincter reflecting muscle atrophy [12].
Its ability to identify sphincter defects as well as external
sphincter atrophy makes endoanal MRI a likely candidate
for a diagnostic strategy to guide treatment decisions in
patients with faecal incontinence. Yet an optimal diagnos-
tic strategy should also try to minimize patient burden, as
extensive testing may be taxing to patients.
It has been shown in a review that anxiety-related
reactions occur in approximately 4% to 30% of patients
undergoing MRI, ranging from apprehension to severe
reactions that interfere with the performance of the test
[13]. These findings were from studies using non-
invasive MRI techniques. It could be hypothesized that
the use of an endoluminal coil could be even more
bothersome for patients, but there exist no data to either
refute or confirm this hypothesis.
We designed a study to evaluate and compare the
patient burden of diagnostic tests used in the work-up of
patients with faecal incontinence as part of a clinical
cohort study aiming to identify prognostic factors for
treatment success by physiotherapy.
We studied the perceived burden of endoanal MRI,
defecography and an anorectal function test combination
consisting of anorectal manometry, pudendal nerve

94 The British Journal of Radiology, February 2006

Patients’ perception of tests
terminal motor latency, rectal capacity measurement, anal
and rectal sensitivity measurement and endoanal
ultrasound.
Materials and methods
The clinical cohort study had started in December
2001. By February 2004, 240 consenting consecutive
patients with faecal incontinence visiting one of 16
participating medical centres (Dutch) were included in
the cohort study. The medical ethics committees of all
participating hospitals approved the study.
Patients were identified by surgeons, gastroenterolo-
gists and a gynaecologist participating in the large
diagnostic cohort study. Patients were referred to these
physicians by general practitioners or by physicians who
were not participating in the cohort study. After
receiving signed informed consent from patients, data
concerning medical history were collected by physicians.
All participating physicians used the same structured
forms for medical history. Patients were questioned
about the duration of their faecal incontinence com-
plaints. The severity of faecal incontinence was assessed
by means of an incontinence scale developed by Vaizey
[14]. This scale contains items about the type (gas, fluid,
solid) and frequency of incontinence and additional
items addressing social invalidation, the need to wear a
pad or plug, the use of constipating medication and the
presence of urge incontinence. The total score on the
Vaizey scale ranges from 0 (complete continence) to 24
(complete incontinence).
Inclusion criteria for the cohort study were the
existence of faecal incontinence complaints for 6 months
or more, a Vaizey incontinence score of at least 12, and
failure of conservative treatment, based on diet recom-
mendations and/or antidiarrhetics. Excluded were
patients aged below 18 years, patients diagnosed less
than 2 years ago with an anorectal tumour and patients
with a previous ileoanal or coloanal anastomosis. As the
clinical cohort study investigated the treatment effect of
physiotherapy, patients with chronic diarrhoea (always
fluid stools, three or more times a day), overflow
incontinence, proctitis, soiling (leakage of faecal material
out of the anus after normal defecation leading to
perineal eczema) and rectal prolapse were also excluded
from participation.
Some patient categories were excluded from one or
more tests. Defecography was not performed in females
younger than 45 years without sterilization, except on
indication based on clinical symptoms (e.g. lower
abdominal pain and/or false urge to defecate) and
clinical findings (e.g. symptoms of prolapse) because of
the cumulative radiation doses of somatic and genetic
effects. Before the MRI examination, patients were
questioned about claustrophobia and had to complete a
questionnaire comprising exclusion criteria for a MRI
examination, such as a pacemaker, claustrophobia and
other contraindications. Patients with a pacemaker were
excluded from the MRI examination, a pacemaker being
an absolute contraindication for MRI.
Only the data of patients who experienced all three test
sessions were analysed in this study.
The British Journal of Radiology, February 2006
Diagnostic tests
Patients underwent three diagnostic sessions: one with
endoanal MRI, a second with defecography and a third
with a combination of anorectal function tests consisting
of anorectal manometry, pudendal nerve terminal motor
latency, rectal capacity measurement, anal and rectal
sensitivity measurement, and endoanal ultrasound. None
of the patients received sedation for any of the tests. The
decision to evaluate the burden of the latter anorectal
function test combination was made because these tests
are usually performed in a single testing session and we
expected that patients would find it difficult to differ-
entiate the tests. Logistical considerations prevented us
from randomizing the order of the tests.
All diagnostic tests were performed according to a
standard procedure that had been established during joint
meetings of the research group members of all participat-
ing hospitals. Not all centres were equipped to perform all
tests, therefore not all patients could be tested at a single
site. Prior to testing, all patients received standard written
information concerning the tests.
MRI
Endoanal MRI visualizes the muscles of the pelvic
floor. Endoanal imaging was performed with 1 T or 1.5 T
MR (General Electric Horizon Echospeed; General
Electric, Milwaukee, IL; Philips Gyroscan ACS-NT;
Philips Medical Systems, Best, The Netherlands) clinical
closed bore units and a dedicated endoanal coil with a
diameter of 18 mm. All patients were asked to fast 4 h
prior to the MR examinations to minimize artefacts from
bowel peristalsis. In all hospitals except one, the patients
were injected intramuscularly with an antiperistaltic
drug to reduce bowel motion before the start of imaging.
No intravenous contrast medium was used. The endoa-
nal coil was covered with a condom and, after lubrica-
tion, inserted into the anal canal with the patient in a left
lateral position. After positioning of the endoanal coil,
the patients were turned to the supine position and
moved into the magnet. The patient was instructed not to
squeeze to prevent artefacts of movement. The scan
period took on average 20 min. As this test was
performed as part of a larger study, patients were also
studied with a phased array coil in the same session after
removal of the endoanal coil. No intravenous contrast
medium was used and no dynamic sequences were
performed with external phased array coil MRI. The
burden expressed by the patients for MRI was the
burden for the combination of endoanal and phased-
array MRI. The total duration for this combination was
around 40 min.
Defecography
Defecography allows an evaluation of the movements of
the rectum, insufficiency of the sphincter, presence or
absence of rectoceles, enteroceles and intussusceptions.
Patients were instructed to drink contrast medium diluted
in water prior to the examination. The test started with the
patient in left decubital position. Through an injection,
pistol barium paste (200–300 ml barium sulphate prepared
95

by the hospital pharmacy or Evacu-Paste (E-Z-EMH Inc.,
Westbury, NY)) was injected manually into the rectum. In
female patients, amidotrizoide acid 50% gel was also
injected via a syringe into the vagina. The perineum was
located with amidotrizoide acid 50% gel solution or located
by a catheter with leadmark. Subsequently, the entire X-ray
table was tilted upright 90˚ and the patient was seated on a
specially developed radiolucent defecography chair.
Defecography took approximately 15 min (room time).
After the test was performed, the patient was instructed to
drink extra to eliminate the contrast.
Anorectal function test combination
All tests were performed in left lateral position with
hips flexed to 90 ˚. Anal manometry evaluates the
muscular contraction and relaxation of the anal sphinc-
ters by the measurement of pressures in the anal canal.
Anal manometry took place according to the solid-state
or water perfused technique, without or with sleeve. The
catheter (Konigsberg Instrument Inc., Pasadena, CA;
Medtronic, Skolvunde, Denmark; Dentsleeve Pty Ltd,
Parkside, Australia) was introduced and stabilized in the
anal sphincter complex. After positioning of the catheter,
the basal sphincter pressure, maximum squeeze pressure
and rectal anal inhibitory reflex were measured.
Pudendal nerve terminal motor latency determines the
integrity of the pudendal nerve. The finger with a glove-
mounted St Mark’s Hospital electrode (Dantec;
Skovlunde, Denmark) was inserted into the rectum.
The pudendal nerve was electrically stimulated (supra
maximum stimulus of 0.05 ms) on each side near the
ischial spine.
With rectal and anal sensitivity measurements the
threshold sensation of the rectum and anus was
determined, respectively. The stimulation electrode
(Dantec Keypoint, Skovlunde, Denmark) was mounted
on a catheter and introduced into the rectum. A constant
current was increased gradually to a maximum of
20 mA. The same procedure was performed in the anus
to determine the threshold sensation of the anus.
The capacity measurement of the rectum was per-
formed by introducing a single use urinary catheter
(female, 14 Ch) with a latex balloon tied to the end,
covered with a lubricant and connected to a 50 ml
syringe, into the rectum. The balloon catheter was
inflated with air in gradual increments of 50 ml until
the maximum tolerable volume was reached. The
minimal rectal sensation perceived (sensory threshold),
the volume associated with the initial urge to defecate
(urge sensation) and the volume at which the patient
experienced discomfort and an intense desire to defecate
(the maximal tolerated volume) were determined.
Endoanal ultrasound was performed with an ultra-
sound scanner (3535 Bruel and Kjaer, Gentfofte,
Denmark; SDD-2000 Multiview Aloka, Tokyo, Japan)
with radial endoscopic probe and a 7.5 MHz transducer.
The probe was covered with a condom and, after
application of a lubricant, introduced into the anal canal
with the patient in left lateral or prone position. The
probe was slightly withdrawn so all the different levels
of the anal sphincter complex could be visualized. The
total duration of the anorectal function test combination
was between 30 min and 55 min.
96
M Deutekom, M P Terra, M G W Dijkgraaf et al
Test questionnaire
The self-administered questionnaire was handed out
by a physician before the first test was performed.
Patients were requested to take the questionnaire home
and to complete the questionnaire after their last test.
One researcher (MD) collected all completed question-
naires and contacted patients when no questionnaires
were returned. When necessary, extra questionnaires
were sent out. The questionnaire consisted of three
modules. First, a standard formatted Likert scoring
module was used with four items concerning pain,
embarrassment, discomfort, and anxiety. The first three
items have previously been used in a study of the
acceptance of CT colonoscopy by patients [15]. Based on
literature data, we added anxiety as the fourth item [13,
16, 17]. Responses were scored on a five-point scale with
1 indicating ‘‘none’’ and 5 indicating ‘‘extreme’’. By
adding the item scores, an overall burden score was
determined. Second, a comparative assessment module
was used, forcing patients to rank the different tests from
least to most inconvenient. Finally, a behavioural intent
module was used by asking patients whether or not they,
if opportune, would recommend each test to friends or
relatives. The different modules were collated into one
comprehensive questionnaire.
Statistical analyses
The general linear model for related samples was used
to compare the burden of the different tests. When a
statistical difference was found, paired t-tests were used
as post hoc tests. Subgroup analyses were performed
based on age, duration of faecal incontinence (using a
median split) and gender. Unpaired t-tests were used to
test for differences between groups of patients with
respect to sum burden scores. We also investigated with
unpaired t-tests if burden values were different depend-
ing on whether patients received their tests in a single
centre or in multiple centres. We analysed whether test
order and the time lag between date of last test and date
of completed questionnaire affected experienced burden,
using Pearson correlation coefficients.
Additional analyses were performed to study the
association between the subjective ranking of a test and
the burden variables. For each of the three tests, patients
were categorized according to position of that test in their
inconvenience ranking. We used analysis of variance to
examine differences in the amount of burden between the
different patient groups. p-values below 0.05 were con-
sidered to represent a statistically significant difference.
Results
Patient characteristics
From the 270 questionnaires distributed, 240 ques-
tionnaires were returned during the study period
(response rate: 89%). Data were missing for one or more
tests in 29 patients for various reasons: last test occurred
after completion of the questionnaire (n515), contra-
indication (n58), claustrophobia (n53) or unknown
(n53). For 211 (23 male; 188 female) patients all test
The British Journal of Radiology, February 2006
Patients’ perception of tests
Table 1. Order of testing
Test order Frequency
Anorectal function combination, 18 (17%)
defecography, MRI 14 (13%)
Anorectal function combination, MRI,
defecography 32 (30%)
Defecography, anorectal function
combination, MRI 32 (30%)
MRI, defecography, anorectal function
combination
Other 12 (11%)
data were available and could be analysed. These
patients had a mean age of 59.2 (SD¡12.2) years,
duration of incontinence 8.5 (SD¡8.4) years and Vaizey
incontinence score of 18.0 (SD¡3.1).
Order and timing of the tests
Information on the date of testing was available in 157
patients (74%). The mean duration between the first and
last test was 62 days (SD¡92). Many tests were
performed on the same day. The mean time between
last test and completion of the questionnaire was 27 days
(SD¡50). As the exact testing times were absent from a
number of patient records, the exact test order could be
derived for 108 patients (Table 1).
Test burden
The reported burden of testing was low for all three
tests, with average burden scores in the 1 to 2 range on
all four items (Figure 1).
Significant between-test differences were noted for
embarrassment, pain and discomfort as well as for the
total burden sum score. For embarrassment, discomfort
and total burden, MRI had the lowest average score
(1.56, 1.62 and 6.16, respectively) and defecography the
The British Journal of Radiology, February 2006
highest (1.92, 2.00 and 6.85, respectively) (all p,0.001).
MRI also scored lowest regarding pain (1.38), whereas
the highest pain score was observed for the anorectal
function combination (1.73) (all p,0.001). The level of
anxiety between tests also reached significance (p50.013)
with higher values for MRI (1.6) compared with for
defecography (1.4) and anorectal function tests (1.4).
Younger patients (below 59 years) had a significantly
higher total burden sum score for MRI (6.6 versus 5.7),
defecography (7.4 versus 6.3) and anorectal function tests
(7.1 versus 6.1) than older patients (Table 2). No differences
with respect to the total burden sum scores of the three tests
were observed between subgroups characterized by
gender or duration of incontinence. Whether patients
received their tests in a single centre (63%) or in multiple
centres (37%) did not influence experienced burden.
Despite these low average scores, a group of patients
(24%, n551) reported on at least one test item (anxiety,
embarrassment, pain, or discomfort) a high burden score
(4 or 5). One or more items of MRI were given a high
burden score by 12% (n525) of all patients; this
percentage was 16% (n533) for defecography and 12%
(n526) for anorectal function test combination.
Patients reporting a high burden score for at least one of
the items of MRI and anorectal function combination were
significantly younger then patients who did not report a
high burden score (55 years versus 60 years (p50.045) and
54 years versus 60 years (p50.015), respectively). There
were no other significant associations between medical
history and the group of patients that gave a high burden
score on at least one item of a test.
In a subset of 137 we analysed the effect of time lag
between last test and completed questionnaire. No
relationship was observed between total burden of any
of the tests and time-lag (MRI (r520.13; p50.14);
anorectal function tests (r520.07; p50.44); defecography
(r520.002; p50.98)).
Order of testing was present in 108 patients. Analysis
showed that test order did not influence the amount of
Figure 1. Burden scores of the three
tests in faecal incontinence with
respect to pain, embarrassment,
discomfort, anxiety and sum
burden. *Difference between 3 tests
(p,0.05). **Difference between 3
tests (p,0.001). Values indicate
mean and 95% confidence interval;
n5211. Post hoc tests showed
significantly lower burden scores for
MRI compared with the
combination of anorectal function
tests (pain and sum burden) and
defecography (embarrassment,
discomfort and sum burden).
97
 M Deutekom, M P Terra, M G W Dijkgraaf et al

Table 2. Subgroup analyses on total burden

MRI Anorectal function c. Defecography
Total burden Total burden Total burden

Gender Male 6.1 7.0 6.4

Female 6.1 p50.68 6.6 p50.73 6.9 p50.28
Age (years) , 59 6.6 7.0 7.4
. 59 5.5 p,0.01 6.3 p50.01 6.3 p50.02
Duration of ,5 6.2 6.7 6.9
incontinence (years) .5 6.0 p50.67 6.5 p50.57 6.9 p50.92
Site Single 6.0 6.9 6.9
Multiple 6.2 p50.71 6.2 p50.16 6.9 p50.94

experienced burden (MRI (p50.36); anorectal function
tests (p50.83); defecography (p50.63)).
Only a small number of patients would not recommend
one of the tests to a friend or relative: 7 for MRI (3.3%), 12
for defecography (5.6 %), and 6 for the anorectal function
test combination (2.8%). Reasons for not advising MRI
were possibly anxious reactions (n54), fear of loss of stool
(n51), headache (n51) and unknown (n51).
Defecography was not advised for various reasons: the
dislike of the ingestion of the contrast medium before
defecography (n53), experienced pain (n55), unclearness
aboutuseofresults(n52),anxiety(n51)andthelackofprivacy
(n51).Reasons fornotrecommendingtheanorectalfunction
testcombinationwerepain(n55)andlongduration(n51).
Not all patients responded to the ranking question,
therefore analyses were done on the 174 respondents
(82%). On the ranking question MRI scored best, with
120 (69%) patients scoring MRI as least inconvenient
(Figure 2). Further analysis of all three tests revealed an
association between the position in the ranking question
(from least to most inconvenient) and the reported
burden (Figure 3). Higher rankings (more inconvenient)
corresponded with a significant higher burden sum score
for that test (MRI: p,0.001, anorectal function combina-
tion: p50.03, and defecography: p,0.001).
Discussion
This study investigated and compared the burden
of endoanal MRI to defecography and the anorectal
function test combination. Although endoanal MRI was
associated with the highest level of anxiety, MRI was
found to have the lowest average scores for pain,
embarrassment, discomfort and total burden. MRI also
did well on the ranking question, with almost 70% of all
patients scoring MRI as least inconvenient.
Despite the significant differences in burden between
tests, we should note that the differences were small and
that absolute levels of burden were low for all tests. There
existed a group of patients (24%) that reported a high
score on at least one aspect of a test. These patients were
on average younger, but it appeared to be impossible to
identify on this group of patients basis of medical history.
Only a small percentage of patients would not recom-
mend one of the tests to a friend or relative.
It could be hypothesized that larger differences in
perceived burden exist across subgroups. We therefore
performed a series of subgroup analyses defined by
gender, age, and duration of incontinence. Younger
patients (below 59 years) reported a significantly higher
total burden sum score for all three tests. This finding
could possibly be explained by a diminished pelvic floor
sensory enervation in the older patient population or to
less anxiety or embarrassment related to the procedure
itself. Gender, duration of faecal incontinence com-
plaints, order of tests, single or multiple site testing
and location played no significant role in the amount of
perceived burden of the tests.
A number of potential limitations of this study should
be taken into account. The obtained results were derived
from data of patients voluntarily seeking help. It is
possible that there exists a group of patients who do not
request medical care, as they are less willing to undergo
diagnostic testing, probably having higher burden scores

Figure 2. Inconvenience ranking of

three tests in faecal incontinence.
Proportion of patients reporting the
test to be most inconvenient (black),
least inconvenient (white) or in
between (light grey).

98 The British Journal of Radiology, February 2006

Patients’ perception of tests
if ever tested. As patients seeking medical attention may
suffer more severely from their complaints they could
downgrade the burden of the testing in comparison with
the burden of their illness.
The observed imbalance between men and women in
our study is not due to a form of selection bias but is
inherent to the disorder of faecal incontinence [18].
We were forced to combine five tests into an anorectal
function combination, as these tests are usually per-
formed in a single testing session. When designing the
study we learned that patients found it difficult to
differentiate the tests during the testing sequence. We
anticipated that in between measurement might interfere
with the experience of the testing sequence and decided
to rely on a post hoc assessment of the overall burden of
the combination of tests. Unfortunately, this prevents us
from making separate statements on endosonography.
As endosonography and endoanal MRI produce com-
parable information, it would be interesting to compare
the burden of these two imaging modalities.
Although some have questioned the role of defeco-
graphy in the diagnostic work up of faecal incontinence,
our research group had decided to include this diag-
nostic modality within the cohort study. Some authors
have underscored the importance of the role of defeco-
graphy for accurately diagnosing intussusceptions and
anterior rectoceles [19] or for determining the aetiology
of outlet obstruction symptoms in patients with com-
bined faecal incontinence [20]. In a suggested work-up of
faecal incontinent patients by Felt [21], defecography was
one of the components of the diagnostic procedures. MR-
defecography is primarily employed in patients with
prolapse or constipation, while the role of MR-defeco-
graphy in incontinent patients is unclear. For this reason
MR-defecography was not part of this diagnostic cohort
study evaluating current practice.
Another possible limitation is the non-random test
order. The order in which tests were offered to patients
varied considerably, but the results of our analysis
showed that test order did not significantly affect
experienced burden.
The British Journal of Radiology, February 2006
Figure 3. Burden sum scores and
inconvenience ranking. Values
indicate mean and 95% confidence
interval.
It has been shown that past experience with testing can
influence the perception of patients of a test [22–24]. It is
unusual for patients to undergo repeated testing, so we
expected very few of our patients to have undergone one
or more of these tests previously. The low prevalence
and the lack of data prevent us from exploring explicitly
any bias due to prior experience. We believe that in our
questionnaire study total, bias is kept to a minimum.
Non-response bias is negligible as we achieved a
response rate of 89%. The questionnaires were self-
administered so there is no potential for interviewer bias.
Response bias was minimized by assuring anonymity of
the patient. Furthermore, the questionnaires were
handed out by a physician, but patients were requested
to complete the questionnaire at home. Finally, due to
the subject of the questionnaire we did not
expect patients to respond in a sociably desirable
manner.
We have tried to standardize the information given to
patients by handing out a written information sheet prior
to testing. However, we cannot claim that all patients
received exactly the same oral information by their
specialists.
To our knowledge this is the first study to investigate
the patient burden of endoanal MRI, defecography and
anorectal function test combination. The burden of MRI
has been studied before, mostly with respect to patient
anxiety. In a review by Melendez et al [13] it has been
shown that anxiety-related reactions occur in approxi-
mately 4–30% of patients undergoing MRI. In this study,
three patients did not undergo an MRI because of
claustrophobia. None of the other patients became
anxious up to a level that the test could not be executed,
and none of the examinations had to be discarded
because of motion artefacts, also associated with high
patient anxiety in the past [17, 25–27].
The percentage of patients reporting high anxiety
levels was low in comparison with other studies. One
possible reason for this could be that anxiety was
measured after the test had been performed. Various
studies have reported lower anxiety levels post-MRI
99

compared with pre-MRI [16, 25, 27]. Another possible
explanation for the lower number of patients with
anxiety reactions could lie in the fact that patients
suffered from their incontinence for a long duration
and were not afraid that MRI would reveal a certain
malignant disease. Studies have shown that test anxiety
could result from insecurity about what the test would
reveal [16, 28]. Although the amount of anxiety of MRI in
this study was low in comparison with earlier studies,
the observed values were slightly higher than those for
defecography and the anorectal function combination.
MRI scored better with respect to other variables than
defecography and the anorectal function combination.
Small but significantly lower scores for MRI were seen
for pain, embarrassment and discomfort. Total burden
sum score was also significantly lower for MRI. Because
all tests were performed as part of a larger study
designed to evaluate the diagnostic performance of these
tests, patients received phased-array MRI as well as
endoanal MRI in a single session. The present study
addressed the burden of that total MRI session. It can be
expected that the burden for a diagnostic session with
endoanal MRI only would be even somewhat smaller,
because of the shorter duration of this single procedure.
Overall, MRI was preferred more often than defecogra-
phy and functional testing, with 120 (69%) patients scoring
MRI as least inconvenient. For every test we observed a
significant relationship between the given inconvenience
rank and the burden sum score. Patients ranking a test as
least inconvenient reported significantly less burden than
patients who ranked this test as most inconvenient. We feel
confident in concluding that the burden sum score, based
on a combination of embarrassment, pain, anxiety and
discomfort, is a reflection of relative inconvenience. The
observed relation between the burden values and the
ranking question supports the construct validity of this
short and apprehensive questionnaire.
In summary, in this study, set up to investigate the
burden of diagnostic tests used in the assessment of
faecal incontinence, we found significant differences
between tests, with MRI scoring significantly better than
defecography and the anorectal function combination.
As the differences were small and the average burden
values were low for all tests, we find it safe to say that
the role of burden of testing in the search for an optimal
strategy in faecal incontinence will be limited. The
preferred diagnostic pathway will most likely be based
on maximizing diagnostic accuracy at acceptable costs.
Efforts to collect more information on test accuracy and
costs are underway.
References
1. Thomas TM, Egan M, Walgrove A, Meade TW. The
prevalence of faecal and double incontinence. Com Med
1984;6:216–20.
2. Perry S, Shaw C, McGrother C, Matthews RJ, Assassa RP,
Dallosso H, et al. Prevalence of faecal incontinence in adults
aged 40 years or more living in the community. Gut
2002;50:480–4.
3. Borrie MJ, Davidson HA. Incontinence in institutions: costs
and contributing factors. CMAJ 1992;147:322–8.
4. Johanson JF, Irizarry F, Doughty A. Risk factors for fecal
incontinence in a nursing home population. J Clin
Gastroenterol 1997;24:156–60.
100
M Deutekom, M P Terra, M G W Dijkgraaf et al
5. Mavrantonis C, Wexner SD. A clinical approach to fecal
incontinence. J Clin Gastroenterol 1998;27:108–21.
6. Jorge JM, Wexner SD. Etiology and management of fecal
incontinence. Dis Colon Rectum 1993;36:77–97.
7. Kamm MA. Obstetric damage and faecal incontinence.
Lancet 1994;344:730–3.
8. Toglia MR. Pathophysiology of anorectal dysfunction.
Obstet Gynecol Clin North Am 1998;25:771–81, vi.
9. Prather CM. Physiologic variables that predict the outcome
of treatment for fecal incontinence. Gastroenterology
2004;126 (Suppl.):s135–s140.
10. Briel JW, Stoker J, Rociu E, Lameris JS, Hop WCJ, Schouten
WR. External anal sphincter atrophy on endoanal magnetic
resonance imaging adversely affects continence after
sphincteroplasty. Br J Surg 1999;86:1322–7.
11. Stoker J, Rociu E. Endoluminal MR imaging of anorectal
diseases. J Magn Reson Imaging 1999;9:631–4.
12. Rociu E, Stoker J, Eijkemans MJ, Schouten WR, Lameris JS.
Fecal incontinence: endoanal US versus endoanal MR
imaging. Radiology 1999;212:453–8.
13. Melendez JC, McCrank E. Anxiety-related reactions asso-
ciated with magnetic resonance imaging examinations.
JAMA 1993;270:745–7.
14. Vaizey CJ, Carapeti E, Cahill JA, Kamm MA. Prospective
comparison of faecal incontinence grading systems. Gut
1999;44:77–80.
15. Van Gelder RE, Birnie E, Florie J, Schutter MP, Bartelsman
JF, Snel P, et al. A comparison of patient preference of CT
colonography and colonoscopy: a five-week follow-up
study. Radiology 2004;233:328–37.
16. Katz RC, Wilson L, Frazer N. Anxiety and its determinants
in patients undergoing magnetic resonance imaging. J
Behav Ther Exp Psych 1994;25:131–4.
17. MacKenzie R, Sims C, Owens RG, Dixon AK. Patients’
perceptions of magnetic resonance imaging. Clin Radiol
1995;50:137–43.
18. Henry MM. Pathogenesis and management of fecal incon-
tinence in the adult. Gastroenterol Clin North Am
1987;16:35–45.
19. Jorge JM, Habr-Gama A, Wexner SD. Clinical applications and
techniques of cinedefecography. Am J Surg 2001;182:93–101.
20. Wiersma TG, Mulder CJ, Reeders JW. Dynamic rectal
examination: its significant clinical value. Endoscopy
1997;29:462–71.
21. Felt-Bersma RJF, Cuesta MA. Fecal incontinence 1994 -
which test and which outcome. Neth J Med 1994;44:182–8.
22. Mueller PR, Biswal S, Halpern EF, Kaufman JA, Lee MJ.
Interventional radiologic procedures: patient anxiety, per-
ception of pain, understanding of procedure, and satisfac-
tion with medication–a prospective study. Radiology
2000;215:684–8.
23. Murphy KJ, Brunberg JA. Adult claustrophobia, anxiety
and sedation in MRI. Magn Reson Imaging 1997;15:51–4.
24. Thorp D, Owens RG, Whitehouse G, Dewey ME. Subjective
experiences of magnetic resonance imaging. Clin Radiol
1990;41:276–8.
25. Dantendorfer K, Amering M, Bankier A, Helbich T, Prayer
D, Youssefzadeh S, et al. A study of the effects of patient
anxiety, perceptions and equipment on motion artifacts in
magnetic resonance imaging. Magn Reson Imaging
1997;15:301–6.
26. Sarji SA, Abdullah BJ, Kumar G, Tan AH, Narayanan P.
Failed magnetic resonance imaging examinations due to
claustrophobia. Australas Radiol 1998;42:293–5.
27. Quirk ME, Letendre AJ, Ciottone RA, Lingley JF. Anxiety in
patients undergoing MR imaging. Radiology 1989;170:
463–6.
28. Gebbensleben B, Rohde H. Anxiety before gastrointestinal
endoscopy–a significant problem? Dtsch Med Wochenschr
1990;115:1539–44.
The British Journal of Radiology, February 2006
The British Journal of Radiology, 79 (2006), 101–109

Enhanced visualization and quantification of magnetic resonance

diffusion tensor imaging using the p:q tensor decomposition
1,2,3
A PEÑA, PhD, 1,3H A L GREEN, MBChB, 3T A CARPENTER, PhD,
1,2,3
S J PRICE, FRCS,
1,3
J D PICKARD,
2,3
MChir, FRCS, FMedSci and J H GILLARD, BSc, MD, FRCR

Departments of 1Neurosurgery, 2Radiology and the 3Wolfson Brain Imaging Centre,

Addenbrooke’s Hospital and the University of Cambridge, Cambridge CB2 2QQ, UK

ABSTRACT. Many scalar measures have been proposed to quantify magnetic resonance
diffusion tensor imaging (MR DTI) data in the brain. However, only two parameters are
commonly used in the literature: mean diffusion (D) and fractional anisotropy (FA). We
introduce a visualization technique which permits the simultaneous analysis of an
additional five scalar measures. This enhanced diversity is important, as it is not known a
priori which of these measures best describes pathological changes for brain tissue. The
proposed technique is based on a tensor transformation, which decomposes the diffusion
tensor into its isotropic (p) and anisotropic (q) components. To illustrate the use of this
technique, diffusion tensor imaging was performed on a healthy volunteer, a sequential
study in a patient with recent stroke, a patient with hydrocephalus and a patient with an
intracranial tumour. Our results demonstrate a clear distinction between different
anatomical regions in the normal volunteer and the evolution of the pathology in the
patients. In the normal volunteer, the brain parenchyma values for p and q fell into a
narrow band with 0.976,p,1.063 6 1023 mm2 s21 and 0.15,q,1.08 6 1023 mm2 s21.
The noise appeared as a compact cluster with (p,q) components (0.011, 0.141) 6
1023 mm2 s21, while the cerebrospinal fluid was (3.320, 0.330) 6 1023 mm2 s21. In the
stroke patient, the ischaemic area demonstrated a trajectory composed of acute, sub- Received 22 April 2003
acute and chronic phases. The components of the lesion were (0.824, 0.420), (0.884, 0.254), Revised 24 May 2005
(2.624, 0.325) at 37 h, 1 week and 1 month, respectively. The internal capsule of the Accepted 1 June 2005
hydrocephalus patient demonstrated a larger dispersion in the p:q plane suggesting
DOI: 10.1259/bjr/24908512
disruption. Finally, there was clear white matter tissue destruction in the tumour patient.
In summary, the p:q decomposition enhances the visualization and quantification of MR ’ 2006 The British Institute of
DTI data in both normal and pathological conditions. Radiology

Magnetic resonance (MR) diffusion tensor imaging
(DTI) is a technique which allows the in vivo measure-
ment of water diffusion in biological tissues from which
tissue microstructure can be inferred [1–5]. It has been
used successfully to investigate a number of neurological
disorders that involve the disruption of white matter
fibres including schizophrenia [6], head injury [7],
multiple sclerosis [8] and stroke [9, 10]. In addition,
DTI data can be used with a set of computational
techniques called ‘‘tractography’’ [11] to reconstruct in
vivo white matter tracts in the human brain, which is a
very promising field, for example, to investigate their
disruption due to an expanding tumour [12].
Diffusion is properly described by a high-dimensional
mathematical quantity called a tensor. A tensor represents
the generalization of scalars and vectors and as such, it
contains more information than these. In three dimensions
a scalar has one element, a vector three elements and a
tensor nine elements. In order to quantify pathological
changes in the diffusion tensor, a transformation is
required which reduces the dimensionality of the tensor
and to this end a number of tensor scalar measures have
Address correspondence to: Dr Jonathan H Gillard, University
Department of Radiology, Addenbrooke’s Hospital, Cambridge CB2
2QQ, UK.
been proposed [1, 13]. From a theoretical point of view,
tensor calculus establishes that many such measures exist.
These include the lattice index (LI), relative anisotropy
(RA), fractional anisotropy (FA), the volume ratio (VR) and
ratios of the various eigenvalues (li), the mean diffusivity
(D), the Euclidean length of the tensor (L), its anisotropy
angle (w) and any algebraic combination of the first, second
and third invariants of the tensor [14].
From a practical point of view, however, only a limited
number of these measures are actually used in clinical
studies. In the MR DTI literature, the most common of
these measures are FA and D. Out of 30 recent studies on
clinical applications of DTI, encompassing diseases such
as schizophrenia, Alzheimer’s disease, stroke, multiple
sclerosis and head injury, 26 reported their results using
both FA and D [7, 9, 12, 15–36], while only four reported
D alone [37–40].
The caveat with exclusively using D and FA to
characterize pathology in clinical applications is that it is
not known a priori which tensor scalar measure is the most
appropriate to quantify pathological changes in brain tissue.
It is conceivable, for example, that a study might fail to
show significant changes when the diffusion tensor is
measured using FA but it may show differences when
using RA or L or some other measure. We have previously

The British Journal of Radiology, February 2006 101


shown this to be the case in acute stroke [34]. The
identification of which is the ‘‘best’’ measure of the
diffusion tensor is an empirical process, which will only
be resolved after a large number of experiments are
conducted and corroborated with external empirical infor-
mation, such as histology. In these circumstances it seems
reasonable to analyse as many scalar measures as possible,
and not rely solely on D and FA.
This article explores the novel application of a
mathematical technique to enhance the visualization
and quantification of brain tissue in MR DTI, which we
will term ‘‘p:q decomposition’’. The technique is based on
a tensor transformation, which decomposes the diffusion
tensor into its isotropic (p) and anisotropic (q) compo-
nents. In contrast to the standard practice in the literature
where only D and FA are analysed, this technique
permits the visualization simultaneously of seven scalar
measures. These are D, p, q, RA, FA, w, and L. This
technique is based on a classical decomposition used in
tensor calculus, already observed by the major contribu-
tions of Basser et al [13] and Pierpaoli et al [5], but which
has not been applied yet to visualize and quantify the
diffusion tensor in MR DTI.
In the following sections we will describe the
technique and apply it to data from a control volunteer
and three clinical examples.
Materials and methods
Theory
Diffusion in tissue can be mathematically represented as
a second-order Cartesian tensor, which in matrix form is:
2 3
Dxx Dxy Dxz
6 7
Dij ~4 Dyx Dyy Dyz 5 ð1Þ
Dzx Dzy Dzz
Given that the tensor is symmetric along its principal
diagonal, i.e. Dyx5Dxy, it has only six independent
components. From the tensor, the eigenvalues li are
calculated using a standard methodology, such as
singular value decomposition [41], as l1, l2 and l3.
According to tensor calculus, based on the eigenvalues
many possible scalar measures of the diffusion tensor
can be devised [42, 43].
The standard methodology in the DTI literature,
however, is to calculate only two scalar measures of
Dij. These are the mean diffusion (D) defined as:
1 l1 zl2 zl3
D~ tr(Dij )~
3 3
where tr represents the trace of the tensor, and the
fractional anisotropy (FA) or the relative anisotropy (RA)
defined as:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rffiffiffi 2 2
3 (l1 {D) z(l2 {D) z(l3 {D)
FA~ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 subtended between the p axis and a line originating in
l21 zl22 zl23
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rffiffiffi 2 2
1 (l1 {D) z(l2 {D) z(l3 {D)
RA~
3 D
102
A Peña, H A L Green, T A Carpenter et al
The proposed technique is based on a classical tensor
decomposition, already observed by Basser et al [13] and
Pierpaoli et al [5]. Our contribution is to construct a
graphical representation of the diffusion tensor based on
this decomposition. This transformation has its conceptual
roots in the mathematical theory of continuum mechanics
[44, 45]. We will term the technique p:q decomposition.
Using this methodology, the first step is to decompose
the diffusion tensor from Equation (1) according to the
next equation:
Dij ~DIij z½Dij
DIij  ð4Þ
into two tensors P and Q, i.e. Dij5Pij+Qij. Here Iij is the
identity tensor Iij5diag(1,1,1). The first term on the right
hand side of Equation (4) is the isotropic tensor, while
the second term (in brackets) represents the deviatoric
tensor. The magnitude of these tensors can be denoted by
its isotropic (p) and anisotropic (q) components. The
values of p and q can be computed as:
pffiffiffi
p~ 3D ð5Þ
and
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
q~ (l1 {D)2 z(l2 {D)2 z(l3 {D)2 ð6Þ
According to these definitions p is therefore a scaled
measure of the mean diffusion in the tensor, while q is a
measure of the variance or deviation of the eigenvalues
with respect to the mean diffusion of the tensor.
The second step is to plot each tensor as a point in a
Cartesian plane with p taken as the x-axis and q as the y-
axis, as in Figure 1a. This plane will be denoted as the p:q
plane. The effect of this transformation is to reduce the
dimensionality of the tensor from six dimensions to two.
The third step is to use the p:q plane to deduce the five
additional tensor scalar measures: D, RA, FA, L and w.
Four of these seven tensor measures (q, RA, FA, w) are
anisotropy measures, while D and p are measures of the
magnitude of diffusion and L is a measure of the total
diffusion of the tensor. D, p, q and L have units of
1023 mm2 s21, RA and FA are dimensionless, and w has
units of degrees.
These scalar measures can be deduced either analyti-
cally or graphically. The analytical method is to directly
compute the measures based on the p,q components
using the formulae:
rffiffiffi  
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi q 3q
2 2 {1 q
L~ p zq , RA~ , FA~ , w~ tan ð7Þ
p 2L p
ð2Þ
However, the real advantage of using the p:q plane is that
we can obtain these values directly from the graph as
follows. Consider a tensor A from which we can com-
pute its location in the p:q plane using Equations (5) and
(6) as po and qo. Therefore, Lo is the distance between the
2
origin of coordinates and the point (po,qo); wo is the angle
ð3Þ the centre of coordinates and passing through point
2 (po,qo), i.e. the segment Lo; RAo is the ratio between qo and
po;ffiffiffiffiffiffi
q FAo is the ratio between qo and Lowith scale factor
3= &1:22; D is p1ffiffi &0:57 of the value in the p axis.
2 3
The British Journal of Radiology, February 2006
Enhanced visualization and quantification in MR DTI
Figure 1b illustrates the geometrical relationship
between these various quantities.
Data acquisition
In order to illustrate the use of the p:q decomposition
with clinical data, four representative cases were selected.
The first comprised five regions of interest (ROIs) in a
healthy 27-year-old volunteer (to illustrate spatial varia-
tion in the tensor field, as in Figure 1c). The second reports
the findings in a 76-year-old hypertensive woman who
presented with a sudden onset of expressive dysphasia
and right-sided hemiparesis. Imaging of her left middle
cerebral artery territory stroke was undertaken at 37 h, 1
week and 3 months from stroke onset to illustrate the
temporal variation in the tensor field (Figure 1d). The third
case is a 85-year-old female hydrocephalus patient with a
history of gait ataxia, falls and memory problems. And
finally, the fourth case is a 46-year-old male patient with a
WHO Grade II oligodendroglioma.
The Local Research Ethics Committee approved the
study and informed consent was obtained. The diffusion
tensor data sets were acquired using a 5 mm slice
thickness. Imaging was performed on a 3 Tesla magnetic
resonance machine (Bruker Medspec S300; Bruker
Medical, Ettlingen, Germany). A single shot spin echo,
echo planar imaging technique, with Stejskal-Tanner
diffusion sensitizing pulses [46] was used. Imaging
parameters were: repetition time (TR)55070 ms, echo
time (TE)5107 ms, a590˚, d521 ms and D566 ms. Eight
interleaved supratentorial slices were acquired with a
phase template in a near axial plane, using a 128 6 128
matrix, field of view of 25 cm 6 25 cm. For each slice,
images were collected from 12 non-collinear gradient
The British Journal of Radiology, February 2006
Figure 1. (a) A point representing a
sample of tissue in the p:q plane.
The x axis corresponds to the
isotropic component of diffusion (p)
and the y axis the anisotropic
component of diffusion (q). Any
tensor can be decomposed into its p
and q components po and qo, which
will correspond to a point in the p:q
plane. (b) Starting from a point in
the p:q plane, we can deduce the
standard anisotropy measures RA
and FA using simple geometry. Both
of these measures will be
proportional to the angle w, in fact
RA is proportional to the tangent
and FA proportional to the sine. (c)
Two tissues will, in general, have
different p and q components. Thus
a tissue A with components pA and
qA, will have a different location
from a tissue B with components pB
and qB. (d) A tissue A in general will
have different p and q components
at different times. By plotting these
different components in the p:q
plane we can obtain a trajectory
that illustrates the evolution of
tissue in time. In this example we see
a trajectory demonstrating three
time points for tissue A.
directions [47]. For each gradient direction an unweighted
bo image and five diffusion weighted images were
collected at equally spaced b-values in the range
bmin5318 s mm22 to bmax51541 s mm22. Using a spe-
cially-written program in MATLAB (The MathWorks Inc.,
Natick, MA) the diffusion tensor was computed on a voxel
by voxel basis, using a singular value decomposition
algorithm to fit the signal intensities to the Stejskal-Tanner
equation, following the method proposed by Basser et al
[1, 2]. From the tensor, the p and q components were
calculated based on Equations (5) and (6), and D, RA, FA,
w, and L using Equations (2) and (7).
Results
Normal volunteer
D and FA maps of the volunteer were used to select
square anatomical ROIs of 5 6 5 voxels, which were
subsequently averaged to obtain a mean value for the
ROI. These were placed in the corpus callosum (CC),
occipital cortex (Cx), cerebrospinal fluid (CSF), internal
capsule (IC) and noise regions, as illustrated in
Figure 2a,b. When these ROIs were plotted in the p:q
plane, they formed clearly segregated clusters (Figure 2c).
The spherical diffusion and deviatoric diffusion fell
within a narrow band with 0.976,p,1.063 6
1023 mm2 s21 for the three structures in brain parench-
yma, i.e. CC, IC, Cx, D50.607, 0.563, 0.613 6
1023 mm2 s21 and p51.052, 0.976, 1.063 6
1023 mm2 s21, respectively. In contrast, all the other
measures varied substantially, with a range bounded by
maximum values for the CC (q51.086, L51.032) 6
1023 mm2 s21 and RA50.879, 45.9 ˚, FA51.512, and
103
 A Peña, H A L Green, T A Carpenter et al

(a) (b)

(c)

Figure 2. (a) Map of the mean diffusion (D) for a horizontal slice of the normal volunteer investigated, demonstrating the
regions of interest used in the study, which are (from top to bottom) noise (N), cerebrospinal fluid (CSF), internal capsule (IC),
splenium corpus callosum (CC) and occipital cortex (Cx). The scale on the right indicates the magnitude of D. (b) Map of the
fractional anisotropy (FA) for the same horizontal slice of the normal volunteer investigated, demonstrating the location of the
same regions of interest. The scale on the right indicates the dimensionless magnitude of FA. (c) p:q plane illustrating the
defined regions of interest (ROIs) in the normal volunteer. Three clusters are observed for the noise (N), with small components
for both p and q. The three parenchyma ROIs (CC, IC, Cx) are located along a line with approximately the same value of p, but
significantly different values of q. The CSF has a much larger dispersion and a larger value of mean diffusion.

minimum values for the Cx (q50.150, L51.073) 6
1023 mm2 s21 and RA50.141, 8.06 ˚, FA50.171.
Noise appeared as a cluster close to the origin of the
coordinates (D50.007, p50.011, q50.141) 6 1023 mm2 s21,
but whose additional scalar measures were amongst the
highest (L51.073 6 1023 mm2 s21, RA512.145, 85.29˚,
FA51.220. CSF presented the opposite characteristics,
being the most distant to the origin of coordinates
(D51.917, p53.320, q50.330) 6 1023 mm2 s21, and having
small additional scalar measures (L51.073 6
1023 mm2 s21, RA50.099, 5.68 ˚, FA50.121).
Statistically significant differences between the various
ROIs were investigated using unpaired Student’s t-tests.
The isotropic diffusion (p) between the noise and the
parenchyma (CC, IC, Cx) was significantly different (p-
value,0.01), and between the CSF and the parenchyma
(p-value,0.01). It was significantly different between the
CC and the IC (p-value,0.01), between the IC and the Cx
(p-value,0.05), but not between the CC and the Cx (p-
value50.5473). The deviatoric diffusion (q) was different
for the three parenchyma ROIs (CC, IC, Cx). It was
significantly different between the CC and IC (p-
value,0.01), between the CC and Cx (p-value,0.01),
and between the IC and Cx (p-value,0.01).
Results for the seven scalar measures are presented in
Table 1.
Stroke patient
Lesion and contralateral control square anatomical
ROIs of 5 6 5 voxels, were selected in the stroke patient
using the FA and D maps at 37 h, 1 week and 1 month

104 The British Journal of Radiology, February 2006

Enhanced visualization and quantification in MR DTI

Table 1. Data from the MR diffusion tensor imaging (DTI) acquisitions for the normal volunteer (upper set) and the stroke
patient (lower set)

Normal volunteer

D p q RA FA w L

Corpus callosum 0.607 1.052 1.086 1.032 0.879 45.92 1.512

Internal capsule 0.563 0.976 0.684 0.701 0.703 35.05 1.192
Cortex 0.613 1.063 0.150 0.141 0.171 8.06 1.073
Noise 0.007 0.011 0.141 12.145 1.220 85.29 0.142
Cerebrospinal fluid 1.917 3.320 0.330 0.099 0.121 5.68 3.336
Stroke patient
D p q RA FA w L

37 h (lesion) 0.476 0.824 0.420 0.510 0.556 27.03 0.925

37 h (control) 0.666 1.154 0.475 0.412 0.466 22.39 1.248
1 week (lesion) 0.511 0.884 0.254 0.287 0.338 16.02 0.920
1 week (control) 0.709 1.229 0.777 0.632 0.654 32.29 1.454
1 month (lesion) 1.515 2.624 0.325 0.124 0.150 7.07 2.644
1 month (control) 0.743 1.288 0.586 0.455 0.507 24.48 1.415
Mean diffusion (D), isotropic component of diffusion (p), anisotropic component of diffusion (q), anisotropy angle (w), fractional
anisotropy (FA), relative anisotropy (RA), Euclidean length (L).

(Figure 3a). To clarify presentation and due to the Tumour patient

multiple number of ROIs used, all variables analysed
were averaged within each ROI to obtain a mean value
and standard deviation. For each ROI the mean is in the
crossing of the bars which represent the magnitude of
standard deviation. The lesion (ischaemic region) in the
stroke patient described a trajectory in the p:q plane
composed of three phases, which occupied the regions:
0.824,p,2.624 6 1023 mm2 s21 and 0.25,q,0.42 6
1023 mm2 s21. The (p,q) components of the lesion ROI
were (0.824, 0.420), (0.884, 0.254), (2.624, 0.325) at 37 h, 1
week and 1 month, respectively. All these results are
shown in Figure 3b.
The corresponding contralateral control ROIs
(Figure 3a, shown in blue), in contrast, demonstrated
only a small degree of change, remaining in the region
1.154,p,1.288 6 1023 mm2 s21 and 0.47,q,0.77 6
1023 mm2 s21. The (p,q) components of the control ROI
were (1.154, 0.475), (1.229, 0.777), (1.288, 0.586) at 37 h, 1
week and 1 month, respectively.
Results for the seven scalar measures are presented in
Table 1.
Hydrocephalus patient
We have investigated microstructural changes in the
internal capsule associated with the ventricular dilatation
in this patient. For this analysis, four ROIs have been
selected (two in the patient bilaterally and two in the control
volunteer bilaterally) in one axial slice corresponding to the
posterior limb of the internal capsule (IC) at the level of the
foramen of Monro. Each ROI was composed of nine voxels.
As a comparison, the same regions were selected in a
control volunteer in the same manner. The results were: for
the patient p51.29¡0.466, q50.85¡0.054 for the left IC and
p50.96¡0.195, q50.86¡0.147 for the right IC; and for the
control, p51.01¡0.023, q50.80¡0.084 for the left IC, and
p50.96¡0.034, q50.84¡0.159 for the right IC. All the p, q
units are in 1023 mm2 and are illustrated in Figure 4. These
results demonstrate that while the mean values of the four
ROIs are roughly similar, there is a marked increase in the
dispersion of the voxels in the IC of the patient.
In this patient we investigated the hypothesis that
there are significant differences between the microstruc-
ture in the tumour region and an equivalently-located
contralateral control. For this analysis, two ROIs have
been selected in the patient in one axial slice: one in the
tumour region (the posterior pericallosal region) and
another in the equivalent normal-appearing contralateral
control region. For the tumour: p52.26¡0.210, q5
0.37¡0.049, while for the contralateral control region p
51.331¡0.128, q50.71¡0.134. All the p, q units are in
1023 mm2 and are illustrated in Figure 5. These results
demonstrate both an increase in the isotropic component
of diffusion and a decrease in the deviatoric component.
We can interpret these changes as suggesting a loss in the
microstructure of tissue, as a decrease in the density of
fibres (more intercellular space) and a loss in fibre
coherence. These changes are consistent with previous
reports [48].
Discussion
We have presented a technique that permits the
simultaneous visualization of multiple tensor scalar
measures from MR DTI data. In particular, we have
shown how, from a single graph (the p:q plane), it is
possible to deduce seven scalar measures of the diffusion
tensor, including D, p, q, RA, FA, w, and L. This
represents an improvement on the standard methodol-
ogy in the MR DTI literature in which only two scalar
measures (typically FA and D) are displayed.
There have already been a number of studies in the
literature that have considered plotting simultaneously
two tensor scalar measures, particularly FA and D. These
include Pierpaoli et al [5] who distinguished various
brain regions based on decomposing the tensor in terms
of D and the volume ratio (VR). Werring et al [8] in an
investigation of normal-appearing white matter lesions
in multiple sclerosis, and Wieshmann et al [49] and Jones
et al [50] have also demonstrated the potential of plotting
FA and D simultaneously. Plotting D vs FA, however,

The British Journal of Radiology, February 2006 105

 A Peña, H A L Green, T A Carpenter et al

(a) (b)

Figure 3. (a) Mean diffusion (D) and fractional anisotropy maps (FA) for a stroke patient at three time points: 37 h (left column),
1 week (central column) and 3 months (right column). FA is lower row and D is upper row. These maps demonstrate the regions
of interest (ROIs) used in this study. The lesion ROIs are presented in orange and the control ROIs in green. Each of these ROIs
consisted of 5 6 5 6 1 voxels. (b) This figure illustrates the p:q plane for the stroke patient, with lesion and control ROIs at (a)
37 h, (b) at 1 week and (c) at 3 months. The control ROIs are denoted in blue and the lesion ROIs in red. The arrows demonstrate
the trajectory followed by the lesion in this patient and show schematically how, while the control ROIs remain in roughly the
same region in the p:q plane, the ischaemic lesion demonstrates a trajectory composed of acute (reduction in p, reduction in q),
subacute (normalization of p while q remains low) and chronic (increased p while q remains low) phases. The inset shows
schematically the location of the lesion ROI with respect to the control ROI and a line of constant fractional anisotropy. As FA is
function of the angle w, the figure indicates that at 37 h the lesion has a higher FA than the control, while at 1 week it has a
lower FA than the control.

does not allow visualization or quantitative analysis of
the other scalar measures (e.g. p, q, RA, w, L) from a single
graph, while our study does.
To illustrate the use of our method for clinical data, we
have applied it to a healthy volunteer, a sequential study
in a patient with recent stroke, a patient with hydro-
cephalus and a patient with an intracranial tumour.
In all cases the p:q plane offers the analyst a concise
and easy-to-use representation of the diffusion tensor.
The first case illustrates the spatial variation in the tensor
field and statistically significant differences between
different tissue types (e.g. grey matter, white matter),
while the second case illustrates the temporal variation
in the tensor field and thus the evolution of the lesion
(e.g. lesion, contralateral control).
We propose that the p:q decomposition is a powerful
aid not only in the visualization of the data, but also in its
analysis, by offering a unique opportunity to assess the
additional non-standard tensor scalar measures (p, q, w,
L) and their relationship with the standard measures (D,

Figure 4. This is the p:q diagram for

the internal capsule (IC) of a
hydrocephalus patient. Regions of
interest (ROIs) have been selected
on the IC bilaterally at the level of
the foramen of Monro. The same
ROIs have been selected in a control
subject. ROI location is shown in the
insets (patient, above; control,
below). The p:q diagram
demonstrates an increased
dispersion (disorganization) of the
white matter tracts of the IC in the
hydrocephalus patient as compared
with the control.

106 The British Journal of Radiology, February 2006

Enhanced visualization and quantification in MR DTI
RA, FA). In this context, several interesting observations
from both the normal volunteer and the stroke patient
have been possible by using the p:q technique.
Normal volunteer
The first observation is that the p:q plane provides a
graphical means to understand the complex equations
that describe RA and FA. From Figure 1b and Equation
(7), one can easily observe that both RA and FA are
composite measures of other more basic tensor quan-
tities. In particular, RA is simply the ratio between q and
p, q ffiffiffiffiffiffi FA the ratio between q and L scaled by a factor
and
of 3=2&1:22.
The second observation is that the p:q plane explains
some anomalies when using FA and RA as measures of
anisotropy. For example, if we take the value of FA for
noise from Table 1, we obtain the theoretical maximum
FA value of 1.22. This is confusing, as one would not
expect empty space to have a large degree of organiza-
tion (anisotropy). This paradoxical result is in fact a
methodological artefact in using FA as an anisotropy
measure, and is due to the presence of L in the
denominator of FA. Given that the diffusion of empty
space should be zero (or close to zero due to experi-
mental error), a very small L will imply a very large FA.
Thus FA is a measure of tissue anisotropy, but weighted
by its total diffusion. The same argument applied to the
value of RA, which gives the enormous value of 12
(while the corpus callosum, for instance, is 1.032).
The third observation is the insight that q might offer
as a measure for the background noise in the data.
Equation (6) can be interpreted in statistical terms such
that q is a measure of the variance of the eigenvalues of
the tensor with respect to the mean diffusion D.
Therefore, isotropic elements in the data (such as the
empty space, Cx and the CSF) should theoretically have
all the eigenvalues equal and thus a q equal to zero.
However, due to experimental error, background noise
and other MR acquisition influences, there is a small
The British Journal of Radiology, February 2006
Figure 5. This is the p:q diagram of
a patient with a Grade II
oligodendroglioma in the posterior
pericallosal white matter. The
location of the regions of interest
(ROIs) is indicated in the inset
(above), and the patient’s MR fluid
attenuation inversion recovery
(FLAIR) image (below). Compared
with the control region, the tumour
region demonstrates both an
increase in isotropic diffusion (p)
and a decrease in deviatoric
diffusion (q). This tissue signature is
consistent with the destruction of
white matter tracts in the tumour
region.
discrepancy and the eigenvalues are not exactly the
same. Our results show that Cx and the noise have
approximately the same q values (0.141 and 0.15), while
the CSF presented a larger dispersion (q50.330), which
might be attributed to the contribution of diffusion and
bulk flow during the acquisition time.
Stroke patient
The first observation is the ability of the p:q planes to
visually convey simultaneous changes in the isotropic
and anisotropic components of the diffusion tensor as
they change in time, in other words the ‘‘trajectory’’ of
the tensor. In addition to the qualitative nature of the
trajectory, the magnitude of tensor changes can be read
directly from the p and q axes of the plots. In our
example, the trajectory describing the lesion evolution is
composed of three segments or phases (Figure 3b),
which can be interpreted in biological terms as the acute
(reduction in p, reduction in q), sub-acute (pseudonor-
malization of p, while q remains reduced) and chronic
(increase in p, while q remains reduced) phases that have
been well-documented in association with stroke [9].
A second observation demonstrates another methodo-
logical artefact or anomaly of FA and RA. Close
inspection of Figure 3b demonstrates that the lesion
clusters (shown in red) with respect to the control
clusters (shown in blue) are displaced first above a line of
constant FA in the acute phase (37 h) and subsequently
below this line in the sub-acute phase (1 week), as shown
in the inset. As both RA and FA are functions of the
angle w, this would imply that they are increased in the
lesion as compared with the control, which is absurd.
This paradox of increased tissue anisotropy (as mea-
sured in terms of RA or FA) was reported by Nusbaum
and colleagues [51] in normal ageing. As we have
explored in more detail in the case of acute stroke [34],
the p:q technique provides a graphical explanation of
why this can be the case, and that this apparent increase
in anisotropy (as measured in terms of RA or FA) can be
107

purely a graphical consequence of the manner in which
FA and RA are calculated and thus a methodological
artefact. Quantitatively, the anisotropy measured with
FA and RA of the lesion’s acute phase (37 h): RA
increased from 0.412 to 0.510 (or +24%) and FA increased
from 0.466 to 0.556 (or +19%). In contrast, q decreased
from 0.475 to 0.420 (or 212%). This behaviour suggests,
albeit tentatively, that theoretically q may detect early
changes in tissue anisotropy that are misrepresented by
RA and FA.
A third observation is that during the sub-acute phase
(1 week after the stroke), the best sensitivity to the
pathology is offered by q rather than by FA or RA. q was
reduced from 0.777 to 0.254 (267%), In contrast, FA and
RA decreased by the smaller amounts of 0.632 to 0.287
(255%), and 0.654 to 0.338 (248%), respectively.
Based on the previous observations, we can conclude
that, at least in some circumstances, some non-standard
anisotropy measures (such as q) can provide a higher
sensitivity to detect pathological conditions than stan-
dard measures such as RA and FA. We have also shown
that RA and FA have the potential to give ‘‘paradoxical’’
results and thus must be used with caution. However,
this analysis does not resolve what is perhaps the most
important question in MR DTI: from all the various
tensor measures, which one is the best one to character-
ize damage to brain tissue? As has been recently noted
by Pierpaoli et al [23], the fact remains that we do not
know a priori which is the best measure because this is
not a theoretical question but an empirical one. It is
equivalent to asking which statistical measure, e.g. the
mean or the variance for example, will better describe a
population. They describe different aspects of a popula-
tion and therefore will be useful in answering different
questions. Tensor calculus can only help by defining
which measures can be used in our analysis. Which one
best describes some aspect of the brain (be it a tissue type
or a pathological condition, such as oedema or necrosis)
can only be answered empirically, by relating the
observed tensor measures with independent biological
data such as histology, other imaging modalities and/or
cognitive tests.
Hydrocephalus patient
The disruption observed in IC using the p:q diagram
from the hydrocephalus patient is encouraging. In
patients with hydrocephalus it is common to observe
clinical symptoms that are thought to be associated with
the disruption of deep white matter tracts. Similar
findings were observed in this patient at the level of the
internal capsule (IC). It has been suggested that during
ventricular dilatation, these tracts are being stretched and
thus become mechanically compromised. Our results
support this notion by demonstrating that the MR DTI
diffusion signature of the IC is altered. In particular, this
disruption is not due to changes in the anisotropy of tissue
but to changes in its mean diffusion. This suggests that the
white matter tracts have been disrupted.
Tumour patient
In the case of the tumour patient, the p:q decomposi-
tion was useful to illustrate simultaneously changes in
108
A Peña, H A L Green, T A Carpenter et al
both the isotropic and the anisotropic components of the
diffusion tensor. There was a decrease in anisotropy (q)
and an increase in mean diffusion (p). These changes are
thought to be associated with white matter destruction.
There are a number of limitations in using the p:q
decomposition. Just like D, RA and FA, the location of
tissue in the p:q plane does not give information about
the directionality of diffusion. Also, there are other
important tensor scalar measures that are not directly
conveyed by the p:q plane, such as the eigenvalues and
the tensor invariants. Finally, from a practical point of
view, the p:q decomposition must be applied to other
brain pathologies in order to establish how beneficial it
might be in those situations.
Conclusion
The p:q tensor decomposition enhances the visualiza-
tion and quantification of MR DTI data in both normal
and pathological conditions. In particular it is an aid to
visualize simultaneously seven scalar tensor measures.
We have also shown the pitfalls of using FA and RA
exclusively, and the potential of using other tensor
measures, particularly q. However, it is important to note
that, despite the enhanced visualization and quantifica-
tion provided by our technique, the choice of which
tensor scalar measure best describes brain tissue and its
changes remains an empirical matter. We hope that the
enhanced repertoire of analysis tools that we propose
might enable improved categorization of tensor abnorm-
alities in pathology.
Acknowledgments
AP is in receipt of a Wellcome Trust Fellowship in
Mathematical Biology. The Cambridge Commonwealth
Trust supports HALG. The Medical Research Council
Technology Foresight grant and the Wolfson Foundation
support the Wolfson Brain Imaging Centre. We acknowl-
edge the help of radiographers Tim Donovan, Victoria
Lupson and Ruth Bisbrown-Chippendale, the many useful
discussions with Dr Neil G Harris, Dr Brian K Owler, Dr
Luzius A Steiner and Dr Shahan Momjian, as well as the
excellent computing support of Mr Julian Evans.
References
1. Basser PJ. Inferring microstructural features and the
physiological state of tissues from diffusion-weighted
images. NMR Biomed 1995;8:333–44.
2. Basser PJ, Mattiello J, LeBihan D. Estimation of the effective
self-diffusion tensor from the NMR spin echo. J Magn
Reson B 1994;103:247–54.
3. Basser PJ, Mattiello J, LeBihan D. MR diffusion tensor
spectroscopy and imaging. Biophys J 1994;66:259–67.
4. Pierpaoli C, Basser PJ. Toward a quantitative assessment of
diffusion anisotropy. Magn Reson Med 1996;36:893–906.
5. Pierpaoli C, et al. Diffusion tensor MR imaging of the
human brain. Radiology 1996;201:637–48.
6. Foong, J, et al. Neuropathological abnormalities of the corpus
callosum in schizophrenia: a diffusion tensor imaging study.
J Neurol Neurosurg Psychiatry 2000;68:242–4.
7. Arfanakis K, et al. Diffusion tensor MR imaging in diffuse
axonal injury. AJNR Am J Neuroradiol 2002;23:794–802.
The British Journal of Radiology, February 2006
Enhanced visualization and quantification in MR DTI
8. Werring DJ, et al. Diffusion tensor imaging of lesions and
normal-appearing white matter in multiple sclerosis.
Neurology 1999;52:1626–32.
9. Sorensen AG, et al. Human acute cerebral ischemia:
detection of changes in water diffusion anisotropy by using
MR imaging. Radiology 1999;212:785–92.
10. Gillard JH, et al. MR diffusion tensor imaging of white
matter tract disruption in stroke at 3 T. Br J Radiol
2001;74:642–7.
11. Mori S, et al. Imaging cortical association tracts in the
human brain using diffusion-tensor-based axonal tracking.
Magn Reson Med 2002;47:215–23.
12. Wieshmann UC, et al. Diffusion tensor imaging demon-
strates deviation of fibres in normal appearing white matter
adjacent to a brain tumour. J Neurol Neurosurg Psychiatry
2000;68:501–3.
13. Basser PJ, Pierpaoli C. Microstructural and physiological
features of tissues elucidated by quantitative-diffusion-
tensor MRI. J Magn Reson B 1996;111:209–19.
14. Skare S, et al. Noise considerations in the determination of
diffusion tensor anisotropy. Magn Reson Imaging 2000;18:
659–69.
15. Foong J, et al. Neuropathological abnormalities of the
corpus callosum in schizophrenia: a diffusion tensor
imaging study. J Neurol Neurosurg Psychiatry 2000;68:
242–4.
16. Werring DJ, et al. Diffusion tensor imaging of lesions and
normal-appearing white matter in multiple sclerosis.
Neurology 1999;52:1626–32.
17. Wieshmann UC, et al. Reduced anisotropy of water
diffusion in structural cerebral abnormalities demonstrated
with diffusion tensor imaging. Magn Reson Imaging
1999;17:1269–74.
18. Werring DJ, et al. Diffusion tensor imaging can detect and
quantify corticospinal tract degeneration after stroke. J
Neurol Neurosurg Psychiatry 2000;69:269–72.
19. Rugg-Gunn FJ, et al. Diffusion imaging shows abnormal-
ities after blunt head trauma when conventional magnetic
resonance imaging is normal. J Neurol Neurosurg
Psychiatry 2001;70:530–3.
20. Tievsky AL, Ptak T, Farkas J. Investigation of apparent
diffusion coefficient and diffusion tensor anisotrophy in
acute and chronic multiple sclerosis lesions. AJNR Am J
Neuroradiol 1999;20:1491–9.
21. Jones DK, et al. Characterization of white matter damage in
ischemic leukoaraiosis with diffusion tensor MRI. Stroke
1999;30:393–7.
22. Rugg-Gunn FJ, et al. Diffusion tensor imaging of crypto-
genic and acquired partial epilepsies. Brain
2001;124:627–36.
23. Pierpaoli C, et al. Water diffusion changes in Wallerian
degeneration and their dependence on white matter
architecture. Neuroimage 2001;13:1174–85.
24. Rovaris M, et al. Cognitive dysfunction in patients with
mildly disabling relapsing-remitting multiple sclerosis: an
exploratory study with diffusion tensor MR imaging. J
Neurol Sci 2002;195:103–9.
25. Bozzali M, et al. Quantification of tissue damage in AD
using diffusion tensor and magnetization transfer MRI.
Neurology 2001;57:1135–7.
26. Sinha S, et al. Diffusion tensor MR imaging of high-grade
cerebral gliomas. AJNR Am J Neuroradiol 2002;23:520–7.
27. Guo AC, Jewells VL, Provenzale JM. Analysis of normal-
appearing white matter in multiple sclerosis: comparison of
diffusion tensor MR imaging and magnetization transfer
imaging. AJNR Am J Neuroradiol 2001;22:1893–900.
28. Cercignani M, et al. Mean diffusivity and fractional
anisotropy histograms of patients with multiple sclerosis.
AJNR Am J Neuroradiol 2001;22:952–8.
The British Journal of Radiology, February 2006
29. O’Sullivan, M, et al. Evidence for cortical ‘‘disconnection’’
as a mechanism of age-related cognitive decline. Neurology
2001;57:632–8.
30. Ciccarelli O, et al. Investigation of MS normal-appearing
brain using diffusion tensor MRI with clinical correlations.
Neurology 2001;56:926–33.
31. Guo AC, MacFall JR, Provenzale JM. Multiple sclerosis:
diffusion tensor MR imaging for evaluation of normal-
appearing white matter. Radiology 2002;222:729–36.
32. Mukherjee P, McKinstry RC. Reversible posterior leukoen-
cephalopathy syndrome: evaluation with diffusion-tensor
MR imaging. Radiology 2001;219:756–65.
33. Lim KO, et al. Reduced frontal white matter integrity in
cocaine dependence: a controlled diffusion tensor imaging
study. Biol Psychiatry 2002;51:890–5.
34. Green HA, et al. Increased anisotropy in acute stroke: a
possible explanation. Stroke 2002;33:1517–21.
35. Rovaris M, et al. Assessment of normal-appearing white
and gray matter in patients with primary progressive
multiple sclerosis: a diffusion-tensor magnetic resonance
imaging study. Arch Neurol 2002;59:1406–12.
36. Rocca, MA, et al. A diffusion tensor magnetic resonance
imaging study of brain tissue from patients with migraine. J
Neurol Neurosurg Psychiatry 2003;74:501–3.
37. Molko N, et al. Monitoring disease progression in
CADASIL with diffusion magnetic resonance imaging: a
study with whole brain histogram analysis. Stroke
2002;33:2902–8.
38. Inglese M, et al. Magnetization transfer and diffusion tensor
MR imaging of acute disseminated encephalomyelitis.
AJNR Am J Neuroradiol 2002;23:267–72.
39. Bozzali M, et al. Quantification of brain gray matter damage
in different MS phenotypes by use of diffusion tensor MR
imaging. AJNR Am J Neuroradiol 2002;23:985–8.
40. Codella M, et al. A preliminary study of magnetization
transfer and diffusion tensor MRI of multiple sclerosis
patients with fatigue. J Neurol 2002;249:535–7.
41. Golub GH, Van Loan CF. Matrix computations 3rd edn. The
Johns Hopkins University Press, Baltimore, MD, 1999.
42. Alexander AL, et al. A geometric analysis of diffusion
tensor measurements of the human brain. Magn Reson Med
2000;44:283–91.
43. Bahn MM. Invariant and orthonormal scalar measures
derived from magnetic resonance diffusion tensor imaging.
J Magn Reson 1999;141:68–77.
44. Mase GE. Continuum Mechanics. Schaum Outline Series.
McGraw-Hill Publishing Company, 1970.
45. Fung YC. A first course in continuum mechanics.
Englewood Cliffs, NJ: Prentice Hall, 1994.
46. Stejskal EO, Tanner JE. Spin diffusion measurements: spin
echoes in the presence of a time dependent field gradient. J
Chem Phys 1965;42:288–92.
47. Papadakis NG, et al. A comparative study of acquisition
schemes for diffusion tensor imaging using MRI. J Magn
Reson 1999;137:67–82.
48. Price SJ, et al. Diffusion tensor imaging of brain tumours at
3T: a potential tool for assessing white matter tract
invasion? Clin Radiol 2003;58:455–62.
49. Wieshmann UC, et al. Reduced anisotropy of water
diffusion in structural cerebral abnormalities demonstrated
with diffusion tensor imaging. Magn Reson Imaging
1999;17:1269–74.
50. Jones DK, et al. Characterization of white matter damage in
ischemic leukoaraiosis with diffusion tensor MRI. Stroke
1999;30:393–7.
51. Nusbaum AO, et al. Regional and global changes in
cerebral diffusion with normal aging. AJNR Am J
Neuroradiol 2001;22:136–42.
109
 The British Journal of Radiology, 79 (2006), 110–115

PET-CT findings in surgically transposed ovaries

1,2,3
R ZISSIN, MD, 1U METSER, MD,
1
H LERMAN, MD,
1
G LIEVSHITZ, MD,
4
T SAFRA, MD and
1,3
E EVEN-SAPIR, MD, PhD

Department of 1Nuclear Medicine and 4Oncologic Surgery Unit, Tel-Aviv Sourasky Medical Center
and the Department of 2Diagnostic Imaging, Sapir Medical Center, Kfar Saba, affiliated to the
3
Sackler Faculty of Medicine, Tel-Aviv, Israel

ABSTRACT. The aim of this study is to present the PET/CT findings of surgically
transposed ovaries. PET/CT studies and associated abdominal imaging studies of seven
women, aged 28–43 years, with 11 transposed ovaries were retrospectively reviewed.
Attention was directed to the location and the 18F-Fluorodeoxyglucose (FDG) avidity of
the transposed ovaries. On the CT part of the PET/CT, location of the transposed ovaries
was in the ipsilateral iliac fossa or paracolic gutter abutting the anterior aspect of the
ipsilateral colon (n56), posterolateral to the cecum (n54) and in the anterior
abdominal cavity (n51). Ovaries were of soft-tissue density (n510 with a hypodense
region in two) and one was cystic. In three patients, the transposed ovary was
associated with increased FDG uptake with standard uptake values ranging from 2.4 to
4.8. Two of the latter patients had more than one PET/CT study. FDG uptake altered
between studies, probably related to the performance of the study on different phases Received 17 February 2005
of the cycle. Menstrual history in one of the patients confirmed that the study was Revised 21 May 2005
performed at the ovulatory-phase of the cycle. To conclude, a transposed ovary may Accepted 15 June 2005
appear on a PET-CT study as a mass with occasionally increased FDG uptake that may be
DOI: 10.1259/bjr/33143536
related to its preserved functionality. Physicians interpreting PET/CT should be aware of
surgically transposed ovaries in young female patients to avoid misdiagnosing it as ’ 2006 The British Institute of
tumour. Radiology

Pelvic radiation therapy for cervical, vaginal or colo-
rectal cancer often leads to ovarian failure. Ovarian
transposition outside the radiation field, to the paracolic
gutter or iliac fossa, is a surgical procedure performed to
preserve ovarian function mainly in young females with
early stages of cervical carcinoma [1]. On imaging, the
transposed ovary may appear as a small soft-tissue mass,
often with one or more tiny cysts, or alternatively as a
larger intraperitoneal cystic mass which may show
functional, periodic changes on follow-up studies, accord-
ing to the expected changes in the ovary during the
different phases of the menstruation cycle. Surgical clips
are usually placed to permit identification of
the transposed ovary [2–5]. In oncologic patients, the
recognition of the position and the appearance of the
transposed ovary are crucial to avoid misinterpreting it as a
tumour. We have encountered 10 18F-Fluorodeoxyglucose
(FDG) PET/CT studies in 7 females with 11 surgically
transposed ovaries and we present their imaging findings
on PET/CT and conventional abdominal imaging.
Materials and methods
We reviewed the clinical data and imaging studies of
seven female oncologic patients (aged 28–43 years) after
ovarian transpositions who were referred for PET/CT
studies. Ovarian transposition was bilateral in four
Address correspondence to: Einat Even-Sapir, Department of
Nuclear Medicine, Tel-Aviv Sourasky Medical Center, 6 Weizman
Street, Tel-Aviv, 64239 Israel.
patients and unilateral in the other three. Five women
had carcinoma of the cervix, one had a rectovaginal cleft
mucinous adenocarcinoma and one had uterine non-
Hodgkin’s lymphoma. Six patients reported amenor-
rhoea after hysterectomy and one was menstruating.
Two of the study patients had more than one PET/CT
study, at different time points in the menstruation cycle,
available for assessment. One patient had two studies
and the other had three. Five PET/CT studies were
performed for findings suggestive of recurrence that
were detected on physical examination and/or seen
on MRI or diagnostic CT performed for follow-up. In two
patients, five follow up PET-CT studies were performed,
for re-staging and for monitoring response to treatment.
PET-CT scan was performed following the adminis-
tration of iodinated oral contrast material and after
intravenous injection of 370–666 MBq (10–18 mCi) of
18
FDG. Low-dose CT scanning was performed (140 kV,
80 mA, 0.8 s per CT rotation, pitch of 6, and table speed
of 22.5 mm s21) during normal respiration. PET scanning
was performed immediately following the CT without
changing the patient position. Images were interpreted at
a work-station (Xeleris Elgems, Haifa, Israel) equipped
with fusion software that enables the display of PET, CT
and fused PET/CT images.
Results
The clinical and imaging findings of the patients are
summarized in Table 1. All 11 transposed ovaries were

110 The British Journal of Radiology, February 2006

PET-CT findings in surgically transposed ovaries
Table 1. The clinical data and imaging findings of 7 patients with transposed ovaries
Patient no., age Medical history
(years), primary tumour
1. 35, carcinoma of 6 months after Lt. SO,
cervix Rt. OT and 3 months
after combined
chemo-radiotherapy
2. 43, carcinoma of 18 months after radical
cervix hysterectomy, Rt. SO
and Lt. OT
3. 32, carcinoma of 10 months after radical
cervix hysterectomy, pelvis
lymphadenectomy and
bilateral OT
4. 30, carcinoma of 10 years after radical
cervix hysterectomy, pelvis
lymphadenectomy
and bilateral OT
-S/P fluid aspiration from
a Lt. ovarian cyst,
5 years earlier
5. 43, carcinoma of 6 years after radical
cervix hysterectomy, pelvis
lymphadenectomy and
bilateral OT
6. 39, uterine 5 years after radical
non-Hodgkin’s hysterectomy, pelvis
lymphoma lymphadenectomy
and bilateral OT
7. 28, rectovaginal cleft 4 months after limited
mucinous surgical excision of
adenocarcinoma, S/P the tumour, Rt OT +
breast cancer chemo-radiotherapy
SO, salpingo-oophorectomy; Lt., left; Rt., right; LLQ, left lower quadrant; RLQ, right lower quadrant; ST, soft tissue; AC,
ascending colon; DC, descending colon; OT, ovarian transposition.
The British Journal of Radiology, February 2006
Indication for PET/CT PET/CT findings
A Rt. gutter ST mass A 2.3 cm63.2 cm ST mass,
on CT – suspicion of with central hypodensity,
recurrence near surgical clips, in Rt.
gutter, posterolateral to
AC, cranially to a normal
appendix. No FDG uptake
An intra-abdominal ST A 1.7 cm63.7 cm ST mass,
mass on CT – suspicion near surgical clips, in
of recurrence the anterior mid-abdomen,
between bowel loops and
Lt. rectus abdomini muscle.
No FDG uptake
Suspected mesenteric Rt. A 2.1 cm63.3 cm ST mass,
lymphadenopathy near surgical clips, in the
on CT Rt. iliac fossa, posterolateral
to the cecum. No FDG
uptake
Lt. A 2.4 cm62 cm ST mass,
near surgical clips, in the
Lt. iliac fossa, anterior
to DC. No FDG uptake
A 5 cm cystic (necrotic?) Rt. A 2.7 cm61.5 cm ST mass,
RLQ mass on near surgical clips, in the
MRI – suspicion Rt. iliac fossa, anterior to
of recurrence the cecum. No FDG uptake
Lt. A 3.8 cm63 cm hypodense
mass, near surgical clips,
in the Lt. iliac fossa,
anterior to DC.
Mild FDG uptake (SUV-2.4)
Clinical suspicion of Rt. A 1.9 cm61 cm ST mass,
recurrence near surgical clips, in the
Rt. gutter posterolateral
to AC. No FDG uptake
Lt. A 2.6 cm60.6 cm ST mass,
near surgical clips, in the
Lt. gutter posterolateral
to DC. No FDG uptake
1st: Clinical suspicion 1st: Rt. A 2.9 cm61.7 cm ST mass
of recurrence near surgical clips, in the
Rt. gutter, posterolateral
to the cecum with FDG
uptake (SUV-4.8)
Lt. A 2.1 cm62.4 cm ST mass
with hypodense centre,
near surgical clips, in the
Lt. gutter, lateral to DC.
2nd: follow-up 6 No FDG uptake 2nd.
months later Rt. A 2.9 cm62.7 cm ST mass.
No FDG uptake
Lt. Same as in the
previous study
3rd: follow-up 6 3rd: no change from
months later previous study
1st: Suspected local 1st: A 2.3 cm61.8 cm
recurrence in the hypodense mass, near
Rt. pararectal space surgical clips, in the
on MRI Rt. gutter anterior to
AC. No FDG uptake
– Pararectal local recurrence
2nd: 3 months later 2nd: A 3 cm61.9 cm ST
(on mid-cycle) – to mass with FDG uptake
monitor response to (SUV-3.6)– Progression of
therapy local pelvic disease
111
 R Zissin, U Metser, H Lerman et al

recognized on the CT part of the PET/CT study, adjacent
to surgical clips. Their location was in the ipsilateral iliac
fossa or paracolic gutter (n510), either abutting the
anterior or lateral aspect of the ipsilateral colon (n56)
(Figure 1) or posterolateral to the cecum (n54)
(Figure 2), and in the anterior abdominal cavity between
small bowel loops and the left rectus abdomini muscle at
the level of L3 vertebra (n51). Ten ovaries were of soft-
tissue density, with a hypodense region in two of them,
while the remaining one showed periodic CT changes,
related to the menstruation cycle, which varied from a
‘‘cystic’’ to a soft-tissue attenuating mass.
In three patients, the transposed ovary was associated
with increased FDG uptake. One patient, with bilateral
ovarian transposition, was referred for PET/CT for the
assessment of a ‘‘necrotic’’ mass demonstrated on MRI
(Figure 1a). On PET/CT, performed 1 month later, the
lesion showed significant diminution in size without
FDG uptake, while minimal uptake (standard uptake
value of 2.4) was seen in the contralateral transposed
ovary (Figure 1b,c). As the patient was amenorrhoeic
following hysterectomy, we could only assume that the
MRI and PET/CT findings represented periodic changes
in bilaterally transposed ovaries. The second patient,
with bilateral ovarian transposition after hysterectomy,
had three PET-CT studies. On the first study, the right
transposed ovary presented as a soft-tissue mass with
increased FDG uptake (standard uptake value of 4.8)
(Figure 2a,b). On the second study, 6 months later,
without any treatment in the interim, the same ovary
presented as a soft-tissue mass with no uptake (Figure
2c). These findings remained unchanged on a third
follow-up study. In the third patient, who was still
menstruating as she had an intact uterus, a rim of FDG
uptake (standard uptake value of 3.6) was detected in the
transposed ovary on a study performed 14 days after
menstruation. That ovary was demonstrated on the
CT part of the study as a soft-tissue mass. Based on
the menstrual history of the patient, it appeared that
the patient was in the ovulatory-phase of the cycle.

(a)

(b)

Figure 1. A 30-year-old woman, 10 years after radical hysterectomy and bilateral ovarian transposition for carcinoma of the
cervix, referred for PET/CT for suspected recurrence on MRI (patient no. 4). (a) An axial T2 weighted MR image at the pelvic inlet
shows the transposed right ovary (RO) anteriorly to the ascending colon (AC) as a 5 cm hyperintense mass with a thin
hypointense rim, suspected to be a necrotic tumour recurrence. Note also the left transposed ovary (LO), abutting the anterior
aspect of the descending colon (DC), as a hypointense lesion. That ovary was not reported on the MRI. (b) Axial PET/CT images
(from left to right: CT, PET and fused PET/CT images). On the CT, the bilateral transposed ovaries are seen (thin white arrows).
Note the diminution in size of the right ovary in comparison with the previous MRI performed 1 month earlier, most likely
related to its periodic functional changes. The left transposed ovary shows increased FDG uptake on the PET and on fused
images (thin arrows). Additional physiological sites of FDG uptake are seen, including bowel (arrowhead), bone marrow
(medium-size arrow) and iliac blood vessels (large arrow). (Continued)

112 The British Journal of Radiology, February 2006

PET-CT findings in surgically transposed ovaries
(c)
Figure 1. (Cont.) (c) Coronal PET-CT images (from left to right: CT, PET and fused PET/CT images) show mild increased FDG uptake
in the left transposed ovary (arrows). Physiological FDG uptake is seen in the brain, myocardium, bowel and liver.
This increased ovarian uptake was not detected on a
previous PET/CT study, performed not at the ovulatory
phase, associated with an altered appearance of the
ovary, seen on the CT part of that study, as a hypodense
mass.
Discussion
Ovarian transposition was described by McCall et al
in 1958 for young (,40 years old) females with an
early-stage cervical carcinoma planned for pelvic
radiosurgical treatment, to maintain ovarian function
[6]. The procedure may be unilateral or bilateral,
performed at the time of the radical hysterectomy or
staging lymphadenectomy [1]. The repositioning of the
ovary outside the radiation field may be above the iliac
crest, into the ipsilateral paracolic gutter or lower down,
below the iliac crest lateral to the iliopsoas muscle [2].
The normal transposed ovary may appear on abdom-
inal CT as a soft-tissue mass, sometimes with small cysts
or as a predominant cystic lesion, mimicking a peritoneal
or retroperitoneal tumour implants. The location of the
transposed ovary on CT is generally either adjacent to
The British Journal of Radiology, February 2006
the ascending or descending colon, or in the upper pelvis
lateral to or anterolateral to the psoas muscle [2–5].
However, in one of our patients, the transposed ovary
was in an atypical location, i.e. in the anterior abdominal
cavity between the abdominal wall musculature and the
small bowel loops, mimicking a peritoneal implant.
Adjacent surgical clips assisted in identifying it as a
transposed ovary.
Lack of familiarity with the procedure as well as with
the CT features of a transposed ovary may lead to a
diagnostic error in the interpretation of abdominal CT or
MR imaging, misdiagnosing the transposed ovary as a
metastatic deposit. It was the case in five of our patients,
who were referred for a PET/CT study for a ‘‘suspected’’
tumoural recurrence on either CT or MRI. A right-sided
transposed ovary should also be differentiated from a
mucocele of the appendix, although an appendectomy is
usually performed at the time of the surgical procedure
[5]. In one of our patients, the appendix was not removed
and was identified separately from the ovary on the CT
part of the study, obviating such an interpretation
mistake.
Recently, hybrid systems composed of PET and
CT have been introduced in addition to conventional
113
 R Zissin, U Metser, H Lerman et al

(a)

(b)

(c)

114 The British Journal of Radiology, February 2006

PET-CT findings in surgically transposed ovaries

Figure 2. A 39-year-old patient with a previous hysterectomy for uterine non-Hodgkin’s lymphoma, referred for suspected
central nervous system recurrence (patient no. 6). (a) On PET data (from left to right: coronal, sagittal and transaxial images)
increased FDG uptake was detected in the right abdomen (arrows). (b) On PET-CT (from left to right: CT, PET and fused PET/CT
images) the increased uptake corresponded in location to a soft-tissue mass posterolateral to the cecum, adjacent to surgical
clips, identified as a transposed ovary (arrows). (c) PET/CT study performed 6 months later, without treatment in the interim. The
transposed ovary shows no increased uptake, confirming the functional aetiology of FDG uptake on the first study (arrow).

cross-sectional imaging methods in the routine practice
of oncologic patients for staging, monitoring response to
treatment and assessment of recurrence. PET and CT
data, acquired at the same clinical setting, with genera-
tion of fused PET/CT images, provide both functional
and anatomical information [7]. A transposed ovary may
show increased FDG uptake on the PET part of the study
due to functional changes, as was seen in three of our
patients. FDG uptake in normal ovaries was reported in
pre-menopausal patients without a known ovarian
malignancy at mid-menstrual cycle. In oligomenorrhoeic
patients too, FDG uptake may be high and resemble the
uptake values found at mid-cycle [8]. In menstruating
patients, the physiological cause of uptake may be sorted
out by discussing the menstruation history with the
patient. However, as ovarian transposition is carried out
primarily in patients with gynaecological malignancies
that are often post-hysterectomy, their ovulatory-phase
cannot be determined by history alone. Therefore, when
detecting a focal increased abdominal uptake on PET in a
young female patient, the possibility of a functional
uptake in a transposed ovary should be born in mind
and adjacent surgical clips should be looked for on the
CT part of the study. Directly interviewing the patient
may also assist, as unfortunately the information of
ovarian transposition is often omitted from the referral
sheath for a PET/CT study. It was, indeed, not provided
in any of our patients.
In conclusion, the physician interpreting a PET/CT
study should be familiar both with the clinical history
and the imaging findings of ovarian transposition.
Increased FDG uptake in a transposed ovary may be
related to its preserved functionality.
References
1. Morice P, Juncker L, Rey A, El-Hassan J, Haie-Meder C,
Castaigne D. Ovarian transposition for patients with cervical
carcinoma treated by radiosurgical combination. Fertil Steril
2000;74:743–8.
2. Newbold R, Safrit H, Cooper C. Surgical lateral ovarian
transposition: CT appearance. AJR Am J Roentgenol
1990;154:119–20.
3. Goldberg RE, Sturgeon JF. Surgically transposed ovary
presenting as an intraperitoneal mass on computed tomo-
graphy. Can Assoc Radiol J 1995;46:229–30.
4. Kier R, Chambers SK, Kier R, Chambers SK. Surgical
transposition of the ovaries: imaging findings in 14 patients.
AJR Am J Roentgenol 1989;153:1003–6.
5. Bashist B, Friedman WN, Killackey MA. Surgical transposi-
tion of the ovary: radiologic appearance. Radiology
1989;173:857–60.
6. McCall ML, Keaty EC, Thompson JD. Conservation of
ovarian tissue in the treatment of the carcinoma of the
cervix with radical surgery. Am J Obstet Gynecol
1958;75:590–600.
7. Rohren EM, Turkington TG, Coleman RE. Clinical
applications of PET in oncology. Radiology 2004;231:
305–32.
8. Lerman H, Metser U, Grisaru D, Fishman A, Lievshitz G,
Even-Sapir E. Normal and abnormal 18F-FDG
endometrial and ovarian uptake in pre- and postmeno-
pausal patients: assessment by PET/CT. J Nucl Med
2004;45:266–71.

The British Journal of Radiology, February 2006 115

 The British Journal of Radiology, 79 (2006), 116–122

An audit of imaging test utilization for the management of

lymphoma in an oncology hospital: implications for resource
planning?
A SCHWARTZ, BSc, M K GOSPODAROWICZ, MD, K KHALILI, MD, M PINTILIE, MSc, S GODDARD, BSc,
A KELLER, MD and R W TSANG, MD

University of Toronto, Princess Margaret Hospital, University Health Network, 610 University
Avenue, Toronto, Ontario, M5G 2M9 Canada

ABSTRACT. The purpose of this study was to assist with resource planning by examining
the pattern of physician utilization of imaging procedures for lymphoma patients in a
dedicated oncology hospital. The proportion of imaging tests ordered for routine
follow up with no specific clinical indication was quantified, with specific attention to
CT scans. A 3-month audit was performed. The reasons for ordering all imaging
procedures (X-rays, CT scans, ultrasound, nuclear scan and MRI) were determined
through a retrospective chart review. 411 lymphoma patients had 686 assessments (sets
of imaging tests) and 981 procedures (individual imaging tests). Most procedures were
CT scans (52%) and chest radiographs (30%). The most common reasons for ordering
imaging were assessing response (23%), and investigating new symptoms (19%).
Routine follow up constituted 21% of the assessments (142/686), and of these, 82%
were chest radiographs (116/142), while 24% (34/142) were CT scans. With analysis
restricted to CT scans (296 assessments in 248 patients), the most common reason for
ordering CT scans were response evaluation (40%), and suspicion of recurrence and/or
new symptom (23%). Follow-up CT scans done with no clinical indication comprised 8% Received 26 April 2005
(25/296) of all CT assessments. Staging CT scans were under-represented at 6% of all Revised 1 June 2005
assessments. Imaging with CT scans for follow up of asymptomatic patients is Accepted 22 June 2005
infrequent. However, scans done for staging new lymphoma patients were
DOI: 10.1259/bjr/27372198
unexpectedly low in frequency, due to scans done elsewhere prior to referral. This
analysis uncovered utilization patterns, helped resource planning and provided data to ’ 2006 The British Institute of
reduce unnecessary imaging procedures. Radiology

Lymphoma clinicians rely heavily on imaging techni-
ques to determine the stage of disease at initial
presentation, to assess the response to treatment and to
follow the disease over time [1]. CT scans remain the
standard for evaluation of nodal disease [2], while MRI
gives additional information for some extranodal sites.
Gallium scans and/or 18FDG-PET scans are also useful
tools in the staging and follow-up where they help to
distinguish residual fibrotic mass from viable lymphoma
[3, 4]. After treatment has been completed and providing
a complete remission has been achieved, the goal of
follow-up investigations is to identify recurrent disease
before symptoms develop [5]. However, routine CT
imaging for follow up has not been shown to be cost-
effective, as investigation of symptoms is the most cited
reason for finding recurrent disease [6–13]. Several
studies documented that only 5–9% of relapses were
imaging-detected before the development of symptoms
[5, 11, 13].
At a dedicated oncology hospital, the policy for
lymphoma patients has not been to perform routine CT
imaging for follow up of asymptomatic patients beyond
Address correspondence to: Dr Richard W Tsang,
the attainment of complete remission of disease. Taking a
chest radiograph has been left to the discretion of the
attending physician. This is in agreement with published
studies that suggested CT scan should be performed
according to clinical indications, not strictly as routine
actions [8, 10]. Edelman et al also proposed that
‘‘eliminating unnecessary testing would decrease the
risk of further physical and psychological harm from
the inevitable occurrence of false-positive tests’’ [9]. The
subject of whether clinicians are optimally utilizing the
available imaging modalities has been seldom studied
[10]. This is particularly important in an environment of
limited resources, as there is usually a waiting list to
access certain imaging procedures such as CT and MRI
scans. Concerns were expressed by the imaging depart-
ment that the lymphoma group may be ordering an
excessive number of scans unnecessarily for routine
follow ups, hence making the resource less available for
staging or other urgent reasons in a timely fashion.
Therefore, in this study, the pattern of physician
utilization of imaging investigations in the management
of lymphoma was examined. The goal was to ascertain
the indications for each imaging examination, for
example: staging, evaluating response to treatment,
suspected or confirmed recurrence, and routine

116 The British Journal of Radiology, February 2006

Imaging utilization in lymphoma patients
follow-up monitoring. The aim was to determine the
relative frequencies of the utilization of various imaging
modalities, for follow-up monitoring versus for staging
and response assessment. There was specific interest in
determining if CT scans were often requested for routine
follow-up in asymptomatic patients, to understand if this
resource was overutilized, possibly at the expense of
patients who may require the scans more urgently for
assessment of disease.
Methods
A 3-month audit of imaging procedures performed on
lymphoma patients from January 1st to March 31st 2003
at a dedicated oncology hospital was performed. The
Research and Ethics Board of the hospital approved the
study. Patients were identified from the Imaging
Department database and all were listed with a diagnosis
of lymphoma. For this study, patients with a diagnosis of
myeloma, leukaemia (acute and chronic) and benign
haematological conditions were excluded. A record of all
plain radiographs, CT scans, MRI, gallium, mammo-
grams, bone scans, and ultrasound examinations were
kept for the 3-month period. Patient demographics,
disease extent, and treatment information were collected
on each patient through a chart review. Details of
histology, Ann Arbor stage, treatment, and response
were abstracted. The oncologist responsible for each
patient was recorded.
For the purpose of this study, a ‘‘procedure’’ was
considered a single imaging examination. For example,
CT thorax, CT abdomen/pelvis, gallium scan and
ultrasound were each counted as separate procedures
(total: four procedures). An ‘‘assessment’’ was defined as
a set of imaging examinations all carried out over a
2 weeks period and requested for the same purpose. By
definition, CT thorax, CT abdomen/pelvis, gallium scan
and ultrasound, if all done for staging, were counted as
one assessment. A ‘‘new’’ patient in this study was a
patient referred with a new diagnosis of lymphoma.
Imaging tests performed up to 6 months after comple-
tion of definitive therapy were counted as performed for
a ‘‘new patient’’. An ‘‘old’’ patient was defined as one
who had had imaging performed more than 6 months
after completing initial therapy. A patient with a
previous diagnosis of lymphoma, but referred for
The British Journal of Radiology, February 2006
management of relapse beyond 6 months of completing
initial therapy was considered an ‘‘old’’ patient. If a
patient was under observation, for example in asympto-
matic advanced stage follicular lymphoma, and the
observation period lasted for more than 6 months from
the time of referral, they would then be considered an
‘‘old’’ patient as well. The purpose of this distinction was
to separate imaging utilization between patients referred
with a new diagnosis of lymphoma for management
(new), and those beyond the stage of initial treatment
and attainment of complete remission (old).
The indication for ordering each assessment was
determined based on the physician’s clinical notes in
the medical record. Reasons for ordering imaging
assessments were categorized into: staging, response
assessment, evaluation of residual disease, investigation
of new symptoms, suspicion of recurrence, routine
follow-up with no specific clinical indication, procedure
related assessments such as biopsies, assessments man-
dated by study protocol, assessments performed for
unrelated medical problems, assessments recommended
by radiologists, surveillance for a secondary malignancy
and assessments done for treatment complications.
Questionable cases were reviewed by additional clini-
cians and a reason assigned by consensus.
Results
411 patients were included in the study. Patient
characteristics are shown in Table 1. The most common
non-Hodgkin’s lymphoma histologies were diffuse large
B-cell lymphoma (22%), follicular lymphoma (20%), and
others (23%). The initial diagnosis date is shown in
Figure 1. 50% of the patients were diagnosed before
Table 1. Patient characteristics at initial diagnosis (n5411)
Characteristic Number (%)
Age Median 47 years (range 8–97 years)
Gender Male 222 (54.0%)
Female 189 (46.0%)
Diagnosis Hodgkin’s disease 140 (34.1%)
NHL 271 (65.9%)
Ann Arbor stage I–II 237 (57.7%)
III–IV 174 (42.3%)
NHL, non-Hodgkin’s lymphoma.
Figure 1. Patient’s initial diagnosis
date.
117

2000, and 24% in 2002, which had the greatest proportion
of patients diagnosed in a single year.
The group of 411 patients generated a total of 686
imaging assessments and 981 procedures within the 3-
month period. Of these assessments, 25% (171/686) were
for ‘‘new’’ patients and 75% (515/686) were for ‘‘old’’
patients. Most patients had one assessment (72%,
Figure 2), while the majority of assessments (70.1%)
consisted of one procedure (Figure 3). The total number
of procedures performed per patient within the 3-month
period is shown in Figure 4. CT scans constituted 52% of
imaging procedures performed on lymphoma patients,
followed by chest radiographs (30%), while others
account for ,10% each (Figure 5).
The most common indications for assessments were
response assessment (23%), investigation of new symp-
toms (19%), and routine follow-up (21%) (Figure 6).
Staging constituted only 4% of assessments (Figure 6).
Other indications for scans accounted for 18% of
118
A Schwartz, M K Gospodarowicz, K Khalili et al
Figure 2. Number of assessments
per patient within the 3 month
period.
assessments, but could be broken down into procedure
related (5%), study protocol (4%), unrelated medical
problem (4%), recommended by radiologist (3%), sur-
veillance for a secondary malignancy (1%) and investiga-
tion of treatment complications (1%). A comparison of
the indications for assessments between ‘‘new’’ and
‘‘old’’ patients is shown in Figure 7. The largest
differences between the new and old patients are in
staging with a 14.2% difference, response assessment
with a 25.6% difference and routine follow-up with a
26% difference.
Imaging requested for routine follow-up
Within the 3-month period, 140 patients had routine
follow-up imaging with no specific clinical indication.
These patients received 142 assessments and 152
procedures, which comprised 16% of all procedures.
Figure 3. Number of imaging
procedures per assessment.
The British Journal of Radiology, February 2006
Imaging utilization in lymphoma patients
The types of imaging procedures are shown in Table 2.
Chest radiographs accounted for the majority (82%) The
follow-up chest radiographs constituted 116 assessments
of 295 performed over the 3-month period (39% of chest
radiograph assessments).
Utilization of CT scans
CT scans were performed on 248 patients. These
patients received 296 assessments and 513 procedures
(Table 3). 30% of these patients were ‘‘new’’ and 70%
were ‘‘old’’. Over the 3-month period, the majority of
patients (85%) received one CT assessment (Table 3), but
Figure 5. Imaging procedures performed during the
3-month study on lymphoma patients. This graph shows the
percentage of procedures that each imaging examination
comprises.
The British Journal of Radiology, February 2006
Figure 4. Total number of
procedures per patient within the 3
month period.
each assessment may involve 1–3 CT procedures
(Table 4).
The indications for CT scans included 40% for
response assessment, 13% for suspicion of recurrence,
11% for residual disease, 10% for investigation of new
symptoms, 8% for routine follow up, 6% for staging and
12% for other reasons (Figure 8). When comparing the
indications for CT between ‘‘new’’ and ‘‘old’’ patients,
the largest differences were 38% for response assess-
ment, and 17% for staging, and a difference of 17% for
suspicion of recurrence and 12% for routine follow up
(Figure 9). All the routine follow up CT assessments
(n525) were done for ‘‘old’’ patients, and accounted for
12.3% of the CT scan assessments done for ‘‘old’’
patients. Medical oncologists requested 92% (23/25)
and radiation oncologists requested 8% (2/25) of the
routine follow up CT scans. Of the follow-up CT scans,
72% (18/25) were in patients diagnosed in 2000–2003.
Discussion
At the time when this study was initiated, there were
two main concerns at the hospital regarding the
utilization of imaging resources by the lymphoma group.
A first concern for clinicians was that scans ordered as
staging investigations might overwhelm the imaging
resource, as it is known that all new patients must be
staged with imaging examinations [14], specifically CT
scans of head and neck, thorax, abdomen and pelvis [1].
The results showed that staging accounts for only 4% of
all the imaging assessments, and for an analysis
restricted to CT scans it was 6% of assessments. The
differences between the ‘‘new’’ patients and ‘‘old’’
patients showed an expected trend of ‘‘new’’ patients
receiving more assessments for staging and evaluation of
response. However, even for ‘‘new’’ patients the utiliza-
tion of imaging for staging is low, and since all patients
are staged with imaging, this implied that the majority
had initial imaging performed prior to referral. The audit
was conducted at a tertiary oncology hospital, with
119
 A Schwartz, M K Gospodarowicz, K Khalili et al

Figure 6. Reason for ordering

assessment for cohort of patients
during 3-month period.

Figure 7. Comparing reasons for

assessments between ‘‘new’’ and
‘‘old’’ patients.

Table 2. Follow up imaging assessments with no specific Table 4. CT scan procedures (n5513) per CT assessment
clinical indication (140 patients with 142 assessments)
Number of CT procedures Number of assessments (%)
Procedure No. of Percentage per assessment
assessments
1 130 (43.9%)
Chest radiograph 116 81.7% 2 115 (38.9%)
CT scan 25a 17.6% 3 51 (17.2%)
Ultrasound 1 0.7%
Total 142 100%
40% (35/88) received radiation therapy. Only 26% (9/35)
of those who received radiation therapy were staged
with imaging procedures at the study hospital and
Table 3. Utilization of CT scans (248 patients with 296 included in the audit. This infers that 74% of new
imaging assessments and 513 individual procedures) patients who received radiation therapy had staging
Number of assessments Number of patients scans performed elsewhere prior to their referral and
performed per patient (%) were not even included in this study. ‘‘Old’’ patients had
1 211 (85.1%)
proportionately more assessments for residual disease,
2 30 (12.1%) suspicion of recurrence and routine follow up. These
.2 7 (2.8%) trends are easily understood by the definition used for
‘‘new’’ patients, as those actively receiving their primary
Number of CT scan Number of patients treatment, or those within 6 months of treatment
procedures per patient (%)
completion when scans are performed to document
1 83 (33.5%) response.
2 95 (38.3%) A second concern stems from the waiting time for
3 54 (21.7%) accessing CT scans, which was up to 2–3 weeks from the
.3 16 (6.5%) time of the request at the time this study was conducted.
It was important to determine if there was a dispropor-
tionately large number of patients being scanned for
many patients seen by external specialists and hence routine follow-up with no specific clinical indication,
were fully assessed with imaging procedures prior to thereby making the resource less available to requests for
their referral. This is especially true for patients with more urgent reasons. It was anticipated that a reduction
stage I–II disease referred for radiation therapy. The in routine follow-up scans would free up resources and
Radiation Oncology Department saw 88 new patients in therefore reduce the waiting time for scans. In this study,
the same 3-month period of this study. Of these patients, imaging assessments performed as part of routine

120 The British Journal of Radiology, February 2006

Imaging utilization in lymphoma patients
follow-up with no discernable clinical indication
accounted for 21% of the total assessments and 16% of
the total procedures. Follow-up represents a large
proportion of assessments when all imaging procedures
are grouped together, but chest radiographs account for
the majority of follow-up procedures (82%), and only
24% consisted of CT scan procedures. A chest radiograph
is less costly and more widely available compared with a
CT scan [8, 9, 11, 15, 16]. Indeed, for routine follow-up,
chest radiographs were more frequently ordered by
clinicians compared with CT scans in this study, but it is
less sensitive compared with a thorax CT scan. Studies
have found that CT scans are minimally effective for
follow-up in identifying relapses [1, 10, 17] as relapses
are most often detected by patients developing disease-
related symptoms [6–9, 11–13, 18]. Therefore, this implies
that the practice of ordering routine chest radiographs
has questionable clinical benefit, although one study did
suggest a role in following Hodgkin’s disease in the first
3 years after treatment [12]. Perhaps the common
practice of using chest radiographs as follow up is more
due to its wide availability, low cost, and minimal X-ray
exposure.
The majority of CT scans were performed for ‘‘old’’
patients. Most of these patients received one assessment
(85%), but for two-thirds of patients this assessment
constituted more then 1 CT procedure. The reason for
ordering CT scans (Figure 8) differed slightly from the
reason for ordering all of the imaging procedures when
considered together (Figure 6). CT scans were used more
to evaluate the response to treatment, rather than to
evaluate new problems or for routine follow up. Perhaps
this reflected a high degree of success in the initial
primary management of lymphomas. The data showed
The British Journal of Radiology, February 2006
Figure 8. Analysis of CT scans, rea-
sons for ordering assessments.
Figure 9. Comparing reasons for CT
scans between ‘‘new’’ and ‘‘old’’
patients.
that although follow-up accounted for 21% of all imaging
assessments, only 25 CT assessments were done for
routine follow-up within the 3-month period (i.e. 8% of
all CT assessments). Of these, 92% were ordered by
medical oncologists and 8% by radiation oncologists. The
reason for this discrepancy could be due to differences in
the patient factors seen by the two specialties (a higher
patient volume, with a higher risk of relapse in more
advanced stage patients seen by medical oncologists) or,
alternatively, patients seen by radiation oncology may be
more likely to have follow up imaging performed
elsewhere.
Many studies have suggested surveillance follow-up
routines based on effectiveness, both in terms of the
ability to detect relapse and cost [2, 5, 7–9, 11–13]. Not all
of the recommended standards are in agreement,
ranging from basic history, physical examination and
serum lactic dehydrogenase (LDH) tests for follow up
[8], which is similar to the practice at our hospital, to a
combination of physical examination, blood work, chest
radiograph and additional imaging tests such as CT and
gallium scans left to the discretion of the investigator
[11]. The United States National Comprehensive Cancer
Network 2004 practice guidelines for Hodgkin’s
disease, which are based on consensus rather than
published data, outline an even more intensive follow
up routine that entails chest imaging (CT scan or
radiograph), to be performed every 3–6 months during
the first 3 years post-treatment and annually from the
4th year after treatment [19]. Abdominopelvic imaging
was recommended every 6–12 months in the first 3 years
post treatment, and annually in years 4 and 5. For
patients treated with radiation therapy to the chest,
mammographic screening was suggested 8–10 years
121

post-therapy. The follow up guidelines for NHL were
not detailed, except for follicular lymphoma where
follow up imaging was regarded as necessary but
ordered as clinically indicated, about every 6 months
[20]. An international workshop that established stan-
dardized response criteria in NHL stated that ‘‘imaging
studies may be added for relevant clinical indications,
but specific tests cannot be currently recommended’’, yet
acknowledges that the issue is still controversial and that
good clinical judgement is the most important compo-
nent of patient follow up [2]. In one study of Hodgkin’s
disease, among patients with recurrence of disease,
imaging-detected cases did not have a better overall
clinical outcome with salvage therapy compared
with patients whose recurrence were detected by
symptoms [12].
A limitation of this study is the retrospective nature of
the review, as it is possible that a clinician had ordered
an imaging procedure with a legitimate indication, but
did not document this in the medical record either before
or after the imaging procedure was performed. Such a
situation will be misclassified under the category of
‘‘routine follow up with no specific clinical indication’’.
Given this limitation, the 8% rate of utilization of CT
scans for this reason is probably acceptable within a
practice environment where there had been significant
variation in the follow up recommendations as cited
above. In addition, other institutions with different
referral pattern and case-mix will invariably find a very
different spectrum for the reasons behind imaging
utilization. However, it is the feasibility of the methodol-
ogy and the potential usefulness of the auditing
procedure in assessing and assigning resource utilization
in this study that should be emphasised. Increasingly,
stakeholders of the healthcare delivery system
such as government, health authorities and hospital
boards demand accurate utilization data to assign
resources, and the type of information requested is
often along the same vein as that provided in this
study.
In conclusion, in the lymphoma patients seen at a
tertiary oncology hospital, imaging assessments
requested for staging are under-represented. A substan-
tial proportion of patients having had imaging tests
completed elsewhere prior to referral explained this.
Imaging requested for routine follow up of asympto-
matic lymphoma patients is infrequent, apart from chest
radiographs. This study reflects the utilization patterns
of imaging within a disease group and would assist in
planning the assignment of imaging resources based on
case-mix. It also reassured the physicians and the
institution that the majority of CT scans were ordered
for valid indications. It is hoped that the study also
raised the awareness of clinicians in the importance of
continually adhering to proper indications for ordering
imaging tests.
References
1. Schiepers C, Filmont JE, Czernin J. PET for staging of
Hodgkin’s disease and non-Hodgkin’s lymphoma. Eur J
Nucl Med Mol Imaging 2003;30 Suppl. 1:S82–8.
2. Cheson BD, Horning SJ, Coiffier B, Shipp MA, Fisher RI,
Connors JM, et al. Report of an international workshop
to standardize response criteria for non-Hodgkin’s
122
A Schwartz, M K Gospodarowicz, K Khalili et al
lymphomas. NCI Sponsored International Working
Group. J Clin Oncol 1999;17:1244–53.
3. Naumann R, Vaic A, Beuthien-Baumann B, Bredow J,
Kropp J, Kittner T, et al. Prognostic value of positron
emission tomography in the evaluation of post-treatment
residual mass in patients with Hodgkin’s disease and non-
Hodgkin’s lymphoma. Br J Haematol 2001;115:793–800.
4. Even-Sapir E, Israel O. Gallium-67 scintigraphy: a corner-
stone in functional imaging of lymphoma. Eur J Nucl Med
Mol Imaging 2003;30 Suppl. 1:S65–81.
5. Weeks JC, Yeap BY, Canellos GP, Shipp MA. Value of
follow-up procedures in patients with large-cell lymphoma
who achieve a complete remission. J Clin Oncol
1991;9:1196–203.
6. Biasotti S, Garaventa A, Padovani P, Faraci M, Fioredda F,
Hanau G, et al. Role of active follow-up for early diagnosis
of relapse after elective end of therapies. Pediatr Blood
Cancer 2005;45:781–6.
7. Radford JA, Eardley A, Woodman C, Crowther D. Follow
up policy after treatment for Hodgkin’s disease: too many
clinic visits and routine tests? A review of hospital records.
BMJ 1997;314:343–6.
8. Elis A, Blickstein D, Klein O, Eliav-Ronen R, Manor Y,
Lishner M. Detection of relapse in non-Hodgkin’s lym-
phoma: role of routine follow-up studies. Am J Hematol
2002;69:41–4.
9. Edelman MJ, Meyers FJ, Siegel D. The utility of follow-up
testing after curative cancer therapy. A critical review and
economic analysis. J Gen Intern Med 1997;12:318–31.
10. Cohen MD, Siddiqui A, Weetman R, Provisor A, Coates T.
Hodgkin disease and non-Hodgkin lymphomas in children:
utilization of radiological modalities. Radiology
1986;158:499–505.
11. Dryver ET, Jernstrom H, Tompkins K, Buckstein R, Imrie
KR. Follow-up of patients with Hodgkin’s disease following
curative treatment: the routine CT scan is of little value. Br J
Cancer 2003;89:482–6.
12. Torrey MJ, Poen JC, Hoppe RT. Detection of relapse in
early-stage Hodgkin’s disease: role of routine follow-up
studies. J Clin Oncol 1997;15:1123–30.
13. Guppy AE, Tebbutt NC, Norman A, Cunningham D. The
role of surveillance CT scans in patients with diffuse large
B-cell non-Hodgkin’s lymphoma. Leuk Lymphoma
2003;44:123–5.
14. Vinnicombe SJ, Reznek RH. Computerised tomography in
the staging of Hodgkin’s disease and non-Hodgkin’s
lymphoma. Eur J Nucl Med Mol Imaging 2003;30 Suppl.
1:S42–55.
15. Valk PE, Pounds TR, Tesar RD, Hopkins DM, Haseman
MK. Cost-effectiveness of PET imaging in clinical oncology.
Nucl Med Biol 1996;23:737–43.
16. Joines JD, McNutt RA, Carey TS, Deyo RA, Rouhani R.
Finding cancer in primary care outpatients with low back
pain: a comparison of diagnostic strategies. J Gen Intern
Med 2001;16:14–23.
17. Rankin SC. Assessment of response to therapy using
conventional imaging. Eur J Nucl Med Mol Imaging
2003;30 Suppl. 1:S56–64.
18. Oh YK, Ha CS, Samuels BI, Cabanillas F, Hess MA, Cox JD.
Stages I-III follicular lymphoma: role of CT of the abdomen
and pelvis in follow-up studies. Radiology 1999;210:483–6.
19. NCCN. The NCCN Clinical Practice Guidelines in
Oncology. Hodgkin’s Disease. www.nccn.org: National
Comprehensive Cancer Network; 2004. Report No.:
V.1.2004.
20. NCCN. The NCCN Clinical Practice Guidelines in
Oncology. Non-Hodgkin’s Lymphoma. www.nccn.org:
National Comprehensive Cancer Network; 2005. Report
No.: v.1.2005.
The British Journal of Radiology, February 2006
The British Journal of Radiology, 79 (2006), 123–129

Image quality and breast dose of 24 screen–film combinations for

mammography
1
A D DIMAKOPOULOU, MSc, 2I A TSALAFOUTAS, PhD,
1
E K GEORGIOU, MD, PhD and
1
E N YAKOUMAKIS, PhD

1
Medical Physics Department, Medical School, University of Athens, 75 Mikras Asias, 115 27, Athens
and 2Medical Physics Unit, Konstantopoulio-Agia Olga Hospital, 3-5 Agias Olgas, Nea Ionia, 142 33,
Athens, Greece

ABSTRACT. In this study the effect of different mammographic screen–film

combinations on image quality and breast dose, and the correlation between the
various image quality parameters, breast dose and the sensitometric parameters of a
film were investigated. Three Agfa (MR5-II, HDR, HT), two Kodak (Min-R M, Min-R
2000), one Fuji (AD-M), one Konica (CM-H) and one Ferrania (HM plus) single emulsion
mammographic films were combined with three intensifying screens (Agfa HDS, Kodak
Min-R 2190 and Fuji AD-MA). The film characteristics were determined by sensitometry,
while the image quality and the dose to the breast of the resulting 24 screen–film
combinations were assessed using a mammography quality control phantom. For each
combination, three images of the phantom were acquired with optical density within
three different ranges. Two observers assessed the quality of the 72 phantom images
obtained, while the breast dose was calculated from the exposure data required for
each image. Large differences among screen–film combinations in terms of image
quality and breast dose were identified however, that, could not be correlated with the
film’s sensitometric characteristics. All films presented the best resolution when
combined with the HDS screen at the expense of speed, and the largest speed when
combined with the AD-MA screen, without degradation of the overall image quality.
However, an ideal screen–film combination presenting the best image quality with the Received 26 February 2005
least dose was not identified. It is also worth mentioning that the best performance for Revised 23 May 2005
a film was not necessarily obtained when this was combined with the screen provided Accepted 7 June 2005
by the same manufacturer. The results of this study clearly demonstrate that
DOI: 10.1259/bjr/84646476
comparison of films based on their sensitometric characteristics are of limited value for
clinical practice, as their performance is strongly affected by the screens with which ’ 2006 The British Institute of
they are combined. Radiology

The main concern in mammography screening is the
detection of features characteristic of breast disease.
These features often have sizes of the order of 1 mm and
differ from the normal tissue only slightly in composi-
tion, thus setting high requirements for the resolution
and contrast that an imaging system must offer in order
to be appropriate for mammography [1]. On the other
hand, given the high radiosensitivity of the breast and
the large number of women examined many times
during their life, it is evident that the doses during
mammography should be kept as low as possible.
While digital mammography may look promising, the
vast majority of mammography examinations are still
carried out with screen–film systems. In recent years,
most film manufacturers have presented new films and
intensifying screens for mammography that reduce the
dose to the breast and produce the image quality
required to maintain the diagnostic sensitivity and
specificity of mammography at high levels. However,
while the design is the major factor in determining the
performance of a film, this may be affected by the
processing conditions, such as the chemicals used, their
temperature and the processing time [2]. Inappropriate
chemicals or a developing temperature lower than
recommended may result in unacceptable mammograms
and this is why some films have been modified to be less
dependent on processing conditions [3, 4].
Film characteristics can be determined and monitored
for changes due to processing by sensitometry. However,
film performance will be dependent on the screen with
which it is combined and thus for clinical practice the
characteristics of the screen–film combination rather than
those of film or screen separately are of interest [5]. The
screen–film characteristics can be determined and
monitored using an appropriate quality control (QC)
phantom, with which changes in image quality due to
processing or other reasons can be identified.
One of the parameters routinely monitored with the
QC phantom is the background or reference optical
density (OD) of the mammographic images. Apart from
the personal preferences of radiologists, it has been
shown that for a given screen–film combination, subtle
details and small contrast differences are best accentu-
ated when the film OD is within a certain range [6–8]. For

The British Journal of Radiology, February 2006 123

 A D Dimakopoulou, I A Tsalafoutas, E K Georgiou and E N Yakoumakis
this reason it has been recommended that each institu-
tion should determine the optimum OD for the screen–
film used and the processing conditions specific to it [7].
In this study, eight films were combined with three
intensifying screens and the resulting 24 screen–film
combinations were compared in terms of image quality
and breast dose. Film characteristics were determined by
sensitometry, whereas the image quality and speed of
the screen–film combinations were assessed using a QC
phantom to obtain images within three different OD
ranges. Our main objective was to investigate the effect
of different screens on a certain film and search for any
correlation between the image quality, breast dose and
the sensitometric parameters of a film.
Materials and methods
The eight single emulsion mammographic films tested
in this study were the MR5-II, HDR, HT (Agfa-Gevaert
N.V., Mortsel, Belgium), Min-R M, Min-R 2000 (Eastman
Kodak Company, New York, NY), AD-M (Fuji Photo Film
Co. Ltd, Tokyo, Japan), CM-H (Konica Corporation,
Tokyo, Japan) and HM plus (Ferrania Sp A, Ferrania
(SV), Italy). One box from each film type was used to
avoid little differences that may exist among different film
batches or films that may have been stored for different
times and under different storage conditions [9]. The three
intensifying screens used in this study were the HDS
(Agfa), the Min-R 2190 (Kodak) and the AD-MA (Fuji).
For each film a 21-step sensitometric strip was
produced, using an X-Rite 334 sensitometer (X-Rite,
Grandville, MI) operated in the green spectrum. All films
were processed in a daylight processor (Curix Capacity,
Agfa) with nominal processing time 90 s (22 s developing
time) and with the developer temperature set to 36 ˚C. The
developer type was the Eos Dev (Agfa) and the fixer type
was the G334i (Agfa). All films were processed sequen-
tially, immediately after exposure and on the same day, to
avoid day-to-day variations in processing conditions
caused by the ageing of the chemicals that may have
variable effects on the characteristics of each film [9, 10].
The OD of the 21 steps of the sensitometric strips was
measured using a calibrated optical densitometer (RMI
331, X-Rite). For each film the Hurter-Driffield (H&D)
curve was plotted and the following sensitometric
parameters were derived: OD of base plus fog (ODb+f),
maximum OD (ODmax), average gradient (AG), film
gamma (c) and film speed. The AG and c are the slopes
of the H&D curve for ODs from 0.25+ODb+f to 2.0+ODb+f
and from 1.0+ODb+f to 2.0+ODb+f, respectively. AG and c
are both used as indices of film contrast, however, only c
can be used to reproduce the linear part of the H&D
curve. The film speed was defined as the reciprocal of
the relative light exposure required to obtain an OD of
1+ODb+f. Using this definition, the higher the film speed
the less exposure is needed for a given OD. In order to
illustrate the expected increase in breast dose – according
to sensitometry – when a film other than the fastest one is
used, the sensitometric relative dose index (SRDI) was
defined as the reciprocal of the relative speed value. The
SDRIs were expressed as percentages of the smallest
SRDI value (highest speed) that was taken as 100%. It
must be noted that from preliminary sensitometric tests
124
it has been confirmed that for different sheets of the same
film type processed within the same day, variations of
less than ¡0.01 in ODb+f, ¡10% in speed and ¡0.1 in
ODmax, AG and c should be expected.
To evaluate the characteristics of screen–film combina-
tions, a mammography QC phantom was employed
(breast phantom, Model 18-222; Nuclear Associates,
Division of Victoreen Inc., NY). This phantom is
realistically shaped and equivalent to an average firm
breast of 4.5 cm compressed thickness, consisting of 50%
adipose and 50% glandular tissue. It includes 12 groups
of calcium carbonate specks (simulating microcalcifica-
tions), 7 hemispheric masses composed of 75% glandular
and 25% adipose equivalent tissue (simulating tumours)
and a wax insert with 5 embedded nylon fibres
(simulating glandular tissue fibrils). The phantom also
contains a five-step stepwedge, simulating breast areas
with compositions 100% adipose, 70% adipose–30%
glandular, 50% adipose–50% glandular, 30% adipose–
70% glandular and 100% glandular tissue. Finally, two
line-pair test targets (5–20 lp mm21 each), one parallel
and one perpendicular to the anode–cathode axis and a
central area where the background OD is measured, are
included. A similar phantom (without the nylon fibres
and with only one line-pair test target) was used by
Nassivera and Nardin [11].
The phantom was exposed using a Senographe 500T
mammography unit (CGR, Buc, France). All exposures
were made with the Mo/Mo target filter combination,
constant tube potential (28 kVp), large focal spot (0.3 mm
nominal size) and without the breast compression paddle.
Using manual mAs selection technique, images of the
phantom were acquired until for each screen–film
combination three films with OD as close as possible to
the central OD of three different optical density ranges
(0.70–1.10, 1.11–1.50, 1.51–2.00) were produced to account
for the wide range of ODs that can be encountered within
actual mammographs. Phantom images were processed
in the same processor and on the same day as the
sensitometric strips, so both film characteristics and
screen–film performance were determined under the
same processing conditions. It must be noted that as the
Kodak and Fuji cassettes were not compatible with
the Agfa daylight processor, the films exposed with these
cassettes had to be transferred manually to the Agfa
cassette in order to be fed into the processor.
For each one of the resulting 72 phantom images, the
OD of the central area was measured with the densi-
tometer, as well as the OD of the areas simulating 100%
adipose and 100% glandular tissue. The OD difference of
these two areas can be used as an index of screen–film
contrast (CI) and, according to the phantom manufac-
turer, it should be ¢0.28.
All phantom images were examined using a viewing
box especially designed for mammography, featuring
adjustable brightness, masking shutters and a magnify-
ing glass. The shutters were closed down to the phantom
image size and the brightness was adjusted as necessary
to obtain the best possible conditions for viewing each
type of simulated lesion, while for speck groups the
magnifying glass was also used. A magnifying glass
supplied with the QC phantom was used to inspect the
line pair object. The above details are mentioned, as
The British Journal of Radiology, February 2006
Image quality and breast dose of screen–film mammography
Figure 1. The Hurter-Driffield (H&D) curves of the eight
films included in this study are given. The optical densities
(ODs) of the 21 steps of the sensitometric strips correspond
to log relative exposure values (LogE) that range from 0 to 3,
in steps of 0.15 each. In these figures, only the ODs for
LogE¢ge;0.9 (steps 7 to 21) are presented in order to
enhance the visibility of the differences in the linear part and
the shoulder of the H&D curves.
viewing conditions are very important for interpreting
mammograms or scoring phantom images [12].
Two observers scored the images independently and
any disagreements were resolved by consensus. Five
scores were recorded for each film: one for the speck-
groups, one for masses, one for fibres and two for the
two line-pair test targets. For ambiguous decisions
concerning not clearly visualized structures, a 0.5 mark
was assigned. In order to have a single index character-
izing the screen–film performance, a total score (TS) was
calculated using the following weighting coefficients: 0.4
for specks, 0.35 for masses and 0.25 for fibres. These
coefficients were selected after consulting with five
radiologists about the clinical importance of each
simulated structure for diagnosis. Since the two scores
for the line-pair test targets have no straightforward
clinical relevance, their mean value was calculated for
reference only (resolution score).
For screen–film combination comparisons in terms of
breast dose, the entrance surface air-kerma (ESAK) at the
phantom surface was calculated from the mAs selected
for each exposure. For the range of mAs selections
Table 1. The results of the sensitometric evaluation of the eight mammographic films included in this study. The lowest speed
was arbitrarily defined as 100%. The sensitometric relative dose index (SRDI) was defined as the reciprocal of the relative speed,
considering the lowest value (highest speed) as 100%
Film Parameter MR5-II HT HDR
(Agfa) (Agfa) (Agfa)
ODb+f 0.20 0.23 0.21
ODmax 3.50 3.84 3.98
AG 2.46 3.12 3.09
c 2.80 4.36 3.98
Rel. Speed (%) 115 121 143
SRDI (%) 145 138 117
Agfa-Gevaert N.V., Mortsel, Belgium; Eastman Kodak Company, New York, NY; Fuji Photo Film Co. Ltd, Tokyo, Japan; Konica
Corporation, Tokyo, Japan; Ferrania USA Inc., USA.
OD, optical density; AG, average gradient; c, film gamma.
The British Journal of Radiology, February 2006
utilized in this study, the output at 28 kVp defined at the
phantom entrance surface was 98¡2 mGy mAs21.
Furthermore, the ESAK required to achieve a net OD of
1 was calculated by interpolation from the ODs and the
ESAKs of the three films acquired for each combination.
The resulting ESAKs for a net OD51 were used to derive
the relative dose index (RDI), expressed as a percentage
of the smallest observed value, that was considered as
100%. Using this definition, the larger the RDI, the larger
the dose to the breast and the smaller the speed of a
given screen–film combination.
To investigate the correlation between the various
image quality parameters, sensitometric parameters, OD
and dose indices linear regression analysis was used. A
correlation coefficient (r) larger than 0.7 was taken as an
indication of good correlation. Specifically, the correla-
tions of all the image quality scores (TS, specks, masses,
fibres and resolution) with AG, c, SRDI, CI, OD and
ESAK were investigated. Furthermore, the correlation of
TS with resolution, the correlations of CI with AG, c,
SDRI, OD and ESAK and the correlations of SDRI with
RDI and ESAK were also investigated.
Results
The H&D curves for all the films studied are plotted in
Figure 1, while their sensitometric parameters are given in
Table 1. In Figure 1, the large differences among the H&D
curve shapes, the high speed of Min-R 2000, the low speed
of AD-M and the non-typical but similar H&D curve
shapes of HDR and HM plus should be noted. From
Table 1, it can be seen that the Min-R 2000 presents the
highest speed and c, while the CM-H has the highest AG.
Concerning the screen–film comparisons, for the 72
phantom images evaluated in this study, speck scores
ranged from 6 to 11, mass scores from 2.5 to 6, fibre scores
from 2 to 4, total scores from 4.3 to 7.3, resolution
scores from 10 to 14.5 lp mm21 and CI from 0.26 to 0.61.
For each of the three OD ranges the respective ESAK ranges
were: 2.0–4.9 mGy, 2.4–6.2 mGy and 3.1–8.8 mGy. The
results for the 24 combinations studied in terms of the
ESAK at the phantom surface, the background OD, the TS,
the CI and the resolution score are presented in Tables 2a,
2b and 2c for the three different OD ranges, respectively. In
the two last columns of each table, the screen–film RDI and
the mean value of TS in the three OD ranges (TSm) are also
given. From these tables it can be seen that the screen
affected the image quality of a given film as well as the dose
Min-R Min-R CM-H AD-M HM plus (Ferrania)
M (Kodak) 2000 (Kodak) (Konica) (Fuji)
0.17 0.20 0.20 0.17 0.22
3.88 4.03 3.80 3.58 4.13
3.13 2.80 3.64 3.50 2.93
4.13 4.46 4.23 4.24 3.68
133 167 126 100 143
125 100 132 167 117
125
 A D Dimakopoulou, I A Tsalafoutas, E K Georgiou and E N Yakoumakis

Table 2a. The results of the evaluation of the 24 film–screen combinations, from the quality control (QC) phantom images with
optical density (OD) in the range 0.7–1.1. The screen–film combination relative dose index (RDI) has been expressed in terms of the
entrance surface air-kerma (ESAK) required to produce a net OD of 1, considering the lowest value observed (2.19 mGy) as 100%
Film Screen ESAK (mGy) OD bgnd TS CI Resolution Screen RDI Film TSm
(lp mm21) (%)

MR5-II HDS (Agfa) 3.9 1.07 4.7 0.31 13.5 204 4.7
(Agfa) Min-R2190 (Kodak) 3.1 0.98 4.5 0.31 12.0 177 4.7
AD-MA (Fuji) 2.4 1.04 4.7 0.30 11.5 132 5.1
HT HDS (Agfa) 3.9 1.06 5.7 0.39 13.5 207 5.9
(Agfa) Min-R2190 (Kodak) 3.1 0.88 5.4 0.36 12.5 178 5.7
AD-MA (Fuji) 2.4 1.00 5.8 0.36 12.5 127 5.9
HDR HDS (Agfa) 3.9 1.04 5.6 0.42 13.5 200 5.7
(Agfa) Min-R2190 (Kodak) 3.1 0.88 5.9 0.37 12.5 179 6.2
AD-MA (Fuji) 2.4 1.00 5.2 0.41 12.0 128 5.3
MIN-R M HDS (Agfa) 3.1 0.81 5.7 0.33 13.0 177 5.8
(Kodak) Min-R2190 (Kodak) 3.1 0.98 5.4 0.38 13.0 163 5.7
AD-MA (Fuji) 2.4 1.05 5.4 0.41 12.5 120 6.0
MIN-R 2000 HDS (Agfa) 3.1 1.08 6.0 0.38 14.0 156 5.7
(Kodak) Min-R2190 (Kodak) 2.4 0.86 5.0 0.26 12.0 141 5.6
AD-MA (Fuji) 2.0 1.02 5.5 0.35 11.5 100 6.1
CM-H HDS (Agfa) 3.1 1.04 5.7 0.45 13.5 153 5.8
(Konica) Min-R2190 (Kodak) 2.4 0.79 5.5 0.36 11.5 140 5.7
AD-MA (Fuji) 2.0 0.96 6.0 0.41 11.5 104 5.9
AD-M HDS (Agfa) 4.9 0.97 6.0 0.43 12.5 272 6.0
(Fuji) Min-R2190 (Kodak) 4.9 0.88 6.5 0.41 12.0 270 6.6
AD-MA (Fuji) 3.1 0.83 7.0 0.38 12.0 191 7.0
HM plus HDS (Agfa) 3.1 0.79 5.6 0.31 13.5 189 5.0
(Ferrania) Min-R2190 (Kodak) 3.1 0.92 6.3 0.38 12.5 169 6.1
AD-MA (Fuji) 2.0 0.76 4.7 0.31 11.5 120 5.6
Agfa-Gevaert N.V., Mortsel, Belgium; Eastman Kodak Company, New York, NY; Fuji Photo Film Co. Ltd, Tokyo, Japan; Konica
Corporation, Tokyo, Japan; Ferrania USA Inc., USA.
TS, total score; CI, index of screen–film contrast.

Table 2b. The results of the evaluation of the 24 film–screen combinations, from the quality control (QC) phantom images with
optical density (OD) in the range 1.11–1.5. The screen–film combination relative dose index (RDI) has been expressed in terms of
the entrance surface air-kerma (ESAK) required to produce a net OD of 1, considering the lowest value observed (2.19 mGy) as
100%
Film Screen ESAK (mGy) OD bgnd TS CI Resolution Screen RDI (%) Film TSm
(lp mm21)

MR5-II HDS (Agfa) 4.9 1.30 5.0 0.31 12.5 204 4.7
(Agfa) Min-R2190 (Kodak) 4.9 1.44 4.9 0.30 12.0 177 4.7
AD-MA (Fuji) 3.1 1.25 5.2 0.31 12.0 132 5.1
HT HDS (Agfa) 4.9 1.26 6.2 0.45 13.5 207 5.9
(Agfa) Min-R2190 (Kodak) 3.9 1.24 5.8 0.48 12.5 178 5.7
AD-MA (Fuji) 3.1 1.45 6.0 0.48 12.5 127 5.9
HDR HDS (Agfa) 4.9 1.38 5.8 0.50 13.5 200 5.7
(Agfa) Min-R2190 (Kodak) 4.4 1.37 6.1 0.50 13.0 179 6.2
AD-MA (Fuji) 3.1 1.39 5.4 0.50 13.0 128 5.3
MIN-R M HDS (Agfa) 3.9 1.27 6.0 0.52 13.5 177 5.8
(Kodak) Min-R2190 (Kodak) 3.9 1.33 6.1 0.48 13.0 163 5.7
AD-MA (Fuji) 3.1 1.49 6.6 0.50 12.5 120 6.0
MIN-R 2000 HDS (Agfa) 3.9 1.39 5.8 0.47 14.5 156 5.7
(Kodak) Min-R2190 (Kodak) 3.1 1.23 5.7 0.44 13.5 141 5.6
AD-MA (Fuji) 2.4 1.40 6.4 0.46 11.5 100 6.1
CM-H HDS (Agfa) 3.9 1.48 6.0 0.49 14.0 153 5.8
(Konica) Min-R2190 (Kodak) 3.1 1.20 5.5 0.44 13.5 140 5.7
AD-MA (Fuji) 2.4 1.24 5.4 0.47 12.0 104 5.9
AD-M HDS (Agfa) 6.2 1.22 5.8 0.47 14.0 272 6.0
(Fuji) Min-R2190 (Kodak) 6.2 1.37 6.5 0.49 13.5 270 6.6
AD-MA (Fuji) 4.9 1.36 6.6 0.50 13.0 191 7.0
HM plus HDS (Agfa) 4.9 1.47 5.0 0.50 14.5 189 5.0
(Ferrania) Min-R2190 (Kodak) 3.9 1.38 6.3 0.47 12.5 169 6.1
AD-MA (Fuji) 3.1 1.48 6.1 0.49 12.5 120 5.6
Agfa-Gevaert N.V., Mortsel, Belgium; Eastman Kodak Company, New York, NY; Fuji Photo Film Co. Ltd, Tokyo, Japan; Konica
Corporation, Tokyo, Japan; Ferrania USA Inc., USA.
TS, total score; CI, index of screen–film contrast.

126 The British Journal of Radiology, February 2006

Image quality and breast dose of screen–film mammography

Table 2c. The results of the evaluation of the 24 film–screen combinations, from the quality control (QC) phantom images with
optical density (OD) in the range 1.51–2.0. The screen–film combination relative dose index (RDI) has been expressed in terms of
the entrance surface air-kerma (ESAK) required to produce a net OD of 1, considering the lowest value observed (2.19 mGy) as
100%
Film Screen ESAK (mGy) OD bgnd TS CI Resolution Screen RDI (%) Film TSm
(lp/mm)

MR5-II HDS (Agfa) 7.8 1.76 4.3 0.29 10.0 204 4.7
(Agfa) Min-R2190 (Kodak) 6.2 1.71 4.9 0.32 12.0 177 4.7
AD-MA (Fuji) 3.9 1.55 5.2 0.32 11.5 132 5.1
HT HDS (Agfa) 6.2 1.74 5.8 0.55 13.0 207 5.9
(Agfa) Min-R2190 (Kodak) 4.9 1.59 5.7 0.52 12.0 178 5.7
AD-MA (Fuji) 3.9 1.84 6.0 0.52 12.0 127 5.9
HDR HDS (Agfa) 6.2 1.73 5.6 0.54 13.0 200 5.7
(Agfa) Min-R2190 (Kodak) 6.2 1.94 6.4 0.56 11.5 179 6.2
AD-MA (Fuji) 3.9 1.77 5.4 0.49 12.0 128 5.3
MIN-R M HDS (Agfa) 6.2 1.82 5.7 0.52 12.5 177 5.8
(Kodak) Min-R2190 (Kodak) 4.9 1.58 5.6 0.49 13.0 163 5.7
AD-MA (Fuji) 3.9 1.75 6.1 0.51 12.5 120 6.0
MIN-R 2000 HDS (Agfa) 4.9 1.72 5.4 0.61 14.5 156 5.7
(Kodak) Min-R2190 (Kodak) 3.9 1.54 6.2 0.50 12.5 141 5.6
AD-MA (Fuji) 3.1 1.72 6.5 0.57 13.0 100 6.1
CM-H HDS (Agfa) 4.9 1.79 5.8 0.50 14.0 153 5.8
(Konica) Min-R2190 (Kodak) 3.9 1.59 6.2 0.48 13.5 140 5.7
AD-MA (Fuji) 3.1 1.71 6.2 0.50 13.0 104 5.9
AD-M HDS (Agfa) 8.8 1.72 6.0 0.55 13.5 272 6.0
(Fuji) Min-R2190 (Kodak) 7.8 1.56 6.7 0.52 13.0 270 6.6
AD-MA (Fuji) 6.2 1.78 7.3 0.55 14.0 191 7.0
HM plus HDS (Agfa) 6.2 1.88 4.4 0.48 13.0 189 5.0
(Ferrania) Min-R2190 (Kodak) 4.9 1.63 5.9 0.51 13.5 169 6.1
AD-MA (Fuji) 3.9 1.89 6.0 0.47 12.0 120 5.6
Agfa-Gevaert N.V., Mortsel, Belgium; Eastman Kodak Company, New York, NY; Fuji Photo Film Co. Ltd, Tokyo, Japan; Konica
Corporation, Tokyo, Japan; Ferrania USA Inc., USA.
TS, total score; CI, index of screen–film contrast.
required for obtaining a phantom image of a certain OD. is evident that, while for TS and CI there was variety in
Furthermore, it can be seen that for a certain screen–film the screen with which a film was best combined, for
combination the TS, CI and resolution were dependent on resolution and RDI the best screen was common for all
the background OD. The screen–film combinations that films. Indeed, all films presented the best resolution
exhibited the largest dependency with OD were the AD- when combined with the HDS screen and the smallest
MA/HM plus for the TS, the Min-R 2190/Min-R 2000 for RDI (largest speed) when combined with the AD-MA
the CI and the HDS/MR5-II for the resolution score. screen. The largest resolution offered by the HDS screen
In order to highlight some of the major results of this was at the expense of speed, while the largest speed
study, the best screen for a given film and the best film offered by the AD-MA screen was at the expense of
for a given screen in terms of TS, CI, resolution and RDI resolution but not at the expense of TS.
are given in Tables 3 and 4, respectively, where the OD In summary, Tables 3 and 4 clearly demonstrate two
range for which the best score is obtained is also noted. It points that deserve special attention. First, the highest

Table 3. The best screen for a given film in terms of total score (TS), index of screen–film contrast (CI), line pairs per millimetre
(lp mm21) and relative dose index (RDI). Letters in parentheses give reference to Tables 2a, 2b or 2c where each value can be
found. The largest values overall for each parameter are given in bold
MR5-II HT HDR MIN-R M MIN-R 2000 CM-H AD-M HM plus
(Agfa) (Agfa) (Agfa) (Kodak) (Kodak) (Konica) (Fuji) (Ferrania)

TS HDS (Agfa) – 6.2 (b) – – – – – –

Min-R2190 (Kodak) – – 6.4 (c) – – 6.2 (c) – 6.3 (a,b)
AD-MA (Fuji) 5.2 (b,c) – – 6.6 (b) 6.5 (c) 6.2 (c) 7.3 (c) –
CI HDS (Agfa) – 0.55 (c) – 0.52 (b,c) 0.61 (c) 0.5 (c) 0.55 (c) –
Min-R2190 (Kodak) 0.32 (c) – 0.56 (c) – – – – 0.51 (c)
AD-MA (Fuji) 0.32 (c) – – – – 0.5 (c) 0.55 (c) –
lp mm21 HDS (Agfa) 13.5 (a) 13.5 (a,b) 13.5 (a,b) 13.5 (b) 14.5 (b,c) 14 (b,c) 14 (b) 14.5 (b)
Min-R2190 (Kodak) – – – – – – – –
AD-MA (Fuji) – – – – – – 14 (b) –
RDI HDS (Agfa) – – – – – – – –
Min-R2190 (Kodak) – – – – – – – –
AD-MA (Fuji) 132 127 128 120 100 104 191 120
Agfa-Gevaert N.V., Mortsel, Belgium; Eastman Kodak Company, New York, NY; Fuji Photo Film Co. Ltd, Tokyo, Japan; Konica
Corporation, Tokyo, Japan; Ferrania USA Inc., USA.

The British Journal of Radiology, February 2006 127

 A D Dimakopoulou, I A Tsalafoutas, E K Georgiou and E N Yakoumakis

Table 4. The best film for a given screen in terms of total score (TS), index of screen–film contrast (CI), line pairs per millimetre
(lp mm21) and relative dose index (RDI). Letters in parentheses give reference to Tables 2a, 2b or 2c where each value can be
found. The largest values overall for each parameter are given in bold
MR5-II HT HDR MIN-R M MIN-R 2000 CM-H AD-M HM plus
(Agfa) (Agfa) (Agfa) (Kodak) (Kodak) (Konica) (Fuji) (Ferrania)

TS HDS (Agfa) – 6.2 (b) – – – – – –

Min-R2190 (Kodak) – – – – – – 6.7 (c) –
AD-MA (Fuji) – – – – – – 7.3 (c) –
CI HDS (Agfa) – – – – 0.61 (c) – – –
Min-R2190 (Kodak) – – 0.56 (c) – – – – –
AD–MA (Fuji) – – – – 0.57 (c) – – –
lp mm21 HDS (Agfa) – – – – 14.5 (b,c) – – 14.5 (b)
Min–R2190 (Kodak) – – – – 13.5 (b) 13.5 (b) 13.5 (b) 13.5 (b)
AD–MA (Fuji) – – – – – – 14 (b) –
RDI HDS (Agfa) – – – – – 153 – –
Min–R2190 (Kodak) – – – – – 140 – –
AD–MA (Fuji) – – – – 100 – – –
Agfa-Gevaert N.V., Mortsel, Belgium; Eastman Kodak Company, New York, NY; Fuji Photo Film Co. Ltd, Tokyo, Japan; Konica
Corporation, Tokyo, Japan; Ferrania USA Inc., USA.

TS, CI, resolution and the lowest RDI were observed for
different combinations and thus an ‘‘ideal’’ combination
was not identified. Second, the best performance of a
given film or screen was not always obtained when
combined with the screen or film, respectively, from the
same manufacturer.
Finally, concerning the correlation between image
quality parameters, sensitometric parameters and dose,
no correlation coefficient larger than 0.7 was calculated
in any of the correlations tested. The largest correlation
coefficient calculated was that between CI and OD,
which was 0.66 but increased to 0.87 when the MR5-II
combinations were excluded, demonstrating that for
modern high contrast films the screen–film contrast
increases with OD. It must be clarified, however, that this
correlation has been assessed for ODs up to 2.0 and it is
expected that for higher ODs the CI will start to decrease
again as the films become saturated.
The poor correlations of image quality parameters (TS,
specks, masses, fibres, resolution and CI) with film contrast
(c, AG) and the poor correlation of RDI with SRDI,
confirmed that film performance is strongly affected by
the screen. Concerning the lack of correlation of TS with CI,
resolution and dose, the following remarks should be
made. While combinations with high TS generally had high
CI, there were many cases with high CI and low TS. There
were also many combinations with high TS that, however,
exhibited low resolution score and vice versa. Finally, slow
combinations did not always give high TS, as expected
according to the general principle that the higher the dose
the lower the quantum mottle.
Discussion
The major conclusion of this study was that film
characteristics are modified by intensifying screens in
such a significant and variable way, that comparisons
among films based on the manufacturer’s specifications
or sensitometry are of limited value. Indeed, a film with
given technical specifications or sensitometric characteri-
stics, when combined with different screens may exhibit
improved or degraded performance.
Instructive of the variable effect that a screen may have
on a film, it can be seen that while Min-R 2000 was the
fastest film and remained the fastest when combined
with the Fuji screen, the RDI ratio of CM-H and Min-R
2000 combinations with Fuji screen was 1.04 while for
film only the respective ratio of SDRI was 1.32. That
means that the Fuji screen spectral emission better
matched the spectral sensitivity of CM-H compared
with Min-R 2000. The slowest film, according to
sensitometry, was the AD-M, which remained the slow-
est when combined with all screens. The SDRI for AD-M
was 167 while its smallest RDI was 191 and it was
observed when combined with the Fuji screen.
Examples of the largest variations observed in score
and dose when a film was combined with different
screens are: the HM plus where the TS was 36% larger
with the Fuji than with the Agfa screen (Table 2c), the
Min-R 2000 where the CI was 46% and 22% larger with
the Agfa than with the Kodak screen (Tables 2a and 2c,
respectively), the Min-R 2000 where the resolution was
26% larger with the Agfa than with the Fuji screen (Table
2b) and the HT (Agfa) where 63% more dose is required
with the Agfa screen than with the Fuji screen.
As previously mentioned, from Tables 2–4 some
conclusions may be drawn concerning the superiority
of certain combinations over others in terms of image
quality or speed. However, the absolute values of scores
and other screen–film characteristics may be quite
different on other mammographic facilities, given the
strong dependence of film characteristics on processing
conditions [13, 14]. This must be emphasised, as the
objective of this study was not to recommend or
condemn certain films or screens but to investigate the
effect of screens on the performance of films. Although
most screen–film comparisons in the literature have been
carried out using the same processing conditions for all
films, it must be noted that the general notion is that a
film would perform optimally when it is processed
according to the recommendations of the manufacturer.
Even so, this does not annul the fact that the breast dose
and image quality for a film optimally processed will
again vary, depending on the screen with which it is
combined, and that some films will be affected by the
screen more than others.

128 The British Journal of Radiology, February 2006

Image quality and breast dose of screen–film mammography
Even if it were assumed that the processing conditions
were optimal for all films, it would again be difficult to
select the best screen–film combination from those
studied, as there are no established criteria about what
increase in breast dose is justified by a superior image
quality. For example the AD-MA/AD-M presented a TS
of 7.3 (11 specks, 6 masses, 3 fibres) and an ESAK of
6.2 mGy while the AD-MA/Min-R 2000 a TS of 6.4 (9
specks, 5 masses, 4 fibres) with an ESAK of 2.4 mGy. To
conclude which is the best combination, one has to
decide if the 14% increase in TS could justify the 158%
increase in breast dose. The same question still holds
when considering that certain combinations (as the AD-
MA/AD-M) exhibited slightly larger TS for larger ODs
but with disproportional increase in breast dose.
An important remark should also be made concerning
the OD of the films studied. It is obvious that the films
included in Table 2a are of too low OD and few of the
films included in Table 2c are of too high OD, compared
with the target OD range of 1.3 to 1.8 proposed for
mammography [1]. Nonetheless, certain combinations
exhibited better scores in Table 2a than in Tables 2b and
2c, while most of the films of Table 2c with ODs larger
than 1.8 exhibited scores similar to those of Table 2b. In
clinical practice, however, given that the wide OD
variations within a mammogram are not uncommon,
some areas may present similar ODs with those of Table
2a or larger than 1.8 and therefore the performance of a
screen–film combination within all OD ranges is of
interest. In this context, comparisons based on the TSm
may be considered more relevant to the clinical situation
than comparisons based on the TS within only one OD
range. The variability of TS with OD should always be
considered when selecting the central OD setting of the
automatic exposure control (AEC) system based on the
results of phantom scores.
Some final comments should be made concerning the
method used to assess the image quality of screen–film
combinations. Phantom scoring does not always represent
clinical practice, as in actual mammograms the perfor-
mance of a given combination will be also dependent on
the breast type [15]. Furthermore, phantom scoring may
be somewhat biased, as it relies on the detection of
structures known to be present at specific positions [16].
Nevertheless, phantoms are considered as the best way
for the objective evaluation of image quality and various
models with fixed or randomly positioned details are
extensively used. Caldwell et al [17] agreed on the
usefulness of such phantoms for the objective evaluation
of image quality and also reported that a subjective
assessment of image quality is better accomplished with
an anthropomorphic breast phantom than with actual
mammograms, where the variability among radiologists
was higher. However, they noted that no significant
correlation was found between the various methods used
to evaluate image quality and concluded that more work
is required to obtain an index of true image quality
correlated with the probability of correct diagnosis.
In conclusion, image quality and dose in mammogra-
phy are more strongly dependent on screen–film
combination than on film or screen separately. While
sensitometry remains an important tool for determining
and monitoring the film characteristics [18], it is of little
value when the image quality and breast dose in clinical
The British Journal of Radiology, February 2006
mammograms are of concern. Therefore, any change of
film or screen type in a mammographic facility should be
carefully investigated with a phantom, for determining
the performance of the selected screen–film combination
and for adjusting the AEC system to the optimum OD
range for this combination.
References
1. van Woudenberg S, Thijssen M, Young K. European
protocol for the quality control of the physical and technical
aspects of mammography screening. In: Perry N, Broeders
M, de Wolf C, Kirkpatrick A, Tornberg S, editors. European
guidelines for quality assurance in mammography screen-
ing (3rd edn). Luxembourg: Office for Official Publications
of the European Communities, 2001.
2. Brink C, De Villiers JFK, Lötter MG, Van Zyl M. The
influence of film processing temperature and time on
mammography image quality. Br J Radiol 1993;66:685–90.
3. Tabar L and Haus AG. Processing of mammographic films:
technical and clinical considerations. Radiology 1989;173:65–9.
4. Kimme-Smith C, Bassett LW, Gold RH, Zheutlin J,
Gornbein JA. New mammography screen/film combina-
tions: imaging characteristics and radiation dose. AJR Am J
Roentgenol 1990;154:713–9.
5. Kirkpatrick AE, Law J. A comparative study of films and
screens for mammography. Br J Radiol 1987;60:73–8.
6. Robson KJ, Kotre CJ, Faulkner K. The use of a contrast-
detail test object in the optimization of optical density in
mammography. Br J Radiol 1995;68:277–82.
7. McParland BG, Boyd MM, Yousef KAL. Optimizing optical
density of a Kodak mammography film-screen combination
with standard-cycle processing. Br J Radiol 1998;71:950–3.
8. McParland BJ. A comparison of two mammography film-
screen combinations designed for standard-cycle proces-
sing. Br J Radiol 1999;72:73–5.
9. Kimme-Smith C, Bassett LW, Gold RH, Chow S. Increased
radiation dose at mammography due to prolonged expo-
sure, delayed processing and increased film darkening.
Radiology 1991;178:387–91.
10. Fernandez JM, Guibelalde E. Technical note: Physical
evaluation of recent Kodak films for mammography. Br J
Radiol 1993;66:828–32.
11. Nassivera E, Nardin L. Daily quality control programme in
mammography. Br J Radiol 1996;69:148–52.
12. Pisano ED, Britt GG, Lin Y, Schell MJ, Burns CB, Brown ME.
Factors affecting phantom scores at annual mammography
facility inspections by the U.S. Food and Drug
Administration. Acad Radiology 2001;8:864–70.
13. Tsalafoutas IA, Dimakopoulou AD, Koulentianos ED,
Serefoglou AN, Yakoumakis EN. The variation of the
sensitometric characteristics of seven mammographic films
with processing conditions. Br J Radiol 2004;77:666–71.
14. Kimme-Smith C, Rotschild PA, Bassett LW, Gold RH, Moler C.
Mammographic Film-Processor Temperature, Development
Time and Chemistry. AJR Am J Roentgenol 1989;152:35–40.
15. Meeson S, Young KC, Rust A, Wallis MG, Cooke J,
Ramsdale ML. Implications of using high contrast mammo-
graphy X-ray film-screen combinations. Br J Radiol
2001;74:825–35.
16. Jackson VP, Harrill CD, White SJ, Gillespie KR, Mail JT, Katz
BP. Evaluation of a dual-screen, dual-emulsion mammogra-
phy system. AJR Am J Roentgenol 1989;152:483–6.
17. Caldwell CB, Fishell EK, Jong RA, Weiser WJ, Yaffe MJ.
Evaluation of mammographic image quality: pilot study
comparing five methods. AJR Am J Roentgenol
1992;159:295–301.
18. West MS and Spelic DC. Using light sensitometry to
evaluate mammography film performance. Med Phys
2000;27:854–60.
129
 The British Journal of Radiology, 79 (2006), 130–141

The effect of phantom type, beam quality, field size and field
position on X-ray scattering simulated using Monte Carlo
techniques
G McVEY, DPhil

Joint Department of Physics, The Royal Marsden NHS Trust, Fulham Road, London SW3 6JJ, UK

ABSTRACT. Determining the amount of scatter inside and outside a diagnostic X-ray
room is important for evaluating the dose to staff and the public. The amount of
scatter is affected by many physical factors including beam quality and field size.
However, there is little published data on patient scatter and there are large
differences between the available data sets. Hence, a Monte Carlo code was developed
to allow a systematic study of the factors affecting patient scatter. A voxel phantom
was used to provide a realistic model of the patient. The variation of scatter with
different phantom types was investigated to show the effect of patient
inhomogeneities and obliquities. The effect of altering tube voltage, filtration, voltage
ripple, field size and position on patient scatter was studied. A larger than expected
variation in the patient scatter was observed with increasing field area due to the Received 25 October 2004
proximity of the field borders with the patient obliquities. The effect of the tube Revised 7 June 2005
voltage ripple on the patient scatter was also calculated. This showed that there would Accepted 15 June 2005
be little effect on the scatter levels within X-ray rooms if ageing X-ray generators,
DOI: 10.1259/bjr/59998010
which produce substantial voltage ripple, were replaced by X-ray tubes with modern
medium frequency generators. Recommendations are made on the choice of published ’ 2006 The British Institute of
scatter data for X-ray room design. Radiology

Scatter is produced by all materials in a diagnostic X-
ray room, with the main source of scattered radiation
being the interaction of X-rays with the patient [1].
However, there is a limited amount of scatter data
available for use in X-ray room design. The data were
obtained from measurements undertaken with tissue
equivalent slab phantoms [2]; with human-shaped
homogeneous phantoms [3] and with heterogeneous
phantoms such as the RANDO phantom [4–6]. Since
these studies used a variety of phantom types and
technique parameters, there was a large variation in the
scatter values reported. An alternative solution is to use
Monte Carlo computer simulations of the scatter pro-
duced by a model of human anatomy. This has enabled
three systematic studies of the effect of different
parameters on scatter from patients undergoing
chest posteroanterior (PA), lumbar spine anteroposterior
(AP) and lumbar spine lateral (LAT) radiographic
examinations.
For the first study, four voxel phantoms (P1 to P4)
have been used to simulate the patient as shown in
Figure 1. These different approaches have been followed
to investigate the effect of patient obliquities and
inhomogeneities on scatter. The first approach was to
use a voxel phantom (P1) reconstructed from CT data. It
was developed by Zubal et al [7, 8] and was recently
used in a Monte Carlo model to optimize image quality
Current address: North Wales Medical Physics, Glan Clwyd
Hospital, Bodelwyddan, Denbighshire LL18 5UJ, UK.
This work was supported by a grant from Anglia and Oxford
Health Authority.
and patient dose in chest and lumbar spine radiography
[9, 10]. Dance et al [11] showed that Zubal’s voxel
phantom was representative of a patient undergoing
chest and lumbar spine radiographic examinations. The
second approach was to use the voxel phantom devel-
oped by Zubal and change all the voxels inside the
patient contour to be soft tissue and those outside to be
air (P2). The third approach was to use the voxel
phantom with all the voxels within the phantom to be
soft tissue (P3). The fourth voxel phantom was devel-
oped as a block of soft tissue specified by the average
dimensions of Zubal’s voxel phantom (P4).
The second study used the Monte Carlo code to
calculate the effect of varying the imaging parameters on
the scatter from the patient model (P1): the tube voltage
(60–150 kV), tube filtration (2.5–7.0 mmAl) and voltage
ripple (0–50%). By studying the effect of voltage ripple
on patient scatter, it can be observed whether replacing
an old X-ray generator, which has substantial voltage
ripple, with a modern X-ray generator, which has
negligible voltage ripple, will make a significant differ-
ence to the scatter levels inside and outside X-ray rooms.
The third study used the Monte Carlo code to calculate
the effect on the scatter from the patient model (P1)
of varying the field area (25–1225 cm2) and the position
of the field on the patient. This study generalized the
results for the chest and lumbar spine regions so that the
data may be interpreted for other X-ray examinations.
The calculated scatter values obtained in this work may
be used to aid the design of X-ray rooms, but they may
also assist in the analysis of the doses received by staff

130 The British Journal of Radiology, February 2006

Simulating scatter from patients
who undertake and assist with interventional radiolo-
gical examinations.
Methods and materials
Voxel Monte Carlo code
The Monte Carlo code is similar to that used
previously to study image quality and patient dose in
radiographic examinations [9, 10], but was extended to
simulate the scatter surrounding a voxel phantom. The
program transports the photons through the voxel
phantom; a collision density estimator [12] is used to
provide an efficient method of calculating scatter. The
model calculates the air kerma at points 1 m from the
phantom surface for scattering angles between 30 ˚ and
150 ˚. Scatter ratios were determined by the air kerma at
each of these points divided by the incident air kerma
without backscatter. The scatter ratios are expressed as
percentages. A large number of photon histories were
used to calculate this parameter so its uncertainty was
less than ¡1% (1 standard deviation).
The patient model was a voxel phantom (P1) derived
from segmented CT data [7, 8]. Each voxel belonged to 1
of 55 organs [10]. The tissue type of each organ was
specified as one of average soft tissue, healthy lung, bone
The British Journal of Radiology, February 2006
Figure 1. The simulation model
used to calculate the scatter from a
patient undergoing, for example, a
chest posteroanterior X-ray
examination. The position of the
detector is shown at a scattering
angle of 135 ˚. The three other
phantoms used in the calculations
are also shown below.
or bone spongiosa. The calculations used tissue densities
and compositions taken from the International
Commission on Radiation Units and Measurements
(ICRU) Report No. 46 [13], except for bone which was
taken from Kramer [14]. The patient support device, i.e.
the chest stand for the chest examination or couch top for
the lumbar spine examinations, was included in the
voxel phantom by the addition of an extra layer of
voxels. Table 1 shows the thickness and composition of
the chest stand and couch top. The dimensions of the
voxel phantom were 89.9 cm long, 35.6 cm wide and
21.4 cm thick. As the lower limbs were not present in the
phantom, its length was determined to be equivalent to
the height of the average European male in sitting
position. The shoulder width and chest thickness were
determined after an initial study [11, 15] which com-
pared calculations with measurements of patient
entrance air kerma.
Figure 1 shows the computer model of a patient
undergoing a radiographic examination for which the
scatter was calculated. The model included the X-ray
spectrum from the X-ray tube, the patient and the couch
top or chest stand. The X-ray spectra were calculated
using a Birch and Marshall [16] model. The grid and the
screen–film imaging system were not included in the
model. This means that the Monte Carlo model will
produce significantly greater forward scatter than would
131

Table 1. The parameters used for the chest and lumbar spine imaging systems
Imaging system parameters
Parameter type Chest
Tube voltage (kV) 141
Filtration (mmAl) 5.7
Target angle ( ˚ ) 15
Voltage ripple (%) 0 0
Focus film distance (cm) 150
Focus surface distance (cm) 127
Field size at focus surface distance (cm6cm) 30.6621.2
Chest stand/couch top 6.0 mm of wood
AP, anteroposterior; LAT, lateral.
be observed clinically if a grid and film cassette were
present (or grid and image intensifier for fluoroscopic
imaging systems). The scatter from the grid and film
cassette would have been negligible as the patient
significantly attenuates the X-ray beam. Therefore, the
forward directed patient scatter calculated by the Monte
Carlo model is a conservative estimate of the clinical
situation.
Table 1 shows the imaging system parameters. These
parameters was found to provide good image quality in
a recent EU clinical trial [17, 18] and were thus used as a
reference system to observe the differences in patient
scatter when the imaging system parameters were
varied.
In the first study, the effect of the patient heterogene-
ities and obliquities on scatter were investigated.
Therefore, in addition to the patient model (P1)
described above, the other three phantoms shown in
Figure 1 were used to simulate scatter from a patient
undergoing chest PA, lumbar spine AP and lumbar spine
lateral examinations. Phantom P2 was defined with the
voxels inside the patient’s surface set to average soft
tissue and those outside the surface set to air. Phantom
P3 was a slab phantom defined with all voxels, apart
from the chest stand or couch top, set to be average soft
tissue. Phantom P4 was also a slab phantom defined by
the average thickness (z direction) and width (y direc-
tion) of the patient model within the field borders for
each projection. Table 2 shows the dimensions of the P4
phantoms including the chest stand or couch top. This
was undertaken as the shoulder width was considerably
larger than the width further down the phantom’s body
outline. Hence, it was interesting to determine which
slab phantom (P3 or P4) scatter approximated the scatter
from the patient model most closely. The patient model
Table 2. The dimensions of the homogeneous voxel phan-
toms (P4) defined by the average patient model dimensions
Phantom dimensions (cm)
Examination Length Width Thickness
Chest PA 89.9 22.0 16.0
Lumbar spine AP 89.9 24.0 18.0
Lumbar spine LAT 89.9 17.6 24.4
PA, posteroanterior; AP, anteroposterior; LAT, lateral.
132
G McVey
Lumbar spine
72 (for AP projection),77 (for LAT projection)
4.7
12
146
112 (for AP projection); 98 (for LAT projection)
27.5610.9 (for AP projection); 23.469.2 (for LAT
projection)
1.2 mm Al
(P1) was used for the other two studies described in the
introduction.
Validation of the patient model
Sandborg et al [9] and McVey et al [10] describe the use
of the voxel Monte Carlo code to simulate image quality
and patient dose. As part of this work, Dance et al [11]
and Sandborg et al [15] compared measurements of
optical density behind phantoms and patient entrance air
kerma with calculations using the Monte Carlo code for
both of these situations. The good agreement obtained
from the comparisons showed that the voxel phantom
(P1) was representative of a patient undergoing chest
and lumbar spine X-ray examinations [11, 15].
Simulation of Williams’ scattering experiment
This section describes the method used to compare the
scatter calculated using the voxel Monte Carlo code with
the scatter measured by Williams [5]. This was carried
out to validate the calculations against recent indepen-
dently published values and also to check the reliability
of Williams’ measured values.
Williams measured the scatter from the abdominal
and pelvic sections of a RANDO phantom. Therefore, a
voxelized cylinder of Alderson Muscle A material was
used to simulate the RANDO phantom of dimensions
50.0 cm long, 25.0 cm wide and 21.5 cm thick. Its
composition and density were obtained from ICRU
Report No. 44 [19]. Williams [5] measured the scatter in
terms of air kerma normalized to the dose–area product
(DAP). Therefore, a DAP meter and the air between it
and the phantom surface were included in the voxel
model. McVey [20] showed that these materials produce
a significant amount of scatter. The DAP meter was
modelled as a solid block of Perspex with dimensions:
16.4 cm length, 18.1 cm width and 1.7 cm thickness. An
average density of 0.315 g cm23 was used for the
Perspex as the DAP meter was constructed from layers
of Perspex 0.2 cm thick with an air gap between them.
The DAP meter was assumed to be at a distance of
26.6 cm from the X-ray focus and the focus to surface
distance (FSD) was 80 cm. The incident field at the
phantom surface was 22 cm long and 17.5 cm wide. The
The British Journal of Radiology, February 2006
Simulating scatter from patients
scatter was calculated at points 1 m from the centre of
the phantom for scattering angles between 30 ˚ and 150 ˚.
Results and discussion
Validation of the scatter calculations
The scatter calculations were validated by comparing
the scatter at points surrounding a block of solid water
calculated with the voxel Monte Carlo code to the values
calculated by an EGS4 Monte Carlo code for the same
geometry as described by McVey [20]. Good agreement
(within 2%) was shown between the values calculated by
the two codes for photon energies between 20 keV and
150 keV and for tube voltages between 50 kV and
120 kV.
The voxel Monte Carlo code could not be used to
simulate the scatter measurements previously carried
out in a clinical X-ray room as detailed by McVey and
Weatherburn [1]. This was due to the geometrical
limitations of the code and the size of the simulated
model. McVey and Weatherburn [1] used the EGS4
Monte Carlo code to calculate the scatter from solid
water blocks placed within a simulated X-ray room and
showed reasonable agreement with the measured scatter.
For this simulation, the percentage scatter contribution
from the X-ray room walls to the total calculated scatter
Figure 2. The variation of the percentage scatter with different phantom types (P1 to P4) for (a) the chest posteroanterior
projection, (b) the lumbar spine anteroposterior projection and (c) the lumbar spine lateral projection.
The British Journal of Radiology, February 2006
was determined to be small, being 3.7% [20]. Thus, the
scatter calculated by the voxel Monte Carlo code should
be reasonably representative of the scatter levels found
in clinical X-ray rooms.
Dependence of percentage scatter on phantom
type
Figure 2 shows the scatter from the four phantoms
calculated for the chest PA, lumbar spine AP and lumbar
spine LAT projections, respectively. The scatter for the
patient (P1) and the contoured phantom (P2) lie between
the values for the thick slab phantom (P3) and the slab
phantom with average dimensions (P4) below scattering
angles of 62 ˚ for the lumbar spine AP view, below 108 ˚
for the chest PA view and below 125 ˚ for the lumbar
spine LAT view. Therefore, the phantom with the
average dimensions (P4) can provide a conservative
estimate of the scatter from a patient (P1) for all
scattering angles for the chest PA and lumbar spine
LAT exams and for scattering angles less than 67 ˚ and
greater than 131 ˚ for the lumbar spine AP exam.
By comparing Figures 2a–c, it can be seen that the
chest PA examination produced the largest amount of
scatter as the highest tube voltage and largest field area
were employed. The lungs also attenuated the scattered
X-rays less than soft tissue. The lumbar spine AP
133
 G McVey

Table 3. Percentage scatter for different phantom types (P1 to P4) relative to the percentage scatter for the patient model (P1)
Percentage scatter for a 45 ˚ scattering angle

Phantom type Chest PA exam Lumbar spine AP exam Lumbar spine LAT exam

Thick slab phantom (P3) 0.62 0.56 0.70

Contoured phantom (P2) 1.12 1.29 1.14
Patient model (P1) 1.00 1.00 1.00
Slab phantom with average dimensions (P4) 1.36 1.46 1.22
Percentage scatter for a 120 ˚ scattering angle

Phantom type Chest PA exam Lumbar spine AP exam Lumbar spine LAT exam

Thick slab phantom (P3) 1.16 0.91 0.96

Contoured phantom (P2) 1.17 1.04 1.08
Patient model (P1) 1.00 1.00 1.00
Slab phantom with average dimensions (P4) 1.32 0.99 1.22
PA, posteroanterior; AP, anteroposterior; LAT, lateral.

projection produced the least scatter in the forward
direction as it had the lowest tube voltage and provided
attenuation by a large thickness of soft tissue. The
lumbar spine AP projection produced more scatter in the
backward direction than the lumbar spine LAT projec-
tion as it had a larger field area.
The largest differences between the different phantom
types occurred between scattering angles of 30 ˚ and 87 ˚.
The largest difference was 77% for the lumbar spine AP
projection at 87 ˚. Table 3 shows the scatter for all the
phantoms relative to the patient model for 45 ˚ and 120 ˚
scattering angles as examples of forward and backward
directed scatter. The thick slab phantom (P3) produced
the least scatter in the forward direction as it had the
largest thickness and width and, therefore, greatly
attenuated the scattered X-rays. The contoured phantom
(P2) produced more scatter than the thick slab phantom
in the forward direction. The phantom obliquity attenu-
ated the scattered X-rays less. The contoured phantom’s
thickness and width varied along its length, and in
places these were larger than the phantom with average
dimensions (P4). Thus, the contoured phantom produced
less scatter than the phantom with average dimensions.
The scatter from the patient model (P1) was lower than
from the contoured phantom (P2) in the forward
direction. Bone attenuated the scattered X-rays more
than soft tissue in all the examinations. In the chest
examination, fewer scatter interactions occurred in lung
than soft tissue as its density was lower.
For different examinations, Table 3 shows that the
differences between the phantom types were larger in
the forward direction for the lumbar spine AP projection
compared with those for the chest examination as a
lower tube voltage was used for the lumbar spine AP
projection. For the lumbar spine LAT examination, the
differences were not as large as would be expected for
the large thickness of the patient in the lateral projection.
The field edge was close to the phantom boundary.
Therefore, there was less tissue to attenuate the forward
scattered photons.
Figure 2 shows that the phantom type had less effect in
the backward direction than in the forward direction for
the frontal projections. The scatter was produced near the
entrance surface of a phantom [20]. For the lateral
projection, changes in the phantom width were more
significant. The effect of the tissue inhomogeneities in the

Table 4. The variation of percentage scatter from the patient (P1) for different tube voltages normalized to the scatter values
for reference imaging systems
Percentage scatter relative to the reference imaging system values

45 ˚ scattering angle 120 ˚ scattering angle

kV Chest PA exam Lumbar spine Lumbar spine Chest PA exam Lumbar spine Lumbar spine
AP exam LAT exam AP exam LAT exam

60 0.64 0.87
70 0.91 0.90 0.97 0.94
72 1.00 1.00
77 1.00 1.00
80 1.23 1.05 1.06 1.04
90 0.68 1.51 1.17 0.89 1.15 1.10
95 1.22 1.12
102 0.76 0.92
110 0.83 1.97 1.36 0.94 1.23 1.18
130 0.95 0.99
141 1.00 1.00
150 1.03 1.01
PA, posteroanterior; AP, anteroposterior; LAT, lateral.

134 The British Journal of Radiology, February 2006

Simulating scatter from patients

backward direction was similar to their effect in the
forward direction for the chest PA and lumbar spine LAT
projections. For the lumbar spine AP projection, there was
a larger difference between the forward and backward
directed scatter as bone volume occupied a greater
proportion of the irradiated volume than for the other
projections.
Dependence of percentage scatter on tube voltage
Table 4 shows the variation of scatter normalized to
the reference system values for X-ray tube voltages
between 60 kV and 150 kV for the different examina-
tions. The largest variation was for forward directed
scatter in all projections. The forward directed scatter
became more penetrating with increasing tube voltage.
The majority of backward scattered X-rays were pro-
duced close to the entrance surface [20]. Therefore, there
was less of an effect for increasing tube voltage.
The percentage scatter for the lumbar spine AP
projection had the largest variation with tube voltage.
This projection had the lowest reference system tube
voltage of 72 kV. Therefore, increasing the tube voltage
had a large effect. The percentage scatter in the forward
direction increased by a factor of 2 for an increase in the
tube voltage of 38 kV.
The lumbar spine LAT projection had a smaller
variation with tube voltage than the AP projection. The
LAT projection had a reference system tube voltage of
77 kV and the field was positioned closer to the patient’s
edge than the AP projection. The chest PA projection had
a smaller variation with tube voltage than both the
lumbar spine projections. This was due to the high
reference system tube voltage of 141 kV and because the
lungs attenuated the scattered X-rays less than tissue in
the chest PA projection.
Table 5 compares the variation of the lumbar spine AP
and LAT values with those from McVey and
Weatherburn [1], Trout and Kelley [3] and Williams [5].
There were large differences in the variations of forward
directed scatter with tube voltage due to differences in
the attenuating properties of the phantoms used. The
differences in the variations were less in the backward
direction as the scatter was produced close to the
entrance surface of the phantom [20]. The variation of
Trout’s scatter values was considerably larger than the
other values. This was due to the 70 kV values being
produced by a self-rectified X-ray tube which had a low
beam quality. The variation of Williams’ scatter values
was slightly less than the other values. This was due to
the significant amount of scatter produced by the DAP
meter which was independent of tube voltage [1].
Dependence of percentage scatter on tube
filtration and voltage ripple
Table 6 shows the variation of percentage scatter with
tube filtration and voltage ripple. The filtration was

Table 5. A comparison of the variation of percentage scatter with tube voltage in the literature [1, 3, 5]
Percentage scatter relative to 70 kV values for a 45 ˚ scattering angle

LS AP LS LAT Measured Calculated Trout and Williams

Tube voltage (kV) exam values exam values values [1] values [1] Kelley [3] values [5] values

50 0.52 0.84
60 0.70
70 1.00 1.00 1.00 1.00
80 1.35 1.17
85 1.13
90 1.65 1.30
95 1.35
100 4.76 1.28
110 2.16 1.51
120
125 6.19 1.51
150 8.10
Percentage scatter relative to 70 kV values for a 135 ˚ scattering angle

LS AP LS LAT Measured Calculated Trout and Williams

Tube voltage (kV) exam values exam values values [1] values [1] Kelley [3] values [5] values

50 0.77 0.80 0.79 0.87

60 0.90
70 1.00 1.00 1.00 1.00 1.00 1.00
80 1.07 1.10
85 1.08
90 1.16 1.15
95 1.17
100 1.75 1.10
110 1.22 1.22
120 1.26 1.30
125 2.00 1.23
150 2.08
PA, posteroanterior; AP, anteroposterior; LAT, lateral.

The British Journal of Radiology, February 2006 135

 G McVey

Table 6. The variation of percentage scatter from the patient model (P1) for different filtrations and voltage ripples normalized
to the reference system scatter values
Percentage scatter normalized to reference values for a 45 ˚ scattering angle

Chest PA exam Lumbar spine AP exam Lumbar spine LAT exam

Filtration (mmAl) 2.5 5.7 7.0 2.5 3.5 4.7 2.5 3.5 4.7

% Voltage Ripple

0 0.80 1.00 1.06 0.70 0.84 1.00 0.80 0.92 1.00

5 – 0.99 – – – 0.94 – – 0.98
20 – 0.96 – – – 0.85 – – 0.94
50 – 0.90 – – – 0.78 – – 0.90
Percentage scatter normalized to reference values for a 120 ˚ scattering angle

Chest PA exam Lumbar spine AP exam Lumbar spine LAT exam

Filtration (mmAl) 2.5 5.7 7.0 2.5 3.5 4.7 2.5 3.5 4.7

% Voltage Ripple

0 0.89 1.00 1.03 0.83 0.91 1.00 0.84 0.93 1.00

5 – 1.00 – – – 0.98 – – 0.99
20 – 0.99 – – – 0.94 – – 0.97
50 – 0.98 – – – 0.91 – – 0.94
PA, posteroanterior; AP, anteroposterior; LAT, lateral.

varied between 2.5 mmAl and 7.0 mmAl for the chest PA
examination and between 2.5 mmAl and 4.7 mmAl for the
lumbar spine examinations. A larger range in filtration was
investigated for the chest PA examination due to the higher
tube voltage employed. The voltage rectification was varied
between 0% and 50% ripple for all examinations.
Table 6 shows that the tube filtration affected the
percentage scatter less than the tube voltage due to the
smaller differences in the incident beam qualities
simulated. The voltage ripple had a similar effect on
the percentage scatter as the tube filtration. Therefore,
changing an old X-ray generator with significant voltage
ripple to a medium frequency X-ray generator will not
produce significantly more scatter.
The percentage scatter for the lumbar spine AP
projection had the largest variation with filtration and
voltage ripple. The scatter in the forward direction
decreased by 15% if the filtration decreased by
1.2 mmAl or the voltage ripple decreased by 20%. The
variations in percentage scatter for the lumbar spine LAT
and chest PA projections were less due to their higher
beam qualities.
Similar variations of scatter with different filtrations and
voltage ripples were observed at 102 kV, 90 kV and 95 kV
for the chest PA, lumbar spine AP and LAT projections,
respectively. Therefore, the variations shown in Table 6
are applicable over a large range of tube voltages.
Dependence of percentage scatter on field area
Table 7 shows the percentage scatter for a 100 cm2
square field area which was used to normalize the scatter
values for the different field areas and field positions
shown in Table 8. All the reference system parameters
for each projection, as shown in Table 1, were employed
except for the field area. Table 7 shows the percentage
scatter in the lumbar spine LAT projection calculated
with the field centre at two positions. First, the field
centre was positioned at the same place as the reference

Table 7. The percentage scatter for the chest PA, lumbar spine AP and LAT reference imaging systems with 100 cm2 square field
areas
Percentage scatter

Examination Chest PA exam Lumbar spine AP exam Lumbar spine LAT exam

Imaging system Reference system Reference system Reference system Field centred on centre of patient

Scattering angle ( ˚ )

30 0.044 0.012 0.039 0.014

45 0.035 0.011 0.048 0.021
60 0.027 0.013 0.050 0.026
87 0.024 0.024 0.050 0.031
120 0.049 0.048 0.058 0.041
135 0.059 0.057 0.062 0.049
150 0.067 0.065 0.066 0.056
PA, posteroanterior; AP, anteroposterior; LAT, lateral.

136 The British Journal of Radiology, February 2006

Simulating scatter from patients

Table 8. Percentage scatter for different square field areas relative to the percentage scatter for the 100 cm2 square field areas
given in Table 7
Percentage scatter normalized to the value for a 10 cm610 cm field size at a 45 ˚ scattering angle

Exam Chest PA exam Lumbar spine AP exam Lumbar spine LAT exam Lumbar spine LAT exam

Imaging system Reference system Reference system Reference system Field centred on patient centre

Field size Expected

(cm6cm) values

565 0.25 0.25 0.21 0.20 0.20

10610 1.00 1.00 1.00 1.00 1.00
15615 2.25 2.20 2.86 1.61 3.38
20620 4.00 4.07 7.55
25625 6.25 6.90 11.92
30630 9.00 9.45 14.52
35635 12.25 11.64 17.38
Percentage scatter normalized to the value for a 10 cm610 cm field size at a 120 ˚ scattering angle

Exam Chest PA exam Lumbar spine AP exam Lumbar spine LAT exam Lumbar spine LAT exam

Imaging system Reference system Reference system Reference system Field centred on patient centre

Field size Expected

(cm6cm) values

565 0.25 0.26 0.25 0.26 0.26

10610 1.00 1.00 1.00 1.00 1.00
15615 2.25 2.22 2.25 1.63 2.29
20620 4.00 3.94 3.86
25625 6.25 6.01 5.19
30630 9.00 7.78 6.13
35635 12.25 9.34 7.16
PA, posteroanterior; AP, anteroposterior; LAT, lateral.

system, which was close to the patient’s lateral bound-
ary. Second, the field centre was positioned at the centre
of the patient’s width. In both cases, the scatter was
calculated at points on the side of the patient where the
field was off-centre. Table 7 shows that the position of
field centre had a considerable effect. The percentage
scatter was larger for all scattering angles with the field
centred on the patient’s obliquity as there was less tissue
to attenuate the scattered radiation. The scatter for the
field centred on the patient obliquity was 2.3 times
greater than for the field centred on a thicker part of the
patient for a scattering angle of 45 ˚. Trout and Kelley [3]
found the scatter for the field centred at the phantom’s
edge was between 6.5 and 1.8 times larger than the
scatter for the field positioned at the centre of the
phantom’s width for tube voltages between 50 kV and
150 kV at a 45 ˚ scattering angle.
Shielding reports [6, 21] indicate a linear relationship
between scatter and field area. Table 8 shows that this
relationship is not valid for the variation of patient
scatter with field area. For a 25 cm 6 25 cm field area
incident in the chest PA view, the normalized scatter was
10.5% higher and 3.9% lower than expected for scattering
angles of 45 ˚ and 120 ˚. The forward directed scatter
values for the chest PA view were closest to the expected
variation with field area. The largest differences were
observed for the forward directed scatter in the lumbar
spine AP view. For the same field area incident in the
lumbar spine AP view, the normalized scatter was 90.7%
higher and 17.0% lower than expected for scattering
angles of 45 ˚ and 120 ˚.
The variation of scatter with field area shown in
Table 8 was due to the patient obliquities. A field width
of 25 cm covered the majority of the patient’s trunk.
The field edges were incident on the patient obliquities
and thus, the scattered X-rays were less attenuated.
The scatter was therefore larger than expected in the
forward direction. The patient’s obliquity produced a
small reduction in the amount of scatter produced in
the backward direction due to there being less tissue.
There was a greater difference for the lumbar spine
view than the chest PA view due to the lower tube
voltage.
A smaller number of field areas were investigated for
the lumbar spine LAT exam due to the small width of the
patient in this orientation. Table 8 shows the normalized
scatter values with the field centre close to the patient
obliquity (reference system position) and the field
positioned at the centre of the patient’s width. With the
225 cm2 field centred medially on the patient, the
normalized values were 3.38 and 2.29 for scattering
angles of 45 ˚ and 120 ˚. With the field moved laterally by
3 cm, the normalized scatter values were smaller and
thus, there was a larger variation if the field was centred
over a thick part of the patient than if the field was
centred on the patient obliquity. However, the actual
scatter values tended to be larger for the field positioned
at the patient obliquity than at the patient centre as there
was less attenuation of the scattered X-rays (Table 7). If
the scatter was calculated at points on the other side of
the patient than the field centre, then the increased
attenuation would substantially reduce its amount.

The British Journal of Radiology, February 2006 137

 G McVey

Figure 3. The variation of the normalized percentage scatter at a 45 ˚ scattering angle with square, rectangular and equivalent
square field sizes between 25 cm2 and 900 cm2 for (a) the lumbar spine anteroposterior projection and (b) the chest
posteroanterior projection.

Figure 3 shows the variation of scatter with square,
rectangular and equivalent square field areas between
25 cm2 and 900 cm2 for a 45 ˚ scattering angle for the
lumbar spine AP and chest PA projections. The
rectangular field length was kept constant at 30 cm and
the width increased from 5 cm to 30 cm. In X-ray room
design it is difficult to account for the variation of field
dimensions used in the clinical situation. Therefore, one
method investigated was to calculate the equivalent
square field area (s2) using the equivalent square field
length (s) and the rectangular field dimensions (x and y)
shown in Equation (1) [22].
2xy
s~ ð1Þ
ðxzyÞ
Figure 3a shows large differences in the variation of
scatter between square and rectangular field sizes for the
lumbar spine AP projection as a low tube voltage was
employed. There was an 82% difference between the
rectangular and square fields with areas of 300 cm2 (i.e.
10 cm 6 30 cm for the rectangular field). For the
rectangular field, the scattered X-rays were more
attenuated as the smaller width covers thicker parts of
the patient. For the square field, the scattered X-rays
were less attenuated as the field edge was closer to the
patient obliquity. Figure 3a shows that the equivalent
square field approximated the variation of scatter for
square fields well when the smallest rectangular field
dimension was greater than or equal to 10 cm. For the
smallest field dimension of 5 cm, the equivalent square
field overestimated the scatter for the square field.
Figure 3b shows that for the chest PA projection, the
differences in the scatter between square and rectangular
field sizes were less than for the lumbar spine AP view
due to the lower attenuation of the lung and the higher
tube voltage. For example, for field areas of 300 cm2, a
difference of 11.4% was observed at a 45 ˚ scattering
angle. The scatter for the rectangular fields was closer to
the square field values than the equivalent square field
values where the smallest rectangular field dimension
was less than or equal to 10 cm. For both examinations,
the variation of backward directed scatter with field area
was similar for both square and rectangular fields.
Therefore, the scatter value for the largest appropriate
square field should be employed in X-ray room design to
provide a conservative dose estimate. The scatter from
large rectangular fields and their equivalent square field
sizes tends to be less than the scatter from square fields.
Comparison of scatter values in the literature
Table 9 shows a summary of the imaging parameters
stated in the literature which were used to obtain the
scatter values shown in Table 10. All the imaging system

Table 9. The parameters used for measuring scatter in the different reports in the literature [1–3, 5, 20]
Image parameters

Parameter type Trout and Kelley [3] Bomford and Burlin [2] McVey and Weatherburn Williams [5]
[1, 20]

Tube voltages (kV) 50, 70, 100 100 50, 70, 100a 50, 70, 100
Filtration (mmAl) 0.5, 1.5, 2.5 – 3.4 3.5
Generator type Self-rectified and – Medium frequency, Constant potential
single phaseb 3% ripple
Phantom material Torso-shaped masonite MixD block Solid Water block RANDO phantom
(Alderson A muscle)
Phantom dimensions 72630620 30630622 30.5630.5610 50625621.5
(cm6cm6cm)
Incident field size 20 cm620 cm 400 cm2 (circle) 20 cm620 cm 22 cm617.5 cm
FSD (cm) 100 – 100 80
a
100 kV for calculation only.
b
50 kV and 70 kV were used with a self-rectified generator and 100 kV was used with a single phase generator.

138 The British Journal of Radiology, February 2006

Simulating scatter from patients

Table 10. Comparison of published scatter values [1–3, 5, 20] and the scatter calculated using the voxel Monte Carlo code (MC)
in this paper for tissue equivalent materials
Percentage scatter values for 50 kV X-rays

Source of data 30 ˚ 45 ˚ 60 ˚ 90 ˚ 120 ˚ 135 ˚ 150 ˚

Trout and Kelley [3] 0.044 0.011 0.012 0.024 0.069 0.095 –
Measured values [1] – – – – – 0.181 –
Calculated values [1] 0.081 0.059 0.048 0.041 0.114 0.140 0.222
Williams [5] 0.071 0.070 0.073 0.120 0.221 0.279 0.314
Voxel MC calculated values 0.059 – 0.068 0.100 0.162 – 0.321
Percentage scatter values for 70 kV X-rays

Source of data 30 ˚ 45 ˚ 60 ˚ 90 ˚ 120 ˚ 135 ˚ 150 ˚

Trout and Kelley [3] 0.052 0.021 0.020 0.035 0.091 0.120 –
Measured values [1] – 0.137 – 0.078 0.195 0.237 0.252
Calculated values [1] 0.123 0.090 0.068 0.052 0.143 0.181 0.269
Williams [5] 0.084 0.083 0.086 0.137 0.253 0.322 0.358
Voxel MC calculated values 0.080 – 0.090 0.126 0.196 – 0.363
Lumbar spine AP view values (72 kV)a 0.072 0.084 0.096 0.119 0.185 0.218 0.244
Percentage scatter values for 100 kV X-rays

Source of data 30 ˚ 45 ˚ 60 ˚ 90 ˚ 120 ˚ 135 ˚ 150 ˚

Trout and Kelley [3] 0.127 0.100 0.098 0.110 0.190 0.210 –
Bomford [2] 0.026 0.039 0.052 0.065 0.156 0.221 0.273
Calculated values [20] 0.180 0.133 0.098 0.070 0.176 0.220 0.309
Williams [5] 0.108 0.106 0.108 0.163 0.291 0.355 0.412
Voxel MC calculated values 0.106 – 0.116 0.153 0.228 – 0.395
AP, anteroposterior.
a
20 cm 6 20 cm incident field size.

parameters were selected to be as similar as possible for
this comparison. The range of tube voltages studied was
restricted to be from 50 kV to 100 kV. All scattering
materials used were tissue equivalent. The phantoms
used by Trout and Kelley [3] and Williams [5] had a
human body contour. The voxel Monte Carlo code was
used to model the experimental set up for Williams
scatter measurements (as detailed above).
The scatter values in Table 10 were corrected where
necessary to the same units of percentage scatter.
Bomford and Burlin [2] measured the percentage scatter
for incident air kerma on the surface of the phantom.
Thus, a backscatter factor of 1.3 [22] was used to give the
values in Table 10 as the ratio of scattered air kerma to
incident air kerma without backscatter. Williams [5]
reported scattered air kerma divided by the DAP.
Williams’ values in Table 10 were converted to be in
terms of scattered air kerma divided by the incident air
kerma without backscatter. These values were also
increased to be equivalent to a field area of 400 cm2,
which was the same field area as the other studies,
instead of 385 cm2 as shown in Table 9.
Table 10 shows that there was good agreement
between the calculated and measured scatter values
reported by McVey and Weatherburn [1] and the
calculated values given in the previous section and with
the measured values given by Williams [5]. These
agreements give confidence in both the calculations
and measurements. Both studies [1, 5] employed modern
equipment, including a DAP meter, in the experimental
set up. A FSD of 100 cm was used in the work of McVey
and Weatherburn and a FSD of 80 cm was used in the
work of Williams (Table 9). Williams’ values were
considerably greater than those of McVey and
Weatherburn for scatter in the backward direction. The
difference in FSD produced a difference in the position of
the DAP meter which resulted in large differences in the
scatter in the backward direction. McVey and
Weatherburn’s values were greater than Williams’
values in the forward direction due to the smaller
phantom thickness. Scatter in the backward direction
was less affected by changes in phantom thickness than
scatter in the forward direction [20].
The patient scatter values, calculated for the lumbar
spine AP view at 72 kV, were similar to Williams’ values
in the forward direction and less than Williams’ values in
the backward direction (Table 10) as a DAP meter was
not included in the scatter calculations.
Table 10 shows that there was poor agreement in the
forward directed scatter values reported by Bomford and
Burlin [2] and Trout and Kelley [3]. For example at a 30 ˚
scattering angle, Bomford’s values were 0.2 times smaller
than Trout’s. It is difficult to understand the reason for
the differences between Trout’s and Bomford’s results as
their phantoms had similar thicknesses (Table 9).
Bomford and Burlin had corrected their values for
scatter from the surroundings and leakage from the X-
ray tube head. These contributions were a large propor-
tion of the total reading as the scatter from the phantom
was small in the forward direction. McVey [20] calcu-
lated that the scatter from the collimators, ceiling, floor
and walls varied between 0.03% and 0.05%. This
accounted for some of the differences which were
between 0.045% and 0.101%, but also suggested that

The British Journal of Radiology, February 2006 139


the masonite and MixD phantoms may have substan-
tially different attenuating properties. The scatter values
were similar in the backward direction for the two
phantoms as the scatter was less dependent on the
phantom thickness or density.
Conclusions
Accurate determination of the scatter in X-ray rooms is
important for designing shielding to meet the desired
radiation protection requirements. Previous studies have
used a variety of phantoms to estimate these scatter levels
and, as a review of the literature has shown, there are large
differences in the published scatter values. The work in
this paper has determined the magnitude of scatter from
patients undergoing diagnostic X-ray procedures with the
imaging system parameters varied in a systematic manner
to provide a comprehensive data set.
The voxel Monte Carlo calculations have demon-
strated that the linear relationship between scatter and
field area, as used in shielding reports [6, 21], is not valid
for patient irradiation. For example, the scatter was 91%
larger than the expected value for increasing the area of a
square field from 100 cm2 to 625 cm2 for a patient
undergoing an X-ray examination in the lumbar spine
region. The position of the field on the patient in relation
to the calculation points also had an effect. The scatter
from a patient undergoing a lumbar spine LAT exam
increased by 2.3 times for the centre of the field being
moved from the centre of patient’s width closer to the
patient’s obliquity with the calculation points on the
same side of the patient. If scatter was calculated on both
sides of the patient and the field centre moved laterally,
then the scatter distribution would become asymmetric
i.e. the scatter would be higher on one side compared
with the other. Thus, the calculation points on the same
side of the patient as the lateral shift would provide a
conservative estimate of the scatter. As well as X-ray
room design, this work can be applied to estimate the
doses received by staff who undertake interventional
procedures.
The Monte Carlo calculations have also demonstrated
small variations in patient scatter, in particular for
changing the tube voltage ripple. For example, the
scatter from a patient undergoing a lumbar spine AP
exam increased by 22% if an X-ray generator with a
voltage ripple of 50% was replaced by a constant
potential X-ray generator, whereas the scatter
increased by 97% if the tube voltage increased by
38 kV. Therefore, replacing X-ray generators with sub-
stantial voltage ripple by medium frequency units would
not produce sufficiently more scatter to warrant a change
to the X-ray room design (using the same dose
constraint).
A review of published scatter values [1, 5] has
suggested that there was a FSD dependence on back-
ward directed scatter when a DAP meter was present.
However, Marshall and Faulkner [4] found no FSD
dependence for air kerma measured adjacent to the
couch for a 90 ˚ scattering angle, i.e. at a position forward
of the DAP meter. Marshall and Faulkner imply that the
FSD was simply increased, leading to an increase in field
size incident on the phantom, which may explain the
140
G McVey
constancy in their scatter measurements. McVey and
Weatherburn [1] showed that there was a large variation
in backward directed scatter from the DAP meter.
Therefore, further work is necessary to investigate these
effects, their influence on patient scatter and their
possible impact on shielding barrier calculations.
For X-ray room design, the largest scatter values
provided by either McVey and Weatherburn [1] or
Williams [5] are recommended to provide a conservative
dose estimate at any FSD. The scatter from the patient
detailed in this paper may also be considered. In this
case, the significant scatter from the surroundings (e.g.
the DAP meter and X-ray collimators) [1] should be
taken into account. The inconsistent forward scatter
values given by Trout and Kelley [3] and Bomford and
Burlin [2] are not recommended for use even for X-ray
units which have significant voltage ripple. In this case,
the scatter data from this work, McVey and Weatherburn
[1] or Williams [5] should be used and modified by the
ratios detailed in this paper depending on the amount of
voltage ripple. The work in this paper can also be used to
study the effect of changing tube voltage, filtration,
voltage ripple, field area and field position on the patient
scatter. For example, in X-ray room design, these factors
can be used to increase the recommended scatter values
or independently measured scatter values to provide a
conservative dose estimate as appropriate to the clinical
situation.
Acknowledgments
I would like to thank Dr David Dance, Prof. Gudrun
Alm Carlsson and Dr Michael Sandborg for providing the
voxel Monte Carlo code which was the basis of the scatter
calculations. I also acknowledge the use of the computer
facilities at the Physics Department, Royal Marsden
Hospital, London, where all the Monte Carlo simulations
described in this paper were undertaken.
References
1. McVey G, Weatherburn H. A study of scatter in diagnostic
X-ray rooms. Br J Radiol 2004;77:28–38.
2. Bomford CK, Burlin TE. The angular distribution of
radiation scattered from a phantom exposed to 100–
300 kVp X-rays. Br J Radiol 1963;36:426–39.
3. Trout ED, Kelley JP. Scattered radiation from a tissue
equivalent phantom for X-rays from 50 to 300 kVp.
Radiology 1972;104:161–9.
4. Marshall NW, Faulkner K. The dependence of scattered
radiation dose to personnel on technique factors in
diagnostic radiology. Br J Radiol 1992;65:44–9.
5. Williams JR. Scatter dose estimation based on dose area
product and the specification of radiation barriers. Br J
Radiol 1996;69:1032–7.
6. Sutton DG, Williams JR. Radiation shielding for diagnostic
X-rays. Joint report of the BIR/IPEM working party, 2000.
7. Zubal G, Harrell CR. Voxel based Monte Carlo calculations
of Nuclear Medicine images and applied variance reduction
techniques. Image Vision Computing 1992;10:342–8.
8. Zubal G, Harrell CR, Smith EO, Rattner Z, Gindi G, Hoffer
PB. Computerised three dimensional segmented human
anatomy. Med Phys 1994;21:299–302.
9. Sandborg M, McVey G, Dance DR, Alm Carlsson G.
Schemes for the optimization of chest radiography using a
The British Journal of Radiology, February 2006
Simulating scatter from patients
computer model of the patient and X-ray imaging system.
Med Phys 2001;28:2007–19.
10. McVey G, Sandborg M, Dance DR, Alm Carlsson G. A
study and optimization of lumbar spine X-ray imaging
systems. Br J Radiol 2003;76:177–88.
11. Dance DR, McVey G, Sandborg M, Persliden J, Alm
Carlsson G. Calibration and validation of a voxel phantom
for use in the Monte Carlo modeling and optimization of X-
ray imaging systems. Proc. SPIE Medical Imaging
1999;3659:548–59.
12. Persliden J, Alm Carlsson G. Calculation of the small-angle
distribution of scattered photons in diagnostic radiology
using a Monte Carlo collision density estimator. Med Phys
1986;13:13–24.
13. International Commission on Radiological Units and
Measurements. Photon, electron, proton and neutron
interaction data for body tissues. ICRU Report No. 46,
Bethesda, MD: ICRU, 1992.
14. Kramer R. Determination of conversion factors
between body dose and relevant radiation quantities for
external X- and c-radiation. GSF Bericht-S-556, Neuherberg:
GSF, 1979.
15. Sandborg M, McVey G, Dance DR, Persliden J, Alm
Carlsson G. A voxel phantom based Monte Carlo computer
program for optimisation of chest and lumbar spine X-ray
imaging. Radiat Prot Dosim 2000;90:105–8.
The British Journal of Radiology, February 2006
16. Birch R, Marshall M. Computation of bremsstrahlung X-ray
spectra and comparison with spectra measured with a
Ge(Li) detector. Phys Med Biol 1979;24:505–17.
17. Almén A, Tingberg A, Mattsson S, Besjakov J, Kheddache S,
Lanhede B, et al. The influence of different technique
factors on image quality of lumbar spine radiographs as
evaluated by established CEC image criteria. Br J Radiol
2000;73:1192–9.
18. Lanhede B, Båth M, Kheddache S, Sund P, Björneld L,
Widell M, et al. The influence of different technique factors
on image quality of chest radiographs as evaluated by
modified CEC image quality criteria. Br J Radiol
2002;75:38–49.
19. International Commission on Radiation Units and
Measurements. Tissue substitutes in radiation dosimetry
and measurement. ICRU Report 44, Bethesda, MD: ICRU,
1989.
20. McVey GH. Monte Carlo computing applied to X-ray room
design. D.Phil. Thesis, University of Oxford, 2002.
21. British Standards Institution. Recommendations for data
on shielding from ionizing radiation: part 2. shielding from
X-radiation. British Standard 4094, Part 2, London: BSI,
1971.
22. British Institute of Radiology. Central axis depth dose data
for use in radiotherapy: 1996. BJR Supplement 25, London:
British Institute of Radiology, 1996.
141
 The British Journal of Radiology, 79 (2006), 142–147

Techniques for measurement of dose width product in panoramic

dental radiography
P DOYLE, MSc, C J MARTIN, PhD and J ROBERTSON, PhD

Health Physics, Department of Clinical Physics and Bio-Engineering, Gartnavel Royal Hospital,
Glasgow G12 OXH, UK

ABSTRACT. Dose width product (DWP) is the quantity recommended for assessment of
patient dose for panoramic dental radiography. It is the product of the absorbed dose
in air in the X-ray beam integrated over an exposure cycle and the width of the beam,
both measured at the receiving slit. A robust method for measuring the DWP is
required in order to facilitate optimization of practices and enable comparison of dose
levels at different centres. In this study, three techniques for measuring the DWP have
been evaluated through comparison of results from 20 orthopantomographic units.
These used a small in-beam semiconductor detector and X-ray film, a pencil ionization
chamber and an array of thermoluminescent dosemeters (TLDs). The mean results
obtained with the three techniques agreed within ¡6%. The technique employing a
pencil ionization chamber of the type used for dose assessment of CT scanners is the
simplest and most reliable method. The in-beam detector and film method has larger
errors both from positioning the radiation detector and from measurement of X-ray Received 7 March 2005
beam width, which should be the full width at half maximum obtained from a scan of Revised 7 June 2005
the film optical density. The TLD array method was accurate, but more time consuming Accepted 15 June 2005
to carry out. The mean DWP for the units studied was 65 mGy mm and the mean dose–
DOI: 10.1259/bjr/33207232
area product was 89 mGy cm2. The DWP for 30% of the units tested exceeded the
diagnostic reference dose of 65 mGy mm, recommended by the National Radiological ’ 2006 The British Institute of
Protection Board. Radiology

Panoramic radiography is a technique used in den-
tistry to show the mandibular joints with the teeth laid
out between them. The X-ray tube and film holder both
rotate during the exposure. The film is exposed to a
narrow X-ray beam through a secondary collimator slit,
across which the film moves as the radiographic image is
built up. The assessment of patient dose in panoramic
radiography is difficult because of the dynamic nature of
the imaging process and the narrow width of the X-ray
beam. The dose quantity used is the product of the
absorbed dose in air and the horizontal width of the
beam, both measured at the front side of the secondary
collimator slit, and integrated over a standard exposure
cycle. This is referred to as the dose width product
(DWP) with units of mGy mm. The DWP provides a
measurement related to the total amount of radiation to
which the patient is exposed. It can be derived either
from the product of the peak dose at the centre of the X-
ray beam and the width of the beam, or from an
incremental summation of the dose across the beam. The
aim of this study has been to compare and evaluate
results obtained from different techniques available for
measuring DWP.
As part of the National Radiological Protection Board
(NRPB) dental X-ray protection service, Napier reported
DWPs for 387 panoramic dental X-ray sets derived from
a technique that employed film to assess both dose and
beam width [1]. Based on results from this survey, the
Dental Guidance Notes recommend that a DWP of
65 mGy mm should be adopted as a local diagnostic
reference level (DRL) for a standard adult panoramic
radiograph [2]. Results from this study have been
compared with this value.
Method
The dose measurement techniques used in this study
were:
(a) ‘‘In-beam’’ detector and film: Measurement of peak
dose within the X-ray beam at the receiving slit
using a small solid state detector and determination
of the beam width using X-ray film [3]. The DWP is
calculated from the product of peak dose and beam
width.
(b) Partial volume detector: Direct measurement of the
summation of dose across the beam obtained from
the partial volume irradiation of a pencil ionization
chamber [4, 5].
(c) Thermoluminescent dosemeter (TLD) array: Measure-
ment of dose at the receiving slit using a linear array
of TLDs. This method can evaluate the DWP either
from the incremental summation of dose across the
beam or from the peak dose multiplied by the beam
width [3, 6].
More details of the techniques used and the measure-
ments made in evaluating them are given in the
following paragraphs.

142 The British Journal of Radiology, February 2006

Dose measurement for paranormic dental radiography
‘‘In-beam’’ detector and film
A solid state detector that has an active width of
1.5 mm, which is marketed for measurement of the
DWP, was used with an Unfors 511 Mult-O-Meter
(Unfors Instruments AB, Billdal, Sweden).
Measurement showed that the length of the sensitive
region was approximately 4 mm. The detector was
attached to the front side of the secondary collimator
parallel to the slit and aligned visually with the slit. It is
important that the detector is aligned accurately with the
X-ray beam and is sufficiently narrow to enable it to lie
entirely within the region of the dose peak in order to
give an accurate result.
An assessment of the spatial response across the solid
state detector was made using an X-ray beam from a
radiographic unit collimated by a 0.2 mm wide lead slit.
The detector was moved perpendicular to the slit in
0.2 mm steps by means of a micromanipulator with a
vernier scale.
Images of the X-ray beam at the receiving slit of each
orthopantomographic (OPT) X-ray unit were obtained by
exposing Kodak T-mat L film. When this technique is
employed, care is needed to avoid saturating the film.
The width of the beam was obtained by measuring the
film optical density with a microdensitometer (MKIII CS;
Joyce-Loebl Ltd, Gateshead, UK) and determining the
full width at half maximum (FWHM) (technique A1).
However, a simple measurement using a ruler with a
light box (technique A2) has been recommended [3] and
this was also used in order to determine whether the
errors involved were significant.
Partial volume detector
A pencil ionization chamber commonly used for CT
dose index measurements (model No. 20X5-CT with a
MDH 2025 electrometer; Radcal Corporation, Monrovia,
NY) was attached in front of the secondary collimator,
perpendicular to the slit. The DWP was taken as the
product of the partial volume irradiation reading and the
active length of the chamber (100 mm) (technique B).
TLD array
This technique involved measuring the dose profile at
the receiving slit using an array of 34 TLDs mounted in a
Perspex jig with 1 mm thick walls and lid. The TLDs
used were high sensitivity LiF:Mg:Cu:P TLD-100H chips
(0.38 mm thick and 3.2 mm diameter), calibrated in a
70 kVp X-ray beam against air kerma in air, measured
using a 6 cm3 chamber and a Radcal 9010 electrometer.
The TLDs were placed on their edge, side by side in the
jig, which was then positioned in the centre of the
secondary collimator perpendicular to the slit. The TLDs
were read out using a Harshaw 5500 TLD reader (Qados,
Sandhurst, UK). The dose that each TLD received was
obtained by correcting the readout for background
radiation and applying a batch calibration factor. The
spacing of the TLDs in the jig was determined from a
measurement of the length of the arrays and found to be
0.40 mm. The DWP was calculated from the sum of the
The British Journal of Radiology, February 2006
doses received by all the TLDs d1–d34, multiplied by the
spacing w (technique C1) i.e.:
DWP 5 w(d1 + d2 +...+ d34)
The doses recorded were plotted against position in
the jig to give a profile of the dose distribution across the
slit (Figure 1). The DWP was also calculated from the
product of the maximum dose at the centre of the beam
and the FWHM value (technique C2). Comparison of
techniques C1 and C2 was used to confirm that the dose
summation and the product of peak dose and FWHM
gave similar results for the DWP.
Study method
Detectors and TLDs were all calibrated with respect to
an ionization chamber with a Keithley Triad 35050A
dosemeter system, which had a calibration traceable to a
national standard. Relative responses were measured
with the detector free in air and lying on a steel plate
with a slit overlying a cassette to simulate actual
exposure conditions. Based on these measurements,
results obtained with the CT chamber and the TLDs
were reduced by 5% to allow for the effect of backscatter.
The Unfors detector is shielded from backscatter to a
greater extent because of the metal its plate to its rear.
Measurements of DWP using the three techniques
were made on 20 different OPT X-ray units from eight
manufacturers (Table 1). The OPT units had been
installed at various times over the previous 25 years
and had an average age of 10 years. Successive
measurements were made using each of the three
techniques at the standard adult settings, typically; tube
potential 70 kV, tube current 10 mA and exposure time
16 s, and values for the DWP derived for each technique.
Experimental errors for the different techniques were
estimated to be A1 ¡16%, A2 ¡19%, B ¡7% and C1
¡8% and C2 ¡8%. Errors are expressed as percentages
for each result, combining errors from individual
components. The largest contributions were from the
measurement of beam width using a ruler and the
positioning of the Unfors detector.
The dose–area products (DAPs) for each unit were
calculated from the product of the DWP and the beam
length L [6]. The mean value for the DWP derived from
measurement techniques A1, B and C2 was employed.
The beam length L was measured with a ruler on a light
box using the film exposed for technique A.
Results
The OPT units studied had a range of beam widths
and examples of dose profiles obtained using TLDs for
OPT units with average beam widths of 2.5 mm and
4 mm are shown in Figure 1. DWP results obtained
using the different techniques are shown in Table 1. The
two sets of values for the DWP obtained from the TLD
data using different calculation methods, i.e. from dose
summation and from the product of the peak and the
FWHM, are compared in Figure 2. The ratio of the DWPs
measured using the two techniques is 0.96¡0.02 (mean
¡ standard error of the mean (sem)), confirming that the
143
 P Doyle, C J Martin and J Robertson

Figure 1. Dose profiles from

orthopantomographic (OPT) unit with
beam widths of 2.5 mm and 4 mm,
measured using thermoluminescent
dosemeters (TLDs), compared with the
measured sensitivity profile across the
width of the Unfors detector.

two methods give results which agree within experi-
mental error. There is also reasonable agreement
between the DWP results obtained using the TLDs,
technique C1 and those from the pencil ionization
chamber, technique B (0.91¡0.014, mean ratio ¡ sem)
and the in-beam and detector method using the FWHM,
technique A1 (0.85¡0.034, mean ratio ¡ sem), see
Figure 3. The average DWP given by the different
techniques are A1 61 mGy mm, A2 80 mGy mm, B
65 mGy mm, C1 72 mGy mm and C2 69 mGy mm. The
standard deviation between the techniques A1, B and C2
averaged for the 20 units was 13%.
The Unfors detector should provide a reasonably
accurate measurement of the peak dose, if it is positioned
at the centre of the X-ray beam. The measured sensitivity
response across the width of the Unfors detector is
compared with two X-ray beam profiles in Figure 1.
Misalignment of the detector and the centre of the beam
by distances of 0.5 mm, 1 mm and 2 mm would give
measurements lower by 2%, 16% and 59%, respectively,
for a 4 mm beam width, and by 5%, 27% and 73%,
respectively, for a 2.5 mm beam width. Results from
technique A were more scattered than those from
techniques B and C. Any misalignment between the
detector and the beam would give a lower value for the
peak dose and results for the DWP from technique A1
were slightly lower than for the other techniques.
The FWHM measured from films using the micro-
densitometer on five of the OPT units selected randomly,
agreed to within 3% with the FWHM derived from TLD
measurements. DWP measurements calculated using the
data measured by the Unfors detector multiplied by the
TLD profile FWHM, rather than the film FWHM, are
included in Table 1 (technique A1). This was because a
few of the films were saturated at the centre of the X-ray
beam and so could not be used. The average beam

Table 1. Dose width product (DWP) measurements taken on different panoramic X-ray models at the standard adult setting
Code Model/commissioned kV mA Time Nominal Beam width DWP (mGy mm):
(s) Film/screen speed (mm)
Technique Technique Technique
A1 B C2

I Sirona Orthophos (2002) 68 8 14 DR 3.5 43 38 45

II Instrumentarium OPT (2002) 73 8 18 DR 2.5 32 51 48
III Siemens Orthophos (1993) 69 15 14 Kodak T-mat L 400 3.0 40 53 48
IV Morita Inc. Panex EC (1982) 70 7 14 Kodak T-mat L 400 4.5 58 43 41
V Siemens Orthophos (1999) 74 14 13 Kodak T-mat L 250 4.1 44 52 65
VI Instrumentarium OPT (1998) 66 10 18 Kodak Ekta 400 2.8 54 56 53
VII Siemens Orthophos (1989) 66 15 15 Kodak T-mat L 400 2.3 45 57 61
VIII Orion Cranex DC2 (1982) 69 6 19 Kodak T-mat L 400 5.5 65 54 49
IX Planmeca PM2002CC (1990) 68 6 18 Kodak T-mat L 400 3.7 59 51 62
X Siemens Orthophos (1990) 60 16 14 Kodak T-mat L 400 3.2 54 55 63
XI Planmeca Proline (2002) 68 7 18 Kodak T-mat L 400 2.4 43 66 63
XII Planmeca PM2002CC (1990) 68 6 18 Kodak T-mat L 400 3.7 62 52 63
XIII Siemens Palomex (1987) 60 14 15 Kodak T-mat L 400 3.0 53 62 63
XIV Planmeca Proline (2002) 68 7 18 Kodak T-mat L 400 3.0 55 61 66
XV Morita Inc. Panex EC (1979) 70 8 16 Agfa Curix 250 4.5 60 66 75
XVI Planmeca Proline (2002) 68 7 18 CR 3.8 58 64 79
XVII Siemens Palomex (1976) 65 15 14 Kodak T-mat L 250 6.3 81 83 76
XVIII Soredex Cranex (1993) 81 10 16 Agfa HTG Ortho 250 3.6 65 93 93
XIX Yoshida Panoura (1990) 85 10 12 Ceahiplus 200 7.6 114 119 122
XX Morita Inc. Panex EC (1980) 90 10 16 Kodak T-mat L 250 6.9 128 128 152
DR, digital radiography; CR, computed radiography.

144 The British Journal of Radiology, February 2006

Dose measurement for paranormic dental radiography
FWHM of the OPT units included in this study is
4.0¡0.3 mm. The average beam width measured from
the film with a ruler and a light box was 20% higher
than the FWHM, and the DWP results calculated using
this (technique A2) were higher than those obtained
using the other techniques (Table 1, Figure 3). The
overestimation of the beam width was partially offset
by the lower dose resulting from misalignment of the
detector.
Values of DAP were calculated for each unit from the
product of the average DWPs derived from techniques
A1, B and C1 and measurements of the slit length L
(Figure 4). The average beam length L was 136¡2 mm
and the average DAP was calculated to be 89¡8 mGy
cm2. Values of the mean and third quartile DWP
and DAP are compared with results from other studies
in Table 2. The average DWPs were greater than
the proposed DRL of 65 mGy mm [2] for 30% of the
units. 400 speed index systems are recommended by
the European Guidelines on Quality Criteria for
diagnostic radiographic images [7]. The five units
with the highest DWPs and DAPs all used films
with speed indices of 200–250, while the two lowest
both used direct digital radiography (DR), (Table 1,
Figure 4).
The British Journal of Radiology, February 2006
Figure 2. Plot of dose width
products (DWPs) derived from
thermoluminescent dosemeters
(TLDs) showing the DWP derived
from the product of the peak dose
and full width half maximum
(FWHM) against the DWP from the
summation of the doses for all the
TLDs across the beam. The line of
identity is a 45 ˚ trendline.
Discussion
All three techniques gave results within reasonable
agreement, but the errors associated with the in-beam
detector and film technique A are larger than those for
techniques B and C. For technique A, a microdensit-
ometer was required for the measurements and it was
important that exposures were limited to avoid satura-
tion of the film emulsion in order that accurate results
could be obtained. Measurement of the beam width with
a ruler gave a result 20% greater than the FWHM and
this method is therefore not appropriate. Use of a 35 mm
film scanner (PrimeFilm 1800u; Pacific Image Electronics,
Torrance, CA) linked to a PC with appropriate software
(e.g. Scion Image; Meyer Instruments Inc., Houston, TX)
and a calibrated film test strip to allow optical densities
to be determined provides an inexpensive method for
film scanning if a microdensitometer is not available,
although this requires further limitation to be placed on
the exposure because the measurable range in optical
density is more limited. Another potential source of error
in technique A is the visual positioning of the detector.
The active area of the detector is 1.5 mm in diameter,
which is similar to the width of the dose peak (Figure 1)
with seven of the units studied having beam widths of
Figure 3. Plot of dose width
product (DWP) measurements using
an Unfors detector (techniques
A1 – with full width half maximum
(FWHM) derived from
thermoluminescent dosemeter (TLD)
profile and A2 – FWHM measured
with a ruler and film) and a pencil
ionization chamber (technique B),
against the DWP derived from
summation of doses across the beam
from TLDs (technique C1). The line
of identity is a 45 ˚ trendline.
145
 P Doyle, C J Martin and J Robertson

Figure 4. Bar chart showing

dose–area product (DAP) values for
a standard adult exposure for the 20
units studied. Data from film/screen
combinations with indices of
200–250 (F250, dark) and 400 (F400,
light), and from computed
radiography (horizontal lines) and
digital radiography (angled lines)
systems are indicated by different
shading.

3 mm or less (Table 1). The detector’s active area must be for the summation of the dose across the beam were
positioned within 0.5 mm of the centre of the beam to similar to the product of the peak dose and FWHM. The
keep errors to within ¡5%. A dedicated holder agreement is closer than that reported in a previous
incorporating a phosphor screen to facilitate alignment study [6], probably because the earlier study used TLDs
that were 0.85 mm thick. As a result, a limited number
of the detector with the X-ray beam is available from the
would lie within the X-ray peak, and this is likely to
supplier of the detector, although it was not used in this affect the accuracy of measurements of both the peak
study. dose and the FWHM.
The partial volume chamber method (B) is the most The average DWP from the three techniques assessed
direct and simplest of the three techniques. Errors in the in this study is similar to the reference dose recom-
technique result from the calibration of the ionization mended by the NRPB [1] and to the mean values
chamber and the magnitude of the backscatter. When reported in other studies by medical physics depart-
using this and technique A, care must be taken to ensure ments [6, 8] (Table 2). In the present study, the third
that the length of cable attached to the detector is quartile was not significantly different from the pro-
sufficient to account for the rotational movement posed DRL [2] because there were a significant number
involved in the scan. of units with similar DAPs (Figure 4). The third quartile
Method (C) using the TLD array is the most time values in other studies tended to be higher than the
consuming of the three techniques because of the DRL. This could reflect differences resulting from the
handling and processing of the high sensitivity TLDs, sample size or distribution, measurement technique
which are brittle and need to be handled with care. The or poorer optimization. It will also be influenced by
technique is accurate and so provides a useful method the 5% correction applied to account for backscatter in
for dose comparisons, but is not recommended for techniques B and C in the present study. DAP measure-
routine use. It was useful for confirmation that values ments were slightly less than results from other studies
(Table 2) [5, 6, 9].
Table 2. Comparison of results of this study with published Six of the units tested in this study had a DWP greater
data than the DRL of 65 mGy mm. Five of these were using a
film/screen combination with a nominal speed index of
Sample DWP (mGy mm): DAP (mGy cm2): 200 or 250, so adoption of a 400 speed system, which
size
Mean 3rd Mean 3rd
could potentially reduce these doses by 40–50%, has been
Quartile Quartile recommended with a proportionate reduction in expo-
sure levels. Two of the units had DWPs that were close to
This study 20 65 67 89 90
the recommended suspension level of 150 mGy mm [10]
Napier [1] 387 57 67
Isoadri and 5 74 84
and these units also had beam widths of 7–8 mm which
Ropolo [4] were significantly broader than the maximum recom-
Perisinakis and 6 113 mended value of 5 mm [2]. Investigation of the operation
Damilakis [5] of the units has been recommended in order to optimize
Williams and 16 65 76 113 139 the system set ups and so reduce the exposures.
Montgomery
[6]
Oduko [8] 26 69 80
Tierris et al 62 101 117 Conclusion
[9] (male) This study has measured the DWP using three
(female) 62 85 97
different techniques. The method using a semiconductor
DWP, dose width product. detector and film required the slit width to be assessed

146 The British Journal of Radiology, February 2006

Dose measurement for paranormic dental radiography
from the FWHM of the exposure peak, measured using a
microdensitometer, as use of a simple ruler measurement
[3] gave a result 20% greater than the true one.
Uncertainty in alignment of the detector with the X-ray
beam of more than 0.5 mm could result in a significant
error. If this technique is employed, a microdensitometer
and a dedicated alignment tool are recommended.
Use of a partial volume ionization chamber technique
(B) provides a simple, robust method for direct measure-
ment of the DWP, and is recommended as the technique
of choice. The measurements are simple to record, avoid
errors from positioning the radiation detector and do not
require a measurement of beam width. The pencil type
ionization chamber is also widely available and com-
monly used in diagnostic radiology departments for the
measurement of CT dose index.
Acknowledgments
The authors would like to thank Louise Lindsay
and Navneet Dulai for their assistance with the DWP
measurements.
10. Institute of Physics and Engineering in Medicine, College of
References
1. Napier D. Reference doses for dental radiography. Br
Dental J 1999;186:8.
The British Journal of Radiology, February 2006
2. NRPB. Guidance notes for dental practitioners on the safe
use of X-ray equipment. NRPB, Department of Health,
Chilton, UK: NRPB, 2001.
3. British Institute of Radiology. Assurance of quality in the
diagnostic imaging department (2nd edn). London, UK:
BIR, 2001:51.
4. Isoardi P, Ropolo R. Measurement of dose–width product
in panoramic dental radiology. Br J Radiol 2003;76:
129–31.
5. Perisinakis K, Damilakis J, Neratzoulakis J, Gourtsoiannis
N. Determination of dose–area product from panoramic
radiography using a pencil ionization chamber: normalized
data for the estimation of patient effective and organ doses.
Med Phys 2004;31:4.
6. Williams JR, Montgomery A. Measurements of dose in
panoramic dental radiology. Br J Radiol 2000;73:1002–6.
7. Commission of the European Communities. European
Guidelines on Quality Criteria for diagnostic radiographic
images. EUR 16260 EN. Brussels, Belgium: CEC, 1996.
8. Oduko J. Optimisation of patient dose and image quality in
dental radiology– Over 65 time to retire your OPG? IPEM
Meeting, York, March 2001.
9. Tierris CE, Yakoumakis EN, Bramis GN, Georgiou E. Dose
area product reference levels in dental panoramic radi-
ology. Radiat Protection Dosim 2004;111:283–7.
Radiographers, NRPB. Recommended Standards for the
Routine Performance Testing of Diagnostic X-ray Imaging
Systems, IPEM Report No. 77. IPEM: York, 1997.
147
 The British Journal of Radiology, 79 (2006), 148–157

A comparison of three-field and four-field techniques in different

clinical target volumes in prostate cancer irradiation using dose
volume histograms: a prospective three-dimensional analysis
A HILLE, MD, N TÖWS and C F HESS, PhD, MD

Department of Radiotherapy, University of Göttingen, Göttingen, Germany

ABSTRACT. The purpose of the current study was to quantitatively assess differences
between irradiation techniques on normal tissue exposure in different clinical target
volumes (CTVs) in irradiation of prostate cancer. 14 patients with prostate cancer
undergoing external beam radiotherapy were investigated. The prostate and prostate
+ proximal/entire seminal vesicles were delineated as CTVs. A three-field and two
different four-field plans were generated and compared concerning rectum, bladder
and femoral head dose–volume histograms (DVHs). The exposure of the rectum
exposed to 40–60 Gy was significantly lower for all CTVs with the three-field technique
compared with both four-field techniques. The exposure of the rectum to 70 Gy was
significantly lower for all CTVs with the weighted four-field technique compared with
the unweighted four-field and three-field techniques. The weighted four-field
technique was worst in bladder dose sparing for the three CTVs. Comparing the three-
field and the unweighted four-field technique for irradiation of the prostate and
prostate + entire seminal vesicles, no technique provided a clear advantage or
disadvantage in bladder dose sparing. For irradiation of the prostate + proximal
seminal vesicles the unweighted four-field technique provided the best bladder dose
sparing. Concerning the exposure of the femoral heads, the three-field technique was Received 11 April 2005
significantly worse for the three CTVs compared with both four-field techniques. No Revised 10 June 2005
difference was found between the unweighted and the weighted four-field Accepted 21 June 2005
techniques. In conclusion, none of the studied techniques consistently proved superior
DOI: 10.1259/bjr/10206556
in different CTVs in prostate cancer irradiation with respect to sparing all organs at risk.
The absolute differences between the three techniques were small and the clinical ’ 2006 The British Institute of
relevance of these findings is uncertain. Radiology

Three-dimensional (3D) conformal radiation treatment
with the use of individual multileaf collimators (MLCs)
has become the standard treatment technique for
localized prostate cancer [1–5]. The number of beams
and their orientation vary from one department to
another. The simplest techniques use three or four fields
[2, 4, 6–10], others use techniques with over five fields
[11, 12]. However, the published data do not indicate
that more sophisticated techniques increase the thera-
peutic index [13–19]. It is known that rectal toxicity
following external beam irradiation of prostate cancer
correlates with radiation dose and the percentage of
rectal volume included in the intermediate and high
dose-volumes [1, 4, 10, 12, 20]. Recently, the impact of
inclusion of the seminal vesicles in the clinical target
volume (CTV) on rectal dose has been recognized and a
risk-adapted CTV with exclusion of seminal vesicles or
inclusion of the proximal 2–2.5 cm of the seminal vesicles
was suggested to reduce the risk of rectal toxicity [21–
24].
Few studies compared different techniques concerning
irradiation of the prostate only [3, 14, 18], the prostate +
Address correspondence to: Dr Andrea Hille, Klinik für
Strahlentherapie, Robert-Koch-Str. 40, 37075 Göttingen, Germany.
base of the seminal vesicles [19], or the prostate + entire
seminal vesicles [14–16]. These studies draw differing
conclusions concerning the best irradiation technique,
which may partly be due to different definitions of the
CTV in these studies. None of these studies investigated
systematically whether there is a difference between
techniques concerning irradiation of different CTVs of
the prostate.
The purpose of the current study was to quantitatively
assess the differences between a simple three-field and
two different four-field techniques on irradiated normal
tissue exposure in irradiation of the prostate only, the
prostate + proximal and the entire seminal vesicles. The
evaluation was based on three-dimensional treatment
planning including dose–volume histograms (DVH). To
our knowledge, this is the first prospective systematic
analysis for the effect of treatment technique on normal
tissue exposure concerning three different CTVs in
prostate cancer irradiation.
Methods and materials
14 consecutive patients with localized prostate cancer
stage T1–2 undergoing external beam radiotherapy with

148 The British Journal of Radiology, February 2006

Treatment techniques in prostate cancer irradiation
curative intent to 72 Gy were investigated prospectively.
3D conformal computer-based planning was carried out
on CadPlan treatment planning system (Varian, Palo
Alto, CA). The prostate (P), the prostate + entire seminal
vesicles (PESV), or the prostate + proximal (PPSV) 2–
2.5 cm (approximately 60% in longitudinal direction) of
the seminal vesicles were taken as CTV and a planning
target volume (PTV) margin of 1 cm was added. The
definition of the proximal seminal vesicles was taken
from the literature [24]. The prostate, the entire and the
proximal seminal vesicles were delineated on each axial
slice on the planning computer. The external wall of the
rectum was contoured. The craniocaudal rectal extension
was defined as the first CT slice above the anal verge
(caudal border) and the cranial limit was defined as the
first slice below the sigmoid flexure. This definition is
consistent with definitions reported in the literature [7, 8,
15, 25]. The external wall of the bladder was contoured.
One planning CT scan (5 mm continuing, 5 mm slice)
was carried out with patients in supine position and a
comfortably filled bladder. Irradiation technique
included individual optimization with conformal treat-
ment planning and the use of individual blocks. Nine
plans were produced for each of the 14 patients.
Three different irradiation techniques using 20 MV
photons were evaluated.
(1) Four-field box technique with equally weighted
fields (so-called unweighted four-field technique for
simplification);
(2) Four-field box technique with unequal weighted
fields (so-called weighted four-field technique for
simplification);
(3) Three-field technique with one anterior and two
lateral fields with 90 ˚ and 270 ˚ wedges.
For technique 2 the weight of the ventral field versus
dorsal field was 1.3:0.7, and 1:1 for the lateral fields; for 3 it
was 1.3:0.85:0.85 with the highest weight for the anterior
field. For techniques 2 and 3, minor modification of the
beam weights were performed in order to homogenise the
dose distribution inside the PTV. Dose was specified
according to the ICRU 50 report [26]. For all techniques the
reference point for dose specification was the same. Dose
was specified at the centre of the treatment field in
projection of the central axes. Concerning dose homo-
geneity, at least 95% of the PTV was covered by 95% of the
prescribed dose as minimum. Field size was adjusted to
reach this dose homogeneity criterion. Dose calculation
included tissue density correction.
To determine the amount of the rectum exposed to
ionizing radiation, the percentage of the irradiated
rectum to 40 Gy, 50 Gy, 60 Gy and 70 Gy were calcu-
lated by the treatment planning system. Several investi-
gations indicate a relationship between DVHs and the
development of chronic rectal toxicity [1, 6, 7, 9, 10, 20,
27–31]. The rectal contouring varies from study to study
with some investigators outlining the whole rectum,
others the rectal wall. Concerning the rectal borders,
some studies outline the anatomic rectum, others the
rectum over the length of the fields. It is known that
there is a high variability of volume fractions of rectal
DVHs depending on how the rectal borders are defined,
and it is difficult to compare the results of different
The British Journal of Radiology, February 2006
studies concerning rectal DVHs [32, 33]. Studies analys-
ing either dose–volume relationships of the rectum with
or without the craniocaudal definition which we have
used in our study, or studies using an identical or similar
craniocaudal definition of the whole rectum, are sum-
marized in Table 1. The above mentioned values were
chosen following the rectal dose constraints given in
these publications [6–8, 27–31].
To determine the amount of the bladder exposed to
ionizing radiation, the volume of V100 (defined as the
percentage of bladder volume receiving 100% of the
prescribed dose) and the percentage of the irradiated
bladder to 40 Gy, 50 Gy, 60 Gy, 65 Gy and 70 Gy were
calculated by the treatment planning system. The data on
the tolerance of the bladder to radiation as a function of
the irradiated volume is limited. This may be due to a
large variation in the bladder DVHs when considering
the modifications of the organ due to different filling [9].
The above mentioned values were chosen following the
data in the literature about a relationship between
bladder toxicity and the irradiated bladder volume [4,
34–37]. The incidence of acute bladder toxicity increased
when more than 30% of the bladder received more than
65 Gy [4, 34, 35]. Late complications, such as bladder
contracture and volume loss, are described in 5% to 10%
at doses of 40 Gy delivered to the majority of the
bladder, at doses of 50–65 Gy delivered to about 30%
of the bladder volume and at doses of 65–75 Gy applied
to below 20% of the bladder volume [36]. Emami et al
estimated the TD5/5 to 65 Gy irradiated to the whole
bladder, and 66% of the bladder volume irradiated to
80 Gy [37].
To determine the amount of the femoral heads
exposed to ionizing radiation the volume of V50 and
V100 (defined as the percentage of femoral head volumes
receiving at least 50% and 100% of the prescribed dose)
were calculated by the treatment planning system. The
available data on the dose–effect relationship for femoral
heads are also limited [37, 38]. The clearest proposal is
that of Emami et al who indicated that a dose of 52 Gy
can be given to the whole femoral head with a risk for
chronic toxicity in 5 years to be 5%. The V100 value was
chosen following Emami et al’s study. However, the dose
to the whole femoral head is always lower than 52 Gy in
practice. Therefore, we additionally estimated the V50
value.
Statistical analysis
Analysis was performed using the program
STATISTICA 6.1 (Stat Soft, Palo Alto, CA). To evaluate
the statistical significance of differences, Friedman’s
ANOVA was performed followed by Wilcoxon matched
pairs test. Closure principle was used for multiple tests.
The level chosen for significance was .0.05.
Results
PTV
The dose distributions in the PTV for all three CTVs
were between 99.8% and 102% and typical standard
149
 A Hille, N Töws and C F Hess

Table 1. Relationship between dose–volume and late rectal toxicity reported in the literature
Authors Craniocaudal definition Rectal Rectal Risk for chronic Toxicity Used Median
of the whole rectum volume (%) dose (Gy) rectal grade toxicity follow up
bleeding (%) score

Cozzarini [6] Anal verge to the sigmoid ¢ 52 50 14 ¢2 RTOG 3 years

flexure (patients with large , 63 50 6.6
air/faecal content in the rectum ¢ 39 60 13.3
were excluded from analysis) , 39 60 7
Fiorino [7] Anal verge to the sigmoid 50 . 53 14.2 ¢2 RTOG 30
flexure (patients with large 50 , 53 4.3 months
air/faecal content in the rectum 60 . 42 14
were excluded from analysis) 60 , 42 5.2
70 . 22 12.9
70 , 22 6.3
Fiorino [8] Anal verge to the sigmoid 50 . 53 14.2 ¢2 RTOG 2 years
flexure (patients with large 50 , 53 4
air/faecal content in the rectum 60 . 42 14
were excluded from analysis) 60 , 42 5
70 . 22 13
70 , 22 6
Greco [27] Anal verge to the sigmoid flexure . 65 40 18 ¢2 RTOG 28
months
, 65 40 0
. 30 60 18
, 30 60 0
. 25 70 18
, 25 70 0
Huang [28] 11 cm in length starting at 2 cm . 26 70 54 ¢2 RTOG/ 6 years
below the inferiormost aspect , 26 70 13 LENT-SOMA
of the ischial tuberositas
Wachter [29] From the lower to the upper ¢ 57 60 31 2 RTOG/ 30
border of the fields , 57 60 11 EORTC months
Zapatero [30] Anus to the sigmoid flexure . 42 . 60 7.7 ¢2 RTOG 4 years
, 42 , 60 0
Storey [31] Rectum was identified with . 25 70 37 ¢2 RTOG/ 2 years
rectal contrast , 25 70 13 ¢2 LENT-SOMA
. 30 ¢ 70 40 3
, 30 ¢ 70 3

deviations ranged from 1% to 2.5% for all patients and all
considered CTVs and techniques.
No statistically significant differences were found
between the three techniques for all three CTVs.
Details are demonstrated in Table 2. The values of
40 Gy, 50 Gy, 60 Gy and 70 Gy for all three techniques
are demonstrated graphically for the prostate only in
Figure 1, for the PPSV in Figure 2 and for the PESV in
Figure 3.

Rectum volume
Bladder volume
The median volume of the rectum was 84 cm3 (mean
value 90 cm3, standard deviation 31 cm3). The median bladder volume was 154 cm3 (mean value
146 cm3, standard deviation 40 cm3).

Exposure of the rectum with different techniques

The exposure of the rectum to 40–60 Gy was sig-
nificantly lower for all CTVs with the three-field
technique compared with both four-field techniques.
The exposure of the rectum to 70 Gy was significantly
lower for all CTVs with the weighted four-field
technique compared with the unweighted four-field
and three-field techniques.
The differences between the rectal volume receiving
40 Gy, 50 Gy, 60 Gy and 70 Gy, respectively, were
significant for all three CTVs between the three-field
technique and both four-field techniques, and between
both four-field techniques.
Exposure of the bladder with different techniques
P
The bladder volume receiving 50 Gy, 60 Gy and
65 Gy, respectively, was significantly higher with the
weighted four-field technique compared with both
the unweighted four-field and the three-field techniques.
The bladder volume receiving 40 Gy was significantly
lower with the unweighted four-field technique com-
pared with the weighted four-field technique. No
significant difference was found in the bladder volume
receiving 40 Gy, 50 Gy, 60 Gy, 65 Gy and 70 Gy,

150 The British Journal of Radiology, February 2006

Treatment techniques in prostate cancer irradiation

p-value
respectively, between the unweighted four-field and the
Table 2. Mean and median values, standard deviations (SD) and p-values for the whole rectum exposed to 40–70 Gy in different techniques and clinical target volumes (CTVs)

0.003
0.001

0.02
0.02
three-field techniques. No significant difference

,
was found in the bladder volume receiving 70 Gy and
the proportion of the bladder volume receiving 100%

43

25
9

8
53

9
62

10

43

25
55
65
3
(V100) of the prescribed dose between the three
techniques.

44

22
56

9
9
69

0.001 44

0.01 19
0.02 56
0.001 71
9
p-value 2

PPSV

,
,

The bladder volume receiving 40 Gy was significantly

8
9
43

25
53
62

43

25
10

55
65

higher with the weighted four-field and three-field

3

techniques compared with the unweighted four-field

9

9
9
46

27
57
74

0.001 45

0.001 26
10

0.001 58
0.001 78
p-value 1

technique. The bladder volume receiving 50 Gy was

significantly higher with the weighted four-field techni-
que compared with the three-field technique.
,

,
,
,

The bladder volume receiving 60 Gy, 65 Gy and

9

9
9
44

22
9
56

44

19
69

56
71
2

70 Gy, respectively, was significantly higher with the

PESV

9
9
46

27
57

45

26
74

10

58
78

weighted four-field technique compared with both the

1

unweighted four-field and the three-field techniques.

p-value

0.004
0.001
0.001

Comparing the bladder volume receiving 60 Gy and

0.01

65 Gy, respectively, the three-field technique resulted in

,
,

a significantly lower value for 60 Gy and in no

6

7
7
36

20
8
45

39

21
54

48
56

significantly different value for 65 Gy compared with

3

the unweighted four-field technique. Concerning the

7

7
6
37

18
7
48

0.001 39

0.007 17
60

0.001 50
0.001 62

proportion of the bladder volume receiving 100% (V100)

p-value 2

of the prescribed dose, no significant difference was

found between the three techniques.
,
,
,

7
7
36

20
8
45

39

21
48
54
56

PESV
3

The bladder volume receiving 40 Gy was significantly

7

7
7
39

23
7
49

0.001 41

0.001 22
0.001 52
64
0.001 65

higher with the weighted four-field and the three-field

p-value 1

techniques compared with the unweighted four-field

technique.
,

,
,
,

The bladder volume receiving 60 Gy was significantly

higher with the weighted four-field compared with the
7

7
6
7

37
39

18
17
48
50
62
60
2

three-field technique. The bladder volume receiving

1, 2, 3 relates to the radiation techniques described in Methods and materials.
PPSV

50 Gy and 65 Gy was significantly higher with the

7

7
7
7

39
41

23
22
49
52
65
64
1

weighted four-field technique compared with the

p-value

unweighted four-field technique. The bladder volume

0.002
0.001
0.001

0.01

receiving 70 Gy was significantly lower with the three-

,
,

field technique compared with both the weighted four-

5

3
8
7
6
7

0.001 24 23

18 17
0.001 18 17
0.001 32 28

25 22
34 28

field and the unweighted four-field techniques.

3

The bladder volume receiving 100% (V100) of the

5
6
0.008 6
3
7
8
p-value 2

prescribed dose was significantly higher with the

weighted four-field technique compared with both
the unweighted four-field and the three-field techniques.
,
,
,

All other values showed no significant differences

17
23

17

5
8
7
3
28

22

6
28

between the techniques.

3

Details for the bladder dose exposure for the three

0.001 19
0.001 25

19

3
0.001 35
10
26

5
9
0.001 9
6
39

CTVs are given in Table 3.

p-value 1

,
,
,
,

Exposure of the femoral head with different

techniques
18
24

18

6
32

5
6

3
25

7
8
34
2

The femoral head DVHs were averaged over the left

and right sides to give a single value.
19
25

19

9
35
10

5
9

3
26

6
39
P

For all three CTVs the radiation dose to the femoral

heads was below 50 Gy with all three techniques. The
Median

Median
Median
Median

70 Gy Mean
60 Gy Mean
50 Gy Mean

mean V50 and V100 values for all three CTVs were below
40 Gy Mean
(%)

53% and 10% of the prescribed dose with the different

SD

SD
SD
SD

techniques. Concerning the V50 values, the three-field

technique was significantly worse for all CTVs compared
with both four-field techniques. No differences were

The British Journal of Radiology, February 2006 151


found between the unweighted four-field and the
weighted four-field techniques. Concerning the V100
values no significant difference was found for all CTVs
between the different techniques.
Discussion
Our study shows that none of the studied three-field and
four-field techniques consistently proved superior in
irradiation of the prostate, prostate + proximal seminal
vesicles and prostate + entire seminal vesicles with respect
152
A Hille, N Töws and C F Hess
Figure 1. Mean values, standard
error and standard deviations (SD)
for the rectum exposed to 40–70 Gy
in irradiation of the prostate only
with different treatment
techniques.
Figure 2. Mean values, standard
error and standard deviations (SD)
for the rectum exposed to 40–70 Gy
in irradiation of the prostate +
proximal seminal vesicles with
different treatment techniques.
to sparing all organs at risk. Published data do not indicate
that more sophisticated techniques increase the therapeutic
index [13–19]. Techniques with more than five fields have a
very high burden for daily routine treatment planning, and
an optimal radiation technique should not only provide the
best sparing for all organs at risk (rectum, bladder, femoral
heads), but also be safely implemented without undue
burden and reduce the risk of any error. For example,
simple, and therefore safe, verification by portal imaging
during radiation treatment is given with simple radiation
fields. For these reasons, a study, investigating only simple
radiation techniques was performed.
The British Journal of Radiology, February 2006
Treatment techniques in prostate cancer irradiation
P bladder and the rectum. However, they found a better
Bedford et al [14] concluded that for irradiation of the
prostate only a four-field technique with two oblique
anterior and lateral fields would be optimal for rectal
sparing. Koswig et al [3] found that, for irradiation of the
prostate, only, the best rectal sparing was with a six-field
technique. Khoo et al [16] concluded for prostate
irradiation, a three-field technique would bring the best
rectal sparing with acceptable bladder and femoral head
doses. They performed plans with 6 MV photons. We
compared in our study a three-field technique with two
different four-field techniques using 20 MV photons and
the results of Khoo et al [16] concerning rectal dose
sparing can be confirmed by our results. Another
recently published study investigated three-field techni-
ques versus four-field techniques in irradiation of the
prostate only and found that the three-field technique
using an anterior and two lateral (270 ˚ and 90 ˚) fields
with 20 MV photons provides the best rectal protection
[18]. This can also be confirmed by our results. The
recent study mentioned found no difference between the
techniques in bladder exposure and discussed that this
may be due to the CTV (prostate only) [18]. Our data
indicate that the three-field technique provides, for
irradiation of the prostate only, the best rectal dose
sparing with no significant differences in bladder dose
sparing compared with the unweighted four-field
technique. The weighted four-field technique was worst
in bladder dose sparing. Concerning the femoral head
doses, the three-field technique was worst.
PPSV
Neal et al compared a three-field, four-field and six-
field technique for irradiation of the prostate + the base
of seminal vesicles [19]. They found no significant
differences considering the irradiated volume of the
The British Journal of Radiology, February 2006
Figure 3. Mean values, standard
error and standard deviations (SD)
for the volumes 50 Gy, 60 Gy and
65 Gy of the rectum in case of
irradiation of the prostate + entire
seminal vesicles with different
treatment techniques.
sparing of the rectum with a weighted four-field
technique and a better sparing of the bladder with the
six-field technique. Our data indicate for irradiation of
the prostate + proximal seminal vesicles the three-field
technique to be optimum in rectal dose sparing. The
weighted four-field technique was worst in bladder dose
sparing and the unweighted four-field technique pro-
vided a better sparing of the rectum compared with the
three-field technique. Concerning the femoral head
doses, the three-field technique was worst.
PESV
For irradiation of the prostate + seminal vesicles,
several authors compared different techniques and all of
these studies conclude that no single technique is
superior when considering all organs at risk (rectum,
bladder, femoral head). Fiorino et al compared various
coplanar techniques for conformal irradiation of the
prostate and seminal vesicles [15]. A three-field techni-
que with an anteroposterior and two lateral 30 ˚ wedged
fields gave the best sparing of the rectum. The bladder
was best spared with a six-field technique. The mean
dose of the bladder was significantly better against the
three-field technique and the four-field technique.
However, considering V95, no significant difference
was found between the techniques. The unweighted
four-field technique gave the worst sparing of the
bladder for Fiorino et al. In our study, the weighted
four-field technique gave the worst sparing of the
bladder.
Bedford et al [14] compared various four-field techni-
ques with a three-field technique which had lateral
oblique fields. They concluded that for irradiation of the
prostate and prostate + seminal vesicles, four-field
techniques with two oblique anterior and lateral fields
with individual field wedges for the different CTVs to be
153
154

Table 3. Mean and median values, standard deviations (SD) and p-values for the radiation exposure to the bladder in different techniques and clinical target volumes (CTVs)
(%) P PPSV PESV

1 2 p-value 1 3 p-value 2 3 p-value 1 2 p-value 1 3 p-value 2 3 p-value 1 2 p-value 1 3 p-value 2 3 p-value

40 Gy Mean 44 48 44 46 48 46 57 69 57 65 69 65 66 80 , 66 81 80 81
Median 45 49 0.003 45 45 n.s. 49 45 n.s. 57 68 0.001 57 62 0.02 68 62 n.s. 68 81 0.001 68 84 0.005 81 84 n.s.
SD 15 17 15 16 17 16 15 16 15 17 16 17 16 15 16 15 15 15
50 Gy Mean 36 37 36 35 37 35 45 46 45 44 46 44 46 48 46 47 48 47 n.s.
Median 35 36 0.005 35 31 n.s. 36 31 0.002 46 47 n.s. 46 44 n.s. 47 44 0.005 47 49 0.003 47 47 n.s. 49 47
SD 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 13 14 13
60 Gy Mean 28 29 28 27 29 27 36 37 36 36 37 36 39 39 39 38 39 38
Median 27 27 0.003 27 24 n.s. 27 24 0.005 36 36 0.02 36 36 0.007 36 36 0.003 40 41 n.s. 40 38 n.s. 41 38 0.01
SD 12 12 12 12 12 12 12 13 12 13 13 13 12 12 12 12 12 12
65 Gy Mean 22 23 22 22 23 22 30 32 30 29 32 29 33 33 33 32 33 32
Median 20 21 0.008 20 21 n.s. 21 21 0.01 30 31 0.003 30 29 n.s. 31 29 0.001 33 34 0.03 33 33 n.s. 34 33 n.s.
SD 11 11 10 10 11 10 11 11 11 11 11 11 11 10 11 11 10 11
70 Gy Mean 14 15 14 15 15 15 22 24 22 20 24 20 25 26 25 24 26 24
The British Journal of Radiology, February 2006

Median 12 12 n.s. 12 13 n.s. 12 13 n.s. 22 25 0.002 22 20 n.s. 25 20 0.003 26 27 0.03 25 24 0.04 27 24 0.01
SD 8 9 8 10 9 10 9 9 9 9 9 9 9 10 9 9 10 9
V 100 Mean 4 4 4 3 4 3 12 12 11 10 12 10 24 27 24 25 27 25
Median 3 3 n.s. 3 3 n.s. 3 3 n.s. 5 6 n.s. 3 5 n.s. 6 5 n.s. 24 27 0.04 24 21 n.s. 27 21 0.02
SD 8 3 8 2 3 2 15 15 15 15 15 15 19 21 19 21 21 21
1, 2, 3 relates to the radiation techniques described in Methods and materials.

A Hille, N Töws and C F Hess

Treatment techniques in prostate cancer irradiation
optimal for rectal sparing. However, such a technique
has a very high burden for daily routine treatment
planning. The simple three-field plan in this study with
an anterior and two lateral fields using 6 MV photons
showed a comparable level with the four-field technique
in rectal dose sparing, in case of irradiation of the
prostate + entire seminal vesicles. However, the dose to
the superficial body and femoral heads was found to be
very high [14]. Khoo [16] concluded for both the prostate
and seminal vesicles irradiation a three-field technique
would bring the best rectal dose sparing with acceptable
bladder and femoral head doses. They performed plans
with 6 MV photons also, but the superficial body dose
was not mentioned. In our study, we compared a three-
field technique with two different four-field techniques
using 20 MV photons, and the results of Khoo et al [16]
concerning rectal dose sparing can be confirmed by our
results. The weighted four-field technique was worst in
bladder dose sparing. Comparing the unweighted four-
field with the three-field technique in bladder dose
sparing, no clear advantages or disadvantages were
found. Concerning the femoral head doses, the three-
field technique was worst.
The studies investigating various techniques draw
differing conclusions concerning the best irradiation
technique. Some studies, comparing four and three-field
techniques concluded the three-field technique to be best
in rectal dose sparing [16, 18]. Others did not confirm
these results [14, 19]. The reasons for these differing
findings are unclear; PTV margins and PTV coverage
which have both an impact on radiation exposure of the
organs at risk were comparable among these studies and
comparable with our study. The different CTVs in theses
studies could have been responsible for the different
findings, but our study shows for all three CTVs the best
rectal dose sparing with the three-field technique.
Concerning the bladder dose, the differing conclusions
may be due to different bladder fillings which is known
to have an impact on bladder DVHs [9]. In most of the
studies, only few endpoints concerning dose volume
histograms had been chosen and the investigated points
were not associated with doses given in dose-constraint
studies. Analysing many dose endpoints, as we have
done in our study, can lead to unclear, or even
conflicting results. This could be an explanation for the
differing results compared with our study.
Although the differences between the three techniques
were small in our study they were significant, and we
conclude from our data that a three-field technique
provided the best rectal but the worst femoral dose
sparing with inconsistent results regarding the bladder
dose sparing for all three CTVs.
Estimated risk for chronic normal tissue exposure
Rectum
To associate the rectal DVHs in our study with an
estimated risk for chronic rectal toxicity, the results were
compared with studies analysing relationships between
dose–volume and rectal toxicity (Table 1).
Although the definitions of the rectum differ in some
of these studies from our definition, and although the
The British Journal of Radiology, February 2006
cut-off levels and the resulting risks for chronic rectal
toxicity ¢ grade 2 also differ in these studies, we can
draw cautious conclusions from our results regarding an
estimated risk for chronic rectal toxicity.
The obtained values from the three different techni-
ques were close together. In patients treated with
irradiation of the prostate only, the values for the rectum
exposure were below 5% for chronic rectal bleeding ¢
grade 2; for irradiation of the prostate + proximal
seminal vesicles between 5% and 15%; and for irradia-
tion of the prostate + entire seminal vesicles over 15%
with all three techniques. Nevertheless, the three-field
technique provided the best rectal dose sparing, except
for the rectal volume exposed to 70 Gy. The weighted
four-field technique provided a significantly better rectal
dose sparing than the unweighted four-field technique.
Whether the small differences between the various
techniques would have an impact on chronic rectal
toxicity is uncertain. Furthermore, published data sug-
gest increased local control with lower normal toxicity
with new technologies such as intensity modulated
radiation therapy (IMRT) [39–42]. IMRT allows the
increase of dose in part of the prostate while continuing
to protect normal tissue. However, until new technolo-
gies such as IMRT are introduced as a widespread
clinical routine treatment, 3D conformal radiation ther-
apy should be optimized to reduce toxicity while
inreasing local control.
Bladder
To estimate the risk for bladder toxicity we tried to
compare our data with the clinical relationship between
DVHs and the development of bladder toxicity, as
reported in the literature [4, 34–37].
The exposed bladder volumes in our study were
similar with the three different techniques. The risk of
chronic bladder toxicity in our study can be estimated to
be less than 5% to 10% in irradiation of the prostate only
with all three techniques. In irradiation of the prostate +
proximal/entire seminal vesicles, the risk for chronic
toxicity can be estimated to be above 10% with all three
techniques. For all three CTVs the weighted four-field
technique provided the worst bladder dose sparing.
Whether the small differences between the various
techniques would have an impact on chronic bladder
toxicity is uncertain.
Femoral heads
The three-field technique provided the worst radiation
exposure to the femoral heads. The differences for the
V50 value were significant; for the V100 value no
significant difference was found between the techniques
for all three CTVs. As the values for the three techniques
were below 52 Gy to the whole femoral heads, the risk of
chronic toxicity can be estimated to be below 5% in
5 years for the three CTVs [41].
Conclusion
In conclusion, none of the studied techniques consis-
tently proved superior in different CTVs in prostate
cancer irradiation with respect to sparing all the organs
155

at risk. The absolute differences between the three
techniques were small and the clinical relevance of these
findings remains uncertain.
References
1. Boersma LJ, van den Brink M, Bruce AM, Shouman T, Gras
L, te Velde A, et al. Estimation of the incidence of late
bladder and rectum complications after high-dose (70–
78 Gy) conformal radiotherapy for prostate cancer, using
dose-volume histograms. Int J Radiat Oncol Biol Phys
1998;41:83–92.
2. Koelbl O, Schwab F, Bratengeier K, Vordermark D, Flentje
M. Radiotherapy of prostate cancer with multileaf collima-
tors (MLCs) optimization of the undulating dose distribu-
tion at the MLC edge. Strahlenther Onkol 2005;181:108–12.
3. Koswig S, Dinges S, Buchali A, Bohmer D, Salk J, Rosenthal
P, et al. Comparison of different 3-dimensional irradiation
techniques in local radiotherapy of prostatic carcinoma.
Strahlenther Onkol 1999;175:10–6.
4. Michalski JM, Purdy JA, Winter K, Roach M 3rd,
Vijayakumar S, Sandler HM, et al. Preliminary report of
toxicity following 3D radiation therapy for prostate cancer
on 3DOG/RTOG 9406. Int J Radiat Oncol Biol Phys
2000;46:391–402.
5. Wachter-Gerstner N, Wachter S, Goldner G, Nechvile E,
Potter R. Biochemical response after 3-D conformal radio-
therapy of localized prostate cancer to a total dose of 66 Gy
4-year results. Strahlenther Onkol 2002;178:542–7.
6. Cozzarini C, Fiorino C, Ceresoli GL, Cattaneo GM,
Bolognesi A, Calandrino R, et al. Significant correlation
between rectal DVH and late bleeding in patients treated
after radical prostatectomy with conformal or conventional
radiotherapy (66.6–70.2 Gy). Int J Radiat Oncol Biol Phys
2003;55:688–94.
7. Fiorino C, Cozzarini C, Vavassori V, Sanguineti G, Bianchi
C, Cattaneo GM, et al. Relationships between DVHs and
late rectal bleeding after radiotherapy for prostate cancer:
analysis of a large group of patients pooled from three
institutions. Radiother Oncol 2002;64:1–12.
8. Fiorino C, Sanguineti G, Cozzarini C, Fellin G, Foppiano F,
Menegotti L, et al. Rectal dose-volume constraints in high-
dose radiotherapy of localized prostate cancer. Int J Radiat
Oncol Biol Phys 2003;57:953–62.
9. Lebesque JV, Bruce AM, Kroes AP, Touw A, Shouman RT,
van Herk M. Variation in volumes, dose-volume histo-
grams, and estimated normal tissue complication probabil-
ities of rectum and bladder during conformal radiotherapy
of T3 prostate cancer. Int J Radiat Oncol Biol Phys
1995;33:1109–19.
10. Koper PC, Heemsbergen WD, Hoogeman MS, Jansen PP,
Hart GA, Wijnmaalen AJ, et al. Impact of volume and
location of irradiated rectum wall on rectal blood loss after
radiotherapy of prostate cancer. Int J Radiat Oncol Biol Phys
2004;58:1072–82.
11. Patel RR, Orton N, Tome WA, Chappell R, Ritter MA.
Rectal dose sparing with a balloon catheter and ultrasound
localization in conformal radiation therapy for prostate
cancer. Radiother Oncol 2003;67:285–94.
12. Skwarchuk MW, Jackson A, Zelefsky MJ, Venkatraman ES,
Cowen DM, Levegrun S, et al. Late rectal toxicity after
conformal radiotherapy of prostate cancer (I): multivariate
analysis and dose-response. Int J Radiat Oncol Biol Phys
2000;47:103–13.
13. Akazawa PF, Roach M 3rd, Pickett B, Purser P, Parkinson
D, Rathbun C, et al. Three dimensional comparison of
blocked arcs vs. four and six field conformal treatment of
the prostate. Radiother Oncol 1996;41:83–8.
14. Bedford JL, Khoo VS, Oldham M, Dearnaley DP, Webb S.
A comparison of coplanar four-field techniques for
156
A Hille, N Töws and C F Hess
conformal radiotherapy of the prostate. Radiother Oncol
1999;51:225–35.
15. Fiorino C, Reni M, Cattaneo GM, Bolognesi A, Calandrino
R. Comparing 3-, 4- and 6-fields techniques for conformal
irradiation of prostate and seminal vesicles using dose-
volume histograms. Radiother Oncol 1997;44:251–7.
16. Khoo VS, Bedford JL, Webb S, Dearnaley DP. An evaluation
of three-field coplanar plans for conformal radiotherapy of
prostate cancer. Radiother Oncol 2000;55:31–40.
17. Lennernas B, Rikner G, Letocha H, Anscher MS. External
beam radiotherapy of localized prostatic adenocarcinoma.
Evaluation of conformal therapy, field number and target
margins. Acta Oncol 1995;34:953–8.
18. Milecki P, Piotrowski T, Dymnicka M. The comparison of
radiotherapy techniques for treatment of the prostate
cancer: the three-field vs. the four-field. Neoplasma.
2004;51:64–9.
19. Neal AJ, Oldham M, Dearnaley DP. Comparison of
treatment techniques for conformal radiotherapy of the
prostate using dose-volume histograms and normal tissue
complication probabilities. Radiother Oncol 1995;37:29–34.
20. Fenwick JD, Khoo VS, Nahum AE, Sanchez-Nieto B,
Dearnaley DP. Correlations between dose-surface histo-
grams and the incidence of long-term rectal bleeding
following conformal or conventional radiotherapy treat-
ment of prostate cancer. Int J Radiat Oncol Biol Phys
2001;49:473–80.
21. Diaz A, Roach M 3rd, Marquez C, Coleman L, Pickett B,
Wolfe JS, et al. Indications for and the significance of
seminal vesicle irradiation during 3D conformal radio-
therapy for localized prostate cancer. Int J Radiat Oncol Biol
Phys 1994;30:323–9.
22. Feldmann HJ, Breul J, Zimmermann F, Wachter S, Wiegel T.
Probability of seminal vesicle involvement in localized
prostatic carcinoma. Significance in conformal radiother-
apy. Strahlenther Onkol 1998;174:566–70.
23. Katcher J, Kupelian PA, Zippe C, Klein EA, Sohn JW.
Indications for excluding the seminal vesicles when treating
clinically localized prostatic adenocarcinoma with radio-
therapy alone. Int J Radiat Oncol Biol Phys 1997;37:871–6.
24. Kestin L, Goldstein N, Vicini F, Yan D, Korman H, Martinez
A. Treatment of prostate cancer with radiotherapy: should
the entire seminal vesicles be included in the clinical target
volume? Int J Radiat Oncol Biol Phys 2002;54:686–97.
25. Sanguineti G, Castellone P, Foppiano F, Franzone P,
Marcenaro M, Tognoni P, et al. Anatomic variations due
to radical prostatectomy. Impact on target volume defini-
tion and dose-volume parameters of rectum and bladder.
Strahlenther Onkol 2004;180:563–72.
26. International Commission on Radiation Units and
Measurements. ICRU Report 50: Prescribing, Recording,
and Reporting Photon Beam Therapy. International
Commission on Radiation Units and Measurements,
Washington, DC, 1993.
27. Greco C, Mazzetta C, Cattani F, Tosi G, Castiglioni S, Fodor
A, et al. Finding dose-volume constraints to reduce late
rectal toxicity following 3D-conformal radiotherapy (3D-
CRT) of prostate cancer. Radiother Oncol 2003;69:215–22.
28. Huang EH, Pollack A, Levy L, Starkschall G, Dong L, Rosen
I, et al. Late rectal toxicity: dose-volume effects of conformal
radiotherapy for prostate cancer. Int J Radiat Oncol Biol
Phys 2002;54:1314–21.
29. Wachter S, Gerstner N, Goldner G, Potzi R, Wambersie A,
Potter R. Rectal sequelae after conformal radiotherapy of
prostate cancer: dose-volume histograms as predictive
factors. Radiother Oncol 2001;59:65–70.
30. Zapatero A, Garcia-Vicente F, Modolell I, Alcantara P,
Floriano A, Cruz-Conde A, et al. Impact of mean rectal dose
on late rectal bleeding after conformal radiotherapy for
The British Journal of Radiology, February 2006
Treatment techniques in prostate cancer irradiation
prostate cancer: dose-volume effect. Int J Radiat Oncol Biol
Phys 2004;59:1343–51.
31. Storey MR, Pollack A, Zagars G, Smith L, Antolak J, Rosen I.
Complications from radiotherapy dose escalation in
prostate cancer: preliminary results of a randomized trial.
Int J Radiat Oncol Biol Phys 2000;48:635–42.
32. Geinitz H, Zimmermann FB, Narkwong L, Kneschaurek P,
Wehrmann R, Kuzmany A, et al. Prostatic carcinoma:
problems in the interpretation of rectal dose-
volume histograms. Strahlenther Onkol 2000;176:
168–72.
33. Foppiano F, Fiorino C, Frezza G, Greco C, Valdagni R. The
impact of contouring uncertainty on rectal 3D dose-volume
data: results of a dummy run in a multicenter trial
(AIROPROS01-02). Int J Radiat Oncol Biol Phys
2003;57:573–9.
34. Pollack A, Zagars GK, Cole CJ, Dinney CP, Swanson DA,
Grossman HB. Significance of downstaging in muscle-
invasive bladder cancer treated with preoperative radio-
therapy. Int J Radiat Oncol Biol Phys 1997;37:41–9.
35. Zelefsky MJ, Fuks Z, Wolfe T, Kutcher GJ, Burman C, Ling
CC, et al. Locally advanced prostatic cancer: long-term
toxicity outcome after three-dimensional conformal radia-
tion therapy–a dose-escalation study. Radiology
1998;209:169–74.
36. Marks LB, Carroll PR, Dugan TC, Anscher MS. The
response of the urinary bladder, urethra, and ureter to
The British Journal of Radiology, February 2006
radiation and chemotherapy. Int J Radiat Oncol Biol Phys
1995;31:1257–80.
37. Emami B, Lyman J, Brown A, Coia L, Goitein M,
Munzenrider JE, et al. Tolerance of normal tissue to
therapeutic irradiation. Int J Radiat Oncol Biol Phys
1991;21:109–22.
38. Chuba PJ, Sharma R, Yudelev M, Duclos M, Shamsa F,
Giacalone S, et al. Hip stiffness following mixed conformal
neutron and photon radiotherapy: a dose-volume relation-
ship. Int J Radiat Oncol Biol Phys 1996;35:693–9.
39. Brabbins D, Martinez A, Yan D, Lockman D, Wallace M,
Gustafson G, et al. A dose-escalation trial with the adaptive
radiotherapy process as a delivery system in localized
prostate cancer: analysis of chronic toxicity. Int J Radiat
Oncol Biol Phys 2005;61:400–8.
40. De Meerleer GO, Vakaet LA, De Gersem WR, De Wagter C,
De Naeyer B, De Neve W. Radiotherapy of prostate cancer
with or without intensity modulated beams: a planning
comparison. Int J Radiat Oncol Biol Phys 2000;47:639–48.
41. Gershkevitsh E, Clark CH, Staffurth J, Dearnaley DP,
Trott KR. Dose to bone marrow using IMRT techniques in
prostate cancer patients. Strahlenther Onkol 2005;181:
1172–8.
42. Li XA, Wang JZ, Jursinic PA, Lawton CA, Wang D.
Dosimetric advantages of IMRT simultaneous integrated
boost for high-risk prostate cancer. Int J Radiat Oncol Biol
Phys 2005;61:1251–7.
157
 The British Journal of Radiology, 79 (2006), 158–161

SHORT COMMUNICATION

A comparative evaluation of two head and neck immobilization

devices using electronic portal imaging
1 1
K DONATO, BSc (Hon), K LESZCZYNSKI, PhD, FCCPM and 2K FLEMING, MHSc, MRT(T)

1
Northeastern Ontario Regional Cancer Centre, Hôpital régional de Sudbury Regional Hospital, 41
Ramsey Lake Road, Sudbury, Ontario, P3E 5J1 and 2Grand River Regional Cancer Centre, 835 King
Street West, Kitchener, Ontario, N2G 1G3, Canada

ABSTRACT. A study was performed to compare the positioning reproducibility and the
cost efficiency for two head and neck immobilization devices: the Uvex’ (Uvex Safety,
Smithfield, USA) plastic mask system and the Finesse Frame with Ultraplast System’
(PLANET Medical, Svendborg, Denmark). 20 patients treated with 3D conformal
radiation therapy for head and neck cancers were randomly selected (10 for each of the
two different immobilization systems) and electronic portal images acquired during
their course of treatment were saved and used in this study. The anatomical landmark
coordinates and their shifts in the anteroposterior (AP) and craniocaudal (CC) directions
with respect to the digitized simulator films for lateral fields were analysed using an in-
house developed portal image registration system. Statistically, no evidence was found
to indicate that the systematic components of the displacement for the Uvex’ system
and the Finesse Frame with Ultraplast System’ were different from each other or from
zero. The random component of displacement was slightly smaller in the AP direction
for the Uvex’ than the Ultraplast’ system (s51.9 mm and 2.9 mm, respectively,
p50.007), but larger in the CC direction (s53.8 mm and 2.2 mm, respectively, p,1029).
Production time and required materials for a radiation therapy department were also Received 16 March 2005
quantified to assess costs for each system. The overall costs per patient were estimated Revised 27 June 2005
at $141.50 (CAD) and $82.10 for the Uvex’ and Ultraplast’ systems, respectively. The Accepted 8 August 2005
Finesse Frame with Ultraplast System’ of immobilization for head and neck cancer
DOI: 10.1259/bjr/32191494
treatment provides a field placement reproducibility that is equal to, or greater than,
that of the Uvex’ plastic mask immobilization system and, while it requires more ’ 2006 The British Institute of
expensive materials, the workload and consequently overall cost is greatly reduced. Radiology

Radiation treatment to the head and neck region is
delivered with accurate and precise placement of pre-
scribed portal fields. Reproducible alignment is increas-
ingly important as we apply high-dose three-dimensional
conformal radiation therapy (3D-CRT) techniques and
intensity-modulated radiation therapy (IMRT) in conjunc-
tion with the need for smaller clinical target volume (CTV)
margins. The consequences of field placement errors have
been described in various publications; failure to treat the
entire planning target volume (PTV) may be responsible
for local failure, and irradiation outside of the PTV may
result in normal tissue complications to important organs
[1–4], such as the spinal cord or the eye, in the case of head
and neck cancers. In order to increase the reproducibility
of portal field placement, various immobilization devices
are used to stabilize the position of the patient’s head
while treatment is delivered.
Previous publications have compared two or three
different systems of immobilization [5–7], and have
assessed treatment field position reproducibility with
Address correspondence to: K Leszczynski.
similar results: the standard deviations of field place-
ment errors, s, were found to be between 1.7 mm and
3.3 mm, for both anteroposterior (AP) and craniocaudal
(CC) directions. In this study, an immobilization system
involving a Uvex’ (Uvex Safety, Smithfield, USA) plastic
mask was compared with a low temperature thermo-
plastic mask system, the Finesse Frame with Ultraplast
system’ (PLANET Medical, Svendborg, Denmark) using
off-line electronic portal imaging. The costs of both
systems in terms of production time and materials were
also calculated since the clinical introduction of a low
temperature thermoplastic mask appears to be less time-
consuming, less costly, and more convenient for the
patient [7] and, therefore, beneficial in general.
Materials and methods
Clinical setup
This study retrospectively selected 20 consecutive
head and neck patients treated at our centre between

158 The British Journal of Radiology, February 2006

Short communication: An evolution of two head and neck immobilization devices
Table 1. A summary of selected demographic data for
patients included in both immobilization groups
Demographics Immobilization device group
Uvex Ultraplast
Age: mean¡SD 65¡10 60¡11
Sex: Females/Males2/8 1/9
Diagnosis Ca tongue – 4 Ca supraglottis – 2
distribution Ca tonsil –2 Ca floor of mouth – 2
Ca hypopharynx – 1Ca tongue –1
Ca pharynx – 1 Ca oropharynx – 1
Ca oral cavity – 1 Ca gingiva – 1
Ca larynx –1 Ca glottis –1
Unknown primary – 2
Weight change 4.8¡3.6 3.4¡5.7
during
treatment:
Mean¡SD
April 2001 and September 2001. These patients were
randomly drawn from two groups representing different
immobilization devices used in treatment setup; 10
patients with a Uvex’ mask system, and 10 patients
with a FinesseFrame with Ultraplast System’. A sum-
mary of selected demographic data for both immobiliza-
tion groups is provided in Table 1. Review of the
summary confirms that there was no apparent signifi-
cant demographical bias between the groups. One
patient’s results were eliminated from the study due to
poor quality imaging. All patients attended the Mevasim
simulator (Siemens Medical Solutions, Concord, USA)
where lateral portal positioning was marked on the mask
for alignment purposes. A planning CT scan using
Somatom Plus (Siemens Medical Solutions) was per-
formed at 5 mm intervals for 3D conformal treatment
planning on the Helax TMS system (Nucletron B.V.,
Veenendaal, The Netherlands), and all patients were
treated on a Mevatron KD-2 or Primus linear accelerator
(Siemens Medical Solutions) with 6 MV beams. The
treatment beam arrangement consisted of two parallel
opposed lateral fields covering the head and neck target.
The nodes in the supraclavicular region were irradiated
with an anterior field. Portal images were acquired and
stored for daily fractions using Beamview Plus’
(Siemens Medical Solutions) video camera based electro-
nic portal imaging. For the purpose of this study, only
right lateral portal views were used. An average of 14
(range 9–19) portal images for each patient were
acquired for a total retrospective analysis of 272 images.
Immobilization devices
Two different thermoplastic masks and their respective
immobilization accessories were evaluated in this study.
To form the Uvex’ mask, two radiation therapists stabilize
the patient’s head on a Timo head rest and form a plaster
impression of the patient’s head and neck. One therapist
later drapes the negative impression and fills it to form a
positive impression. The 1/160 Uvex’ plastic is then
heated in a vacuum former and then moulded around
the plaster positive. The patient returns approximately 2
days later for the fitting process where the plastic mask is
fitted directly onto the patient’s head and neck and
The British Journal of Radiology, February 2006
mounted onto a Perspex’ (Lucite International Canada
Inc., Mississauga, Canada) acrylic headboard at three
fixation points on each side of the head (six fixation points
in total, distributed evenly from the lower neck to the top
of the head). The treatment field area is cut from the Uvex’
mask once a radiation oncologist approves the first-day
portal image in order to allow for increased skin sparing.
To form the Ultraplast’ mask, the Ultraplast’ material
is dipped into a hot water bath (<75 ˚C) while the patient
is positioned on a Silverman head rest that attaches to a
‘‘Quick Snap and Lock’’ carbon fibre headboard by a
Finesse Frame system’ which is fixed at three points (at
the top and on either side of the head). Two therapists
then stretch the material and mould the mask directly
onto the head of the patient. After approximately 8 min,
the mask has hardened. The field area is not cut out from
the Ultraplast’ masks since they offer better skin sparing
than the Uvex’ masks [8] and, while reducing the dose in
the first few millimetres of skin, field cutouts may affect
positioning reproducibility [9].
Image analysis
Right lateral portal images were captured using the
Beamview Plus’ portal imaging system. These images
were then imported into a Portal ViewStation software
system developed in-house [10]. Corresponding simulator
films were digitized and imported into the Portal
ViewStation system. While the field borders of the
simulator films were defined manually, the field borders
of the portal images were extracted automatically by
applying an edge detection algorithm [11]. Adaptive
histogram equalization [12] was applied to the portal film
in order to enhance contrast. Two experienced radio-
therapists delineated bony landmarks (the vertebrae) on
both the simulator film and all portal images. The portal
images were transformed onto the simulator film coordi-
nate system and, using the chamfer matching registration
algorithm [13, 14], the corresponding anatomical land-
marks were aligned. Next, the borders of the simulated
field were matched with the borders of the portal images
by applying a polygon-matching algorithm [15] and the
displacement between the simulator film treatment field
borders and the portal image borders was recorded and
analysed to determine the translational field placement
error in the AP and CC directions, as well as the rotational
error measured in the plane orthogonal to the lateral beam.
The accuracy of the field placement measurement
method was previously assessed to be within 1.5 mm
and 1 ˚ for translational and rotational displacements,
respectively [10].
Statistical analysis
The reproducibility of treatment field placement is
reflected by a combination of both systematic and
random error. The systematic placement error for one
patient is that component of the field displacement that
was constant throughout the treatment period and can,
therefore, quantify the accuracy of patient positioning.
This error is defined as the mean of all displacements in
either the AP or CC direction. The random error, or the
precision with which the patient is positioned daily, is
159

determined by subtracting the systematic displacement
from the total displacement for one fraction. For the
Uvex’ or Ultraplast’ group, the systematic error is
quantified by the range and standard deviation of the
mean field displacements for all individuals and the
random error for the group is the standard deviation of
all individual random errors.
Cost calculations
The average time of mask production was determined
by taking an average for the production of five masks of
each type. The production time included patient educa-
tion, mask production and, for the Uvex’ masks, time
spent in plastering the positive mould. The mid-range
salary rate for a Radiation Therapist at the Northeastern
Ontario Regional Cancer Centre was multiplied by the
time of production. In 2004, the salary range for radiation
therapists was between $48 750 and $67 500 (CAD) per
annum, therefore, the average salary of $58 125 was used
in this analysis. Full-time employees working 37.5 h per
week, work an average of 1950 h per year, resulting in a
cost per minute of $0.50 (CAD).
The total non-reusable materials costs were assessed
for masks of each type and the cost of materials for 10
masks was added to the labour cost for 10 masks of that
type. An approximate yearly cost was also calculated for
both the Uvex’ and Ultraplast’ immobilization systems.
Results
Patient reproducibility
Whether immobilized with a Uvex’ mask system or a
Finesse Frame with Ultraplast System’, no difference in
the accuracy with which the patients were placed in
position for treatment (or systematic error) was detected.
A t-test applied to the magnitudes of systematic errors
measured in both groups yielded p.0.2 for translational
and rotational errors. For both immobilization systems,
using a Z-test, the average systematic errors were not
statistically different from zero (p.0.2 for translations in
AP and CC directions, and also for rotations). For the
Uvex’ mask system, the range of systematic errors was
[22.8, 4.4] mm in the AP direction with a standard
deviation of 2.6 mm, [22.3, 3.2] mm in the CC direction
with a standard deviation of 1.6 mm, and [21.7 ˚, 2.7 ˚] in
rotation with a standard deviation of 1.3 ˚. For the
Ultraplast’ mask system, the range of systematic errors
was [22.5, 2.7] mm in the AP direction with a standard
deviation of 1.9 mm, [23.6, 2.8] mm in the CC direction
with a standard deviation of 1.8 mm, and [23.6 ˚, 2.1 ˚] in
rotation with a standard deviation of 1.6 ˚ (Table 2).
Table 2. Systematic and random errors for Uvex’ and Ultraplast’ immobilization systems
Anteroposterior
Systematic (S) or SD (mm) Positioning error
Mask Type Random (R) .5 mm
Uvex S 2.6 16 (11%)
R 1.9
Ultraplast S 1.9 11 (8%)
R 2.3
160
K Donato, K Leszczynski and K Fleming
There was, however, a difference in the random error, or
the precision, in field positioning for the two immobiliza-
tion setup systems: for the Uvex’ mask system, the
standard deviation of the random error in the AP direction
was found to be 1.9 mm compared with 2.3 mm for the
Finesse Frame with Ultraplast System’ (p50.007) while in
the CC direction, the standard deviation of the random
error for the Uvex’ mask system was 3.8 mm compared
with 2.2 mm for the Finesse Frame with Ultraplast
System’ (p,1029). The standard deviations of the random
error in rotation were 1.8 ˚ and 2.1 ˚ for the Uvex’ and
Ultraplast System’ mask immobilizations, respectively,
and the difference was not statistically significant (p.0.05).
The frequency of translational field placement errors
larger than 5 mm was 11% for Uvex’ mask systems
and 8% for Finesse Frame with Ultraplast Systems’ in
the AP direction and was 8% for Uvex’ mask systems
and 5% for Finesse Frame with Ultraplast Systems’ in
the CC direction.
Costs
The average time taken by a radiation therapist to
produce a Uvex’ mask was 134 min. This time measure-
ment included patient contact as well as the production
of a plaster positive and vacuum-forming the mask. The
average time spent by a second radiation therapist was
47 min for the patient impression and mask fitting. The
total time was therefore 181 min, at a cost of $90.50
(CAD) per patient for labour alone.
The average time taken by a radiation therapist to
produce an Ultraplast’ mask, including room prepara-
tion and patient education, was 23.7 min. A second
radiation therapist spent an average time of 18.7 min
assisting with the production of the mask. The total time
was therefore 42.4 min, or a cost of $21.20 (CAD) per
patient for labour alone.
The cost of mask materials, excluding start-up costs,
base plates, or head rests, was $510 CAD for 10 Uvex’
masks and $609 CAD for 10 Ultraplast’ masks.
Therefore, the total materials and labour costs for the 10
Uvex’ masks were $1415.00 (CAD) and $821.00 (CAD) for
the 10 Ultraplast’ masks. At the NEORCC, there are
approximately 92 patients requiring immobilization
masks per year. At a cost difference per patient of $59.40
(CAD), the total cost difference per year is $5464.80 (CAD)
in favour of the Finesse Frame with Ultraplast System’.
Discussion
Setup reproducibility
With the use of IMRT for treatment of head and neck
cancers, it is highly desirable to outline smaller margins
Craniocaudal Rotational
SD (mm) Positioning error SD ( ˚)
.5 mm
1.6 11 (8%) 1.3
3.8 1.8
1.8 6 (5%) 1.6
2.2 2.1
The British Journal of Radiology, February 2006
Short communication: An evolution of two head and neck immobilization devices
that are added to the CTV and, thus, to more effectively
avoid normal tissue complications while still increasing
the total dose to the PTV in order to gain local control of
the tumour. Effective immobilization of the area to be
treated is essential for accurate and precise delivery of the
treatment plan and dose prescription. While the systematic
component of field placement error is primarily due to the
transfer of patient setup from treatment planning to
delivery, any mispositioning or moving of the patient
within the mask may cause the random component of
error.
The systematic component of error for both the Uvex’
and Ultraplast’ mask systems was less than 3 mm (1 SD)
for both the AP and CC directions of motion. This value
is consistent with previous studies [6, 17, 18] indicating
that the patient setup errors are often determined by
transfer errors from the simulator to the treatment unit
and are not necessarily affected by mask type.
In other publications, the use of thermoplastic masks, as
opposed to masks made of other materials such as plastics
or polycarbonate, appears to result in a comparable
random error [5, 7]. In this study, it was determined that
a decrease in random field placement error in the CC
direction by using the Finesse Frame with Ultraplast
system’ could be clinically significant since, according to
the formula used by Stroom et al [16], the CTV-PTV margin
may be reduced by 1.1 mm in the CC direction. It is not
apparent, however, whether this improvement in field
placement in the CC direction is due to the rigidity of the
thermoplastic mask material or the fixation of the head and
neck in the Finesse Frame’ which has a fixation point at the
vertex of the head. Since the random component of
displacement in the AP direction was actually slightly
smaller for the Uvex’ system, there is no clear evidence that
our practice of cutting out the treatment field area from
Uvex’ masks for better skin sparing, had a significant
detrimental influence on the rigidity of immobilization.
There also appear to be fewer field placement errors
larger than 5 mm with the use of the Finesse Frame with
Ultraplast System’. This may be attributed to a better fit
to the patient’s anatomy since the Ultraplast’ material
can be fitted directly onto the patient and there is no
intermediary cast necessary. The Finesse Frame with
Ultraplast System’ provides equal, or perhaps better,
immobilization for head and neck cancer radiation
treatment.
While the costs of materials for the Finesse Frame with
Ultraplast System’ are initially higher than those for the
Uvex’ mask and accessories, the increased time commit-
ment required to produce the Uvex’ masks and the
corresponding labour costs, in addition to the inconve-
nience for the patient of having to attend a mould room
fitting twice, make the Finesse Frame with Ultraplast
System’ a preferable option for a radiation therapy
department.
Conclusion
The Finesse Frame with Ultraplast system’ of immo-
bilization for head and neck cancer treatment provides a
field placement reproducibility that is equal to, or greater
than, that of the Uvex’ plastic mask immobilization
system. The Ultraplast’ system is also cheaper and more
The British Journal of Radiology, February 2006
time-efficient, making it a superior product for use in a
radiation therapy department.
Acknowledgments
We are grateful to Louise Beausoleil, Susan Boyko,
Scott Cosby and Janice O’Brien, as well as all the
Radiation Therapy Program staff who provided assis-
tance to this study.
References
1. Bentel G, Marks L, Hendren K, et al. Comparison of two
head and neck immobilization systems. Int J Radiat Oncol
Biol Phys 1997;38:867–73.
2. Bentel G, Marks L, Sherouse G, et al. A customized head
and neck support system. Int J Radiat Oncol Biol Phys
1995;32:245–8.
3. Goitein M. Calculation of the uncertainty in the dose
delivered to the patient. Med Phys 1985;12:606–12.
4. Goitein M, Busse J. Immobilization errors: Some theoretical
considerations. Radiology 1975;117:407–12.
5. Weltens C, Kesteloot K, Vandevelde G, Van Den Bogaert W.
Comparison of Plastic and Orfit’ masks for patient head
fixation during radiotherapy: Precision and costs. Int J
Radiat Oncol Biol Phys 1995;33:499–507.
6. Gilbeau L, Octave-Prignot M, Loncol T, Renard L, Scalliet P,
Gregoire V. Comparison of setup accuracy of three different
thermoplastic masks for the treatment of brain and head
and neck tumors. Radiother Oncol 2001;58:155–62.
7. Lord L, May S, Bailey M, Smith L. Is one head and neck
immobilization system as good as another? One center’s
experience. Med Dosim 2003;28:39–43.
8. Carl J, Vestergaard A. Skin damage probabilities using
fixation materials in high-energy photon beams. Radiother
Oncol 2000;55:191–8.
9. Halm E, Tamri A, Bridier A, Wibault P, Eschwege F.
Influence of thermoplastic masks on absorbed skin dose for
head and neck radiotherapy. Cancer Radiother
2002;6:310–9.
10. Leszczynski K, Loose S, Boyko S. An image registration
scheme applied to verification of radiation therapy. Br J
Radiol 1998;71:413–26.
11. Leszczynski K, Shalev S, Cosby S. The enhancement of
radiotherapy verification images by an automated edge
detection technique. Med Phys 1992;19:611–21.
12. Leszczynski K, Shalev S. Digital contrast enhancement for
on-line portal imaging. Med Biol Eng Comput
1989;27:507–12.
13. Leszczynski K, Loose S, Dunscombe P. Segmented chamfer
matching for prescription-treatment image registration in
radiotherapy. Phys Med Biol 1995;40:83–94.
14. Borgefors G. Hierarchical chamfer matching: a parametric
edge matching algorithm. IEEE Trans Pattern Anal Mach
Intell 1988;10:849–65.
15. Leszczynski K, Loose S. A polygon matching algorithm and
its applications to verification of radiation field placement
in radiotherapy. Int J Biomed Comput 1995;40:59–67.
16. Stroom JC, de Boer HC, Huizenga H, et al. Inclusion of
geometrical uncertainties in radiotherapy treatment plan-
ning by means of coverage probability. Int J Radiat Oncol
Biol Phys 1999;43:905–19.
17. Bel A, Keus R, Vijlbrief R, Lebesque J. Setup deviations in
wedged pair irradiation of parotid gland and tonsillar
tumors, measured with an electronic portal imaging device.
Radiother Oncol 1995;37:153–9.
18. Gildersleve J, Dearnaley D, Evans P, Swindell W.
Reproducibility of patient positioning during routine radio-
therapy as assessed by an integrated megavoltage imaging
system. Radiother Oncol 1995;35:151–60.
161
 The British Journal of Radiology, 79 (2006), 162–164

SHORT COMMUNICATION

Excessive leakage radiation measured on two mobile X-ray units

due to the methodology used by the manufacturer to calculate
and specify the required tube shielding
I A TSALAFOUTAS, PhD

Medical Physics Unit, ‘Konstantopoulio-Agia Olga’ Hospital, 3-5 Agias Olgas, Nea Ionia, 142 33,
Athens, Greece

ABSTRACT. During the quality control (QC) procedure of a new mobile X-ray unit, it was
revealed that the leakage radiation was well in excess of the current limit of 1 mSv h21.
As a result, this unit was returned to the vendor company and it was replaced by a new
unit of the same brand and model. Leakage measurements revealed that the second unit Received 16 May 2005
presented the same problem. After consulting the vendor company and the tube Revised 3 August 2005
manufacturer, it was discovered that the excessive leakage identified in these two X-ray Accepted 10 August 2005
units was not due to a defective construction, but due to the methodology with which
DOI: 10.1259/bjr/17920806
the maximum permissible leakage and therefore the tube shielding had been
determined. In this study, the implications of using such methods to the radiation ’ 2006 The British Institute of
protection of personnel and public are discussed. Radiology

Case history even larger if the measurements were made at the

Upon the arrival of a new mobile X-ray unit, a
thorough quality control (QC) procedure was carried
out in order to measure its performance. For leakage
testing, the tube head was initially covered by radio-
graphic cassettes and, with the collimator diaphragms
completely shut, an exposure was performed with a
tube potential of 100 kVp and a tube loading of 10 mAs.
After the films were developed, large areas of maxi-
mum optical density were identified in all films, except
for one film positioned on the top of the tube. By
positioning a dosemeter on various points on the front
face of the tube and the collimator, it was verified that
the leakage was arising from the tube and not the
collimator.
In an effort to measure the leakage radiation, a survey
meter with measuring range from 0.5 mSv h21 to
1000 mSv h21 was initially employed and exposures with
a tube potential of 100 kVp and a tube loading of 50 mAs
(2 s) were performed. However, even at a distance of 3 m
from the tube the leakage radiation was exceeding the
maximum measurable dose rate of the instrument.
In order to determine as accurately as possible the
actual leakage, a solid state dosemeter with measuring
dose range from 20 nGy to 10 Gy andmeasuring dose
rate range from 40 nGy s21 to 185 mGy s21 was posi-
tioned by a nearby wall at a distance 1 m away from the
tube. For an identical exposure (100 kVp and 50 mAs
(2 s)) the dosemeter reading was 8.2 mGy. By reducing
the leakage measured in this single exposure to that
assuming continuous operation for 1 h with tube current
5 mA, a value of 2.95 mGy was obtained. This is about
three times the current limit for leakage and it would be
maximum tube potential of the unit (i.e. 115 kVp).
As a result of these measurements, the mobile X-ray
unit was returned to the vendor company, which a few
weeks later supplied us with a new unit (the same brand
and model) that unfortunately presented the same
problem. For the second unit, the leakage was measured
with tube potential 110 kVp and tube loading 50 mAs
(2.5 s). The dosemeter reading at 1 m from the tube was
12.1 mGy and thus, assuming continuous operation for
1 h with tube current 4 mA, a leakage of 3.5 mGy was
derived.
In view of these results, the available certificates and
documentation concerning this X-ray unit were scruti-
nized and an anomaly was apparent. The tube QC
certificate stated that the maximum value measured for
leakage was 1750 mR h21 (15.3 mGy h21), with the limit
set at 3.4 R h21 (30 mGy h21). Furthermore, the technical
specifications section of the operator manual stated that
the leakage radiation limit had been defined as 3448 mR
h21 (30 mGy h21) for a duty cycle of 12 exposures per
hour, with a 5 min time interval between exposures. On
the other hand, in the collimator certificate (given with
the serial number of the specific tube), the maximum
leakage (presumably from the collimator alone) had been
measured at less than 34 mR h21 (0.3 mGy h21) with
exposure factors 125 kVp and 4 mA, while the limit has
been set at 40 mR h21 (0.35 mGy h21).
The methodology used by the tube manufacturer to
determine the maximum permissible leakage was as
follows. The limit for the leakage radiation in 1 h had
been set to 1 mGy (air kerma) and this had been
converted to its equivalent of 115 mR. The X-ray unit
had been assumed to have a duty cycle of 12 radiographs

162 The British Journal of Radiology, February 2006

Short communication: Excessive leakage radiation in mobile X-ray units
per hour and thus the maximum leakage for each
exposure had been calculated as 115 mR/1259.58 mR
(0.084 mGy). By assuming a tube loading of 200 mAs
(20 mA 6 10 s) per exposure, the manufacturer calcu-
lated that the maximum leakage would be (9.58 mR/
10 s) 6 (3600 s h21)53448.8 mR h21 (30 mGy h21). In
this way, the limits of 3.4 R h21 and 30 mGy h21 given in
the operator manual were derived. If the value of 3.4 R
h21 (30 mGy h21) is multiplied by 120 s, that is the
cumulative exposure time in 1 h for the duty cycle
assumed, the limit of 115 mR (1 mGy) in 1 h is obtained
and, therefore, according to the manufacturer the X-ray
tube was complying with the relevant norms.
This methodology, although seeming reasonable at
first glance, is incorrect. It could lead to significant doses
being received by both employees and members of the
public, as is illustrated in the following discussion. For a
quick overview of the differences between the two
methodologies for calculating leakage radiation, the
results of the measurements made at the hospital and
the factory are summarized in Table 1.
Discussion
In the recently published report No. 147 of the NCRP
[1] it is stated that ‘‘manufacturers are currently required
by regulation to limit the leakage radiation to 0.876 mGy
h21 (100 mR h21) at 1 m. Compliance with this require-
ment is evaluated using the maximum X-ray tube
potential and the maximum beam current at that
potential for continuous tube operation’’. These max-
imum tube potential and current ratings (kVpmax and
Imax, respectively) are usually quoted as leakage
technique factors. Imax depends on kVpmax and the
values typically assumed for Imax are 3.3 mA, 4 mA and
5 mA for kVpmax of 150 kVp, 125 kVp and ¡100 kVp,
respectively [1, 2].
Since in radiation protection a number of dosimetric
quantities are simultaneously used and often confused
(as it was seen in the manufacturer calculations where a
limit of 115 mR instead of 100 mR was used), their
relationships should be clarified. An exposure of 1 R
corresponds to an air kerma of 8.76 mGy that for
shielding calculations is traditionally assumed to result
in an absorbed dose in tissues of 10 mGy and an
equivalent dose of 10 mSv [1, 3, 4]. Within this context
the leakage limit is also given as 1 mSv h21 [1], in terms
of ambient dose equivalent. Thus, while the quantity
usually measured with dosemeters is the exposure (in R)
or the air kerma (in Gy), the aforementioned correspon-
dence between units is used to convert the measured or
Table 1. The measurement conditions and assumptions used at the hospital and the factory for calculating the leakage
radiation. Measured and calculated dose values are for a distance of 1 m from the tube focus
Mobile Unit Tube potential Tube loading
(kVp) (mAs)
Unit 1 (at Hospital) 100 50 (25 mA62 s)
Unit 2 (at Hospital) 110 50 (20 mA62.5 s)
Unit 2 (at Factory) 115 200 (20 mA610 s)
The British Journal of Radiology, February 2006
theoretically estimated values of exposure and air kerma
to equivalent dose (ambient dose equivalent). To find the
required shielding thickness or assess the adequacy of
the existing shielding, the resulting value of ambient
dose equivalent is compared with the respective limits of
effective dose for the personnel and public and, in certain
cases, with the equivalent dose limits for the skin and the
lens of the eye.
To determine the shielding requirements of the given
diagnostic tube, the methodology of Tsalafoutas et al [4]
was employed, assuming leakage technique factors of
4 mA and 115 kVp and using published data on the X-
ray output [2] and the attenuation properties of lead [5].
To conform to the current limit for leakage, lead
shielding of 2.07 mm Pb is required, whereas according
to the manufacturer-derived limit of 30 mGy h21 the
respective value is only 0.72 mm Pb. According to
the measurements made at the hospital, the leakage for
the second unit was 3.5 mGy h21 and thus the shielding
of the tube should be about 1.44 mm Pb equivalent.
The implications of the duty cycle concept used by the
tube manufacturer to calculate the tube shielding require-
ments could be made obvious, if one were to initially
accept that the leakage limit is 1 mGy h21 (air kerma) and
only 12 exposures of 200 mAs can be realised in 1 h, and
then assume that the operator of this mobile unit was
performing only these 12 examinations within a ward
during a working day, standing at 1 m away from the
tube. Under these assumptions and even if the scattered
radiation from the patient is ignored, the operator would
be exposed to 1 mGy air kerma corresponding to an
equivalent dose of about 1.15 mSv. Assuming 22 working
days per month and 10 months per year, the cumulative
equivalent dose would be 25 mSv per month and 250 mSv
per year. It is obvious that these values are too high
compared with the annual effective dose limit for
occupationally exposed persons (20 mSv) and the annual
equivalent dose limits for the lens of the eye and the skin
(150 mSv and 500 mSv, respectively).
Good practice requires that the operator should wear a
protective lead apron and should be 2 m or more away
from the tube or behind a wall, while the tube potential
and tube loading routinely used are less than that
assumed and therefore the actual dose would be much
less. The above simplistic calculations, however,
illustrate that the duty cycle concept is by definition
dangerous since, except for the operator, one must also
take into account the patients on the nearby beds and the
other medical staff within the ward or within the nearby
unshielded rooms. It is worth also mentioning that since
a 400 speed class screen–film combination obtains a net
optical density of 1 with about 2.5 mGy, special care
Measured air Dose rate Tube current Leakage
kerma (mGy) (mGy h21) and exposure radiation
time used for (mGy h21)
calculating leakage
radiation per hour
8.2 14.7 5 mA660 min 2.95
12.1 17.3 4 mA660 min 3.5
43 15.3 20 mA62 min 0.52
163

would be required to shield the cassettes from leakage
radiation when transporting them with that mobile unit.
Since the radiographic and fluoroscopic mobile units
used in a fixed location or frequently in the same location
may also require structural shielding [1], the implications
of the duty cycle concept on the shielding requirements
should also be mentioned. Whilst stationary X-ray units
are able to operate at higher tube loadings than mobile
units, as far as the required structural shielding is
concerned there is no essential difference between
stationary and mobile units, if the weekly workload,
operating potential etc. are the same. This is because
when calculating the shielding requirements of a room
the weekly workload is assumed in mA min without
differentiating if a workload of 300 mA min, for
example, will be obtained with 1 h continuous fluoro-
scopy and 5 mA tube current, intermittent fluoros-
copy with tube current 1 mA and cumulative fluoro-
scopy time of 300 min, or with radiographic exposures of
300 mA and total beam-on time of 1 min made
up by many short exposures with duration of a few
milliseconds. Thus, it is easily understood that if a tube
were shielded according to the duty cycle concept, the
structural shielding requirements would be considerably
determined by the leakage radiation. Therefore, the tube
shielding should be designed for the maximum possible
duty cycle, not for a typical duty cycle.
Conclusion
Proper shielding of any X-ray tube, using the standard
methodology and leakage limit, is mandatory for the
radiation protection of the operators, medical personnel,
patients and public. Indeed, most tube manufacturers
164
I A Tsalafoutas
shield their tubes so as to comply with stricter limits than
1 mSv h21. Tubes with shielding calculated in ways
similar to that reported in this study should be
considered as a potential radiation hazard and should
be recalled in order to be properly shielded. The proper
shielding of the tube is imperative for mobile radio-
graphic units and fluoroscopic C-arm units used for
interventional procedures, as in these cases the operator
and the rest of the medical staff do not enjoy the
radiation protection offered by the shielded walls of a
common fluoroscopic or radiographic facility.
Concerning this specific mobile X-ray unit, it must be
mentioned that after negotiations the unit was recalled to
the factory in order to be properly shielded. After it was
returned to the hospital, the leakage radiation had been
reduced to about 1/8 of its previous value, thus
conforming to the current leakage radiation limit.
References
1. National Council on Radiation Protection and
Measurements. Structural shielding design for medical X-
ray imaging facilities. NCRP Report 147. Bethesda, MD:
NCRP, 2004.
2. Simpkin DJ, Dixon RL. Secondary shielding barriers for
diagnostic X-ray facilities: scatter and leakage revisited.
Health Phys 1998;74:350–65.
3. Archer BR, Fewell TR, Conway BJ, Quinn PW. Attenuation
properties of diagnostic X-ray shielding materials. Med Phys
1994;21:1499–507.
4. Tsalafoutas IA, Yakoumakis E, Sandilos P. A model for
calculating shielding requirements in diagnostic X-ray
facilities. Br J Radiol 2003;76:731–7.
5. Simpkin DJ. Transmission data for shielding diagnostic X-
ray facilities. Health Phys 1995;68:704–9.
The British Journal of Radiology, February 2006
The British Journal of Radiology, 79 (2006), 165–166

SHORT COMMUNICATION

Improvements in dose homogeneity for tangential breast fields

from a selection of combinations of library compensators
1 1
R J WILKS, BSc, PhD, T CAMMACK, MPhys, MSc and 2P BLISS, MRCP, FRCR

Departments of 1Medical Physics and 2Clinical Oncology, Torbay Hospital, Newton Road, Torquay
TQ2 7AA, UK

ABSTRACT. Individually paired physical compensators are used in our centre to improve
dose homogeneity for radiotherapy to the whole breast. This technical note describes
the further improvements that may be achieved when all possible combinations of
individual compensators within the library are considered. A retrospective study of 78 Received 9 June 2005
patients using a total of 16 (left-sided) and 14 (right-sided) sets of library compensators Revised 16 August 2005
was evaluated, and the results expressed in terms of the standard deviation of the Accepted 31 August 2005
differential dose–volume histogram and the dose range within the breast volume. The
DOI: 10.1259/bjr/53167057
mean of the standard deviations was 3.17% (uncompensated), 2.16% (paired
compensators) and 1.97% (combinations) and the mean homogeneity was 15.3%, ’ 2006 The British Institute of
11.8% and 11.1%, respectively. Radiology

Wilks and Bliss [1] showed that it was possible to use a
library of a relatively small number of reusable compen-
sators as a routine procedure in reducing the dose
variation to the whole breast from tangential field
radiotherapy. This approach reduced the number of
individual compensators requiring manufacture, thus
reducing the workload on the staff involved. It was
found that out of a 100 patients, approximately 50
required compensation and 45 could be treated with one
of the library compensator pairs. The percentage of
patients planned with a library compensator has subse-
quently increased to approximately 70% of all breast
patients. In an attempt to both improve treatment dose
variation and to reduce the number of individual
compensators requiring manufacture, it was decided to
evaluate the likely improvements to be gained from
using the library compensator plates in different combi-
nations. Accordingly, a retrospective study was under-
taken of 78 patients of whom 4 patients were treated with
their own individually made compensators, 15 were
treated without compensators and 59 were treated with
paired library compensators. When these patient treat-
ments were originally planned, a nominal threshold for
the standard deviation of the dose–volume histogram of
2.5% was used to decide whether the uncompensated
dose distribution was acceptable for treatment. Increased
computer processing speeds have since made it practic-
able to investigate compensation for all patients.
Method
Each patient treatment was re-planned using addi-
tional software on the Osiris planning system [1], which
Address correspondence to: T Cammack.
allowed the individual library compensator plates to be
used in combination with all (or none) of the other
possible plates. Only those compensators whose dimen-
sions were the same or greater than the tangential fields
used to treat the patient were chosen for analysis. This
means that the smaller field sizes had more possible
combinations of plates than the larger field sizes. Of N
possible pairs of compensator plates, the number of
combinations was 2N (single plates) + N2.
Results
It was found possible to select a mixed compensator
combination for all 78 patients, whereas only 73 patients’
plans were improved with respect to no compensation
when using the standard paired compensators. Of the
remaining five patients, four were treated using indivi-
dual compensators. The fifth was treated uncompen-
sated as the standard deviation of the dose–volume
histogram was below the nominal threshold (2.5%) for
production of individual compensators. For the other
three patients treated with individual compensators, in
two of them the standard deviations of the differential
dose–volume histograms were improved by 0.1%, using
mixed compensator combinations. The third resulted in
no change. Therefore, using mixed compensator combi-
nations would have meant that no additional individual
compensators would have needed to be made.
Overall, for the 78 mixed compensator combinations
chosen, the dose variation was unchanged for 18 (23%)
patients and improved (i.e. reduced) for 59 (76%)
patients. One patient showed a small increase in dose
variation for a mixed pair over an individually manu-
factured compensator pair. The compensator combina-
tion selected improved on the standard library pairs

The British Journal of Radiology, February 2006 165

 R J Wilks, T Cammack and P Bliss

Table 1. Measures of the improvements achieved by the use of compensator combinations. Each figure is followed by the
standard deviation (SD) of that parameter
Compensator combinations

None Original pairs Mixed and single

Mean SD of dose–volume histogram (%) 3.17¡0.6 2.16¡0.6 1.97¡0.6

Mean dose range (%) 15.9¡4.1 11.9¡3.9 11.2¡3.7
Mean homogeneitya (%) 15.3¡3.8 11.8¡3.9 11.1¡3.6
Mean percentage of PTVb volume with . 1.056reference dose 28.4¡15.9 10.0¡10.1 7.9¡10.1
Mean percentage of PTVb volume with . 1.106reference dose 6.6¡9.4 0.3¡0.7 0.3¡1.0
a
Homogeneity is defined here as 26(maximum dose – minimum dose)/(maximum dose + minimum dose).
b
Planning target volume (PTV), here taken simply as the breast tissue within the tangential fields.

method and would probably have been judged to be
acceptable by the consultant, even though the standard
deviation was greater than the nominal threshold (2.8%
rather than 2.5%).
Some improvements were more marked than others,
since many original plans were already close to their
optimum. Several measures of the dose uniformity were
used to assess the differences in the dose distributions:
(1) the standard deviation (SD) of the differential dose–
volume histogram; (2) the dose range; (3) the breast
volume which had a dose greater than 5% above the
reference dose. The improvements of the SD ranged from
0.0 to 0.6%. Table 1 shows a summary of the improve-
ments obtained. As Table 1 shows, there is a systematic
but not highly significant improvement in dose uni-
formity when using mixed combinations of compensa-
tors. However, mixed compensators have proved to be
applicable to more patients than the standard library
pairs, as expected.
Conclusion
The technique of employing combinations of reusable
compensators has a definite advantage in both increasing
the likelihood of achieving a more uniform distribution
and in reducing the number of additional individual
compensators, which would otherwise require
manufacture.
References
1. Wilks RJ, Bliss P. The use of a compensator library to reduce
dose inhomogeneity in tangential radiotherapy of the breast.
Radiother Oncol 2002;62:147–57.

166 The British Journal of Radiology, February 2006

The British Journal of Radiology, 79 (2006), 167–169

CASE REPORT

Ruptured spinal dermoid cyst with disseminated intracranial fat

droplets
J G CHA, MD, S-H PAIK, MD, J-S PARK, MD, S-J PARK, MD, D-H KIM, MD and H-K LEE, MD

Department of Radiology, Soonchunhyang University Bucheon Hospital, 1174, Jung-dong, Wonmi-

gu, Bucheon-St Gyeonggi-do, 420-021, Republic of Korea

ABSTRACT. Fat droplets in the cerebrospinal fluid (CSF) is a well-known complication of

ruptured intracranial dermoid tumours. We report an unusual case of a ruptured spinal Received 25 February 2005
dermoid tumour. MR images showed a tethered spinal cord and an intramedullary Revised 6 May 2005
Accepted 6 May 2005
fat-containing mass. Fat droplets were revealed in the ventricles and the cisternal
spaces on brain CT and brain MR. In the English literature, a ruptured spinal dermoid DOI: 10.1259/bjr/17232685
tumour accompanying a tethered spinal cord is extremely rare.
’ 2006 The British Institute of
Radiology

Intraspinal dermoid tumours are rare benign, slow-
growing tumours and tend to extend to the subarachnoid
space. Dermoid tumours comprise 1.1% of intraspinal
tumours [1]. There is no communication between the cyst
and the subarachnoid space. Several causes including
spontaneous, iatrogenic or traumatic rupture have been
reported to result in dissemination of lipid material from
the dermoid tumours into the subarachnoid space or
ventricles. We report a case of spontaneously ruptured
spinal dermoid tumour with disseminated intracranial
fat droplets and tethered cord.
Case report
A 44-year-old man presented with a history of voiding
difficulty starting 10 years ago and exacerbation 3
months prior to admission. There was no history of
lumbar puncture or major operation. Physical examina-
tion revealed radiating pain down left S2 dermatome
level (hyperaesthesia). Laboratory findings were normal.
The patient underwent MRI of the spine using a 1.5 T
scanner. T1 weighted (660/10/4 [repetition time/echo
time/excitation]) (Figure 1a) and T2 weighted (4000/
123/4 [repetition time/echo time/excitation]) (Figure
1b) images showed a hyperintense intramedullary mass
at the level of L3–L5 and tethered cord. Brain CT
(Figure 1c) and brain MR (4000/9/4 [repetition time/
echo time/excitation]) (Figure 1d) demonstrated multi-
ple small fat droplets in the intraventricular and cisternal
space suggesting rupture of spinal dermoid cyst. The
patient underwent a laminectomy from L2 to L5.
Discussion
Spinal dermoids are dysontogenetic tumours arising
from inclusion of ectopic embryonic rests of the
ectoderm within the spinal canal at the time of neural
tube closure between the third and the fifth week of
embryonic development [2]. Dermoid tumours show a
slight male predominance, and most dermoid tumours
are revealed during the second and third decades.
Dermoid tumours may be related to bony malforma-
tions, myelomeningocele [3], hypertrichosis and/or a
dermal sinus tract. The lumbosacral region is most
common site (60%) involving the cauda equina and the
cornus medullaris followed by the upper thoracic (10%)
and cervical (5%) regions [4]. They have a thick wall
covered with stratified squamous epithelium containing
dermal appendages such as hair, sebaceous glands,
sweat glands and hair follicle and less commonly, teeth
and nails [5]. Dermoid tumours commonly have areas of
calcification. The relatively high signal from fat on MRI,
especially the bright signal on T1 weighted images,
makes identification of lipid droplets easy, particularly
within the cerebral sulci, fissures, the perimedullary
subarachnoid space and the central canal of the spinal
cord. MRI has also shown more frequent asymptomatic
spillage of lipid material [6].
Dermoid tumours may have two distinct portions, a
lipid one and a more solid or more fluid one [4] as in our
case, showing fat tissue in peripheral portion of the
tumour and fluid content in the central portion. In
addition, especially with leakage of lipid material, the
use of intravenous contrast medium makes diagnosis of
a meningeal inflammation easier.
Although dermoid tumours develop from the embryo-
nic period, symptoms may not occur until adulthood
due to their slow growth [1], symptoms and signs
secondary to the space-occupying lesion are location-
dependent and are due to the irritative effect on and/or
compression of the adjacent structures. When accompa-
nied by a tethered cord, particularly with small lesions,
neurological symptoms can be elicited without mass
effect.

The British Journal of Radiology, February 2006 167

 J G Cha, S-H Paik, J-S Park et al

(a) (b)

(c) (d)

Figure 1. (a) Sagittal T1 weighted (repetition time (TR) 660/echo time (TE) 10) spin-echo image demonstrates central
hypointense and peripheral hyperintense intramedullary mass at the level of L3–L5 and tethered spinal cord. (b) Sagittal T2
weighted (TR 4000/TE 123) spin-echo image shows homogeneous hyperintense mass. (c) Brain CT shows fat droplets in both
frontal and lateral ventricles and cisternal spaces. (d) Axial T1 weighted (TR 400/TE 9) revealed hyperintense lipid materials in left
frontal and both lateral ventricles, and subarachnoid space.

168 The British Journal of Radiology, February 2006

Case report: Ruptured spinal dermoid cyst
Once rupture of the cyst occurs acute symptoms relate
to chemical or aseptic meningitis [3], headache or
seizures may be developed due to dissemination of lipid
droplets in the cerebrospinal fluid (CSF) pathways.
Lumbar arachnoiditis may be developed as a result of
leakage of fat and proteinaceous material into the
subarachnoid space. The highly irritative lipid content
of dermoid tumours may cause severe inflammatory
response, though spread of fat into the CSF may also be
clinically silent [7].
After rupture of dermoid tumour occurs, lipid
droplets float in the CSF and are passively conveyed
by CSF movement. It can therefore spread throughout
the subarachnoid space and ventricular system. Scearce
et al [8] insisted that fat droplets reach the ventricles
from the perimedullary subarachnoid space by retro-
grade flow through the foramina of Luschka and
Magendie. To our knowledge, few cases of rupture of
dermoid spinal tumours have been reported [8–12].
In conclusion, MRI is not only helpful in detecting
intraspinal dermoid tumours and the fat droplets in
CSF space even in an asymptomatic case of rupture of
the tumour, but also diagnosing the associated con-
genital anomalies such as tethered cord, as is seen in our
case.
References
1. Lunardi P, Missori P, Gagliardi FM, Fortuna A. Long-term
results of the surgical treatment of spinal dermoid and
epidermoid tumors. Neurosurgery 1989;25:860–4.
The British Journal of Radiology, February 2006
2. Netsky MG. Epidermoid tumors: review of the literature.
Surg Neurol 1988;29:477–83.
3. Quigley MR, Schinco F, Brown JT. Anterior sacral menin-
gocele with an unusual presentation: case report. J
Neurosurg 1984;61:790–2.
4. Graham DV, Tampieri D, Villemure JG. Intramedullary
dermoid tumor diagnosed with the assistance of magnetic
resonance imaging. Neurosurgery 1988;23:765–7.
5. Amirjamshidi A, Ghodsi M, Edraki K. Teeth in the
cerebellopontine angle: an unusual dermoid tumour. Br J
Neurosurg 1995;9:679–82.
6. Stephenson TF, Spitzer RM. MR and CT appearance of
ruptured intracranial dermoid tumors. Comput Radiol
1987;11:249–51.
7. Funke M. Ruptured intracranial dermoid as an incidental
finding. Aktuelle Radiol 1995;5:232–4.
8. Scearce TA, Shaw CM, Bronstein AD, Swanson PD.
Intraventricular fat from a ruptured sacral dermoid cyst:
clinical, radiographic, and pathological correlation: case
report. J Neurosurg 1993;78:666–8.
9. Calabro F, Capellini C, Jinkins JR. Rupture of spinal
dermoid tumors with spread of fatty droplets in the
cerebrospinal fluid pathways. Neuroradiology
2000;42:572–9.
10. Karadag D, Karagulle AT, Erden A, Erden I. MR imaging of
a ruptured intraspinal dermoid tumour with fat droplets in
the central spinal canal. Australas Radiol 2002;46:444–6.
11. Goyal A, Singh D, Singh AK, Gupta V, Sinha S.
Spontaneous rupture of spinal dermoid cyst with dissemi-
nated lipid droplets in central canal and ventricles. J
Neurosurg Sci 2004;48:63–5.
12. Garg A, Gupta V, Gaikwad S, Deol P, Mishra NK, Suri A,
et al. Isolated central canal rupture of spinal dermoid:
report of two cases. Australas Radiol 2003;47:194–7.
169
 The British Journal of Radiology, 79 (2006), 170–172

CASE REPORT

Colobronchial fistula: a late complication of childhood

radiotherapy
G C MACKAY, MRCP, J HOWELLS, MRCP, FRCR and F W POON, FRCR

Department of Radiology, Glasgow Royal Infirmary, Queen Elizabeth Building, 16 Alexandra

Parade, Glasgow G31 2ER, UK

ABSTRACT. We present the case of a colobronchial fistula in a 41-year-old man who

underwent radiotherapy for nephroblastoma as an infant. He attended for barium Received 15 February 2005
enema, which demonstrated a fistula between colon and bronchial tree. Following Revised 9 May 2005
Accepted 11 May 2005
right hemicolectomy and pathological examination of the resected bowel, no active
disease process was identified to explain the development of this rare fistula. DOI: 10.1259/bjr/27258284
Radiotherapy was deemed the most probable aetiology. We are unaware of this having
been previously described. ’ 2006 The British Institute of
Radiology

Colobronchial fistulae are rare, and have previously extended superiorly towards the right subphrenic space
been reported in adults secondary to Crohn’s disease [1], (Figure 2). The examination was immediately termi-
colonic malignancy [2], tuberculosis [3] and as complica- nated, although the patient remained haemodynamically
tions of gastrointestinal surgery [4–6]. We report the stable with no signs of peritonism. Appearances were
relevant radiological and clinical findings in a case of consistent with a localized colonic perforation, for which
colobronchial fistula as a likely result of radiotherapy an iatrogenic cause was not thought likely. The leak
40 years previously. appeared confined, with no evidence of free intra-
To the best of our knowledge, this aetiology has not peritoneal air nor generalized barium contamination of
previously been described. the peritoneal cavity.
Contrast-enhanced CT of abdomen and pelvis per-
formed on the same day confirmed a broad tract of
Case report
A 41-year-old man was referred to the surgical out-
patient department for investigation of a left sided dis-
charging perianal sinus. Colonoscopy had been normal.
99
Tcm-hexamethylpropyleneamineoxime (HMPAO)
labelled white cell isotope scan demonstrated increased
tracer uptake in the subhepatic space and to the right of
the lumbar spine (Figure 1), the significance of which
was initially unclear. Subsequent examination under
anaesthetic (EUA) and endoanal ultrasound diagnosed a
complex extrasphincteric fistula. A barium enema was
arranged to try to demonstrate any fistulous connection
which may have been missed by previous colonoscopy.
His relevant past medical history included a right
nephrectomy and subsequent intensive radiotherapy
for nephroblastoma at the age of 1 year.
On presentation for barium enema, the patient
complained of a 6 week history of general malaise,
weight loss, right upper quadrant pain, dyspnoea and
cough with associated malodorous ‘‘chocolate-coloured’’
sputum.
Barium enema demonstrated a tract of extraluminal
barium arising from the proximal transverse colon which
Figure 1. 99Tcm hexamethylpropyleneamineoxime (HMPAO)
Address correspondence to: Dr Gillian MacKay, 7 Rosevale Road, labelled isotope scan demonstrating high uptake in the right
Bearsden, Glasgow G61 2RX, Scotland, UK. paravertebral region (illustrated by arrow).

170 The British Journal of Radiology, February 2006

Case report: Colobronchial fistula
Figure 2. Fistulous tract of barium arising from proximal
transverse colon (arrows delineate the extent of the tract).
Marked deformity of the vertebral bodies is noted consistent
with previous radiotherapy induced damage.
Figure 3. Contrast enhanced CT chest demonstrating a large
thick-walled cavity, containing barium and air, within the
lower lobe of the right lung. Barium is seen within the right
lower lobe bronchial tree.
barium extending superiorly from the transverse colon,
passing behind the liver to enter the right hemithorax. It
communicated directly with a thick walled cavity within
right posterior hemithorax, measuring 7.5 cm 6 4.5 cm,
containing air and barium. Barium was also seen to enter
adjacent bronchi (Figure 3). Overall, appearances were
consistent with an established colobronchial fistula.
Marked enlargement of the azygous and hemi-azygous
venous systems was noted, together with a reduction in
calibre of the inferior vena cava. Deformities of the upper
thoracic vertebrae were also present.
The British Journal of Radiology, February 2006
The patient was admitted to the general surgical ward
on the day of presentation, and underwent extended
right hemicolectomy with ileo-transverse anastomosis.
The thoracic cavity was not examined or drained by the
cardiothoracic surgeons, who were present at the time of
surgery. Dense adhesions in the right hypochondrium,
between colon and liver, and duodenum and colon were
identified. The operative findings were thought to have
been radiotherapy related. Pathology of the right hemi-
colon revealed fibrous adhesions and serosal fibrosis
with no evidence of active inflammation or malignancy.
The patient endured a stormy post-operative course,
but made a slow recovery. He has been found to be unfit
for thoracotomy, and his pulmonary sepsis has therefore
been managed conservatively.
A subsequent barium follow-through examination
identified a hitherto clinically silent duodenal stenosis
(Figure 4). In the absence of other significant history, this
series of findings were felt to be consistent with late
sequelae of wide-field, high-dose radiotherapy.
It is worth noting that the presenting complaint of
the ischiorectal fistula was unrelated to the eventual
diagnosis.
Discussion
Nephroblastoma (Wilms’ tumour) is the most common
cancer of the urinary tract in children. In the pre-
chemotherapy era, post-operative radiotherapy was
shown to increase patient survival. Wide-field radio-
therapy was administered to the side of the abdomen on
which the tumour occurred, concurring with the findings
in this patient.
Figure 4. Barium follow-through examination showing a
significant duodenal stenosis (illustrated by arrow).
171

With regard to the patient’s symptoms on presenta-
tion, it now seems likely that the brown, malodorous
sputum he described was actually faeculent. The organ-
isms grown from the endotracheal tube secretions were
Coliform bacilli, consistent with this conclusion.
In retrospect, and with the benefit of further imaging,
it is likely that the area of high white cell uptake adjacent
to the right lumbar area represented inflammation at the
site of the fistulous track within the abdomen.
The patient has had a normal colonoscopy, negative
Gram-staining of sputum and blood cultures, and non-
specific pathology of the right hemicolon (in particular,
there is a lack of features to suggest Crohn’s disease). The
most likely aetiological conclusion has therefore been
reached by a process of elimination.
Colobronchial fistula is an uncommon finding. It has
been described in relation to Crohn’s disease, in which
fistula formation is a classical manifestation, as a late
complication of appendicitis, and following laparoscopic
biliary surgery. Colonic malignancy and tuberculosis are
other documented causes. However, we can find no
reference to a fistula of this nature ever having been
described secondary to abdominal radiotherapy. Not
only is the extent of this fistula highly unusual, but the
length of time taken for symptoms to develop is
exceptional.
Although it is extremely rare, fistula formation many
years following radiotherapy has been previously
described in a patient who developed an enterocuta-
neous fistula 27 years after radiotherapy for carcinoma of
the penis [7]. Given that the oncological treatment of this
patient took place in the early 1960s, the field of
radiotherapy would have been right-sided but wide,
with larger fractions being administered in comparison
172
G C MacKay, J Howells and F W Poon
with today’s treatment doses. Fistulae more localized to
the site of disease and area of treatment are known to be
potential direct complications of radiotherapy and have
been well-documented. However, one would expect the
onset of related symptoms to occur within a relatively
short timescale of treatment.
In treating this condition, surgical resection of the
colon is usual, with subsequent lung resection.
Unfortunately, the patient has not, as yet, been deemed
fit enough to undergo lobectomy, and so treatment has
been suboptimal. Long term clinical outcome is therefore
unclear.
References
1. Karmy-Jones R, Chagpar A, Vallieres E, Hamilton S.
Colobronchial fistula due to Crohn’s disease. Ann Thorac
Surg 1995;60:446–8.
2. Tenuchi S, Saku N, Ishii Y, et al. A case of colon cancer with
tension pneumothorax and empyema as a consequence
of colopleural fistula. Nikon Kokyuki Gakkai Zaashi
2000;38:865–9.
3. Crofts TJ, Dalrymple JO, Buhrmann JR. Tuberculous bronch-
ocolic fistula. S Afr Med J 1978;54:795–6.
4. Pochin R, Frizzelle F. Colonic-broncho fistula: a previously
unreported complication following laparoscopic cholecys-
tectomy. Case Rep Clin Pract Rev 2003;4:77–9.
5. Corlett SK, Windle R, Cookson JB. Colobronchial fistula: a
late complication of appendicitis. Thorax 1988;43:420–1.
6. Lucas TA, Reynolds HY. Diagnosis and management of a
colobronchial fistula in a 55 year old male with feculent
sputum. Chest 1999;116:395–6.
7. Chintamani, Badran R, Rk D, Singhal V, Bhatnagar D.
Spontaneous enterocutaneous fistula 27 years following
radiotherapy in a patient of carcinoma penis. World J Surg
Oncol 2003;1:23.
The British Journal of Radiology, February 2006
The British Journal of Radiology, 79 (2006), 173–174

CASE REPORT
Ingested foreign body mimicking an appendicolith in a child
V MOORJANI, MD, C WONG, FRANZCR and A LAM, FRANCZR

Department of Imaging, The Childrens Hospital at Westmead, Locked bag 4001, Westmead,
Sydney, NSW 2145, Australia

ABSTRACT. We describe an unusual case of a child who had ingested sand and stones
and presented with signs and symptoms suggestive of appendicitis. Plain radiographs Received 2 December 2004
revealed calcific opacities in the right iliac fossa simulating appendicoliths. At surgery Revised 10 April 2005
Accepted 23 May 2005
and histopathology a small sealed off perforation of the terminal ileum with hard
concretions in the wall was observed. DOI: 10.1259/bjr/85745053

’ 2006 The British Institute of

Radiology

Case report normal. The ileal segment and the appendix were
resected and an ileocolic anastomosis performed.
An 18-month-old male child was referred to our Histopathology revealed ulcerations with full thickness
hospital with nausea, vomiting abdominal pain and fever necrosis. There was hard debris within the intestinal
lasting 3 days. On examination the child had a non- wall. No intrinsic inflammation was seen in the appendix
distended abdomen and generalized peritonism. An but the serosa and subserosa were actively inflamed with
abdominal radiograph revealed two radiopaque foreign
some purulent exudate on the surface.
bodies in the right iliac fossa with distension of proximal
On further questioning the parents, a history of the child
loops of bowel (Figure 1). Mobility of the calcific opacities
ingesting sand and stones was elicited. A calcific opacity
was seen on the radiographs taken prior to admission and
was still present on the repeat radiograph (Figure 4). The
after being admitted to the hospital (Figure 2). An
child was discharged and parents were asked to follow up
ultrasound examination demonstrated two echogenic
and to check the stools daily for foreign bodies.
areas in the right iliac fossa which appeared to be in a
non-compressible inflamed loop of bowel (Figure 3). The
bowel wall was thickened and the adjacent mesentery
adherent to it. There was minimal free fluid in the
peritoneal cavity. A provisional diagnosis of acute
appendicitis with appendicoliths was entertained, a
comment was made to the referring surgeon that the
overall appearance was unusual for acute appendicitis.
The child underwent an exploratory laparotomy.
At operation the terminal ileum was severely inflamed
with a small sealed perforation. The appendix looked

Figure 1. Plain supine abdominal radiographs. Two calcific Figure 2. Plain supine abdominal radiograph, a day later
opacities are seen in the right iliac fossa. revealed the opacities to be further apart.

The British Journal of Radiology, February 2006 173


Figure 3. Ultrasound of the right iliac fossa. A calcific
opacity is seen in the lumen of the bowel. The walls of the
loop are thickened.
Figure 4. Plain supine radiograph. A calcific opacity is still
seen in the right iliac fossa.
Discussion
Depending on their nature, ingested foreign bodies
may have an uneventful course or be impacted in the
gastrointestinal tract. Most of the foreign bodies pass
through the gastrointestinal tract within a week [1]. The
reported incidence of bowel perforation is less than 1%,
the objects being pointed or sharp in most cases like tooth
picks, sewing needles, dental plates, fish bones and
chicken bones [1]. The most common sites of perforation
174
V Moorjani, C Wong and A Lam
are the ileocecal junction and the rectosigmoid region.
There has been a single case report of plum pits in the
terminal ileum. This resulted in an inflammatory con-
glomerate tumour proximal to the ileocecal valve and
presented years later with intestinal bleeding [2]. Segal [3]
reported a case of ingestion of soft drink bottle top and
stones which resulted in chronic ileitis. Sclerosing
encapsulated peritonitis also called as calcifying peritoni-
tis appears as dilated bowel loops, air fluid levels and
peritoneal calcifications on plain films. Ultrasound reveals
dilated fixed bowel loops matted together and tethered
posteriorly, intraperitoneal echogenic strands and an
echogenic sandwich appearance of the membrane.
The child in our case had ingested stones and sand.
The wall of the terminal ileum was ulcerated either due
the direct action of the stones or due to the inflammation
or both with subsequent necrosis and perforation, and
the sand and small stones spilling out into the
peritoneum. The calcific opacities were not observed on
histology as one of the calcifications must have been
crushed by the microtome and the other could have
moved out of the operative field.
To conclude a common cause of a calcific opacity in the
right iliac fossa in a child who presents with signs and
symptoms of appendicitis would be an appendicolith. The
other differential diagnosis of a calcific opacity in the right
iliac fossa would include calculus in a vesical diverticula,
enterolith, phlebolith in a pelvic vein, ectopic gallstone,
calcification within a teratoma and calcification in a
mesenteric lymph node. A giant appendicolith measuring
3 cm has been reported [4]. Calculi located in urinary
bladder diverticula have a dumb-bell shape with one end
lodged in the diverticulum and the other projecting into
the lumen of the bladder. In a phlebolith usually a central
lucency can be seen. An ectopic gallstone was ruled out
since the gallbladder was normal on ultrasound. The
classic ultrasound appearance of teratoma is a prominent
cystic component and at least one contained mural nodule
that is often echogenic as a result of contained fat, hair,
sebum or calcium. The calcific opacity in our case was
located in a bowel loop on ultrasound.
Retrospective review of the radiographs revealed that
the calcification was very dense in our case and did not
have the typical laminated appearance of an appendico-
lith. The change in the position of the calcifications on
serial radiographs would also be a pointer against an
appendicolith. Rarely a foreign body may simulate
appendicitis with gut inflammation and a radiopacity
should be considered in the differential diagnosis.
References
1. Chintamani, Singhal V, Lubhana P, Durkhere R, Shabnam B.
Liver abscess secondary to a broken needle migration: a case
report. BMC Surgery 2003;3:8–13.
2. Lucker C, Schulte GA, Moldenhauertt. Plum pits in the
terminal ileum. Rontgenblatter 1987;40:191–2.
3. Segal I, Nouri MA, Hamliton DG, Ou Tim L, Giraud RM,
Mirwis J, et al. Foreign-body ileitis: a case report. S Afr Med J
1980;58:421–2.
4. Devalia H, Dhamdhere M, Horner J. A giant appendicolith:
miscellaneous case 2947. In: Gastrointestinal Imaging 2004.
Available via EURORAD. http://www.eurorad.org/case
[Accessed 5 February 2004].
The British Journal of Radiology, February 2006
The British Journal of Radiology, 79 (2006), 175–176

CASE REPORT

Misleading positioning of a Foley catheter balloon

1 1 2
S ABADI, MD, O R BROOK, MD, E SOLOMONOV, MD and 1D FISCHER, MD

1
Department of Diagnostic Imaging, Rambam Medical Center, Haifa, Israel and
2
Department of Surgery, Rambam Medical Center, Haifa, Israel

ABSTRACT. Indwelling catheters in the urinary bladder are associated with numerous
and various complications, e.g. infection, haemorrhage, epididymo-orchitis and
perforation. Abdominopelvic CT is frequently performed in hospitalized patients, with
the bladder being included in the examination. Familiarity with the various bladder
pathologies and a routine and meticulous search for them are indicated in every case. Received 14 January 2005
Revised 9 June 2005
Moreover, an awareness of certain pitfalls may prevent over-diagnosis and over- Accepted 9 June 2005
treatment. We present a case in which a Foley catheter balloon inflated in a bladder
diverticulum simulates sealed bladder perforation with extraluminal location of the DOI: 10.1259/bjr/13576050
balloon. This potentially misleading diagnosis should be considered in the presence of
’ 2006 The British Institute of
apparent extraluminal position of catheter tip or balloon not substantiated by the
Radiology
clinical presentation.

Indwelling catheters in the urinary bladder are
associated with numerous and various complications,
e.g. infection, periurethral abscess, urethral diverticulum,
perineal erosion, bladder atrophy, bladder stones,
haemorrhage, epididymo-orchitis or urinary fistula [1,
2]. Intraperitoneal or extraperitoneal perforation can
occur as a rare but life threatening complication [3]. CT
is a reliable method for evaluating the bladder and
demonstrating possible pathologies such as tumours,
calculi, fistulas or diverticula [4].
Case report
We present the case of an 80-year-old man who
underwent surgery for colon obstruction due to
adhesions. A bladder Foley catheter was placed for
drainage and haemodynamic monitoring.
A CT scan was performed 5 days following surgery,
because of fever, abdominal pain and tenderness. A
collection of fluid was observed in the left lower quadrant.
In addition, on the pelvic axial scan, the Foley catheter was
seen beyond the bladder wall, with the balloon outside its
boundaries (Figure 1). The Foley balloon was not
surrounded by urine and formed an acute angle with
the bladder walls. Two sequential scans were performed
with a 5 h interval between them; this was related to poor
bowel opacification on the initial scan. Both scans showed
the balloon at the same position with the only differences
being that the later scan (Figure 1b) was done with the
bladder more distended with urine than the first (Figure
1a) and opacified with contrast material.

(a) (b)

Figure 1. Axial CT scan of the pelvis at the level of the urinary bladder (a) before and (b) after opacification with contrast
material demonstrates a partially filled bladder with an inflated catheter balloon situated outside and close to its anterior wall.

The British Journal of Radiology, February 2006 175


Figure 2. Axial CT scan at the same level as in Figure 1
performed after repositioning of the catheter in the bladder
demonstrates a distended bladder with a diverticulum
arising from its anterior wall.
This radiological presentation was highly suggestive
of extraluminal location of the balloon, i.e. perforation of
the urinary bladder, and since there was no urine or
contrast material around the catheter, this raised the
suspicion of a sealed perforation.
However, this diagnosis did not concur with the
clinical picture, as the treating surgeon reported normal,
clear urine output from the catheter.
The patient was scheduled for CT guided drainage of
the coincident collection several hours later. Just before
the procedure, the surgeon slightly withdrew the
catheter. A 3 cm bladder diverticulum arising from the
anterior bladder wall was confirmed on a further limited
176
A Abadi, O R Brook, E Solomonov and D Fischer
CT scan (Figure 2). Thus, the diagnosis of intradiverti-
cular placement of the bladder catheter balloon was
made.
Discussion
Abdominopelvic CT is frequently performed in hospi-
talized patients and the bladder is included in the
examination. In many cases the urinary bladder is
catheterized for various indications and durations.
Familiarity with the different bladder pathologies and a
routine meticulous search for them are indicated in every
case. Moreover, a familiarity with the various pitfalls
may prevent over-diagnosis and over-treatment. We
present a case in which a bladder catheter balloon
inflated in a bladder diverticulum simulates sealed
bladder perforation with extraluminal location of the
balloon. This observation is a pitfall that should be
considered in the presence of apparent extraluminal
position of catheter tip or balloon not substantiated by
the clinical presentation.
References
1. Lowthian P. The dangers of long-term catheter drainage. Br J
Nurs 1998;7:366–8, 370, 372.
2. Winson L. Catheterization: a need for improved patient
management. Br J Nurs 1997;6:1229–32, 1234, 1251–2.
3. White SA, Thompson MM, Boyle JR, Bell PR. Extraperitoneal
bladder perforation caused by an indwelling urinary
catheter. Br J Surg 1994;81:1212.
4. Caoili EM, Cohan RH, Korobkin M, et al. Urinary tract
abnormalities: initial experience with multi-detector row CT
urography: Radiology 2002;222:353–60.
The British Journal of Radiology, February 2006
The British Journal of Radiology, 79 (2006), 177–178

CASE OF THE MONTH

An unusual cause and presentation of a pelvic mass

1 2
S HARISH, FRCS, FRCR, A REHM, FRCS and 1P W P BEARCROFT, FRCR

Departments of 1Radiology and 2Orthopaedics, Addenbrooke’s Hospital, Hills Road, Cambridge,

CB2 2QQ, UK

Received 22 November
2004
Revised 15 February 2005
Accepted 14 June 2005

DOI: 10.1259/bjr/65970261

’ 2006 The British Institute of

Radiology

Clinical history
A 19-year-old female was referred to the orthopaedic
clinic with a history of limb length discrepancy. Her only
complaint was that the right leg seemed shorter than the
left. On physical examination, the pelvis was slanting
slightly towards the right with a limited range of
movement of the right hip. There was right-sided hip
pain at the extremes of rotational movements and
flexion. A 10 cm 6 15 cm firm, non-tender mass was
palpable just proximal to the anterior rim of the iliac
crest. According to the patient this mass had been
present for a long time. Clinically, the patient did not
have a leg length discrepancy, but had a fixed pelvic
Figure 2. Axial fat suppressed fast spin echo T2 weighted MR
obliquity causing apparent leg length discrepancy. A image.
plain film was obtained (Figure 1) and the appearance
prompted an MR examination of the pelvis. Axial T1
weighted and fat suppressed axial T2 (Figure 2) and fat
suppressed coronal T2 weighted images (Figure 3) were
obtained. What do these images show?

Figure 1. Plain radiograph of the pelvis.

Address correspondence to: Dr S Harish, Department of Diagnostic

Imaging, St Joseph’s Hospital, 50 Charlton Ave E, Hamilton, Figure 3. Coronal fat suppressed fast spin echo T2 weighted
Ontario, L8N 4A6, Canada. image.

The British Journal of Radiology, February 2006 177


Findings
Plain radiograph of the pelvis (Figure 1) showed a
large irregular, calcified, bony mass arsing from the right
ilium. A diagnosis of an osteochondroma was made.
Because of the size of the lesion, there was a concern that
the lesion may represent a malignant chondroid lesion.
MR of the pelvis (Figures 2 and 3) showed an 8 cm 6
7 cm 6 7 cm lesion arising from the right ilium with
typical features of an osteochondroma. There was a rim
of high signal on T2 weighted imaging in keeping with a
cartilage cap. The cartilage cap measured up to 1 cm in
thickness, but the iliac crests had not fused indicating
that the patient was not skeletally mature. There were no
sinister features to suggest malignant degeneration into a
chondrosarcoma, but nevertheless, with a lesion of this
size, malignancy could not be completely excluded.
Interestingly, the right kidney could not be visualized in
the MR images with some hypertrophy of the left kidney,
and the lesion itself was located close to the expected
anatomical site of the right kidney (Figure 3). Also, there
was hypoplasia of the ipsilateral hemipelvis (Figure 1)
with minor scoliosis concave to the right (Figure 3). At
this point, the possibility of previous right nephrectomy
and radiation to the renal bed causing this osteochon-
droma was invoked. Further history in fact revealed that
the patient did have a right nephrectomy and partial
right hepatectomy for stage 3 type Wilm’s tumour at the
age of 2 years, which was followed by chemotherapy
and radiation of 3000 cGy to the right hemiabdomen and
the tumour bed. Incision biopsy at a tertiary bone
tumour centre confirmed this lesion to be an osteochon-
droma with no pathological features to suggest malig-
nancy. The patient was offered excision of the lesion, but
since she was not symptomatic from the lesion itself, she
preferred to wait for a few years.
Discussion
Radiation-induced osteochondromas arise typically in
patients who received radiotherapy as children, and they
are indistinguishable radiologically and pathologically
from a spontaneous osteochondroma. In a series of 42
patients with Wilm’s tumour treated with radiotherapy
and a minimum follow-up of 5 years, there was a 4.8%
incidence of osteochondromas and a 2.4% incidence of
sarcomas [1]. In another series of 58 children who
received radiotherapy as part of therapy for bone
marrow transplantation, 5 developed osteochondromas
and all of these 5 children were less than 5 years of age at
the time of radiotherapy with no patient who underwent
178
S Harish, A Rehm and P W P Bearcroft
radiotherapy after 5 years of age developing osteochon-
dromas [2]. There is an average latent period of about 5–
8 years before these lesions develop [2, 3] and they may
be found to be enlarging until normal growth ceases and
growth of tumour is maximal at times of patient’s
growth spurt [4]. Malignant degeneration in radiation-
induced osteochondromas is uncommon and the criteria
applied to spontaneous lesions such as increase in size of
the lesion after epiphyses closure, increasing soft tissue
mass and development of pain in the absence of an
alternative explanation should be used for these as well
[3, 5]. A thickened cartilage cap greater than 2 cm raises
the suspicion of malignant change in an osteochondroma
[5]. Only two cases have been adequately described
wherein radiation-induced osteochondromas underwent
malignant change [6]. Other complications of these
lesions include restriction of movement at adjacent
joints, bursitis and pressure on adjacent neurovascular
bundle [5]. In conclusion, this is a very unusual
presentation of a rare case of radiation-induced osteo-
chondroma of the right ilium in a 19-year-old female
secondary to radiotherapy, which had been performed
when the patient was 2 years old. Detection of mass
lesions in such young patients who are survivors of
childhood cancer causes significant anxiety and
psychological impact on the patients’ lives. The
current literature suggests very low incidence of
malignant degeneration in radiation-induced osteochon-
dromas, and awareness of this behaviour of these
lesions should help the clinicians to manage them
appropriately.
References
1. Paulino AC, Wen BC, Brown CK, Tannous R, Mayr NA,
Zehn WK, et al. Late effects in children treated with radiation
therapy for Wilm’s tumour. Int J Radiat Oncol Biol Phys
2000;46:1239–46.
2. Taitz J, Cohn RJ, White L, Russell SJ, Vowels MR.
Osteochondroma after total body irradiation: an age-
related complication. Pediatr Blood Cancer 2004;42:
225–9.
3. Libshitz HI, Cohen MA. Radiation-induced osteochondro-
mas. Radiology 1982;142:643–7.
4. DeSimone DP, Abdelwahab IF, Kenan S, Klein MJ, Lewis
MM. Radiation-induced osteochondroma of the ilium.
Skeletal Radiol 1993;22:289–91.
5. Lee KC, Davies AM, Cassar-Pullicino VN. Imaging the
complications of osteochondromas. Clin Radiol 2002;57:
18–28.
6. Mahboubi S, Dormans JP, D’Angio G. Malignant degenera-
tion of radiation-induced osteochondroma. Skeletal Radiol
1997;26:195–8.
The British Journal of Radiology, February 2006
The British Journal of Radiology, 79 (2006), 179
Book reviews
The Essential Guide to the New FRCR Part 1. By
T Jeswani, J Morlese. pp. 152, 2005 (Radcliffe
Medical Press Ltd, Abingdon, Oxon, UK) £19.95
ISBN 1-85775-616-9
As the title implies, this book is directed at candidates
preparing for the new FRCR Part 1 examination in
physics.
It is presented in six sections, each divided into sub-
sections, covering all aspects of the syllabus in logical
order. The questions are mostly multiple choice, with a
few requiring a short descriptive answer, and simulate
the type of question asked in the examination. In short it
is a ‘‘crammer’’.
Does the book fulfil its purpose? The short answer is
‘‘probably yes’’. There are almost 250 questions here,
with quite a lot of repetition in the MCQs, each with
answers and sometimes a short explanation. Enough of
the answers are correct for a candidate who absorbs the
book assiduously to pass the examination.
I have two main concerns. First, the authors were
extremely unwise not to involve a qualified medical
physicist working in diagnostic radiology as a co-author
or, at the very least, an advisor. As a consequence there
are many mistakes, ambiguities and even errors of
principle – I registered more than 60.
Examples of mistakes are (a) Thallium-67 (sic) (pg 88
A13); (b) how can the risk of inducing fetal cancer be esti-
mated if the dose to the mother is not stated? (pg 75 Q9).
Ambiguities arise because the scope of knowledge of
physics of the authors is clearly limited. For example (a)
‘‘the Compton attenuation of an X-ray beam is inversely
proportional to the beam energy’’ (pg 12 Q7) – not in the
diagnostic range it isn’t; (b) ‘‘the half value layer
increases with photon energy’’ (pg 10 Q4) – it decreases
at absorption edges, a most important practical pheno-
menon in diagnostic radiology.
There are also some major errors of principle, notably
in respect of quantum noise. Noise (strictly quantum
noise) is not inversely to the number of X-rays (strictly
the number of X-ray photons) (pg 117 Q20). It is directly
proportional to the square root of the number of photons.
This whole question, and others, confuse the concepts of
noise and signal to noise ratio and certainly will not help
the reader.
More worryingly, far too many (almost all) the questions
require purely factual recall. How many candidates will
remember, or indeed should need to remember, 10
minutes after the examination that Magnification 5 F/
(F2P), not (F+P)/(F2P)? If this over-emphasis on factual
recall reflects accurately the content of the examination,
and I suspect it does, this is a trend in the wrong direction.
MCQs can be broadly divided into four categories,
knowledge, understanding, application and extension of
knowledge. Physics is important for radiologists because
it helps them to understand the imaging process, the
complex inter-relationships between image quality and
patient dose, and the application of new technology to
their subject. Even this first phase of physics training
should encourage radiologists to think.
The British Journal of Radiology, February 2006
If we persist in an approach to teaching and examina-
tion that is heavily weighted to factual recall (there are
many parallels to the teaching of Latin here) the case for
removing physics from the syllabus may become over-
whelming. This would be a great loss. Radiologists
would have a poorer understanding of the imaging
process and their competence for maintaining the
intellectual high ground over their clinical colleagues in
respect of image interpretation would have been
seriously weakened.
P P DENDY
Imaging diseases of the chest (4th Edn). By D M Hansell,
P Armstrong, D A Lynch, H P McAdams. pp. ix+1220,
2005 (Elsevier Mosby, London, UK) £175.00
ISBN 0-323-03660-0
First, I must inform the reader that there is potential bias
in this review as, we had already agreed to purchase this
text (on my recommendation). It was therefore a
surprise, welcome but daunting, to review this fourth
edition of Imaging of Diseases of the Chest.
This book is designed to meet all the requirements
essential to the interpretation of the disease pathologies
in lung imaging. The text begins with chapters dealing
with technical factors and the anatomy of the thorax.
This is followed by various chapters describing the basic
patterns of lung disease, specific disease process, con-
genital anomalies and chest trauma.
All of the chapters are supplemented by various tables
and boxes providing summaries of information available
in the text, including clinical features, differential
diagnoses, diagnostic criteria and the radiological signs
(in all imaging modalities) relevant to individual
pathologies. The emphasis of this text (as described by
the authors) is to ensure detailed discussion of the more
complex and rare entities but also ensuring comprehen-
sive coverage of the more common conditions.
Each chapter is extensively referenced (record is
chapter on Lung Neoplasms with 1171) and has
extensive well annotated images (chest X-rays were of
variable quality). Although this book has in excess of
1200 pages, this was generally an easy and enjoyable text
to read (in bite size portions!). The authors have
managed to achieve a successful balance between fact,
respected opinion and clinical pragmatism. As expected
HRCT is comprehensively and logically discussed and I
particularly welcomed the chapter on chest trauma. The
index was also reader friendly giving helpful tips on
where else to look when appropriate.
There is no major fault to find with this book and it
would seem churlish to mention any minor inconsisten-
cies, given the dedication of the authors. However,
having said that there is an apparent odd typo, e.g.
Fig3.90 and Fig6.107.
This major work is an essential prerequisite for all
X-ray departments and I am very pleased to add this
fourth edition to the radiology library.
F GARDNER
179
BJR
The British Journal
of Radiology

March
2006
Volume 79
Issue 939
March 2006, Volume 79, Issue 939

● BJR Review of the Year – 2005

● New directions in ultrasound: microbubble contrast
● Enhanced biological effectiveness of low energy X-rays and
implications for the UK breast screening programme
● Comparison of image quality, diagnostic confidence and
interobserver variability in contrast enhancedMR angiography and
2D time of flight angiography in evaluation of carotid stenosis
● Comparison of Radiologists’ confidence in excluding significant
colorectal neoplasia with multidetector-row CT colonography
compared with double contrast barium enema
● The Bristol Hip View: a new hypothetical radiographic projection
for femoral neck fractures
● Visceral and testicular calcifications as part of the phenotype in
pseudoxanthoma elasticum: ultrasound findings in Belgian
patients and healthy carriers
● Life-threatening common carotid artery blowout: rescue
treatment with a newly designed self-expanding covered nitinol
stent
● Quantitative assessment of hip osteoarthritis based on image
texture analysis
● Trends in image quality in high magnification digital specimen
cabinet radiography
● Margins between clinical target volume and planning target
volume for electron beam therapy
● Gold nanoparticles: a new X-ray contrast agent
● Calculation of high-LET radiotherapy dose required for
compensation of overall treatment time extensions
● CT fluoroscopic guided insertion of inferior vena cava filters
● Gastric carcinoma presenting with extensive bone metastases
and marrow infiltration causing extradural spinal haemorrhage
● Diagnosis of myocardial contusion after blunt chest trauma using
18 F-FDG positron emission tomography
● Correspondence
● Book review
● An intranasal mass
● Book review
The British Journal of Radiology, 79 (2006), 183–187
EDITORIAL
BJR Review of the Year – 2005
Introduction
Following the successful launch of the ‘‘Review of the
Year’’ 12 months ago [1], the Editorial Board decided that
a similar review should be compiled for 2005. Once again
a selection of noteworthy papers on currently important
topics has been made by the Honorary and Deputy
Editors.
This has been a good year for the Journal with no
shortage of suitable papers from which to choose. One
of the reasons has been that since the Journal
introduced online submission towards the end of 2004,
there has been a sharp increase in the number of papers
received. Twice as many papers were submitted
during the first 6 months of 2005 as for the same period
in 2004.
This welcome shift has created some problems. The
Editorial Board has been expanded to deal with the
increased work-load and refereeing standards have been
tightened. In spite of this, the number of papers accepted
for publication each month has exceeded the number of
pages available in each issue and this has led to an
increasing interval between acceptance and publication.
We are currently looking at ways to address this
problem.
Another important development this year has been the
introduction of a Young Investigator Award. This Award
is made by the Honorary Editors for the best paper
submitted by an author, or first-named author, under
35 years of age at the time of submission. Only papers
that are accepted by both referees at the first submission,
either without change or with only minimal changes, are
eligible for consideration.
18 authors were eligible for the 2005 Award and
several of them reached the final short list. The Award is
made to Dr T Xiong for his paper entitled ‘‘Incidental
lesions found on CT colonography: their nature and
frequency’’ [2]. Compilation of this comprehensive,
systematic review required careful, painstaking
research and provides valuable data on the potential
benefits and pitfalls of CT colonography. It is always
enlightening to review published research and identify
at first hand some of the weaknesses in the published
data, despite the eminence of the authors and the
reputations of the journals in which they were published.
Hopefully the Award and the experience gained by
reviewing these data will stimulate Dr Xiong to
contribute more to the field of research. A summary of
the main findings of the paper appears in the next
section.
The British Journal of Radiology, March 2006
DOI: 10.1259/bjr/63430650
’ 2006 The British Institute of
Radiology
Diagnostic Radiology and the Young
Investigators Award
The difficulty of staying ‘‘up to date’’ in medicine is of
concern to most of us in practice, and this has been
reinforced this year in the Journal. Whilst our Young
Investigator Award winner Dr Xiong was reporting on
an analysis of published papers on CT colonography, just
2 months later Jardine et al [3] were perhaps showing us
the future of colonography in their report on the
potential problems using MR.
Xiong et al [2] identified additional abnormalities
reported in almost 40% of patients, with the total number
of abnormalities exceeding the number of patients
analysed. Furthermore, as would be expected, the
number of abnormalities and their potential significance
varied dependent on the age of the population studied,
and the definition of important abnormality. The
examination methodologies also varied, and it would
seem reasonable to expect the experts to agree a
standardized technique and reporting terminology for
the future. A more detailed discussion of the pros and
cons of CT colonography may be found in the commen-
tary by Ng and Freeman [4] that accompanied the
Award-winning paper. As a practising radiologist, the
significance and reliable detection of extracolonic disease
may be of critical importance in the future, regarding
decisions on who reports these scans. This is a topic of
increasing concern for UK based radiologists facing the
prospect of the ever increasing use of skill-mix.
Of the many other radiology articles published last
year in the BJR, the reader’s attention is drawn to the
article by Reuben et al [5]. This nicely conducted study
on the interpretation of facial trauma images by surgical
trainees, managed to justify the use of 3D reconstructions
and also suggested that there may still be a role for
conventional radiography. 17 trainee faciomaxillary
surgeons reviewed plain radiographs, conventional and
3D CT images of facial trauma patients. They were asked
to score their ease of interpretation of each type of image.
Their results were compared to the ‘‘subjective’’ gold
standard of a consultant radiologist and a faciomaxillary
surgeon. The trainee surgeons were best at interpreting
3D images, but were less good at interpreting conven-
tional axial CT images compared with plain radiographs.
Finally, in this section on diagnostic radiology, a
commentary by Munro [6] is noted. This article, although
written as a commentary on a radiotherapy paper, is a
‘‘must’’ for all those radiologists plying their trade in
research. This beautifully written piece, with quotations
from J K Galbraith and Francis Bacon, explains the need
183

for well-performed observational reporting, and will be a
fillip to all those who do not achieve the current research
nirvana of a randomized controlled trial.
Computer-aided diagnosis
With the increasing availability of digitized images,
computer-aided diagnosis (CAD) is currently a hot topic
and this was the subject of one of the two Special Issues
of the Journal in 2005. A good review article by Doi
entitled ‘‘Current status and future potential of compu-
ter-aided diagnosis in medical imaging’’ [7] was sup-
ported by six other contributions.
The basic concept is to provide a computer output as a
second opinion to assist image interpretation by radio-
logists by improving the accuracy and consistency of
radiological diagnosis and also by reducing the image
reading time. To achieve this goal it is necessary not only
to develop suitable algorithms but also to quantify and
maximize the effect of computer output on the perfor-
mance of radiologists. Research and development of
CAD has therefore involved a team effort by investi-
gators with different backgrounds – physicists, radiolo-
gists, computer scientists, engineers, psychologists and
statisticians.
The technique of receiver operator characteristic (ROC)
curve analysis can be very helpful in assessing the impact
of input from the computer. ROC methodology makes no
requirement of the positive/negative decision thresholds
of the observer, other than that they remain constant.
Results are readily understood ‘‘at a glance’’ – see, for
example, Figure 8 in the paper by Doi [7].
Potential clinical applications of CAD considered
included: detection of lung nodules on digital chest
radiographs; lung nodule detection based on morpho-
logy and sequential volume changes in CT images;
prompting techniques to assist in the reading of
mammograms and their impact; virtual colonoscopy as
a screening method for colorectal neoplasia.
The overall conclusion was that since CAD can be
applied to all imaging modalities, all body parts and all
kinds of examination, it is likely to have a major impact
on medical imaging and diagnostic radiology in the
21st century. However, further careful evaluation is
required and an awareness of on-going developments is
important.
Diagnostic ultrasound and ultrasound
measurement
A number of papers have been published during the
year on clinical applications of diagnostic ultrasound.
Some have illustrated the use of relatively new techni-
ques. For example Gorg et al [8] used quantitative colour
Doppler ultrasound to evaluate and characterize arterial
supply of chest wall lesions and Dietrich et al [9] used
microbubble contrast-enhanced phase inversion ultra-
sound to differentiate focal nodular hyperplasia and
hepatocellular adenoma.
A paper of particular clinical interest was ‘‘Ultrasound
evaluation of the fibrosis stage in chronic liver disease by
the simultaneous use of low and high frequency probes’’
184
Editorial
by Nishiura et al [10]. Those involved in ultrasound will
realise that for more than 25 years innumerable attempts
have been made at tissue characterization by ultrasound
and whilst this article does not claim to do such it seems
to give useful clinical help which may reduce the
number of invasive liver biopsies required. The authors
looked at 103 patients, examining them with two probes,
one at 2–5 MHz and another at 5–12 MHz. They used a
scoring system evaluating the edge, surface and par-
enchymal texture of the liver. They found that the high
frequency probe was more sensitive for identifying
mildly abnormal changes, whereas the low frequency
probe was more useful for scoring advanced changes.
The accumulated scores of the three parameters was the
most reliable with 100% sensitivity for fibrosis stage 4
and results which were almost as impressive in mild
disease. The authors do not indicate whether this has had
any impact as yet in reducing the number of liver
biopsies but the hope must be that it will. They are
intending to extend the technique to other hepatic
diseases including the potentially more difficult ones
seen in children. Let us hope this proves a fruitful field of
endeavour.
Given the continuing diversification of the application
of diagnostic ultrasound in medicine, the Editorial by
Francis Duck entitled ‘‘Ultrasound exposure measure-
ment: a hidden science?’’ [11] was timely. Under the dual
pressures of clinical users asking for improved perfor-
mance and manufacturers competing to gain commercial
edge, there is evidence from long-term studies of a
general trend towards increased acoustic output.
Doppler techniques frequently use outputs that come
close to the limits set by the thermal (TI) and mechanical
(MI) safety indices [12]. Collapse cavitation of micro-
bubbles, with potentially harmful effects in vivo, can
occur at high intensities. In spite of the above, current
hospital practice makes little or no attempt to evaluate
the ultrasound radiation generated by their equipment
and there is little investment in measurement tools or
commitment of personnel time to such measurements.
Duck makes two general recommendations. First there
is a need for appropriate acoustic measurement devices,
suitable for both laboratory-based and field measure-
ments. For example a well-engineered, portable, com-
mercial power meter capable of measuring acoustic
power down to 10 mW or less would make available a
direct and simple means to confirm the accuracy of the
displayed TI for imaging and to identify critical condi-
tions for transducer heating. Similarly an integrated,
easy-to-use portable package including hydrophone,
positioning, data acquisition and output would enable
on-line measurement of pressure wave-form, acoustic
frequency, MI, and temporal-averaged intensity, all key
quantities for exposure measurement.
Second, there needs to be a culture change in NHS
trusts and hospitals. Measurement of ionizing radiation
is now part of the scientific ethos of all NHS trusts, who
are committed to a regimen of measurement support that
is integrated within their risk management policies. A
similar commitment to ultrasound measurement is
required, thereby ensuring accountability and the ability
to respond robustly to the question ‘‘What evidence do
you have that your ultrasound scanners are safe to use
on patients?’’
The British Journal of Radiology, March 2006
Editorial
Radiobiology – low dose risk
The ‘‘holy grail’’ of radiation protection is ‘‘what
happens at low doses?’’ and the January issue opened
with a fascinating discussion of one of the longest
running debates in radiation biology: Is low dose
irradiation harmful or protective? [13–17]. This was not
simply a rehash of old information. It included the
presentation of previously unpublished data from a very
large cohort of US workers (28 000) on nuclear powered
ships over the period 1980–1988. Their death rates were
compared with a similar number of shipyard workers
who had not been exposed to abnormal radiation levels
and the results were intriguing. The death rate from
cancer of the occupationally exposed workers was
significantly lower than that of the controls and
dramatically lower for non-cancer deaths. This supple-
ments a significant body of epidemiological and radio-
biological data supporting the view that low doses may
indeed induce protective mechanisms including apopto-
sis and enhanced immune competence. Counter argu-
ments were also made and while they did not dispute the
evidence for a protective effect they relied on the
fundamental importance of the double strand break as
the dominant DNA lesion and its relationship to dose. It
is significant that the linear generation of double strand
breaks with dose is not incompatible with mechanisms
that enhance apoptosis and immune surveillance.
However, the undesirable implications of a change of
policy from assuming no safe dose of radiation to one
where low doses are desirable, was a compelling
argument for maintaining the status quo and the linear
no threshold hypothesis seems set to underpin the
legislative framework for the foreseeable future.
Notwithstanding, there is a genuine scientific debate
about mechanisms and risks at low doses, some of which
was aired in subsequent issues of the Journal [18–22].
There is a need for the scientific community to maintain a
truly open mind on this issue.
Another important outcome of low dose radiation
exposure is the delayed expression of chromosomal
aberrations as a consequence of genomic instability. This
phenomenon was first described 15 years ago [23] and
has since been associated with bystander effects [24, 25].
An interesting insight into these processes was reported
in the October issue [26]. The study showed that short
term repopulating cells in the bone marrow are capable
of fully repairing DNA damage, as manifest by chromo-
some aberrations, within a few cell generations, whereas
cells responsible for long term repopulation retain
damage in a form that can be expressed as aberrations
many months after irradiation. This reinforces the view
that radiation induced genomic instability is an early
event in multi-step carcinogenesis.
Patient doses and image quality
The subject of patient dosimetry – whether or not we
are confident about the risk magnitude – was promi-
nently discussed during the year. A screening pro-
gramme which involves the irradiation of asymptomatic
patients is always subject to particular scrutiny and a
comprehensive study of doses received in the UK breast
The British Journal of Radiology, March 2006
screening programme was reported by Young et al [27].
The radiation dose received by large breasts has been
reduced by the use of automatic beam quality selection
and large format film. However, this is offset to some
extent by an increase in the total dose per woman due to
the introduction of two view screening at every visit. The
authors call for a revised definition of the standard breast
used in the UK to reflect better the exposure factors and
doses received in clinical practice.
On the same subject, optically stimulated lumines-
cence was applied in a novel way for in vivo dose
measurements in mammography. The presence of the
small probes required did not significantly interfere with
the diagnostic quality of the images and good agreement
with ionization chamber dosimetry was reported [28]. A
further novel mammographic technique – near-infrared
optical transillumination spectroscopy – has been devel-
oped to determine physiological properties of the breast
tissue and thus hopefully to quantify differences
between women with low and high breast cancer risk
[29].
Because of the relatively high effective doses involved,
patient dose reduction in CT is a topical issue. Lewis and
Edyvean [30] considered the implications of multislice
CT scanners, whose ability to utilize long scan lengths
and narrow slices can lead to increased doses. Although
automatic exposure control and particularly for children
and smaller patients, tailoring of tube current to patient
size, can lead to dose reductions, the establishment of
acceptable levels of image quality for different examina-
tion types is the key to dose optimization.
The assessment of image quality raises many scientific
issues concerned with the perception and cognition of
images and these were outlined in a commentary by
Manning et al [31]. Two groups of factors which
influence the ability of the observer to interpret image
information are those which are image dependent and
relate to the visual conspicuity of relevant features and
those which are image independent and primarily
cognitive. The latter is particularly deserving of further
study. Further complications arise because of the
introduction of digital image technology, softcopy work-
stations and the phasing out of hardcopy images. There
are countless ways in which images may be processed
and manipulated and thus many opportunities to
optimize the link between the image and the visual
systems of the observers. But how can diagnostic
outcomes be measured and compared with hard copy?
Computer-aided detection tools may help here. A
comparison of full-field digital mammography and
film–screen mammography from the point of view of
image quality and lesion detection was reported this year
[32], demonstrating the superiority of the former.
We leave this subject by reflecting that after many
years of reporting patient dose measurements the point
has been reached where we need to think carefully
whether further data add to our overall understanding.
This means that the Journal will be more selective in the
studies accepted for publication and will favour work
which takes into account all aspects of the imaging
process (of which dose is only one component). As
Martin stated in his Editorial this year [33] ‘‘…finding the
appropriate level of image quality is the most important
185

objective. Keeping the dose low should always be
secondary’’.
Image quality, as noted above, is an often ill-defined
feature of the imaging process. In the November edition,
an international collaborative study [34] described a
comparison of conventional X-ray imaging of mice and
rabbit lungs with two types of phase-contrast imaging.
Phase contrast techniques highlight lung boundaries and
provide enhanced lung visibility compared with con-
ventional X-ray imaging, although the wider clinical
applications at present remain uncertain.
Procedures for the radiation protection of staff working
with X-rays are well understood, even if, at low doses, the
resultant risks are not. In general, the regulatory require-
ments for safe practice in this field are not too inhibiting.
Unfortunately, the same may not be the case for MRI, at
least in the European Union. Impending legislation on
electromagnetic field exposure to staff may cause serious
problems in the provision of clinical services, as gradient
field limits will inhibit work close to the magnet bore
during imaging. This has particularly serious conse-
quences for interventional MR procedures [35]. We hope
that the sparse evidence for deleterious effects will prompt
a future UKRC debate on the motion ‘‘There is no risk to
health at low current densities’’.
Radiotherapy and Oncology – adverse effects
of radiation treatment
This was a year for reminding us, if we needed
reminding, of the fundamental paradox that underlies
clinical radiotherapy. We treat a dread disease with an
intervention that is, itself, dangerous. Shakespeare was,
as so often, there before us:
‘‘……, and wish
To jump a body with a dangerous physic,
That’s sure of death without it’’
(Coriolanus: Act 3, Scene 1).
Two papers early in the year reminded us of the late
consequences of treatment [36, 37]. In the first Rowland
Payne et al [36] investigated the efficacy and tolerability of
the hyfrecator (Conmed, Utica, NY), a versatile office-
based electrosurgical instrument, as a treatment for
radiation-induced telangiectasia. The treatment was well
tolerated and generally resulted in a substantial reduction
in telangiectasia with a concomitant improvement
in quality of life. In the second, Power [37] reviewed
the proceedings of a meeting held at the BIR in May 2004 to
discuss issues regarding the recording and analysis of late
effects of radiotherapy treatments. The outcome was that
data on late effects are essential to assess the therapeutic
effect of treatment, but there is a need for international
consensus as to the best methods of data collection – which
should be validated, sensitive, reproducible and user-
friendly. In addition to these two papers, the BJR
Supplement on ‘‘Radiation-induced multi-organ involve-
ment and failure’’ [38] provided an excellent overview for
practising oncologists – keeping in mind that, with the
threat of terrorists bearing dirty bombs, we might be called
upon to deal with the casualties following the use of such a
weapon.
On the more hopeful side, the paper by Amemiya et al
[39] was reassuring, showing once again, that the
186
Editorial
benefits offered by radiotherapy – treatment of early
squamous cell carcinoma of the head and neck in this
work – usually far outweigh the disadvantages (radia-
tion-induced second cancers in this study). Cominos et al
[40] showed that radiotherapeutic technique is important
in avoiding the adverse consequences of treatment.
Using a four-field technique for treating oesophageal
cancer significantly reduced the dose to the heart.
Finally, in a careful meta-analysis of clinical trials
Dayes et al [41] showed that there was no evidence to
support the use of nitroimidazoles as hypoxic cell
sensitizers in patients treated with radiotherapy for
carcinoma of the cervix but there was a significant
increase in the rate of neurotoxicity.
In view of the severe, and largely unexpected adverse
late effects of neutrons, two papers by Jones and others
on particle therapy deserve mention [42, 43]. The role
that particle therapy can, and should, play in the
management of cancer is a matter of some debate. We
hope that, in the coming months, the BJR will make a
useful and authoritative contribution to this debate. We
need informed discussions, based on facts. We cannot
afford to let the agenda be hijacked for political or
commercial reasons. To do so would be to fail in our
scientific obligations and more importantly, to fail in our
duty of care to patients with cancer.
References
1. BJR Review of the Year – 2004. Br J Radiol 2005;78:181–5.
2. Xiong T, Richardson M, Woodroffe R, Halligan S, Morton
D, Lilford RJ. Incidental lesions found on CT colonography:
their nature and frequency. Br J Radiol 2005;78:22–9.
3. Jardine VL, Sala E, Lomas DJ. MR colonography: baseline
appearance of the unprepared rectosigmoid. Br J Radiol
2005;78:202–6.
4. Ng CS, Freeman AH. Incidental lesions found on CT
colonography: their nature and frequency. Br J Radiol
2005;78:20–1.
5. Reuben AD, Watt-Smith SR, Dobson D, Golding SJ. A
comparative study of evaluation of radiographs, CT and 3D
reformatted CT in facial trauma: what is the role of 3D? Br J
Radiol 2005;78:198–201.
6. Munro AJ. The conventional wisdom and the activities of
the middle range. Br J Radiol 2005;78:381–3.
7. Doi K. Current status and future potential of computer-
aided diagnosis in medical imaging. Br J Radiol
2005;78:S3–S19.
8. Görg C, Bert T, Görg K, Heinzel-Gutenbrunner M. Colour
Doppler ultrasound mapping of chest wall lesions. Br J
Radiol 2005;78:303–7.
9. Dietrich CF, Schuessler G, Trojan J, Fellbaum C, Ignee A.
Differentiation of focal nodular hyperplasia and hepatocel-
lular adenoma by contrast-enhanced ultrasound. Br J Radiol
2005;78:704–7.
10. Nishiura T, Watanabe H, Ito M, Matsuoka Y, Yano K,
Daikoku M, et al. Ultrasound evaluation of the fibrosis
stage in chronic liver disease by the simultaneous use of
low and high frequency probes. Br J Radiol 2005;78:189–97.
11. Duck FA. Ultrasound exposure measurement: a hidden
science? Br J Radiol 2005;78:289–91.
12. IEC Standard 60601-2-37. Medical electrical equipment –
particular requirements for the safety of ultrasound medical
diagnostic and monitoring equipment. Geneva:
International Electrotechnical Commission, 2002.
13. Dendy P. Low dose radiation risk: UKRC 2004 debate. Br J
Radiol 2005;78:1–2.
The British Journal of Radiology, March 2006
Editorial
14. Feinendegen LE. Evidence for beneficial low level radiation
effects and radiation hormesis. Br J Radiol 2005;78:3–7.
15. Chadwick KH, Leenhouts HP. Radiation risk is linear with
dose at low doses. Br J Radiol 2005;78:8–10.
16. Cameron JR. Moderate dose rate ionizing radiation
increases longevity. Br J Radiol 2005;78:11–3.
17. Martin CJ. The LNT model provides the best approach for
practical implementation of radiation protection. Br J Radiol
2005;78:14–6.
18. Kendall GM, Muirhead CR. Moderate dose rate ionizing
radiation increases longevity? Br J Radiol 2005;78:573–4.
19. Edwards A, Bouffler S. Abandoning linear no threshold. Br
J Radiol 2005;78:770–1.
20. Feinendegen LE. Author reply. Br J Radiol 2005;78:773.
21. Doll R, Berrington A, Darby SC. Low mortality of British
radiologists. Br J Radiol 2005;78:1057–8.
22. Tubiana M. Linear no-threshold dogma. Br J Radiol
2005;78:1060.
23. Pampfer S, Streffer C. Increased chromosome aberration
levels in cells from mouse fetuses after zygote X-irradiation.
Int J Radiat Biol 1989;55:85–92.
24. Wright EG. Commentary on radiation-induced bystander
effects. Hum Exp Toxicol 2004;23:91–4.
25. Mothersill C, Seymour C. Radiation-induced bystander
effects and adaptive responses–the Yin and Yang of low
dose radiobiology? Mutat Res 2004;568:121–8.
26. Uma Devi P, Satyamitra M. Tracing radiation induced
genomic instability in vivo in the haemopoietic cells from
fetus to adult mouse. Br J Radiol 2005;78:928–33.
27. Young KC, Burch A, Oduko JM. Radiation doses received in
the UK Breast Screening Programme in 2001 and 2002. Br J
Radiol 2005;78:207–18.
28. Aznar MC, Hemdal B, Medin J, Marckmann CJ, Andersen
CE, Bøtter-Jensen L, et al. In vivo absorbed dose measure-
ments in mammography using a new real-time lumines-
cence technique. Br J Radiol 2005;78:328–34.
29. Simick MK, Lilge L. Optical transillumination spectroscopy
to quantify parenchymal tissue density: an indicator for
breast cancer risk. Br J Radiol 2005;78:1009–17.
30. Lewis MA, Edyvean S. Patient dose reduction in CT. Br J
Radiol 2005;78:880–3.
The British Journal of Radiology, March 2006
31. Manning DJ, Gale A, Krupinski EA. Perception research in
medical imaging. Br J Radiol 2005;78:683–5.
32. Fischmann A, Siegmann KC, Wersebe A, Claussen CD,
Müller-Schimpfle M. Comparison of full-field digital
mammography and film-screen mammography: image
quality and lesion detection. Br J Radiol 2005;78:312–5.
33. Martin CJ. 20 years of patient dose studies: where should
we go from here? Br J Radiol 2005;78:477–9.
34. Kitchen MJ, Lewis RA, Yagi N, Uesugi K, Paganin D,
Hooper SB, et al. Phase contrast X-ray imaging of mice and
rabbit lungs: a comparative study. Br J Radiol
2005;78:1018–27.
35. Keevil SF, Gedroyc W, Gowland P, Hill DLG, Leach MO,
Ludman CN, et al. Electromagnetic field exposure limita-
tion and the future of MRI. Br J Radiol 2005;78:973.
36. Rowland Payne CME, Somaiah N, Neal AJ, Glees JP. The
hyfrecator: a treatment for radiation induced telangiectasia
in breast cancer patients. Br J Radiol 2005;78:143–6.
37. Power DA. Late effects of radiotherapy: how to assess and
improve outcomes. Br J Radiol 2005;78:150–2.
38. Fliedner TM, Meineke V, editors. Radiation-induced multi-
organ involvement and failure. Br J Radiol Supplement 27,
2005.
39. Amemiya K, Shibuya H, Yoshimura R, Okada N. The risk of
radiation-induced cancer in patients with squamous cell
carcinoma of the head and neck and its results of treatment.
Br J Radiol 2005;78:1028–33.
40. Cominos M, Mosleh-Shirazi MA, Tait D, Henrys A, Cornes
P. Quantification and reduction of cardiac dose in radical
radiotherapy for oesophageal cancer. Br J Radiol
2005;78:1069–74.
41. Dayes IS, Abuzallouf S. Local tumour control in women
with carcinoma of the cervix treated with the addition of
nitroimidazole agents to radiotherapy: a meta-analysis. Br J
Radiol 2005;78:777–82.
42. Jones B, Rosenberg I. Particle Therapy Co-operative
Oncology Group (PTCOG 40) Meeting, Institute Curie
2004. Br J Radiol 2005;78:99–102.
43. Jones B, Price P, Burnet NG, Roberts JT. Modelling the
expected increase in demand for particle radiotherapy:
implications for the UK. Br J Radiol 2005;78:832–5.
187

COMMENTARY
New directions in ultrasound: microbubble contrast
V R STEWART, MRCP, FRCR and P S SIDHU, MRCP, FRCR
Department of Radiology, King’s College Hospital, Denmark Hill, London SE5 9RS, UK
Ultrasound is the most frequently performed ‘‘tomo-
graphic’’ imaging technique which befits the perception
of ultrasound as being a low cost, safe and accessible
imaging modality. Widely used as a ‘‘screening’’ tool,
particularly in liver imaging, ultrasound is often per-
ceived as inferior to other imaging modalities such as CT
and MRI in diagnostically challenging clinical situations.
Both CT and MR imaging techniques make use of well
established contrast agents to improve the image and
hence the diagnostic potential, but with the burden of
increased radiation with CT and cost with both CT and
MRI. With such widespread utilization of ultrasound, it
remains an enigma that until recently ultrasound had no
effective contrast agent to improve imaging. The advent
of microbubble contrast has brought new possibilities
and, not surprisingly, recent advancement of ultrasound
has been driven by research into the properties and
clinical application of microbubble contrast agents.
An overview of the physical properties of microbubble
agents and the interaction of the microbubbles with
sound waves that produce image contrast improvement
is presented in this commentary. In addition, current
established clinical use is detailed and areas of potential
utilization are discussed.
Microbubble contrast properties
Ultrasound contrast as a concept was first observed in
cardiology practice, rather fortuitously, when it was noted
on echocardiography that small air bubbles surrounding a
catheter tip placed in the left ventricle during cardiac
catheterization produced transient high reflections [1].
Over the decades research into producing ‘‘bubbles’’ for
use as an ultrasound contrast agent was hampered by the
need to produce bubbles that were stable in the circulation,
traversed the pulmonary circulation to allow recirculation
and inert to the recipient. Technological advances over
the last 20 years have allowed microbubbles with the
Address correspondence to Dr Paul S Sidhu
188
The British Journal of Radiology, 79 (2006), 188–194
Received 19 July 2005
Revised 23 September
2005
Accepted 5 October 2005
DOI: 10.1259/bjr/17790547
’ 2006 The British Institute of
Radiology
necessary characteristics to be developed and, impor-
tantly, to be diagnostically useful [2].
In order to achieve transpulmonary recirculation and to
be an effective contrast agent, the microbubbles need to
pass through the smallest vascular component, the
capillary system, intact. The ideal diameter for this to
occur is between 2 mm and 8 mm, below that of red blood
cells. Enhancement life-time of the microbubble, often
several minutes in the circulation, is a manifestation of
microbubble design. Microbubble stability is increased by
external bubble encapsulation (galactose, phospholipids,
denatured albumin or poly-butyl-cyanoacrylate) with or
without surfactants and using gases with a low diffusion
coefficient (perfluorocarbons) or a combination of both [3].
The gas components of the microbubbles are normally
eliminated via the lungs whilst the stabilizing components
are eliminated via the hepato-renal route [4].
Currently the agents in clinical use are LevovistH
(Schering AG, Berlin, Germany; air with a galactose/
palmitic acid surfactant), SonoVueH (Bracco SpA, Milan,
Italy; sulphur hexachloride with a phospholipid shell),
OptisonH (Nycomed/Amersham, Oslo, Norway; octa-
fluoropropane with an albumin shell), ImagentH
(Alliance Pharmaceutical, CA; perflexane lipid micro-
sphere) and DefinityH (Bristol-Meyers-Squibb, NY; octa-
fluoropropane with a lipid shell). A newer agent is
CARDIOsphereH (Point Biomedical, CA; bilayer poly-
mer/albumin shell containing nitrogen) developed for
use in cardiology.
Ultrasound techniques to exploit microbubble
contrast properties
Microbubbles behave as echo enhancers, by expanding
and contracting to create backscatter, on exposure to an
ultrasound beam of any frequency. Microbubbles perform
this task supremely well and increase the backscatter by
.300 fold. Resonance of the microbubble will occur when
there is a specific relationship between the bubble size
and the ultrasound frequency (about 3 MHz). At low
The British Journal of Radiology, March 2006
Commentary: Microbubble contrast

ultrasound beam power the expansion and contraction is
symmetrical, therefore the bubbles oscillate in a ‘‘linear’’
fashion and the frequency of the scattered signal is
unaltered. At higher power the microbubbles behave in
‘‘non-linear’’ fashion as they resist contraction under
positive pressures more than expansion under negative
pressures. The ‘‘non-linear’’ response results in emission of
harmonics which are specific to the microbubbles. These
harmonics occur within the range 1–20 MHz. The micro-
bubble vibration may include both higher harmonics of the
fundamental ultrasound frequency (2f, 3f, 4f etc.) and sub-
harmonics (mostly f/2). The lower limit of 1 MHz relates to
sub-harmonic generation when operating at 2 MHz. The
upper limit is arbitrary, and could in principle extend
much higher than 20 MHz. In practice the upper limit is
constrained by the working bandwidth of the transducer.
The bandwidth extends over a sufficient range of frequen-
cies to enable the generated harmonics to be detected. This
is typically not broad enough to detect more than the
second harmonic (and sub-harmonics if needed), although
other harmonics certainly exist in the backscattered signal.
However, this will enable preferential imaging of micro-
bubbles compared with the surrounding tissues.
Further increases in pressure cause the microbubbles to
burst resulting in a strong non-linear echo, but this effect is
transient and no further diagnostic information can be
obtained until there is reperfusion of the area by intact
microbubbles [2, 5, 6]. By imaging with a low mechanical
index (MI) that allows for a non-linear response the
amount of microbubble destruction is minimized, prolong-
ing the effective period for diagnostic imaging. The MI
(scaled by pulse amplitude and calculated from the peak
rarefaction acoustic pressure and centre frequency) was
conceived as a safety indicator of the potential for
cavitation. It has been found useful as an approximate
indicator to distinguish between high MI and low MI
regimens of microbubble contrast use, although there are
quantitative deficiencies for this application.
To process the resultant signal from the microbubbles,
new techniques are necessary which selectively display
the non-linear response from the contrast microbubbles
preferentially. Pulse inversion harmonic imaging relies
on the different behaviour of microbubbles exposed to
consecutive pulses of inverted phase; linear signals from
normal tissue cancel out whilst non-linear signals from
microbubbles summate to produce an image [7]. Pulse
inversion harmonic imaging requires the use of a
broader transmit and receive bandwidth [8].
Another phenomenon observed with certain microbub-
ble contrast agents (LevovistH and SonazoidH, an agent not
licensed) is the display of a late delayed phase in the liver,
with signal displayed from stationary microbubbles.
Uncertainty surrounds the exact reason for the persistence
of microbubbles in the liver (and in the spleen, where
SonoVueH also demonstrates this phenomenon [9]); spec-
ulation is that the microbubbles are trapped in the liver
sinusoids [10] or actively taken up by the reticuloendothe-
lial system [11]. This phase occurs at approximately 2 min
and lasts for a variable period of time; about 3 further
minutes with LevovistH and is best imaged with a
‘‘destructive’’ mode using high machine power with
velocity 2D colour Doppler. This method, known as
stimulated acoustic emission (or loss of correlation mode),
results in a transient colour mosaic in liver tissue contain-
ing normal cells and a ‘‘black-hole’’ in malignant tissue
containing no normal liver cells [12, 13]. This method of
imaging microbubble contrast in the liver, excellent for
detecting the presence of liver metastasis, is less favoured
by radiologists in comparison with low mechanical MI
techniques (Figure 1).
Clinical applications of microbubble contrast
All the licensed microbubble agents are injected
intravenously and do not cross cell membranes, remaining
in the intravascular compartment, a distinct difference
from other radiological contrast media. Microbubbles
therefore give information on the vascularity and enhance-
ment characteristics of a tissue rather than the functional
properties; application is directed towards this unique
feature. As microbubble contrast can be delivered under

(a) (b)

Figure 1. (a) A B-mode image of the right lobe of the liver obtained at 3 min 6 s following the administration of LevovistH.
There is an indeterminate heterogeneous area (between arrows) that is poorly defined but suspicious of malignancy in a patient
with a known primary tumour outside the liver. (b) Using a late-phase destructive mode (Agent Detection Imaging, ADIH;
Siemens, Mountain View, CA), there is transient destruction of the microbubble contrast agent in normal liver tissue, but
absence of microbubble contrast in the tumour appears as ‘‘two-black holes’’ (arrow).

The British Journal of Radiology, March 2006 189

 V R Stewart and P S Sidhu

(a) (b)

(c)

Figure 2. (a) A well demarcated tumour in the right lobe of the liver (arrow) on B-mode imaging. (b) Image of the same tumour
obtained at 18 s (arterial phase) following the administration of SonoVueH, and using a low mechanical index imaging
technique (Cadence Contrast Pulse Sequencing, CPSH; Siemens, Mountain View, CA), demonstrates prominent arterial signal
within the tumour. (c) Image of the same tumour at 60 s (portal-venous phase) demonstrating complete in-filling; the
appearances are in keeping with a benign tumour and representative of an area of focal nodular hyperplasia.

real time ultrasound observation it may provide additional
information to CT and MRI.
Established clinical applications
The first applications for the use of microbubble
contrast were cardiac, where there are established
clinical practices; these applications are outside the remit
of the current review and will not be discussed.
Microbubble contrast has been widely used in imaging
of solid organs, particularly the liver where it has a
number of established applications. The original applica-
tion for these agents was in ‘‘Doppler rescue’’, with
improvement in detection of colour Doppler signal from
large vessels, particularly the portal vein and hepatic
artery in transplantation [14–16] and in documentation of
abnormal vessels in liver tumours [17]. With the advent
of low MI imaging coupled with pulse inversion
techniques, liver tumour imaging is now relatively
sophisticated, precipitating a consensus publication of
guidelines for identifying contrast enhancement patterns
in various focal liver tumours [18].
In lesions where there are distinctive enhancement
patterns, microbubble contrast enables accurate characteri-
zation of lesions so that more expensive, time-consuming
examinations do not need to be performed. Recent
studies have demonstrated characterization of liver
lesions to be accurate in 85–96% of cases in distinguishing
benign from malignant lesions [13, 19, 20]. Benign lesions
tend to enhance in the arterial phase and retain micro-
bubble contrast through the different vascular phases
(arterial 10–35 s, early portal-venous 30–120 s and late
portal-venous phases .120 s after administration)
(Figure 2) [20].

190 The British Journal of Radiology, March 2006

Commentary: Microbubble contrast
Benign lesions often have characteristic enhancement
patterns, such as peripheral nodular enhancement in
haemangioma and homogeneous arterial enhancement
with a central ‘‘spoke wheel’’ arterial pattern in focal
nodular hyperplasia [20]. Metastases have variable
enhancement patterns and may be hypovascular or
hypervascular on the arterial phase following micro-
bubble contrast administration, often displaying periph-
eral rim enhancement. On the portal-venous phase
images, the enhancement fades and the metastases
become of decreased reflectivity compared with normal
hepatic parenchyma [19]. This appearance is accentuated
by those microbubble contrast agents which display the
late delayed phase of imaging in the liver when, at
between 2 min and 5 min, increased conspicuity of focal
lesions against the enhancing normal liver tissue is
observed with a ‘‘destructive’’ mode using high machine
power [13, 19]. However, confusion may arise if imaging
is performed in the late delayed phase only. Multiple
hepatic abscesses [21] or the rare biliary hamartomas [22]
may present as focal areas of low reflectivity, mimicking
metastases on a microbubble contrast ultrasound study;
the only difference being a complete absence of vessels
centrally in these two conditions if imaged with low MI
techniques through all the vascular phases.
The application of intraoperative ultrasound during
surgery for resection of metastases identifies metastases
that were not seen on any form of pre-operative imaging,
and changes management in 50% of cases [23]. Any
further improvement on this would be useful; prelimin-
ary results suggest that microbubble contrast ultrasound
demonstrates increased sensitivity and a capability of
detecting lesions as small as 2–3 mm in diameter
allowing improved outcome of patients undergoing
‘‘curative’’ metastasis resection [23, 24].
A number of ablation treatments (including ethanol,
cryotherapy, high-intensity focused ultrasound and
radiofrequency ablation) are employed in the manage-
ment of malignant disease within the liver when the
patient is not suitable for surgical resection or transplan-
tation. Currently, radiofrequency ablation is receiving
the most attention. In most situations ultrasound is the
modality of choice for implementing this therapy,
predominantly as it allows real time visualization of
electrode placement. The outcome of radiofrequency
ablation is dependent on attaining a successful ‘‘tumour-
free’’ margin and complete necrosis of the tumour itself
[25]. Performing biphasic CT or contrast enhanced MRI
periprocedure is relatively impractical in delineating this
margin, but microbubble contrast readily demonstrates
residual tumour enhancement. Ablation therapy fol-
lowed by imaging 10 min post-procedure will demon-
strate residual tumour as an irregular margin that
maintains the enhancement pattern seen prior to ablative
therapy, different to the rim of enhancement seen post-
ablation on CT thought to represent reactive hyperaemia
[26]. If performed following ablation, microbubble
contrast allows immediate further therapy if required,
decreasing the number of treatment sessions.
Microbubble contrast originally developed for Doppler
rescue remains invaluable in demonstrating vessel patency,
firmly established in such diverse areas as transcranial
Doppler, echocardiography, liver transplantation and in
the diagnosis of renal artery stenosis [5, 6, 16, 27, 28].
The British Journal of Radiology, March 2006
Microbubble contrast has also found a niche outside
the vascular compartment in the setting of vesico-
ureteric reflux in children where a high sensitivity and
specificity compared with conventional micturating
cystourography (MCUG) has been demonstrated [29].
In a study comparing conventional MCUG with a
microbubble contrast examination, a significant number
of children were up-graded from a grade 1 reflux on
MCUG to a grade 2; with management and prognostic
implications [30]. The advantage of avoiding ionizing
radiation is obvious, although the procedure remains
invasive.
Potential clinical applications
Whilst most applications for ultrasound contrast are
established in the liver, further uses are developing in
other areas.
A number of groups have investigated the utility of
microbubble contrast as an adjuvant to the FAST scan
(Focused Assessment Sonography in Trauma) in blunt
abdominal trauma [31–34]. Non-enhanced FAST scan-
ning is able to ‘‘triage’’ patients with blunt abdominal
trauma accurately; patients with negative imaging
virtually never need surgical intervention [35]. The
addition of microbubble contrast to the examination
would increase the confidence of the operator in the face
of a negative examination. The most likely role of
microbubble contrast in blunt abdominal trauma would
be the ability to assess patients more accurately in order
to expedite the most appropriate management whether
this is surgery, further imaging with CT or observation
alone.
The role of ultrasound, colour Doppler ultrasound and
microbubble contrast ultrasound in detecting breast
carcinomas is yet to be fully established. There is a
suggested role for colour Doppler ultrasound in the
differential diagnosis of breast disease [36, 37]. Studies
have demonstrated an increased sensitivity in vascular-
ity with microbubble contrast, but with conflicting views
on the specificity of differentiating benign and malignant
lesions [38, 39]. These studies were performed with
conventional machine settings, but as the newer harmo-
nic and phase inversion techniques develop, analysis of
breast masses with microbubble contrast may become a
useful tool. A study used microbubble contrast in the
evaluation of radiofrequency ablation in breast tumours
with some success, a similar use as with radiofrequency
ablation of liver tumours [40].
One of the goals of treatment of any cancer is the
identification of disease involvement of the sentinel node
– the first node to drain a tumour into the lymphatic
system. This predicts the need to remove the regional
lymph nodes. Microbubble contrast may play a role in
this scenario; sentinel nodes in swine models with
melanoma demonstrated sentinel node enhancement in
28 of 31 sentinel lymph nodes, some within seconds of
peritumour injection of microbubble contrast. The
authors, using low MI greyscale pulse-inversion imag-
ing, also demonstrated signal voids within the lymph
nodes representative of intranodal metastasis with 95%
sensitivity [41]. This method of sentinel node detection is
as good as alternative techniques without the adverse
191

effects of these established techniques; blue dye is
invasive, has a relatively high rate of allergic reaction
and a technical failure rate of 20%, whereas technetium
99m scintigraphy has a reported failure rate of 12% [42].
Both these techniques may detect non-sentinel nodes
(false-positive), leading to unnecessarily extensive nodal
dissection. Another group have successfully developed a
specific microbubble that targets lymph nodes, using the
stimulated acoustic emission ultrasound imaging
method [43]. Further studies are required to fully
evaluate these techniques with application, particularly
within the axilla of breast cancer patients, an important
potential clinical use [44].
Another interesting area for clinical use of microbub-
ble contrast is in musculoskeletal ultrasound for the
demonstration of synovitis. MRI of joints, although
informative and accurate, is not readily accessible to
provide an ‘‘on-demand’’ clinical service that patients
with inflammatory synovitis require for rapid diagnosis
and disease management. More pertinent is the better
resolution capability of ultrasound in comparison with
MRI especially in the smaller joints. Ultrasound can
accurately differentiate between joint fluid and syno-
vium, except in the presence of echogenic joint fluid,
when the addition of microbubble contrast may help
[45]. The addition of microbubble contrast to the
ultrasound examination of the synovium will demon-
strate ongoing or recurrence of inflammation by asses-
sing the increase of vascular enhancement; which may be
of promise in the small joints of the hand [46].
A further important clinical use for microbubble
contrast is in vascular ultrasound, applicable in parti-
cular to the carotid circulation in the management of
cerebrovascular disease. Numerous studies have estab-
lished the importance of assessing the degree of
narrowing of the internal carotid artery in relation to
symptoms in order to ascertain the need for surgical or
increasingly radiological intervention [47, 48].
Ultrasound is highly accurate in assessing the degree of
stenosis of the internal carotid artery, far more cost-
effective than other imaging modalities and much more
‘‘patient-friendly’’. However, there remain instances of
ultrasound limitation. Ordinary colour Doppler micro-
bubble contrast enhanced examinations are problematic
as a consequence of artefacts, most notably ‘‘blooming’’
[49]. Technical advances with high frequency linear
transducers, coupled with the newer harmonic imaging
techniques, has allowed improvements in lumen deli-
neation without the need to use colour Doppler ultra-
sound. The images produced are likened to ‘‘ultrasound
angiograms’’ as they clearly display the outer and inner
luminal margins of the vessel allowing precise assess-
ment of intima-media thickening, atheromatous plaques,
ulceration and areas of marked stenosis [50].
The use of microbubble contrast in gene therapy and
targeted delivery of drugs is an area of active research,
where microbubbles are engineered to carry antibodies
or DNA to target tissues [51]. With gene therapy, a
particular area that shows promise is skeletal muscle [52,
53]. Ultrasound enhances gene transfer by increasing cell
permeability, termed ‘‘sonoporation’’ a process enabled
by microbubble contrast, believed to occur by lowering
the threshold for ultrasound bioeffects [54]. Interestingly
the type of microbubble may influence the rate of gene
192
V R Stewart and P S Sidhu
transfection, with the perfluorocarbon microbubbles
(OptisonH in this study) the most efficient [55]. There is
even a suggestion that perfluorocarbon microbubbles
may promote gene transfection without the need for
ultrasound [56].
Safety
Microbubble contrast agents approved for clinical use
are well tolerated with serious side-effects rarely
observed, predominantly minor in nature (headache,
nausea) which are invariably self limiting [57].
Generalized allergy-like reactions occur rarely [58].
There is the possibility of bioeffects arising from the
use of microbubble agents; microvascular rupture can
occur where gas bodies are insonated [18]. This may be
problematic in areas of sensitivity such as the retina and
the brain when imaged through the open fontanelle. A
further concern is the development of premature
ventricular contractions when high MI end systolic
triggering is specifically used in echocardiography but
not with other applications [59].
Conclusion
Ultrasound microbubble contrast already has estab-
lished uses in the liver as a Doppler rescue agent and
further applications are constantly being developed. It is
likely that administration will eventually become routine
in day to day practice as is the situation in a number of
European countries (Italy, Germany and Spain) and
Japan (soon to be followed by China) where the
microbubble agents are licensed and there is enthusiasm
among the ‘‘imagers’’. The UK has been slower in the
uptake of using microbubble contrast for a number of
reasons [60]. However, the imagers in the USA can only
admire from a distance the advances made in the clinical
application of microbubble contrast agents by their
European and Asian counterparts; only limited off-
licence use of these agents for abdominal imaging is
endorsed.
References
1. Gramiak R, Shah PM. Echocardiography of the aortic root.
Invest Radiol 1968;3:356–66.
2. Harvey CJ, Blomley MJK, Eckersley RJ, Cosgrove DO.
Developments in ultrasound contrast media. Eur Radiol
2001;11:675–89.
3. Correas JM, Bridal L, Lesavre A, Mejean A, Claudon M,
Helenon O. Ultrasound contrast agents: properties, princi-
ples of action, tolerance, and artifacts. Eur Radiol
2001;11:1316–28.
4. Greis C. Technology overview: SonoVue. Eur Radiol Suppl
2004;14:P11–P15.
5. Blomley MJK, Cooke JC, Unger EC, Monaghan MJ,
Cosgrove DO. Microbubble contrast agents; a new era in
ultrasound. BMJ 2001;322:1222–5.
6. Leen E. Ultrasound contrast harmonic imaging of abdom-
inal organs. Semin Ultrasound CT MRI 2001;22:11–24.
7. Simpson DH, Chin CT, Burns PN. Pulse inversion Doppler:
a new method for detecting nonlinear echoes from
microbubble contrast agents. IEEE Trans Biomed Engin
1999;46:372–82.
The British Journal of Radiology, March 2006
Commentary: Microbubble contrast
8. Whittingham TA. Broadband transducers. Eur Radiol
1999;9:S298–S303.
9. Lim AKP, Patel N, Eckersley RJ, Taylor-Robinson SD,
Cosgrove DO, Blomley MJK. Evidence for spleen-specific
uptake of a microbubble contrast agent: a quantitative
study in healthy volunteers. Radiology 2004;231:785–8.
10. Hauff P, Fritzsch T, Reinhardt M, Weitschies W, Luders F,
Uhlendorf V, et al. Delineation of experimental liver tumors
in rabbits by a new ultrasound contrast agent and
stimulated acoustic emission. Invest Radiol 1997;32:95–9.
11. Forsberg F, Goldberg BB, Liu JB, Merton DA, Rawool NM,
Shi WT. Tissue-specific US contrast agent for evaluation of
hepatic and splenic parenchyma. Radiology
1999;210:125–32.
12. Blomley MJK, Sidhu PS, Cosgrove DO, Albrecht T, Harvey
CJ, Heckemann RA, et al. Different types of liver lesions
differ in their uptake of the microbubble contrast agent SH
U 508A in the late liver phase? Early experience. Radiology
2001;220:661–7.
13. Bryant TH, Blomley MJK, Albrecht T, Sidhu PS, Leen E,
Basilico R, et al. Liver phase uptake of a liver specific
microbubble improves characterization of liver lesions: a
prospective multi-center study. Radiology 2004;232:392–9.
14. Marshall MM, Beese RC, Muiesan P, Sarma DI, O’Grady J,
Sidhu PS. Assessment of portal venous patency in the liver
transplant candidate: a prospective study comparing ultra-
sound, microbubble-enhanced colour Doppler ultrasound
with arteriography and surgery. Clin Radiol 2002;57:377–83.
15. Gebel M, Caselitz M, Bowen-Davies PE, Weber S. A
multicenter, prospective, open label, randomized, con-
trolled phase IIIb study of SH U 508A (Levovist) for
Doppler signal enhancement in the portal vascular system.
Ultraschall Med 1998;19:148–56.
16. Sidhu PS, Shaw AS, Ellis SM, Karani JB, Ryan SM.
Microbubble ultrasound contrast in the assessment of
hepatic artery patency following liver transplantation: role
in reducing frequency of hepatic artery arteriography. Eur
Radiol 2004;14:21–30.
17. Tano S, Ueno N, Tomiyama T, Kimura K. Possibility of
differentiating small hyperechoic liver tumours using
contrast-enhanced colour Doppler ultrasonography: a pre-
liminary study. Clin Radiol 1997;52:41–5.
18. Albrecht T, Blomley MJK, Bolondi L, Claudon M, Correas
JM, Cosgrove DO, et al. Guidelines for the use of contrast
agents in ultrasound. January 2004. Ultraschall Med
2004;25:249–56.
19. Albrecht T, Hohmann J, Oldenburg A, Skork J, Wolf KJ.
Detection and characterisation of liver metastases. Eur
Radiol Suppl 2004;14:P25–P33.
20. Bleuzen A, Tranquart F. Incidental liver lesions: diagnostic
value of cadence contrast pulse sequencing (CPS) and
SonoVue. Eur Radiol Suppl 2004;14:P53–P62.
21. Kim KW, Choi BI, Park SH, Kim AY, Koh YH, Lee HJ, et al.
Pyogenic hepatic abscesses: distinctive features from
hypovascular hepatic malignancies on contrast-enhanced
ultrasound with SH U 508A; early experience. Ultrasound
Med Biol 2004;30:725–33.
22. Berry JD, Boxer ME, Rashid HI, Sidhu PS. Microbubble
contrast enhanced ultrasound characteristics of multiple
biliary hamartomas (von Meyenberg complexes).
Ultrasound 2004;12:95–7.
23. Leen E, Moug S, Horgan P. Potential impact and utilization
of ultrasound contrast media. Eur Radiol Suppl
2004;14:P16–P24.
24. Siosteen AK, Elvin A. Intra-operative uses of contrast-
enhanced ultrasound. Eur Radiol Suppl 2004;14:P87–P95.
25. Solbiati L, Tonolini M, Cova L. Monitoring RF ablation. Eur
Radiol Suppl 2004;14:P34–P42.
26. Wen YL, Kudo M, Zheng RQ, Minami Y, Chung H, Suetomi Y,
et al. Radiofrequency ablation of hepatocellular carcinoma:
The British Journal of Radiology, March 2006
therapeutic response using contrast-enhanced coded phase-
inversion harmonic sonography. AJR Am J Roentgenol
2003;181:57–63.
27. Grant EG. Sonographic contrast agents in vascular imaging.
Semin Ultrasound CT MRI 2001;22:25–41.
28. Satchithananda K, Sellars MEK, Ryan SM, Sidhu PS.
Microbubble ultrasound contrast agents in the visualization
of peripheral vasculature in ‘Doppler rescue’: a review.
Ultrasound 2004;12:176–84.
29. Darge K, Troeger J, Duetting T, Zieger B, Rohrschneider W,
Moehring K, et al. Reflux in young patients: comparison of
voiding US of the bladder and retrovesical space with echo
enhancement versus voiding cystourethrography for diag-
nosis. Radiology 1999;210:201–7.
30. Darge K, Troeger J. Vesicoureteral reflux grading in
contrast-enhanced voiding urosonography. Eur J Radiol
2002;43:122–8.
31. Catalano O, Lobianco R, Raso MM, Siani A. Blunt hepatic
trauma: evaluation with contrast-enhanced sonography. J
Ultrasound Med 2005;24:299–310.
32. Catalano O, Lobianco R, Sandomenico F, Siani A. Splenic
trauma: evaluation with contrast-specific sonography and a
second generation contrast medium. J Ultrasound Med
2003;22:467–77.
33. Oldenburg A, Hohmann J, Skork J, Albrecht T. Imaging of
pediatric splenic injury with contrast-enhanced ultrasono-
graphy. Pediatr Radiol 2004;34:351–4.
34. Poletti PA, Platon A, Becker CD, Mentha G, Vermeulen B,
Buhler LH, et al. Blunt abdominal trauma: does the use of a
second-generation sonographic contrast agent help to
detect solid organ injuries? AJR Am J Roentgenol
2004;183:1293–301.
35. Farahmand N, Sirlin CB, Brown MA, Shragg GP, Fortlage
D, Hoyt DB, et al. Hypotensive patients with blunt
abdominal trauma: performance of screening US.
Radiology 2005;235:436–43.
36. Cosgrove DO, Kedar RP, Bamber JC, Al-Murrani B, Davey
JBN, Fisher C, et al. Breast disease: color Doppler US in
differential diagnosis. Radiology 1993;189:99–104.
37. Cosgrove DO, Bamber JC, Davey JB, McKinna JA, Sinnett
HD. Color Doppler signals from breast tumors. Work in
progress. Radiology 1990;176:175–80.
38. Kedar RP, Cosgrove DO, McCready VR, Bamber JC, Carter
ER. Microbubble contrast agent for color Doppler US: effect
on breast masses - work in progress. Radiology
1996;198:679–86.
39. Kook SH, Kwag HJ. Value of contrast-enhanced power
Doppler sonography using a microbubble echo-enhancing
agent in evaluation of small breast lesions. J Clin
Ultrasound 2003;31:227–38.
40. Lamuraglia M, Lassau N, Garbay JR, Mathieu MC, Rouzier
R, Jaziri S, et al. Doppler US with perfusion software and
contrast medium injection in the early evaluation of
radiofrequency in breast cancer recurrences: a prospective
phase II study. Eur J Radiol 2005) 6.
41. Goldberg BB, Merton DA, Liu JB, Thakur M, Murphy GF,
Needleman L, et al. Sentinel lymph nodes in a swine model
with melanoma: contrast-enhanced lymphatic US.
Radiology 2004;230:727–34.
42. Bostick PJ, Guiliano AE. Vital dyes in sentinel node
localization. Semin Nucl Med 2000;30:18–24.
43. Hauff P, Reinhardt M, Briel A, Debus N, Schirner M.
Molecular targeting of lymph nodes with l-selectin ligand-
specific US contrast agent: a feasibility study in mice and
dogs. Radiology 2004;231:667–73.
44. Goldberg BB, Merton DA, Liu JB, Murphy G, Forsberg F.
Contrast-enhanced sonographic imaging of lymphatic
channels and sentinel lymph nodes. J Ultrasound Med
2005;24:953–65.
193

45. Magarelli N, Guglielmi G, Di Matteo L, Tartaro A, Mattei
PA, Bonomo L. Diagnostic utility of an echo-contrast agent
in patients with synovitis using power Doppler ultrasound:
a preliminary study with comparison to contrast-enhanced
MRI. Eur Radiol 2001;11:1039–46.
46. Szkudlarek M, Court-Payen M, Strandberg C, Klarlund M,
Klausen T, Oÿstergaard M. Contrast-enhanced power
Doppler ultrasonography of the metacarpophalangeal joints
in rheumatoid arthritis. Eur Radiol 2003;13:163–8.
47. North American Symptomatic Carotid Endarterectomy
Trial Collaborators. Beneficial effect of carotid endarter-
ectomy in symptomatic patients with high-grade carotid
stenosis. N Engl J Med 1991;325:445–53.
48. European Carotid Surgery Trialists’ Collaborative Group.
Randomised trial of endarterectomy for recently sympto-
matic carotid stenosis: final results of the MRC European
Carotid Surgery Trial (ECST). Lancet 1998;351:1379–87.
49. Ferrer JME, Samso JJ, Serrando JR, Valenzuela VF, Montoya
SB, Docampo MM. Use of ultrasound contrast in the
diagnosis of carotid artery occlusion. J Vasc Surg
2000;31:736–41.
50. Kono Y, Pinnell SP, Sirlin CB, Sparks SR, Georgy B, Wong W,
et al. Carotid arteries: contrast-enhanced US angiography–
preliminary clinical experience. Radiology 2004;230:561–8.
51. Miller MW. Gene transfection and drug delivery.
Ultrasound Med Biol 2000;26:S59–S62.
52. Taniyama Y, Tachibana K, Hiraoka K, Aoki M, Yamamoto
S, Matsumoto K, et al. Development of safe and efficient
194
V R Stewart and P S Sidhu
novel nonviral gene transfer using ultrasound: enhance-
ment of transfection efficiency of naked plasmid DNA in
skeletal muscle. Gene Ther 2002;9:372–80.
53. Lu QL, Liang HD, Partridge T, Blomley MJ. Microbubble
ultrasound improves the efficiency of gene transduction in
skeletal muscle in vivo with reduced tissue damage. Gene
Ther 2003;10:396–405.
54. Tachibana K, Uchida T, Ogawa K, Yamashita N, Tamura K.
Induction of cell-membrane porosity by ultrasound. Lancet
1999;353:1409.
55. Li T, Tachibana K, Kuroki M, Kuroki M. Gene transfer with
echo-enhanced contrast agents: comparison between albu-
nex, optison, and levovist in mice–initial results. Radiology
2003;229:423–8.
56. Blomley M. Which US microbubble contrast agent is best
for gene therapy? Radiology 2003;229:297–8.
57. Jakobsen JA, Oyen R, Thomsen HS, Morcos SK. Safety of
ultrasound contrast agents. Eur Radiol 2005;15:941–5.
58. Crucinio N, Nacchiero MC, Panella C, Ierardi E. A case of
adverse reaction potentially related to a microbubble
contrast agent for ultrasonography. Eur J Radiol (Extra)
2004;52:25–6.
59. van der Wouw P, Brauns A, Bailey S, Powers J, Wilde A.
Premature ventricular contractions during triggered imag-
ing with ultrasound contrast. J Am Soc Echocardiogr
2000;13:288–94.
60. Sidhu PS. Microbubbles are here to be burst! Radiology
Now 2003;20:2–3.
The British Journal of Radiology, March 2006
The British Journal of Radiology, 79 (2006), 195–200

Enhanced biological effectiveness of low energy X-rays and

implications for the UK breast screening programme
1 2
G J HEYES, PhD, A J MILL, PhD and 2M W CHARLES, PhD, DSc

1
Department of Medical Physics, University Hospital Birmingham NHS Foundation Trust,
Birmingham B15 2TH and 2Radiation Biophysics Group, School of Physics and Astronomy, The
University of Birmingham, Birmingham B15 2TT, UK

ABSTRACT. Recent radiobiological studies have provided compelling evidence that the
low energy X-rays as used in mammography are approximately four times – but possibly
as much as six times – more effective in causing mutational damage than higher energy
X-rays. Since current radiation risk estimates are based on the effects of high energy
gamma radiation, this implies that the risks of radiation-induced breast cancers for
mammography X-rays are underestimated by the same factor. The balance of risk and
benefit for breast screening have been re-analysed for relative biological effectiveness
(RBE) values between 1 and 6 for mammography X-rays. Also considered in the analysis
is a change in the dose and dose-rate effectiveness factor (DDREF) from 2 to 1, women
with larger than average breasts and implications for women with a family history of
breast cancer. A potential increase in RBE to 6 and the adoption of a DDREF of unity
does not have any impact on the breast screening programme for women aged 50–
70 years screened on a 3 yearly basis. Situations for which breast screening is not Received 19 April 2005
justified due to the potential cancers induced relative to those detected (the detection- Revised 23 June 2005
to-induction ratio (DIR)) are given for a range of RBE and DDREF values. It is concluded Accepted 24 June 2005
that great caution is needed if a programme of early regular screening with X-rays is to
DOI: 10.1259/bjr/21958628
be used for women with a family history of breast cancer since DIR values are below 10
(the lowest value considered acceptable for women below 40 years) even for modest ’ 2006 The British Institute of
increases in the RBE for mammography X-rays. Radiology

There is increasing evidence from radiobiology studies
[1–3] that the low energy X-rays used for mammography
breast screening are more effective in inducing biological
damage than higher energy X-rays. Risk estimates for
radiation-induced cancer – principally derived from the
atomic bomb survivor study (ABSS) – are based on the
effects of high energy c-rays and thus the implication is
that the risks of radiation-induced breast cancer arising
from mammography may be higher than that assumed
based on standard risks estimates. As with any clinical
examination, the radiological breast screening pro-
gramme must be justified, in that the risk associated
with the exposure must be greatly outweighed by the
potential gain to a patient as a result of a the procedure.
This is particularly so for breast screening since the
large majority of women undergoing mammography
are asymptomatic. Thus, while the radiation dose to
the breast can be accurately measured and kept as low
as reasonably practicable, the risk associated with this
dose, and therefore the risk-benefit ratio is less well
known.
The most recent and relevant radiobiology studies
have utilized the immortalized human cell line desig-
nated CGL1 in which the transformation frequencies
induced by low energy X-rays were compared with
the effects of higher energy X-rays, c-rays and electrons.
This work was supported by EPSRC grant RRAH07673.
In particular, in one study [1] a direct comparison
between 29 kVp X-rays, generated using a clinical
mammography unit, and radiation simulating the atomic
bomb spectrum at Nagasaki was made. The best estimate
of the limiting (low-dose) relative biological effectiveness
(RBEM) of 29 kVp X-rays compared with the atomic
bomb spectrum radiation was found to be 4.42¡2.02.
While it is recognized that there can be limitations in
directly extrapolating data obtained in vitro to carcino-
genesis in vivo, these results provide strong evidence that
the radiation risks from mammography may be under-
estimated by a factor of approximately four, and possibly
as high as six. In this paper we re-analyse the balance
of risk and benefit for breast screening using the methods
of Law and Faulkner [4, 5] based on these higher values
of the RBE of mammography energy X-rays.
The National Health Service Breast Screening
Programme
The National Health Service Breast Screening
Programme (NHSBSP) was started in the UK in 1988 in
an attempt to reduce mortality from breast cancer. The
programme has been largely successful, with the
International Agency for Research on Cancer (IARC)
concluding that breast screening by mammography of
women aged between 50 years and 69 years reduces
mortality from breast cancer by 35% [6]. It is clear that a

The British Journal of Radiology, March 2006 195


breast screening programme does save lives. However,
such a widespread (two million women screened per
year in the UK in seven screening rounds) use of a
radiological examination in asymptomatic women needs
to be treated with some caution. The risks associated
with such a programme need to be properly evaluated,
especially since there are moves both in the UK and
elsewhere to extend the use of mammography to a wider
section of the population. Current guidelines for the
NHSBSP in the UK are for two-view radiological
examinations for women aged between 50 years and
70 years with a frequency of every 3 years [7]. An
average mean glandular dose (per two-view screening
examination) of 4.5 mGy is used for the purpose of
benefit-risk analysis [8]. The UK is considering expand-
ing the screening programme to include younger
women (currently considering a minimum age of
40 years). The intention of this NHSBSP ‘‘UK Age-trial’’
with women without family history [8] is to screen
women annually from 40 years to 47 years. Such an
increase in the screening lifetime will clearly increase
the cumulative dose. The frequency of this age trial
means that women will in fact receive a higher dose in
the age trial (10 screening rounds) than they would in
the normal screening programme (seven screening
rounds). Just by attending this age trial, women will
therefore more than double the lifetime absorbed dose
to the breast. The effect is magnified, since the
screening of younger, denser breasts can increase the
dose required to produce the required film exposure by
approximately 15–20% [9]. Since the risks of radiation-
induced breast cancer are age-dependent [9], younger
breasts are also more susceptible to radiation-induced
cancers, thus compounding the increased risk even
further.
The NHSBSP report [8] also highlights a ‘‘high dose
subgroup’’ population: a proportion of women who
receive a higher than average mean glandular dose. This
increase in dose may be due to the use of difficulties
with imaging equipment or to higher than average
breast tissue thickness under compression. The mean
glandular dose for this subgroup is 21.4 mGy per two-
view examination [8]. This high dose subgroup is
estimated to account for 0.1% of the screened population.
However, this figure is set to increase with increasing
breast size of the UK population. Over the course of the
screening programme (seven two-view sessions) a
woman in the high dose subgroup can expect to receive
a mean glandular dose of more than 170 mGy. It is likely
that an age trial will be targeted at women who are
judged to be at an increased risk of breast cancer. Such
women are identified if (amongst other reasons) there is
a family history of breast cancer (thought to be the most
important factor) or if they are obese (and are therefore
also likely to be in the high-dose subgroup). Women
with a family history of breast cancer are thought
most at risk of developing the disease, since about 10%
of breast cancers are thought to have a genetic basis.
These women may be deficient in one of the known
breast cancer suppressor genes, BRCA1 and BRCA2.
Such a gene deficiency may well increase the suscepti-
bility of a woman to develop a radiation-induced
tumour, since the number of targets in each cell requiring
damage is reduced. The radiation risk for this sensitive
196
G J Heyes, A J Mill and M W Charles
subgroup may therefore be significantly greater than the
risks associated with the average UK population.
Currently these risks are poorly understood, and this
paper seeks to highlight the need for caution if
mammography breast screening is to be used in this
subgroup.
In the NHSBSP, the risks associated with mammo-
graphy doses are calculated from various epidemiologi-
cal sources [8]. The source includes data from North
American women who were given high doses of
radiation for medical reasons (e.g. X-ray therapy for
acute post-partum mastitis and multiple sessions of
direct fluoroscopy for tuberculosis, mostly during the
1930s and 1940s). In addition to being a high-dose
population, the North American women were exposed to
therapy energies of X-rays, highlighting the importance
of calculating the efficacy in inducing cell damage by
lower energy X-ray radiation. ABSS data are considered,
but not used in the calculations of risk. The justification
for the omission of the ABSS is [8] that since Japanese
women ‘‘have a markedly lower natural incidence of
breast cancer than women from western counties such as
the UK and USA…’’ it is ‘‘difficult to transfer radiation
risks between these two populations.’’ Published risk
figures for potential cancer inductions varying with age
of exposure to X-rays are available [4]. Using these data,
with the published values of breast cancers detected in
the UK screening programme (NHS report) the ratio of
cancers detected to those induced by the mammography
dose (detection to induction ratio (DIR)) can be calcu-
lated. The use of the DIR is not a true measure of the
benefit-risk ratio, but it is an indication of its likely
magnitude. An examination of the relationship of these
two ratios [4] has shown that the difference between
them is likely to be as little as 15–20%. A DIR of 100 is
considered to be ample, whilst a ratio of 10 is considered
to be sufficient in terms of justifying the use of a
radiological exposure [5].
The effect of a higher RBE for mammography
X-rays
The in vitro cell transformation data from Heyes and
Mill [1] suggest a best estimate for the RBE of
mammography X-rays of approximately 4 with 90%
confidence intervals in the range 2 to 6. Table 1 presents
the cancer induction-rates for the UK population that are
assumed in the NHSBSP. Table 2 then presents the
number of induced cancers for a range of RBE values
from 1 (as assumed in the NHSBSP) to 6 (an approximate
upper limit based on the latest in vitro data). In Table 3
the DIR values are given. The data given in Tables 2 and
3 are for women screened once every 3 years in the UK
NHSBSP and include those in the high dose sub group as
well as the ‘‘normal’’ group of women. Using an RBE of 1
for mammography X-rays, as used by the NHSBSP, the
DIR falls below 100 only when the screening age is below
55 years. With an RBE value of 2, the DIR falls below 100
for all screening ages below 65 years. For higher values
of the RBE, all values of DIR lie below 100. For women in
the high dose sub group DIR values are clearly much
lower.
The British Journal of Radiology, March 2006
Implications of an increased RBE for UK mammography screening

Table 1. Total number of breast cancers induced per million NHSBSP. The EPA states that there is epidemiological
women screened per mGy absorbed for women of different evidence that dose-fractionation has little or no effect on
ages (after Law et al [4]) risk to the breast [10], and that a DDREF of unity should
be adopted.
Age (years) Total breast cancers induced per million
women exposed per mGy If the DDREF for breast cancer is unity, as advocated
by the American EPA, the DIR falls below 100 with an
25–29 18.4 RBE of 1 for all screening ages below 65 years. If the RBE
30–34 18.2 is increased, the DIR is below 100 for all ages, and would
35–39 17.8
40–44 16.6
even fall below 10 for a screening age of 50–54 years if a
45–49 15.0 ‘‘worst case scenario’’ RBE of 6 is considered. DIR values
50–54 13.2 for a DDREF of unity are exactly one half of the values
55–59 11.5 given in Table 3.
60–64 9.4
65–70 7.4
Screening of women with a family history of
breast cancer
Dose and dose-rate effectiveness and the breast
The National Institute for Clinical Excellence (NICE)
screening programmes in the UK and USA
[11] considers the efficacy of early screening of women
According to the dual action theory of radiation with a family history of breast cancer. Whilst most of the
damage, the dose response should be linear at low studies listed as research literature evidence (section
doses. At higher doses and dose-rates, multiple track 7.2.2 of [11]) conclude regular, early screening sessions
events become important, thereby bending the dose would lead to an increase in tumour detection, very little
response upwards. As a result, the response per unit mention is made to the potential of women with a
dose at low doses will be overestimated if a linear genetic disorder being more susceptible to radiation-
extrapolation is made from observations at high doses. induced breast tumour. One study [12] even concluded
The degree of overestimation is expressed in terms of a that mammography was relatively insensitive to detect-
dose and dose-rate effectiveness factor (DDREF). A ing tumours in women with BRCA1/2 mutations, and
DDREF of 2 is used in the NHSBSP [6]. However, the that other forms of detection were likely to be more
US Environmental Protection Agency (EPA) advocates a beneficial to this high risk group.
DDREF value of 1 for breast cancer [10] and risks based Law et al [4] have published the detection/incidence
on a DDREF of 1 are twice those calculated by the ratio for younger women (aged 25–49 years) with a

Table 2. Estimated numbers of cancers detected and induced for women in the normal (N) and high dose (H) groups screened
three-yearly in the NHSBSP
Age (years) Group (N5normal; H5high-dose) Number of cancers per 106 women

detected induced for

RBE51 RBE52 RBE54 RBE56

3
50–54 N 5.3610 59 119 238 356
H 290 581 1162 1742
55–59 N 6.26103 52 104 208 311
H 253 506 1012 1518
60–64 N 7.76103 42 85 169 254
H 207 414 827 1241
65–70 N 106103 33 67 133 200
H 163 326 651 977
RBE, relative biological effectiveness; N, normal population; H, high dose subgroup; NHSBSP, National Health Service Breast
Screening Programme.

Table 3. Estimated ratio of cancer detection to induction (DIR) for women in the normal (N) and high dose (H) groups screened
3 yearly in the NHSBSP
Age (years) Ratio detected/induced (DIR)

Normal group (N) High dose subgroup (H)

RBE51 RBE52 RBE54 RBE56 RBE51 RBE52 RBE54 RBE56

50–54 89 45 22 15 18.3 9.1 4.6 3.0

55–59 120 60 30 20 24.5 12.3 6.1 4.1
60–64 182 91 46 30 37.2 18.6 9.3 6.2
65–70 300 150 75 50 61.4 30.7 15.4 10.2
RBE, relative biological effectiveness; DIR, detection to induction ratio; N, normal population; H, high dose subgroup;
NHSBSP, National Health Service Breast Screening Programme.

The British Journal of Radiology, March 2006 197

 G J Heyes, A J Mill and M W Charles

Table 4. Estimated cancer induction for women in the age Discussion

range 25–49 years with a family history of breast cancer
(after Law et al [4]). Annual 2-view screening
Age (years) Total breast cancers induced per 106
women exposed per mGy
25–29 82.8
30–34 81.9
35–39 80.1
40–44 74.7
45–49 67.5
family history of breast cancer screened annually with
two-view mammography. The cancer induction-rates
upon which these values are calculated are shown here
in Table 4. The DIR for women with two index ages are
given in Table 5. When the index patient (mother, sister,
daughter) age is 30–39 years, the ratio of detection to
induction falls below 10 for women with such a family
history if they are screened below the age of 30 years
using an RBE of 2, and below 35 years when the RBE is 4.
The screening age of a woman with a family history
(index age 30–39 years) where the DIR falls below 10 is
45 years for an RBE of 6.
If benefit is to exceed radiation risk in a screening
programme, the ratio of detection to induction (DIR)
should exceed a factor of about 5 [13]. This may not
apply to women below the age of 40 years, for whom the
DIR should exceed 10.
If the index patient is 40–49 years old the detection-
rate is less, and this means the DIR falls below 10 at
higher ages than for the index patient age of 30–39 years.
The DIR falls below 10 in this case if the screening age is
less than 30 years for an RBE of 1, an age less than
35 years for an RBE of 2 and does not exceed 10 for any
screening age if the RBE is greater than or equal to 4.
These results therefore imply that the caution should be
used when considering mammography screening for
women with a history of breast cancer, especially if the
index patient age is greater than 40 years old. Such
women may be carriers of BRCA1 and BRCA2 gene
mutations, and as such may be more susceptible to a
radiation induced tumour.
Using a DDREF of 2 and an RBE of unity in the normal
screening programme, only women aged between
50 years and 54 years will have a DIR of less than 100.
If an RBE of 2 is used then the DIR for women in the 50–
54 year range falls to 45 years, and women up to and
including the 60–64 year age range have a DIR of less
than 100. If an RBE of 4 is considered (as suggested from
the in vitro radiobiology results [1]), the DIR is below 100
for women of all ages considered, and has a minimum
value of 22 for women aged between 50 years and
54 years.
Considering a change in DDREF to unity, such as that
used by the American system, the DIR would fall to 11
for the youngest women with an RBE of 4. In the worst
case scenario (RBE56) the DIR falls below 10 for women
between 50 years and 59 years (Table 3).
If the risk of radiation induced tumours can be
justified in terms of maintaining the DIR above 5 for
women over 40 years, then an increase in RBE, even to 6
and a DDREF of unity, would not therefore have an
impact on the use of mammography as a breast screen-
ing tool for the normal 3 yearly screened population of
women aged between 50 years and 70 years since the
lowest DIR calculated in these cases is 7.4 for women
aged between 50 years and 54 years. Younger (50–
54 years) women in the high-dose subgroup are likely
to have a DIR less than 5 if the RBE is significantly
increased from unity. Unfortunately, there are no cancer
detection data upon which to accurately calculate the
DIR for this population. In these calculations the
detection-rates for the high dose subgroup population
have been assumed to be the same as those in the normal
screened population.
In the cases of women with a family history of breast
cancer, women screened annually with two-views from
the age of 25 years are considered. The DIR for such a
population is already reduced, even with a DDREF of 2
and an RBE of unity. The lowest DIR is 6 for the youngest
age range (25–29 years). The effect on a change to the
RBE is most dramatic in this screening group. For these
younger women, the DIR should exceed 10 if the benefit
is to exceed the radiation risk [13]. In fact, an increase in

Table 5. Effect of an increase in RBE on the cancer detection/induction ratio for UK women screened annually (two views) with
a family history of breast cancer (values in parenthesis are for a DDREF of 1)
Age (years) Number of cancers detected per 106 women Ratio detected/induced (DIR)

RBE51 RBE52 RBE54 RBE56

(a) Patient Index age530–39

25–29 0.506103 6.0 (3.0) 3.0 (1.5) 1.5 (0.8) 1.0 (0.5)
30–34 1.656103 20.1 (10.1) 10.1 (5.0) 5.0 (2.5) 3.4 (1.7)
35–39 3.356103 41.8 (20.9) 20.9 (10.5) 10.5 (5.2) 7.0 (3.5)
40–44 4.106103 54.9 (27.1) 27.4 (13.7) 13.7 (6.9) 9.1 (4.6)
45–49 4.106103 60.7 (30.4) 30.4 (15.2) 15.2 (7.6) 10.1 (5.1)
(b) Patient Index age540–49
25–29 0.506103 6.0 (3.0) 3.0 (1.5) 1.5 (0.8) 1.0 (0.5)
30–34 0.996103 12.1 (6.0) 6.0 (3.0) 3.0 (1.5) 2.0 (1.0)
35–39 1.656103 20.6 (10.3) 10.3 (5.1) 5.1 (2.6) 3.4 (1.7)
40–44 2.506103 33.5 (16.7) 16.7 (8.4) 8.4 (4.2) 5.6 (2.8)
45–49 2.506103 37.0 (18.5) 18.5 (9.3) 9.3 (4.6) 6.2 (3.1)
RBE, relative biological effectiveness; DDREF, dose and dose rate reduction factor; DIR, detection to induction ratio.

198 The British Journal of Radiology, March 2006

Implications of an increased RBE for UK mammography screening

Table 6. Situations when breast screening should not be carried out for different RBE and DDREF values – based on a cut-off DIR
value of 5 (10 for women aged 40 years and under) below which screening should not occur
DDREF RBE Breast screening should not occur in women

2 (as assumed in the
UK programme)
1 (as used in the USA)
RBE, relative biological effectiveness; DDREF, dose and dose rate reduction factor; DIR, detection to induction ratio.
2 with a family history of breast cancer and aged between 25 years and 29 years where
the patient index age is 30–39 years; aged between 25 years and 34 years where
the patient index age is 40–49 years
4 (a) with a family history of breast cancer and aged between 25 years and 34 years
where the patient index age is 30–39 years; aged between 25 years and 49 years
where the patient index age is 40–49 years
(b) in the high dose subgroup below 55 years of age
6 (a) with a family history of breast cancer and aged between 25 years and 39 years
where the patient index age is 30–39 years; aged below 44 years where the patient
index age is 40–49 years
(b) in the high dose subgroup below 60 years of age
2 (a) with a family history of breast cancer and aged between 25 years and 34 years
where the patient index age is 30–39 years; aged between 25 years and 39 years
where the patient index age is 40–49 years
(b) in the high dose subgroup below 55 years of age
4 (a) with a family history of breast cancer and aged between 25 years and 39 years
where the patient index age is 30–39 years; aged below 49 years where the patient
index age is 40–49 years
(b) in the high dose subgroup below 65 years of age
6 (a) with a family history of breast cancer and aged between 25 years and 44 years
where the patient index age is 30–39 years; aged below 49 years where the patient
index age is 40–49 years
(b) in the high dose subgroup below 65 years of age. (and with caution for older
women, since the DIR is 5.1 for women 65–70 years old)
(c) screened annually from the age of 40 years with two-view mammography

the RBE to just 2 means that women aged 25–34 years
have a DIR less than 10 (with a DDREF of 2). If the RBE
of mammography X-rays is 4, then the entire age range
(25–49 years) of women with a family history of breast
cancer (index age 40–49 years) will have a DIR less than
10 (DDREF52). In this case, the youngest women (age
range 25–29 years) have a DIR of less than 2.0. It is still
only 3.0 (i.e. not justifiable) for this age range if the RBE is
decreased to 2.0.
These comments are summarized in Table 6, where a
DIR of 5 is chosen to be a cut-off, below which a
screening programme could not be justified due to the
potential cancers induced relative to those detected. (For
women under the age of 40 years, a DIR of 10 is used as
the cut-off [13]).
Conclusion
This work suggests great caution should be exercised
if a programme of early, regular screening using X-rays,
is to be used in women with a family history of breast
cancer. The effect of an increase in RBE or DDREF is
most pronounced in younger women screened due to a
family history and women in the high-dose subgroup.
Other methods for detecting tumours at an early stage
are available, and have been considered by NICE [11].
Many studies [14–16] have shown that MRI is more
accurate than mammography in screening young women
with a family history of breast cancer. Such a widespread
use of MRI would, however, place a high burden on
limited NHS resources in the UK, due to the significant
increase in screening cost. A number of studies [12, 17]
have observed that mammography surveillance is less
sensitive in younger women, women with a family
history and BRCA1/2 mutant carrier, a point noted by
NICE [11].
It is recognized that in vitro radiobiological experi-
ments using immortalized cell lines cannot be considered
in isolation or used directly as a basis for reviewing the
breast screening programme. The recently observed high
values of RBE for cancer-related in vitro end-points do,
however, strengthen the long accepted evidence, based
largely on non-cancer end-points, that low energy X-
radiations have larger RBE values than higher energy
photons. Such findings have long been accepted by the
International Commission on Radiological Protection
(ICRP). For general radiological protection situations
the ICRP has, however, maintained a radiation weight-
ing factor of 1 for all low LET radiations (including
electrons and photons of all energies). In specific
situations the ICRP recommends the use of more
appropriate assumptions. The possibility of RBE values
for mammography exposures which are in excess of
the ICRP radiation weighting factors focuses attention on
the results of related epidemiology studies. It widens the
range of possible assumptions which can be used in cost-
benefit analyses to inform the debate regarding exten-
sions to existing screening programmes.
References
1. Heyes GJ, Mill AJ. The neoplastic transformation potential
of mammography X rays and atomic bomb spectrum
radiation. Radiat Res 2004;162:120–7.
2. Frankenberg D, Kelnhofer K, Bar K, Frankenberg-Schwager
M. ERRATA: Enhanced neoplastic transformation by
mammography X rays relative to 200 kVp X rays:
Indication for a strong dependence on photon energy of

The British Journal of Radiology, March 2006 199


the RBEM for various end points (vol 157, pg 99, 2002).
Radiat Res 2002;158:126.
3. Goggelmann W, Jacobsen C, Panzer W, Walsh L, Roos H,
Schmid E. Re-evaluation of the RBE of 29 kV x-rays
(mammography x-rays) relative to 220 kV x-rays using
neoplastic transformation of human CGL1-hybrid cells.
Radiat Environ Biophys 2003;42:175–82.
4. Law J, Faulkner K. Two-view screening and extending the
age range: the balance of benefit and risk. Br J Radiol
2002;75:889–94.
5. Law J, Faulkner K. Concerning the relationship between
benefit and radiation risk, and cancers detected and
induced, in a breast screening programme. Br J Radiol
2002;75:678–84.
6. IARC. Press Release N˚ 139: Mammography screening can
reduce deaths from breast cancer. Lyon, France, 2002.
7. Patnick J, editor. Breast Screening Programme Annual
Review. Sheffield, UK: National Health Service Cancer
Screening Programmes, 2002.
8. Young KC, Faulkner K, Wall B, Muirhead C. Review of
radiation risk in breast screening: report by a joint working
party of the NHSBSP National Coordinating Group for
Physics Quality Assurance and the National Radiological
Protection Board. Sheffield, UK: National Health Service
Cancer Screening Programmes, 2003.
9. Law J. Cancers detected and induced in mammographic
screening: New screening schedules and younger women
with family history. Br J Radiol 1997;70:62–9.
10. Puskin JS, Nelson CB. Estimating radiogenic cancer risks
(report 402-R-93-076), Washington DC: US Environmental
Protection Agency, 1994.
200
G J Heyes, A J Mill and M W Charles
11. McIntosh A, Shaw C, Evans G, Turnbull N, Bahar N,
Barclay M, et al. Clinical guidelines and evidence review for
the classification and care of women at risk of familial
breast cancer. London: National Collaborating Centre for
Primary Care, University of Sheffield, 2004.
12. Goffin J, Chappuis PO, Wong N, Foulkes WD. Re: Magnetic
resonance imaging and mammography in women with a
hereditary risk of breast cancer. J Natl Cancer Inst
2001;93:1754.
13. Law J, Faulkner K. Cancers detected and induced, and
associated risk and benefit, in a breast screening pro-
gramme. Br J Radiol 2001;74:1121–7.
14. Warner E, Plewes DB, Shumak RS, Catzavelos GC, Prospero
LS, Yaffe MJ, et al. Comparison of breast magnetic
resonance imaging, mammography, and ultrasound for
surveillance of women at high risk for hereditary breast
cancer. J Clin Oncol 2001;19:3524–31.
15. Stoutjesdijk MJ, Boetes C, Jager GJ, Beex L, Bult P, Hendriks
J, et al. Magnetic resonance imaging and mammography in
women with a hereditary risk of breast cancer. J Natl
Cancer Inst 2001;93:1095–102.
16. Kuhl CK, Schmutzler RK, Leutner CC, Kempe A,
Wardelmann E, Hocke A, et al. Breast MR imaging
screening in 192 women proved or suspected to be carriers
of a breast cancer susceptibility gene: preliminary results.
Radiology 2000;215:267–79.
17. Kerlikowske K, Carney PA, Geller B, Mandelson MT,
Taplin SH, Malvin K, et al. Performance of screening
mammography among women with and without a first-
degree relative with breast cancer. Ann Intern Med
2000;133:855–63.
The British Journal of Radiology, March 2006
The British Journal of Radiology, 79 (2006), 201–207

Comparison of image quality, diagnostic confidence and

interobserver variability in contrast enhanced MR angiography and
2D time of flight angiography in evaluation of carotid stenosis
1
D MITRA, FRCR, 1D CONNOLLY, FRCR, 1S JENKINS, FRCR, 1P ENGLISH, DCR, 1D BIRCHALL, FRCR,
1
C MANDEL, FRCR, 1K SHRIKANTH, MD, 2B GREGSON, PhD and 1A GHOLKAR, FRCR

1
Department of Neuroradiology, Regional Neurosciences Centre, Newcastle General Hospital,
Westgate Road, Newcastle upon Tyne NE4 6BE and 2Academic Department of Neurosurgery, School
of Surgical and Reproductive Sciences, University of Newcastle upon Tyne NE1 7RU, UK

ABSTRACT. The aim of this study was to compare image quality, level of diagnostic
confidence and interobserver agreement in assessment of carotid stenosis with contrast
enhanced MR angiography (CE MRA) in comparison with 2D time of flight MR
angiography (2D TOF MRA). 60 carotid arteries in 30 patients were examined by three
observers. Image quality and diagnostic confidence were assessed on the basis of a
visual analogue scale. Interobserver variability was assessed with the help of intraclass
correlation coefficient. Median values on the visual analogue scale for image quality
and diagnostic confidence were higher for CE MRA compared with 2D TOF MRA for all Received 3 February 2005
three observers. Higher intraclass correlation values were recorded for interobserver Revised 14 June 2005
variability for CE MRA compared with 2D TOF MRA both for visual estimation of carotid Accepted 15 July 2005
stenosis as well as for measurement of carotid stenosis on the basis of North American
DOI: 10.1259/bjr/72842752
Symptomatic Carotid Endarterectomy Trial (NASCET) and European Carotid Surgery
Trial (ECST) criteria. CE MRA provides better image quality, higher level of diagnostic ’ 2006 The British Institute of
confidence and more interobserver agreement compared with 2D TOF MRA. Radiology

An atherosclerotic lesion at the carotid bifurcation is
one of the major causes of ischaemic strokes. The North
American Symptomatic Carotid Endarterectomy Trial [1]
(NASCET) and European Carotid Surgery Trial [2]
(ECST) have demonstrated that surgical intervention is
more beneficial compared with medical management in
symptomatic patients with more that 70% carotid
stenosis. The value of carotid endarterectomy has been
extended to include asymptomatic carotid stenosis
greater than 60% after the Asymptomatic Carotid
Atherosclerosis Study [3] (ACAS). Accurate pre-opera-
tive assessment of the degree of carotid stenosis is
therefore of crucial importance as the benefit of surgery
is not proven in lesser degrees of stenosis.
Conventional catheter angiography (CA) is accepted as
the gold standard in assessment of carotid stenosis and is
the modality used in the measurement of stenosis in
NASCET and ECST trials. However, due to the known
risks of CA (overall 1–2% risk of thromboembolic
complication, risks increasing with age and presence of
generalized atherosclerosis), increasing numbers of
centres are using non-invasive methods for pre-operative
evaluation of carotid stenosis. Doppler ultrasound (DUS)
is routinely used as the screening technique in many
centres. MR angiography (MRA) is another technique,
which is used either to confirm or supplement DUS
findings.
Time of flight MRA (TOF MRA) uses inflow of
unsaturated protons in blood to generate signal within
a blood vessel. However, due to dependence on flow,
TOF MRA is prone to flow related artefacts such as
signal dropout caused by turbulence in a severely
stenosed artery. The technique is also prone to move-
ment artefacts due to relatively long scan time.
Contrast enhanced MRA (CE MRA) uses the T1
shortening effect of intravenous paramagnetic contrast
agent gadolinium to generate the signal. It is, therefore,
less prone to, although not completely free of, the flow
related artefacts in TOF MRA. CE MRA also requires less
scan time and covers a wider field of view, which allows
assessment of the aortic arch and proximal common
carotid arteries.
A number of studies have been carried out to evaluate
specificity and sensitivity of CE MRA using CA as the
gold standard [4–12]. However, there are relatively few
studies [13, 14], which have compared TOF and CE MRA
directly. These latter studies have looked at the compar-
ison of the two techniques in terms of delineation of
morphological details and observer confidence but have
not included interobserver variability assessment. The
aim of the current study is to evaluate the image quality,
diagnostic confidence of the observer and interobserver
variability of the two techniques.
Materials and methods
30 consecutive patients with suspected carotid bifur-
cation disease were prospectively included in this study.
A DUS study was performed in all the patients. DUS

The British Journal of Radiology, March 2006 201


results were available in all but three patients. All
patients had 2D TOF MRA of the carotid bifurcation and
3D CE MRA from the aortic arch to the skull base. All the
MRA studies were performed in a Philips 1.5 Tesla
scanner (Philips, Best, The Netherlands) using flexible
phased array coil.
For CE MRA, an 18-gauge cannula was inserted in the
ante-cubital vein. A power injector (Medrad Spectris;
Medrad Inc., Maastricht, The Netherlands) was used to
administer 15 ml of Magnevist (Gadopentate; Schering
AG, Berlin, Germany) at a rate of 1.5 ml s21. Bolus
tracking technique was used for image acquisition,
whereby a single coronal slice was acquired at a rate of
1.67 frames per second while the contrast was being
injected and acquisition of CE MRA was triggered after
contrast was seen in the aortic arch. A fast gradient echo
sequence (3D FFE; Philips; repetition time (TR)55.2 ms,
echo time (TE)51.8 ms, flip angle 40 ˚, field of view
270 mm and matrix size 336 6 512) in the coronal plane
was acquired using 50% slice interpolation giving a voxel
size of 1 mm 6 0.5 mm 6 0.5 mm. Data in the central k-
space was acquired first using an elliptocentric k-space
filling technique (CENTRA; Phillips). As central k-space
holds data with high amplitude and low spatial resolu-
tion, this technique allows most of the contrast informa-
tion to be obtained while gadolinium was in the arterial
phase. Total acquisition time for CE MRA sequence was
1 min 27 s.
An axial 2D TOF MRA (TR517 ms, TE53.4 ms, flip
angle 60 ˚, field of view 200 mm and matrix size 224 6
512) study incorporating overlapping 3 mm slices to
cover the carotid bifurcation was obtained in the same
session. Venous contamination was prevented by using a
15 mm ‘‘travelling’’ superiorly positioned pre-saturation
pulse. Total acquisition time for the TOF MRA sequence
was 2 min.
Apart from the different scanning parameters men-
tioned above, CE MRA can be visually distinguished
from TOF MRA by the coronal plane of acquisition of the
source images (as opposed to axial acquisition for TOF
MRA), larger anatomical coverage and more background
suppression.
Hard copy images were produced for both CE MRA
and TOF MRA with 9 maximum intensity projection
(MIP) reconstructions at 40 ˚ steps and assessment of
stenosis was made from the hard copies.
In all 120 sets of images (60 carotids in 30 patients
imaged with two MRA technique for each carotid) were
independently assessed by 3 radiologists. Image order
was completely randomized so that images of the left
and the right carotids as well as the images in the two
modalities (i.e. CE and TOF MRA) were scattered
throughout the 120 sets of images thereby reducing the
bias affecting the assessment of the degree of stenosis. It
took several sittings by each radiologist to complete the
assessments.
Image quality was assessed by visual analogue scale
(VAS). The VAS consisted of a 5 cm long line with
maximum quality at 5 and minimum quality at 0. The
images were specifically assessed for slice misregistration,
pulsation artefact, venous flow signal, presence of plaque
ulceration, visualization of external carotid artery (ECA)
branches (superior thyroid and lingual) and signal dropout.
202
D Mitra, D Connolly, S Jenkins et al
Carotid stenosis was assessed both by visual estima-
tion and by measurements on the basis of NASCET and
ECST criteria. Stenosis was measured with the film on a
horizontal viewing box. Electronic callipers (Digimax
Measy 2000; Swissprecision) were used to ensure
accurate measurement of stenosis. Visual estimation
was graded from 1 to 6 on the basis of NASCET criteria
(1, 0%; 2, ,50%; 3, 50–70%; 4, 70–95%; 5, .95%; and 6,
100%). Calliper measurements were carried out at the
level of maximum stenosis, distal normal internal carotid
artery, common carotid artery and estimated carotid
bulb. Percentage of stenosis was then calculated on the
basis of NASCET and ECST criteria.
Level of observer confidence on assessment of stenosis
was scored both for visual estimation and estimation on
the basis of measurements described above. This was
again done on a VAS described above with most
confident at 5 and least confident at 0.
Statistical analysis
Data were transferred to a Microsoft Access database
and statistical analysis performed with SPSS software
(SPSS Inc., Chicago, IL). Differences between the two
techniques in terms of image quality and observer
confidence were assessed using paired t-tests.
Interobserver variability between observers was calcu-
lated with the help of intraclass correlation coefficient.
Mean, median, mode and standard deviation were
calculated on the VAS scores and displayed in box plots.
Results
The study was performed on 30 patients, including 21
males and 9 females. In all 60 carotid arteries were
assessed. Initial screening DUS demonstrated ,50%
stenosis in 24 carotids, 50–70% stenosis in 5 carotids
and .70% stenosis in 25 carotids. DUS results were not
available in 6 carotids (3 patients).
Image quality
The median scores of image quality for CE MRA by the
three raters were 4.0, 3.5 and 2.65 and those for TOF MRA
were 2.05, 2.0 and 1.05, respectively (Figure 1). All three
differences were statistically significant (p,0.00001).
For further assessment of image quality, visualization of
the superior thyroid and lingual branches of external carotid
arteries by the two techniques were assessed (Table 1).
Observer confidence for visual assessment of
stenosis
Median scores for confidence level for visual assess-
ment of stenosis by the three raters were 4.0, 3.7 and 4.0
for CE MRA. Corresponding scores for TOF MRA were
2.8, 1.9 and 2.8 (Figure 2). All three differences were
statistically significant (p,0.00001). Results of visual
assessment of stenosis by three observers by both TOF
MRA and CE MRA are given in Table 2.
The British Journal of Radiology, March 2006
CE MRA and TOF MRA in carotid stenosis

Table 2. Results of assessment of carotid stenosis by three

observers (A, B and C) with the three methods (NASCET, ECST
and visual estimation) for both CE MRA and TOF MRA.
Numbers under columns A, B and C denote number of
carotids under each category of stenosis. Visual estimation
was based on 6 grades depending on the severity of stenosis
CE MRA
NASCET% A B C ECST% A B C Visual % (grade) A B C

0% 1415 8 0% 1113 8 0%(Grade1) 1114 8

,50% 151620 ,50% 151515 ,50%(Grade2) 181817
50–70% 510 8 50–70% 5 8 8 50–70%(Grade3) 7 5 6
70–95% 11 7 9 70–95% 151213 70–95%(Grade4) 91111
95–99% 10 810 95–99% 9 811 95–99%(Grade5) 10 813
100% 5 4 5 100% 5 4 5 100%(Grade6) 5 4 5
TOF MRA
NASCET% A B C ECST% A B C Visual % (grade) A B C

0% 1312 3 0% 1311 3 0%(Grade1) 1311 3

,50% 151918 ,50% 111120 ,50%(Grade2) 142618
Figure 1. Box plots showing distribution of visual analogue 50–70% 131119 50–70% 81111 50–70%(Grade3) 121115
scores for image quality from each of the three raters for 70–95% 13 1 8 70–95% 22 616 70–95%(Grade4) 14 212
contrast enhanced MR angiography (CE MRA) and time of 95–99% 113 7 95–99% 117 7 95–99%(Grade5) 2 2 9
flight MR angiography (TOF MRA). 100% 5 4 5 100% 5 4 3 100%(Grade6) 5 8 3
TOF MRA, time of flight MR angiography; CE MRA, contrast
enhanced MR angiography; NASCET, North American
Table 1. Visualization of ECA branches (all values out of 60)
Symptomatic Carotid Endarterectomy Trial; ECST,
TOF MRA CE MRA European Carotid Surgery Trial.

Superior thyroid artery 13(21%) 37(62%)

Lingual artery 11(18%) 49(81%)
TOF MRA, time of flight MR angiography; CE MRA, contrast
enhanced MR angiography.
Observer confidence for assessment of stenosis by
measurement
The median scores for confidence level for assessment
of stenosis by measurement by the three raters were CE
MRA were 4.0, 3.55 and 3.5. Corresponding scores for
TOF MRA were 2.65, 1.95 and 1.7, respectively
(Figure 3). Again, all the differences were highly
statistically significant (p,0.00001). Results of assess-
ment of stenosis by both NASCET and ECST methods by
three observers with both TOF MRA and CE MRA are
given in Table 2.
Image artefacts
Image artefacts observed in the two techniques were
analysed (Table 3). The figures in the table are out of 60
carotid bifurcations analysed in this study.

Figure 2. Box plots showing distribution of visual analogue
scores for observer confidence (for visual estimation of
stenosis) from each of the three raters for contrast enhanced
MR angiography (CE MRA) and time of flight MR angio-
graphy (TOF MRA).
Figure 3. Box plots showing distribution of visual analogue
scores for observer confidence (for stenosis estimation on the
basis of measurements) from each of the three raters for
contrast enhanced MR angiography (CE MRA) and time of
flight MR angiography (TOF MRA).

The British Journal of Radiology, March 2006 203

 D Mitra, D Connolly, S Jenkins et al

Table 3. MR artefacts in the two techniques Discussion

TOF MRA CE MRA The morbidity associated with carotid endarterectomy
is dependent on the complication rate of surgery as well
Slice misregistration 32 NA
Problematic slice 17 NA
as any complication from pre-operative investigation.
misregistration Therefore, if CA is used in the pre-operative evaluation
Pulsation artefact 29 0 of carotid stenosis, a complication rate of 1–2% is added
Venous signal 3 23 to the surgical complication rate of 1–2%. Reducing the
Problematic venous signal 0 1 risk related to CA would, therefore, improve patient
NA, not applicable; TOF MRA, time of flight MR angiogra-
outcome. Furthermore, although CA was used as a gold
phy; CE MRA, contrast enhanced MR angiography. standard in NASCET and ECST, its position as a gold
standard investigation for carotid stenosis is no longer
incontrovertible. The limited projections of carotid
Table 4. Incidence of signal dropouts in the two techniques bifurcation obtained in CA can underestimate the degree
(all values out of 60) of stenosis caused by eccentric plaques. This may be one
TOF MRA CE MRA of the factors causing the reported overestimation of
stenosis by MRA compared with CA [11]. Current
At stenosis 28 11 practice is moving towards non-invasive evaluation of
Beyond stenosis 35 9
degree of stenosis, with CA reserved only for selected
Tortuosity 32 0
cases. In this scenario, it is of utmost importance to
TOF MRA, time of flight MR angiography; CE MRA, contrast optimize the modality of the non-invasive investigation
enhanced MR angiography. to prevent misclassification of patients and the resultant
inappropriate treatment.
The presence of signal dropouts were analysed DUS has been advocated by some investigators [15] as
separately as this particular artefact caused significant a method of evaluating carotid stenosis prior to
problems in the accurate assessment of the degree of endarterectomy. However, DUS is limited by operator
stenosis (Table 4). Lower incidence of signal dropout dependency, difficulty in identifying sub-total occlusion
was noted with CE MRA both at the level of stenosis with very slow flow as well as difficulty in clearly
(46.7% with TOF MRA, 18.3% with CE MRA) as well as defining the morphology of lesions in the carotid
beyond the level of stenosis (58.3% with TOF MRA, 15% bifurcation. In many centres, therefore, DUS is used in
with CE MRA). tandem with MRA, with the latter often being the
13 patients had normal carotid bulbs. Five of these confirmatory investigation [9].
patients demonstrated signal dropout in TOF MRA but Time of flight imaging is a well-established MRA
none showed signal dropout in CE MRA. technique. This is based on the signal generated from the
Plaque ulceration was detected more frequently with inflowing unsaturated protons. As this technique does
CE MRA compared with TOF MRA. An average of 18.6 not require external contrast injection, the image quality
plaque ulcers in 60 carotids were detected with CE MRA does not depend on factors such as the timing of the
compared with 12.3 plaque ulcers in 60 carotids with bolus injection, volume of contrast injected, etc. This
TOF MRA by the three observers. technique also has improved sensitivity to slow flow [16]
and is more accurate in defining the morphology of the
proximal internal carotid artery compared with DUS. We
Interobserver variability have used 2D TOF technique as it has been validated as
an accurate method [9, 17] in this context and was the
Interobserver agreement was measured with the help standardized technique used in our department at the
of intraclass correlation using a two way mixed effect time of the study.
model for absolute agreement. Measurements were Both 2D and 3D TOF MRA, however, have limitations.
made for visual evaluation of stenosis, NASCET grading The most serious limitation is the loss of signal caused by
of stenosis and ECST grading of stenosis (Table 5). The complex flow pattern in the stenotic segment of the
intraclass correlation values for CE MRA (0.893 for visual artery causing over-estimation of the degree of stenosis.
estimation, 0.890 for NASCET grading and 0.800 for In order to produce a signal the inflowing blood should
ECST grading) were consistently higher compared with be perpendicular to the scan plane. However, severe
TOF MRA (0.730, 0.758 and 0.737, respectively). stenosis results in turbulent flow where many of the
protons in the arterial blood are no longer flowing
Table 5. Intraclass correlation values for agreement perpendicular to the scan plane and therefore do not
between observers produce a signal. Furthermore, signal is only produced
by fresh protons flowing into the scan-plane, which have
TOF MRA CE MRA
not received saturation pulses. If, as in a subtotal
Visual estimation 0.730 0.893 occlusion, the flow is slow enough, these protons lose
NASCET grading 0.758 0.890 their signal due to in-plane saturation. Long scan times
ECST grading 0.737 0.800 also result in movement artefacts (mostly due to
TOF MRA, time of flight MR angiography; CE MRA, contrast swallowing) (Figure 4) as well as slice misregistration.
enhanced MR angiography; NASCET, North American In the present study, problematic slice misregistration
Symptomatic Carotid Endarterectomy Trial; ECST, was seen in 17 carotids with TOF MRA technique but
European Carotid Surgery Trial. none with CE MRA technique.

204 The British Journal of Radiology, March 2006

CE MRA and TOF MRA in carotid stenosis

(a) (b)

Figure 4. (a) Time of flight MR angiography (TOF MRA) and (b) contrast enhanced MR angiography (CE MRA) showing oblique
projections of the carotid arteries. Note the significant degradation of the image in TOF MRA (arrows) due to movement, which
is not seen in the CE MRA image.

(a) (b)

Figure 5. (a) Anteroposterior (AP) and (b) oblique projections of contrast enhanced MR angiography (CE MRA) of carotid
arteries showing presence of venous signal (arrows). Note that despite the presence of venous signal the visualization of the
carotid arteries is not significantly impaired.

CE MRA uses the T1 shortening effect of gadolinium to
produce signal from a vessel. It is, therefore, not directly
dependent on flow to produce a signal and has less of the
flow related artefacts described above. However, as an
external contrast agent is administered, the timing of the
injection, volume injected and the flow rate are of crucial
importance. One of the major problems in CE MRA
technique is presence of venous signal, which can cause
difficulty in image interpretation (Figure 5). Venous
signal was seen in three carotids with TOF MRA
technique compared with 23 carotids with CE MRA in
this study. However, in only 1 out of the 23 carotids did
the venous signal prove to be a problem in image
interpretation. This is in variance with another study [14]

The British Journal of Radiology, March 2006 205


where 27% of the contrast enhanced MRA images of the
carotid bifurcation was deemed non-diagnostic due to
masking of the carotid bifurcation by veins.
Specific techniques have been used in order to avoid
this phenomenon and to optimize signal from the
arteries imaged. One such technique is called time
resolved CE MRA where image is acquired repeatedly
at a certain rate in a method akin to DSA (hence the
technique is also known as MR DSA). In this technique
the timing of the bolus injection is not critically
important. In the present study, a bolus-tracking
technique was used in order to optimize the timing of
the injection. This technique is considered to be an
improvement on the time resolved CE MRA [4].
Furthermore, the central k-space data (i.e. the high
amplitude and low resolution data) was acquired first,
to make sure that high contrast information was acquired
while gadolinium was still in the arterial phase.
CE MRA provides a much wider field of view
compared with TOF MRA (Figure 6) and allows assess-
ment from the arch to the base of the skull and if
necessary up to the circle of Willis. This allows the
coverage from CE MRA to be on par with CA and helps
detect any concomitant intracranial disease, which may
alter the decision to proceed to end-arterectomy.
Quality of images in both TOF MRA and CE MRA depend
on the spatial and contrast resolution as well as on the
presence or absence of artefacts. Loss of signal caused by
non-linear flow and in-plane saturation can result in poor
resolution in TOF MRA. Signal dropouts in the stenotic and
post-stenotic segments were also seen more frequently with
the TOF technique than with the CE MRA technique
(Table 4). Signal voids were also seen in some normal bulbs
as well as a significant number of tortuous carotid arteries in
TOF MRA but not seen with CE MRA (Table 4).
(a)
Figure 6. (a) Time of flight MR angiography (TOF MRA) and (b) contrast enhanced MR angiography (CE MRA) showing oblique
projections of the carotid arteries. Note the much larger field of view of CE MRA demonstrating vessels from the arch of the
aorta to the base of the skull and the better demonstration of the external carotid branches on CE MRA (long arrows) compared
with TOF MRA (short arrows).
206
D Mitra, D Connolly, S Jenkins et al
The higher spatial resolution of CE MRA compared
with TOF MRA is also indicated by the better ability to
demonstrate branches of the external carotid artery
(Table 1 and Figure 6). In addition, plaque ulceration
was also seen more frequently by CE MRA technique
than TOF MRA. This is consistent with findings of an
earlier study comparing the two techniques [13].
Slice misregistration, another known problem of 2D
TOF MRA [16], resulted in difficulty in image inter-
pretation in 17 out of 60 carotid bifurcations in this study.
This problem is not encountered with CE MRA, as it is a
volume acquisition technique. With all the factors
described above, it is not surprising that the level of
diagnostic confidence of all three raters have been
consistently higher with CE MRA than TOF MRA, both
for visual estimation and estimation based on measure-
ment (Figures 2 and 3). This is in keeping with the
findings of a similar study [13] where on a scale of 1 to 3
(1 being the best and 3 being the worst technique) the
mean diagnostic confidence score was 1.10 for CE MRA
1.90 for pooled 2D and 3D TOF images (p,0.01).
Any imaging technique also needs to be assessed for
interobserver variability, particularly a relatively new
technique such as CE MRA. High observer variability in
some imaging techniques such as DUS has resulted in
criticism and lack of wide acceptance. Low observer
variability of DSA [18] on the other hand is one of the
factors favouring the use of this technique for pre-
operative carotid stenosis assessment. In the present
study, interobserver variability has been studied
between all three observers as well as for all three
methods of assessment of stenosis, i.e. visual assessment,
calliper measurement by NASCET and calliper measure-
ment by ECST criteria. Interobserver variability was
measured with the help of the intraclass correlation
(b)
The British Journal of Radiology, March 2006
CE MRA and TOF MRA in carotid stenosis
coefficient as it is considered to be a better test than
kappa statistic when there are more than two observers.
Intraclass correlation values are consistently higher in CE
MRA compared with TOF MRA suggesting better
interobserver agreement. With CE MRA, agreement
was best for visual assessment followed by NASCET
grading with the lowest agreement with ECST grading.
Greatest difference between CE MRA and TOF MRA
with regard to intraclass correlation values was in visual
assessment of stenosis with very high agreement in CE
MRA and only moderate agreement in TOF MRA.
Variability between different observations by the same
observer (i.e. intraobserver variability) can also be an
important tool in assessing the reliability of a technique.
This assessment was not included in the present study
and this may be considered a shortcoming of the study.
In current radiological practice, assessment of stenosis
is made by looking at the reconstructed images in a
workstation. However, it is possible that each observer
would use a different set of projections compared with
other observers for estimation of carotid stenosis.
Therefore, in this study, assessment of images was made
from hard copies so that each observer saw exactly the
same MIP projections and therefore eliminated any bias
in the estimation of interobserver variability.
The lack of comparison of the techniques described
with CA could be considered to be a weakness of the
study. CA was not performed in these patients because
this was not part of the normal diagnostic protocol for
assessment of carotid stenosis in the centre where the
study was carried out. In view of the risks associated
with CA, it would have been difficult to obtain ethical
approval to perform CA just for the purpose of the study.
The technique of performing CE MRA has evolved
from the time of this study. The technique of CE MRA
described in this paper was one that was being used at
the time in the department where the study was carried
out. However, the results show that even with the
technique used, CE MRA method was better than TOF
MRA in terms of higher image quality, higher level of
diagnostic confidence and less interobserver variability.
Conclusion
CE MRA provides better image quality, higher level of
diagnostic confidence and less interobserver variability
compared with 2D TOF MRA. The CE MRA technique
has now replaced TOF MRA and CA as the modality of
choice in pre-surgical evaluation of extracranial carotid
stenosis in the centre where this study was carried out.
References
1. Beneficial effect of carotid endarterectomy in symptomatic
patients with high-grade carotid stenosis. North American
Symptomatic Carotid Endarterectomy Trial Collaborators.
N Engl J Med 1991;325:445–53.
2. Randomized trial of endarterectomy for recently sympto-
matic carotid stenosis: final results of the MRC European
Carotid Surgery Trial (ECST). Lancet 1998;351:1379–87.
3. Endarterectomy for asymptomatic carotid artery stenosis.
Executive Committee for the Asymptomatic Carotid
Atherosclerosis Study. JAMA 1995;273:1421–8.
The British Journal of Radiology, March 2006
4. Borisch I, Horn M, Butz B, Zorger N, Draganski B,
Hoelscher T, et al. Preoperative evaluation of carotid artery
stenosis: comparison of contrast-enhanced MR angiography
and duplex sonography with digital subtraction angiogra-
phy. AJNR Am J Neuroradiol 2003;24:1117–22.
5. Cosottini M, Pingitore A, Puglioli M, Michelassi MC, Lupi
G, Abbruzzese A, et al. Contrast-enhanced three-dimen-
sional magnetic resonance angiography of atherosclerotic
internal carotid stenosis as the noninvasive imaging
modality in revascularization decision making. Stroke
2003;34:660–4.
6. Johnson MB, Wilkinson ID, Wattam J, Venables GS,
Griffiths PD. Comparison of Doppler ultrasound, magnetic
resonance angiographic techniques and catheter angiography
in evaluation of carotid stenosis. Clin Radiol 2000;55:912–20.
7. Kollias SS, Binkert CA, Ruesch S, Valavanis A. Contrast-
enhanced MR angiography of the supra-aortic vessels in 24
seconds: a feasibility study. Neuroradiology 1999;41:391–400.
8. Lenhart M, Framme N, Volk M, Strotzer M, Manke C, Nitz
WR, et al. Time-resolved contrast-enhanced magnetic
resonance angiography of the carotid arteries: diagnostic
accuracy and inter-observer variability compared with
selective catheter angiography. Invest Radiol
2002;37:535–41.
9. Patel MR, Kuntz KM, Klufas RA, Kim D, Kramer J, Polak JF,
et al. Preoperative assessment of the carotid bifurcation.
Can magnetic resonance angiography and duplex ultra-
sonography replace contrast arteriography? Stroke
1995;26:1753–8.
10. Remonda L, Senn P, Barth A, Arnold M, Lovblad KO,
Schroth G. Contrast-enhanced 3D MR angiography of the
carotid artery: comparison with conventional digital sub-
traction angiography. AJNR Am J Neuroradiol
2002;23:213–9.
11. Serfaty JM, Chirossel P, Chevallier JM, Ecochard R, Froment
JC, Douek PC. Accuracy of three-dimensional gadolinium-
enhanced MR angiography in the assessment of extracra-
nial carotid artery disease. AJR Am J Roentgenol
2000;175:455–63.
12. Slosman F, Stolpen AH, Lexa FJ, Schnall MD, Langlotz CP,
Carpenter JP, et al. Extracranial atherosclerotic carotid
artery disease: evaluation of non-breath-hold three-dimen-
sional gadolinium-enhanced MR angiography. AJR Am J
Roentgenol 1998;170:489–95.
13. Willig DS, Turski PA, Frayne R, Graves VB, Korosec FR,
Swan JS, et al. Contrast-enhanced 3D MR DSA of the carotid
artery bifurcation: preliminary study of comparison with
unenhanced 2D and 3D time-of-flight MR angiography.
Radiology 1998;208:447–51.
14. Jager HR, Moore EA, Bynevelt M, Coley S, Mounfield P,
Kitchen N, et al. Contrast-enhanced MR angiography in
patients with carotid artery stenosis: comparison of two
different techniques with an unenhanced 2D time-of-flight
sequence. Neuroradiology 2000;42:240–8.
15. Wasserman BA, Haacke EM, Li D. Carotid plaque forma-
tion and its evaluation with angiography, ultrasound, and
MR angiography. J Magn Reson Imaging 1994;4:515–27.
16. Bosmans H, Marchal G, Lukito G, Yicheng N, Wilms G,
Laub G, et al. Time-of-flight MR angiography of the brain:
comparison of acquisition techniques in healthy volunteers.
AJR Am J Roentgenol 1995;164:161–7.
17. Rasanen HT, Manninen HI, Vanninen RL, Vainio P, Berg M,
Saari T. Mild carotid artery atherosclerosis: assessment by
3-dimensional time-of-flight magnetic resonance angiogra-
phy, with reference to intravascular ultrasound imaging
and contrast angiography. Stroke 1999;30:827–33.
18. Gagne PJ, Matchett J, MacFarland D, Hauer-Jensen M,
Barone GW, Eidt JF, et al. Can the NASCET technique for
measuring carotid stenosis be reliably applied outside the
trial? J Vasc Surg 1996;24:449–55; discussion 455–6.
207
 The British Journal of Radiology, 79 (2006), 208–215

Comparison of Radiologists’ confidence in excluding significant

colorectal neoplasia with multidetector-row CT colonography
compared with double contrast barium enema
S A TAYLOR, MD, MRCP, FRCR, S HALLIGAN, MD, MRCP, FRCR, A SLATER, MRCP, FRCR, M MARSHALL, MRCP,
FRCR and C I BARTRAM, FRCP, FRCS, FRCR

Department of Intestinal Imaging, St Mark’s and Northwick Park Hospitals, Harrow, London HA1
3UJ, UK

ABSTRACT. The aim of this study was to compare the confidence of experienced
radiologists in excluding colonic neoplasia with CT colonography (CTC) compared with
barium enema. 78 patients (median age 70 years, range 61–87 years, 44 women)
underwent same day CTC and barium enema. Two radiologists experienced in
reporting barium enema assessed whether the examination had excluded a polyp 6 mm
or greater as ‘‘yes’’, ‘‘probably’’ or ‘‘no’’ for each of 6 colonic segments. Two different
radiologists experienced in CTC independently performed the same assessment on the
CT datasets. Responses were compared using a paired exact test. Formal barium enema
and CT reports were compared with any endoscopic examination performed within
1 year. Studies reporting polyps 6 mm+ in patients not subsequently undergoing
endoscopy were reviewed by two independent observers. Radiologists stated they had
confidently excluded a significant lesion in 314 (71%) and 382 (86%) of 444 segments
with barium enema and CTC, respectively (p,0.001). Confidence was significantly
higher with CTC in the in the descending and ascending colon (p50.02 and p,0.001,
respectively), and caecum (p,0.001). 22 patients underwent some form of endoscopy. Received 6 June 2005
Of five patients with proven colorectal neoplasia (including two with cancer), CTC and Revised 13 July 2005
barium enema correctly identified five and three, respectively. In 56 patients not Accepted 15 July 2005
undergoing endoscopy, CTC reported 17 polyps 6 mm+, of which 16 were
DOI: 10.1259/bjr/99126323
retrospectively classified as definite or probable. 11 could not be identified on the
barium enema, even in retrospect. Confidence in excluding polyps 6 mm or larger is ’ 2006 The British Institute of
significantly greater with CT colonography particularly in the proximal colon. Radiology

Symptoms of colorectal neoplasia are notoriously non-
specific with the result that the majority of patients
investigated do not harbour significant pathology. Even
when applying defined symptom complexes, such as
those specified in the recent ‘‘2 week wait’’ initiative [1],
the prevalence of significant pathology is increased to no
more than 10–15% [2]. Colonoscopy remains the refer-
ence standard whole colon examination but is technically
demanding, invasive, and associated with a small
morbidity and even mortality. Adverse effects are well
documented, largely related to the cardiorespiratory
effects of sedation [3–6], with some evidence of increased
susceptibility amongst the elderly [7].
Radiological alternatives to colonoscopy, including
both CT colonography and barium enema, are generally
viewed as safer, less invasive investigations [8]. Barium
enema remains the standard radiological investigation,
although the day-to-day diagnostic performance in
comparison with CT colonography has not been assessed
in large-scale clinical trials of symptomatic patients.
Advocates of CT colonography point to increased patient
acceptability [9, 10] and extrapolated higher sensitivity
for significant colonic pathology [11, 12]. However, given
that most symptomatic patients will not harbour
significant colonic neoplasia, one important, but often
neglected, consideration is the degree of confidence with
which the reporting radiologist can confirm normality
and thus spare the patient further expensive and
invasive investigations. If CT colonography is to become
the standard radiological investigation, the incidence of
inconclusive examinations should therefore be at least
equal to, and preferably less than that of the barium
enema in elderly symptomatic patients.
The purpose of this study was to compare the
confidence of experienced radiologists in excluding
significant colonic neoplasia with both CT colono-
graphy and barium enema in patients undergoing both
examinations.

Address correspondence to: Dr Stuart Taylor, Department of

Specialist X-ray, Level 2, University College Hospital, 235 Euston Materials and methods
Road, London NW1 2BU, UK.
This research was supported by a research fellowship from the Our local ethical review committee approved the
Royal College of Radiologists. study and all subjects gave informed written consent.

208 The British Journal of Radiology, March 2006

Radiologists’ confidence in excluding significant colorectal neoplasia
All patients 60 years of age or older referred for double
contrast barium enema between July 2002 and December
2003 were identified from clinical request cards sent to
the Department of Radiology. Only those referred
because of a clinical suspicion of colorectal neoplasia
were eligible for inclusion. All eligible patients were then
invited by letter to additionally undergo CT colonogra-
phy immediately before barium enema. A total of 78
patients (median age 70 years, range 61–87 years, 44
women) were recruited. Reasons for referral were as
follows: change in bowel habit (n560); iron deficiency
anaemia (n510); palpable abdominal mass (n58).
CT colonography
CT colonography was performed immediately prior to
same day barium enema. Patients underwent the standard
bowel purgation regimen used at our institution, consist-
ing of 24 h of a clear liquid diet together with two sachets
of sodium picosulphate/magnesium citrate (Picolax;
Ferring Pharmaceuticals, Berkshire, UK). No tagging
agents were used. All but two patients received 20 mg of
intravenous hyoscine butylbromide (Buscopan;
Boehringer Ingelheim, Bracknell, UK) prior to gas insuf-
flation. The remaining two patients received intravenous
glucagon (Nova Nordisk Pharmaceuticals, Crawley, UK)
because of a contraindication to hyoscine butylbromide
(both due to recent acute cardiovascular events). Colonic
insufflation was performed with carbon dioxide using an
automatic insufflator (Protocol; E-Z-EM, Westbury, NY).
Insufflation occurred at a rate of 1–2 l min21 with a
maximum intracolonic pressure of 25 mmHg, set using the
pump controls, and was continued until patient discom-
fort, or if distension was deemed adequate by the
supervising radiologist from the CT scout image.
Patients were then scanned in the supine position using
a four detector row CT scanner (Lightspeed plus; General
Electric Medical Systems, Milwaukee, WI) and the follow-
ing parameters: 2.5 mm collimation; pitch of 1.5; 120 kVp;
50 mA; 50% slice overlap. Patients were turned prone and
further gas insufflated if a second scout image suggested
areas of collapse. A scan in the prone position was then
performed using identical CT parameters. Intravenous
contrast was not administered.
Barium enema
After CT colonography was complete, patients were
escorted from the CT scanner to the fluoroscopy suite.
Appointment times were such that there was at least 1 h
between completion of CT colonography and commence-
ment of barium enema. Barium enemas were performed
by either one of three experienced radiographers (68
patients), or by a radiology trainee (10 patients) according
to a standard protocol consisting of multiple digital
fluoroscopic spot views of the double-contrasted colon
followed by two lateral decubitus over-couch radiographs.
The barium preparation (94% w/w, PolibarTM; E-Z-EM,
Westbury, NY) was diluted with 700 ml water and
instilled via a rectal catheter. Colonic distension was
achieved with carbon dioxide introduced by manual
compression of the gas-filled enema bag. Patients received
The British Journal of Radiology, March 2006
a second identical dose of the spasmolytic that had been
administered for CT colonography (either hyoscine butyl-
bromide or glucagon) prior to the barium enema.
Image analysis: CT colonography
Image analysis was performed using a dedicated
workstation with proprietary software (Advantage
Windows 4.0 and Colonography; GE Medical Systems,
Milwaukee, WI). Two radiologists experienced in CT
colonography (defined by prior reading of at least 150 CT
colonographic datasets with full endoscopic correlation)
independently analysed the CT datasets. Reader one
read the first 36 patients and reader two the second 42
patients. Analysis was performed using primary analysis
of two-dimensional (2D) axial supine and prone images
with multiplanar reformats and 3D endoluminal views
reserved for problem solving. For the purpose of the
study the colon was divided into six segments using
previously published criteria [13]. Readers noted the
presence of diverticular disease or colonic neoplasia in
each of the six segments on a study sheet designed for
the trial. Colorectal neoplasia and diverticular disease
were defined using previously well-established criteria
[14, 15]. A formal CT report was also generated for the
referring clinician as per usual practice.
Readers additionally independently assessed each
colonic segment as to whether they could answer the
clinical question ‘‘has the test excluded a significant
colonic lesion?’’ For the purposes of the trial, a significant
colonic lesion was defined as a polyp 6 mm or larger.
The readers graded their response for excluding a
significant lesion as ‘‘yes’’, ‘‘probably’’ or ‘‘no’’. If the
response was ‘‘probably’’ or ‘‘no’’, readers listed reason
for non-exclusion as ‘‘fluid’’, ‘‘poor distension’’ or
‘‘faecal residue’’. A significant lesion was by definition
not excluded (i.e. ‘‘no’’) if such a lesion was reported as
being present in that particular segment.
Image analysis: barium enema
All barium enemas were reported on the day they
were performed by one of two experienced readers
(defined as a radiologist with a declared subspecialty
interest in gastrointestinal radiology with at least 5 years
experience of reporting more than 4 barium enema
examinations per week). These readers were different
from the radiologists analysing the CT scans and were
blinded to the CT report. Individual readers single read
the barium enema studies as they appeared on their
clinical lists and a formal report was generated for the
referring clinician as per usual practice. Readers one and
two read 40 and 38 studies, respectively.
For the purposes of the trial, the colon was again divided
into six segments using the same criteria as for CT
colonography. Readers noted the presence of diverticular
disease or colonic neoplasia in each of the six segments on a
study sheet designed for the trial, identical to that for the CT
readers. As for the CT, readers additionally independently
assessed each colonic segment as to whether they could
answer the clinical question ‘‘has the test excluded a
significant colonic lesion?’’ (polyp 6 mm or larger), listing
209
 S A Taylor, S Halligan, A Slater et al

their response as ‘‘yes’’, ‘‘probably’’ or ‘‘no’’. If the response ‘‘probably’’ responses were combined into a single group
was ‘‘probably’’ or ‘‘no’’, readers listed reason for non- and compared with the ‘‘yes’’ responses. The first set of
exclusion as ‘‘poor barium coating’’, ‘‘poor distension’’, analyses were performed for each segment of the colon
‘‘barium pool’’ or ‘‘faecal residue’’. Again, a significant separately using a paired exact test (binomial based exact
lesion was by definition not excluded (i.e. ‘‘no’’) if such a test).
lesion was reported by the radiologist for that particular The effect of patient age (categorised into 65 or less versus
segment. greater than 65) upon radiologist confidence was also
examined using Fisher’s Exact test separately for the two
procedures. The effects of individual readers on confidence
Endoscopic correlation scores were then compared for both barium enema and CT,
and any effect on who had performed the barium enema
After trial completion, a non-observer searched the local (radiographer or radiology registrar) was sought.
endoscopic database to ascertain if patients had undergone Confidence scores from all six segments were then
any form of endoscopy within 1 year of the barium enema combined into a single analysis. Because segments in
and CT scan (either before or after). There was a time individual patients are not wholly independent of each
period of 6 months between the CT/barium enema of the other, logistic regression with robust standard errors was
last patient recruited and the database search. The trial used for the analysis and any effect of patient age, who
study sheets were then correlated with the endoscopic performed test and who reported the test was sought by
report to derive the CT and barium enema sensitivity and adding each factor to the basic regression model.
false positive rate for colorectal neoplasia. A radiologically
detected polyp was deemed true positive if a correspond-
ing polyp was found in the same segment at endoscopy
Results
and if the estimated size of the polyp agreed as follows; for
polyps less than 6 mm at endoscopy radiological measure-
ment was within ¡90%, for polyps 6–9 mm radiological Radiologist confidence
measurement was within ¡70%, and for polyps 10 mm or A total of four patients did not tolerate one of the two
greater radiological measurement was within ¡50%. A tests (four failed barium enema and one also failed CT
radiologically detected polyp was deemed false positive if colonography) and were excluded. Two of the four
either no polyp was found in the corresponding segment patients were intolerant of colonic distension (including
during subsequent endoscopy or if the measured size fell the one who failed CT colonography) and two were
outside the above criteria. If endoscopy had preceded insufficiently mobile to undergo barium enema. A total
imaging, endoscopically removed polyps were excluded of 74 patients were thus left for analysis.
from the comparison. All readers were blinded to the Overall, the reporting radiologists stated they had
endoscopic data. confidently excluded a significant lesion in 314 of 444
segments (71%) with barium enema and in 382 of 444
segments (86%) with CT colonography (p,0.001).
Radiological review Reasons for non-exclusion (other than reporting a
lesion) with barium enema were residue: 41%, poor
All cases where a lesion at least 6 mm had been reported coating: 12%, barium pools: 32% and poor distension:
on either the CT colonography, or barium enema and yet 15%. Reasons for non-exclusion (other than reporting a
the patient had not subsequently been referred for lesion) with CT colonography were residue: 35%, fluid
endoscopy were identified and reviewed. An independent pools: 20% and poor distension: 45%.
observer, experienced in CT colonography with audited The number of individual segments in which a lesion
performance in line with the published literature, reviewed
was confidently excluded is shown in Table 1.
the CT colonography datasets, and another expert gastro- Radiologists reporting CT colonography were signifi-
intestinal radiologist reviewed the barium enema, both cantly more likely to confidently exclude a significant
unblinded to the original study reports. If the lesion(s) had lesion in the descending and ascending colon (p50.02
been reported on CT colonography alone, the abnormality
was found in the CT colonography dataset and classified as
‘‘definite’’, ‘‘probable’’ or ‘‘likely false positive’’ by the Table 1. Radiologist confidence at excluding a significant
independent CT observer. The barium enema was then colonic lesion for barium enema and CT colonography
according to colonic segment
carefully reviewed to see if the lesion was in retrospect
‘‘definitely present’’, ‘‘probably present’’ or ‘‘not identi- Segment Lesion excluded Lesion excluded p-value
fied’’. If the lesion(s) was identified on barium enema on barium enema on CT
[patients (%)]a [patients (%)]a
alone, the same process was undertaken in reverse. Lesions
reported on both CT colonography and barium enema Rectum 64 (86) 69 (93) 0.27
were classified as ‘‘definite’’, ‘‘probable’’ or ‘‘likely false Sigmoid 49 (67) 52 (71) 0.69
positive’’ by the independent observers for each modality. Descending 63 (85) 70 (94) 0.02
Transverse 53 (72) 61 (82) 0.13
Ascending 44 (59) 65 (87) ,0.001
Caecum 41 (55) 65 (88) ,0.001
Statistical analysis
n574.
a
For the purposes of analysis of radiologist confidence at Figure refers to the number of ‘‘yes’’ responses to whether a
excluding a significant colonic lesion, the ‘‘no’’ and significant lesion was confidently excluded.

210 The British Journal of Radiology, March 2006

Radiologists’ confidence in excluding significant colorectal neoplasia

and p,0.001, respectively) and caecum (p,0.001) com-

pared with those reporting barium enema. There was no
significant effect of who had performed the barium
enema (p50.27), or individual reader (p50.35) on overall
confidence scores for the barium enema. Similarly there
was no significant difference between confidence scores
for the two CT colonography readers (p50.72).
Confidence at excluding a significant lesion was not
significantly affected by patient age on an individual
segmental basis for either test, or overall for barium
enema. However, overall confidence was significantly
higher with CT colonography for patients 65 or less
compared with those over 65 (odds of excluding a lesion
0.42 (confidence interval 0.20 to 0.89), p50.02).

Endoscopic correlation
Of the cohort of 78 patients, a total of 22 underwent
some form of endoscopy within 1 year of the CT and
barium enema. Of the 22 patients 10 underwent colono-
scopy as a result of reported positive findings on CT
colonography and/or barium enema. Of these 10, 2
colonoscopies were incomplete proximal to the reported
abnormality (small polyps up to 8 mm) and have not
been repeated. The results of the eight completed
endoscopies in comparison with the radiological find-
ings are shown in Table 2. All radiologically detected
polyps fell within the size criteria listed above for
positive correlation with the endoscopic findings. On a Figure 1. Spot view from a double contrast barium enema
demonstrates a large filling defect (arrows) classified as a
per patient basis, CT colonography correctly identified
polyp by the reader. Subsequent histology confirmed
all four patients with endoscopically proven polyps (one invasive carcinoma.
with a single 12 mm sigmoid polyp, one with a rectal
cancer and 10 mm ascending colon polyp, and two with
several small polyps less than 5 mm) whereas barium patients the limited endoscopy, subsequent barium
enema detected two of the four (missing the two patients enema and CT colonography were all reported as
with polyps up to 5 mm). CT colonography correctly normal. In one patient with colonoscopy complete to
identified a histologically confirmed rectal cancer, the distal transverse colon, a caecal cancer was correctly
although the same lesion was reported as a polyp on diagnosed by both barium enema and CT colonography,
barium enema (Figure 1). In the four patients with the latter revealing multiple liver metastasis. In the
confirmed neoplasia there were two presumed CT false remaining patient with long-standing Crohn’s disease
positives (10 mm and 6 mm). CT colonography did and weight loss, CT colonography and barium enema
however suggested a total of six polyps (three 6–9 mm both confirmed a mid transverse colon stricture.
and three 1–5 mm) in four patients in whom both the Whereas barium enema confidently diagnosed a
barium enema and subsequent colonoscopy were Crohn’s stricture, CT colonography was unable to
reported as normal and were therefore classified as false exclude cancer (Figure 3). Subsequent biopsy excluded
positives for CT (Figure 2). malignancy.
The remaining 12 of the 22 patients underwent either The remaining 56 patients did not undergo any form
an incomplete colonoscopy or a flexible sigmoidoscopy of endoscopy either prior or subsequent to the radio-
prior to the barium enema, which was requested by the logical tests. Diverticular disease was reported in 26 on
clinician to assess the non-visualized colon. In 10 of these CT colonography and in 30 on barium enema.

Table 2. Findings of complete colonoscopy performed as a result of reported abnormal radiological (CT colonography or
barium enema) findings
Pathology Colonoscopic CT detection (%) Barium enema CT false positives Barium enema false
findings detection (%) positives

Cancer 1 1 (100) 1 (100)a 0 0

Polyp 1–5 mm 10 3 (30) 0 (0) 3 0
Polyp 6–9 mm 0 N/a N/a 4 1
Polyp 10 mm+ 2 2 (100) 2 (100) 1 0
N/a, not applicable.
a
Cancer detected by barium enema but reported as a polyp.
n58

The British Journal of Radiology, March 2006 211

 S A Taylor, S Halligan, A Slater et al

(a) (b)

Figure 2. Presumed CT colonographic false positive. (a) Axial view and (b) CT colonographic endoluminal view demonstrates a
6 mm filling defect (arrows) reported as a polyp but not found on subsequent colonoscopy.

(a) (b)

Figure 3. Transverse colonic Crohn’s stricture. (a) Double contrast barium enema demonstrated the stricture (arrows) correctly
classified as benign by the reader. (b) Axial CT colonographic image shows the short thick walled stricture (arrows) reported as a
possible cancer by the reader.

Radiological review Barium enema reported just one 6 mm polyp (not

In 56 patients, CT colonography reported 27 polyps in
19 patients (7: 10 mm+, 10: 6–9 mm and 10: 1–5 mm). Of
the 17 polyps 6 mm+ reported on CTC, 10 were classified
as definite, 6 as probable and 1 as a false positive, on
retrospective review. Of the 16 polyps re-classified as
probable or definite on review, 11 could not be identified
on the barium enema, even in retrospect, including 4 of 7
polyps 10 mm+ (Table 3).
reported on CT, even on review) in the 56 patients.
Discussion
Radiological colonic imaging is generally regarded as
safer and less invasive than total colonoscopy, particu-
larly in patients with attendant comorbidity. Although,
quite rightly, much emphasis is placed on the sensitivity

212 The British Journal of Radiology, March 2006

Radiologists’ confidence in excluding significant colorectal neoplasia
Table 3. Retrospective independent observer classification
of polyps 6 mm plus seen exclusively on CT colonography in
patients without subsequent endoscopy
Polyp size CT classificationa Barium enema classification
Definite Probable Not seen Total
6–9 mm Definite 0 1 4 5
Probable 0 1 3 4
10 mm+ Definite 1 0 4 5
Probable 2 0 0 2
a further investigation in elderly patients who were often
Excluding one 7 mm polyp re-classified as a false positive on
CT colonography.
of any particular technique, the ability of the test to
confidently confirm normality is also an important
consideration given that most symptomatic patients do
not harbour significant pathology. Assuming it is
technically complete, a normal barium enema or CT
colonography is usually sufficient to spare the cost and
risks of additional total colonoscopy in most patients
with non-specific symptoms.
We found that experienced radiologists had signifi-
cantly greater confidence in excluding a lesion 6 mm or
larger with CT colonography than with barium enema,
particularly in the proximal colon. Adequate visualiza-
tion of the ascending colon and caecum is often difficult
with barium enema, particularly in frail, immobile
patients, mainly due to difficulties in barium filling
and achieving the correct balance between adequate
coating and unwanted liquid pools. Incomplete exam-
inations are therefore relatively frequent in this patient
group [16]. This difficultly is less apparent during CT
colonography, when all that is required is gaseous
distension of the proximal colon, something that can
usually be achieved reliably [13, 17, 18]. The data suggest
therefore that CT colonography is technically more
‘‘forgiving’’ than barium enema in older symptomatic
patients. This has direct clinical implications, particularly
as a combination of flexible sigmoidoscopy and barium
enema has frequently been advocated in symptomatic
patients [19, 20]. Our results suggest that experienced
radiologists may be more confident when excluding
significant pathology with CT colonography rather than
barium enema in those undergoing limited endoscopy.
Diagnostic confidence was almost identical in the
sigmoid for barium enema and CT, suggesting this
segment remains problematic, although recent data
suggest CT colonography is as effective as flexible
sigmoidoscopy for detecting significant lesions in
patients presenting with rectal bleeding [21].
Residual fluid/barium or faecal residue generally
decreased diagnostic confidence for CT and barium
enema in similar proportion, although this affected more
patients during barium enema overall. Interestingly,
poor distension was the most common reason for
inability to confidently exclude a significant lesion with
CT colonography. The use of supine and prone imaging
[17, 18] and spasmolytic [13] have all been shown to
improve distension during CT colonography, but it is
clear that any further improvements would still have
significant impact on diagnostic confidence.
We did find evidence that referring clinicians consider
a negative radiological test sufficiently reassuring to halt
The British Journal of Radiology, March 2006
colonic investigation in symptomatic patients: no patient
with a normal CT or barium enema went on to
subsequent endoscopy. Indeed, 19 patients with abnor-
mal CT reports (including 6 with suspected lesions at
least 10 mm in size) have had no further colonic
investigation. By way of explanation, all these patients
had normal barium enema reports and the clinician may
be more comfortable with this rather than the newer
technology. However, it seems likely that two radiologi-
cal investigations negative for cancer was enough to halt
frail.
The number of patients undergoing full colonoscopy
was insufficient for any meaningful analysis of the
diagnostic performance of CT colonography versus
barium enema. CT identified all patients with colonic
neoplasia on endoscopy whereas barium enema missed
two (albeit with small and clinically insignificant
polyps). However, of the two patients proven to have
colorectal cancer (one of which was diagnosed only after
incomplete colonoscopy), CT diagnosed both (and
confirmed metastatic spread in one), whereas barium
enema incorrectly classified a rectal cancer as a polyp. CT
colonography did raise the possibility of a cancer in a
patient with a Crohn’s stricture, which was confidently
reported as benign on barium enema, emphasising the
benefit of barium enema for visualizing mucosal detail
and the problems of using CT colonography in patients
with inflammatory bowel disease.
CT colonography reported polyps in 29 (37%) patients
compared with just 4 (5%) on barium enema. Although
four patients had a subsequent normal colonoscopy
suggesting CT false positives (although colonoscopy is
an imperfect reference standard [22, 23]), the majority
have not undergone endoscopic examination to confirm
or refute the CT findings. All but one of the 17 polyps
6 mm plus initially reported on CT colonography in
patients not undergoing endoscopy were classified as
definite or probable on review by an independent
observer, although only 4 were seen in retrospect on
the barium enema. It could be argued that some of these
additional polyps are false positives and if CT colono-
graphy were used alone could trigger unnecessary
endoscopy. However, the prevalence of polyps even in
an asymptomatic screening cohort over 50 is around 30–
40% [24, 25], which perhaps gives some weight to the
assertion that CT colonography was more sensitive than
barium enema. Furthermore, there is increasing evidence
of CT colonography’s superior performance in polyp
detection compared with barium enema [26]. Of course
in the elderly symptomatic population, even sizeable
polyps (over 1 cm) are likely incidental and a more
sensitive test, such as CT colonography, does not always
improve the outcome for patients (and indeed may
worsen it if clinicians feel duty bound to ‘‘chase’’
incidental polyps reported on CT colonography in this
vulnerable patient group). Conversely a 12 mm caecal
polyp in a fit 70-year-old for example may well be highly
significant for that individual. There needs to be a clear
understanding between radiologist and referring clini-
cian as to the lesion size threshold reported and the
significance of findings in individual patients.
Our study does have weaknesses. It is possible that the
prior CT colonography adversely affected the quality of
213

subsequent barium enema. However carbon dioxide
(which is readily absorbed through colonic mucosa) was
used for CT colonography and there was at least 1 h
between the two tests. Anecdotally those performing the
enemas did not report additional problems in study
patients, but we cannot exclude negative effects on the
quality of the barium enemas. Different radiologists
graded the barium enema and CT colonographic studies
and it is possible that the confidence scores for excluding
neoplasia were merely a reflection of the personalities of
the individual radiologist rather than a comparison
between the two tests. However, all radiologists had
good experience of the technique they were reporting
and importantly there were no statistically significant
differences between levels of confidence for each of the
two separate readers for each test, suggesting the scores
reflected the procedure more than the radiologist. It
would, however, be useful for further work to investi-
gate intraobserver confidence in those trained in both CT
colonography and barium enema. Although within the
context of a trial we demonstrated greater confidence for
excluding lesions 6 mm and over with CT colonography,
it is clear that any level of uncertainty was often not
reflected in the formal report sent to the referring
clinician. In other words, even if the radiologist
reported on the trial sheet that a significant lesion could
not be excluded in, say, the caecum, this reduced
confidence was not portrayed in the issued report. It is
therefore questionable whether improved confidence
with CT colonography would necessarily have a direct
clinical impact in reducing unnecessary colonoscopy
after technically imperfect barium enema. Such data
will come from larger randomized trials currently in
progress.
It is acknowledged that our definition of a significant
lesion as anything 6 mm or larger is relatively wide and
accept that there is a significant difference in not being
able to exclude, say, an 8 mm polyp, compared with a
20 mm polyp. In essence our definition required the
radiologist to be confident he/she could exclude a 6 mm
lesion. However, detection of relatively small lesions is a
reflection of the overall capabilities of the examination,
especially when compared with endoscopy. Finally, as
discussed above, CT colonography reported polyps in
many more patients than barium enema and theoreti-
cally could actually act to increase endoscopic referral if
the results are not viewed wisely in clinical context by
clinicians. However, it must be remembered that a small
but significant number of polyps around the centimetre
mark will harbour cancer [27] and as life expectancy
increases such lesions will assume greater importance in
older individuals.
In conclusion, radiologist confidence in excluding
polyps 6 mm or larger is significantly greater with CT
colonography than barium enema, particularly in the
proximal colon.
Acknowledgments
This research was supported by a research fellowship
from the Royal College of Radiologists.
The authors would like to thank Paul Bassett for his
statistical advice.
214
S A Taylor, S Halligan, A Slater et al
References
1. Thompson M. ACPGBI Referral guidelines for colorectal
cancer. Colorectal Dis 2002;4:287–97.
2. Flashman K, Faux W, O’Leary D, Senapati A, Thompson M.
Diagnostic yields of the six Department of Health (DOH)
higher risk criteria for the ‘‘two week standard’’ clinic
(TWSC). Colorectal Dis 2002;3 (Suppl.) poster 41.
3. Eckardt VF, Kanzler G, Schmitt T, Eckardt AJ, Bernhard G.
Complications and adverse effects of colonoscopy with
selective sedation. Gastrointest Endosc 1999;49:560–5.
4. Holm C, Christensen M, Rasmussen V, Schulze S,
Rosenberg J. Hypoxaemia and myocardial ischaemia dur-
ing colonoscopy. Scand J Gastroenterol 1998;33:769–72.
5. Ristikankare M, Julkunen R, Mattila M, et al. Conscious
sedation and cardiorespiratory safety during colonoscopy.
Gastrointest Endosc 2000;52:48–54.
6. Thompson AM, Park KG, Kerr F, Munro A. Safety of
fibreoptic endoscopy: analysis of cardiorespiratory events.
Br J Surg 1992;79:1046–9.
7. Yano H, Iishi H, Tatsuta M, Sakai N, Narahara H, Omori M.
Oxygen desaturation during sedation for colonoscopy in
elderly patients. Hepatogastroenterology 1998;45:2138–41.
8. Taylor SA, Halligan S, O’Donnell C, et al. Cardiovascular
effects at multi-detector row CT colonography compared
with those at conventional endoscopy of the colon.
Radiology 2003;229:782–90.
9. Gluecker TM, Johnson CD, Harmsen WS, et al. Colorectal
cancer screening with CT colonography, colonoscopy, and
double-contrast barium enema examination: prospective
assessment of patient perceptions and preferences.
Radiology 2003;227:378–84.
10. Taylor SA, Halligan S, Saunders BP, Bassett P, Vance M,
Bartram CI. Acceptance by patients of multidetector CT
colonography compared with barium enema examinations,
flexible sigmoidoscopy, and colonoscopy. AJR Am J
Roentgenol 2003;181:913–21.
11. Sosna J, Morrin MM, Kruskal JB, Lavin PT, Rosen MP,
Raptopoulos V. CT colonography of colorectal polyps: a
metaanalysis. AJR Am J Roentgenol 2003;181:1593–8.
12. de Zwart IM, Griffioen G, Shaw MP, Lamers CB, de Roos A.
Barium enema and endoscopy for the detection of colorectal
neoplasia: sensitivity, specificity, complications and its
determinants. Clin Radiol 2001;56:401–9.
13. Taylor SA, Halligan S, Goh V, et al. Optimizing colonic
distention for multi-detector row CT colonography: effect of
hyoscine butylbromide and rectal balloon catheter.
Radiology 2003;229:99–108.
14. Fenlon HM, Clarke PD, Ferrucci JT. Virtual colonoscopy:
imaging features with colonoscopic correlation. AJR Am J
Roentgenol 1998;170:1303–9.
15. Fletcher JG, Johnson CD, MacCarty RL, Welch TJ, Reed JE,
Hara AK. CT colonography: potential pitfalls and problem-
solving techniques. AJR Am J Roentgenol 1999;172:1271–8.
16. Tinetti ME, Stone L, Cooney L, Kapp MC. Inadequate
barium enemas in hospitalized elderly patients. Incidence
and risk factors. Arch Intern Med 1989;149:2014–6.
17. Chen SC, Lu DS, Hecht JR, Kadell BM. CT colonography:
value of scanning in both the supine and prone positions.
AJR Am J Roentgenol 1999;172:595–9.
18. Yee J, Kumar NN, Hung RK, Akerkar GA, Kumar PR, Wall
SD. Comparison of supine and prone scanning separately
and in combination at CT colonography. Radiology
2003;226:653–61.
19. Mendelson RM, Kelsey PJ, Chakera T. A combined flexible
sigmoidoscopy and double-contrast barium enema service:
initial experience. Abdom Imaging 1995;20:238–41.
20. Rex DK, Mark D, Clarke B, Lappas JC, Lehman GA. Flexible
sigmoidoscopy plus air-contrast barium enema versus
colonoscopy for evaluation of symptomatic patients without
evidence of bleeding. Gastrointest Endosc 1995;42:132–8.
The British Journal of Radiology, March 2006
Radiologists’ confidence in excluding significant colorectal neoplasia
21. Taylor SA, Halligan S, Vance M, Windsor A, Atkin W,
Bartram CI. Use of multidetector-row computed tomo-
graphic colonography before flexible sigmoidoscopy in the
investigation of rectal bleeding. Br J Surg 2003;90:1163–4.
22. Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss
rates of adenomas determined by back-to-back colonosco-
pies. Gastroenterology 1997;112:24–8.
23. Pickhardt PJ, Choi JR, Hwang I, et al. Computed tomographic
virtual colonoscopy to screen for colorectal neoplasia in
asymptomatic adults. N Engl J Med 2003;349:2191–200.
24. Rex DK, Lehman GA, Ulbright TM, et al. Colonic neoplasia
in asymptomatic persons with negative fecal occult blood
The British Journal of Radiology, March 2006
tests: influence of age, gender, and family history. Am J
Gastroenterol 1993;88:825–31.
25. Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H,
Chejfec G. Use of colonoscopy to screen asymptomatic
adults for colorectal cancer. Veterans Affairs Cooperative
Study Group 380. N Engl J Med 2000;343:162–8.
26. Rockey DC, Paulson E, Niedzwiecki D, et al. Analysis of air
contrast barium enema, computed tomographic colonogra-
phy, and colonoscopy: prospective comparison. Lancet
2005;365:305–11.
27. Muto T, Bussey HJ, Morson BC. The evolution of cancer of
the colon and rectum. Cancer 1975;36:2251–70.
215
 The British Journal of Radiology, 79 (2006), 216–220

The Bristol Hip View: a new hypothetical radiographic projection

for femoral neck fractures
M BRADLEY, FRCR, M SHAW, FRCR and D FOX, FRCR

Department of Radiology, Southmead Hospital, North Bristol Trust, Westbury on Trym, Bristol BS10
5NB, UK

ABSTRACT. This experimental study is to evaluate a modified radiographic view of the

femoral neck in the assessment of femoral fractures. A dry femur and pelvis was set up
in a rig to simulate the positioning of a routine anteroposterior (AP) pelvis X-ray view.
Films were exposed to create a routine AP pelvis, AP hip and two views with external
tube angulation of 15 ˚ and 30 ˚. Observers were asked to evaluate the films using a
visual analogue score on two separate occasions. The same films were performed on a
further fractured femoral neck to assess the fracture clarity. There was good Received 6 May 2005
intraobserver and interobserver correlation. Observers ranked the 15 ˚ and 30 ˚ angled Revised 28 June 2005
films as showing the femoral neck most clearly, over and above the traditional views Accepted 15 July 2005
(p,0.001). The fracture was best demonstrated on the 30 ˚ angled film (p,0.001). The
DOI: 10.1259/bjr/31965396
30 ˚ angled view appears to demonstrate the femoral neck anatomy more clearly than
the traditional views but also showed increased fracture sharpness. The authors are ’ 2006 The British Institute of
proceeding to a clinical trial to assess this in trauma practice. Radiology

Our standard departmental policy for radiographs for reasons. Angles of internal and external foot rotation
the patient with a suspected femoral neck fracture is an were studied.
anteroposterior (AP) pelvis with lateral hip of the A disarticulated femur and pelvis was assembled in a
symptomatic side. The geometry of the AP film means rig to closely simulate the AP pelvis with feet in-turned.
that the angle of incidence of the central beam to the The femoral neck angle was set according to the mean
femoral neck is in the order of 70 ˚. An AP hip view data measured from the CT.
centred on the head then reaches approximately 75 ˚. Four films were then exposed to create the standard
Ideally a 90 ˚ angle should be obtained for the optimum AP pelvis (5D), AP hip (5B), 15 ˚ angled beam towards
visualization of the femoral neck. femoral head (5A), 30 ˚ angled beam towards femoral
It has been observed that when the diagnosis is in head (5C) (A and C were centred on the femoral head)
doubt due to difficulty with identifying the fracture, a (Figure 2).
view, similar to the Judet obturator oblique view, can be A second femur was fractured at right angles through
useful in delineating the fracture. In order to assess this the mid femoral neck using an osteotome and then glued
observation an experimental study was set up with together anatomically. The rig was set up in the same
cadaveric bones. The study compared four different fashion with CT confirming the same femoral neck angle.
views, two representing the AP pelvis and AP hip and The same four films were then exposed to demonstrate
two new angled views to obtain angles of incidence of
90 ˚ and 105 ˚ to the femoral neck (Figure 1).
Femoral neck fractures may result in varying degrees
of external rotation of the lower limb due to unopposed
action of the gluteus maximus, piriformis, obturator
internus and gemelli muscles on the femur. The second
aim of our study was to assess the affect of changes in
external rotation of the lower limb on the femoral neck
angle relative to a base line of the anterior inferior iliac
spine (AIIS). This external rotation could have a direct
affect on the angle of incidence of the X-ray beam to the
femoral neck.

Methods
Ethics committee approval was granted. This study
was largely performed experimentally.
CT was used to measure femoral neck angles on
patients who were undergoing CT for valid clinical
Figure 1. Diagrammatic representation of the typical beam
incident angles for a routine anteroposterior (AP) pelvis (70 ˚
to femoral neck), AP hip, (75 ˚ to femoral neck), beam angle
of 15 ˚ (90 ˚ to femoral neck), and 30 ˚ beam angle (105 ˚ to
femoral neck).

216 The British Journal of Radiology, March 2006

The Bristol Hip View

(a) (b)

(c) (d)

Figure 2. (a) Film of femoral neck obtained using a 15 ˚ angled beam. (b) Film of femoral neck simulating an anteroposterior
(AP) hip. (c) Film of femoral neck obtained using a 30 ˚ angled beam. (d) Film of femoral neck obtained simulating an AP pelvis.

The British Journal of Radiology, March 2006 217

 M Bradley, M Shaw and D Fox

(a) (b)

Figure 3. (a) Film (Y) of fracture with least sharpness (view equivalent to an anteroposterior (AP) pelvis). (b) Film Z showing the
greatest fracture sharpness (30 ˚ angulation to the femoral head).

the neck and fracture (Figure 3). These were randomly The CT data measured femoral neck angles relative to
labelled W, X, Y, Z (Table 1). the AIIS with internal and external rotation. This was to
Blinded observers were asked to fill in a questionnaire evaluate whether a correction angle was needed to be
based on the four X-ray views randomly displayed for added to the new views to ensure consistency of 90 ˚
both rigs using a visual analogue scoring scale. The same beam incidence to the femoral neck when patients with
observers repeated the process a month later to show neck fractures present with limb shortening and external
intraobserver consistency. A variety of observers were rotation.
asked including; radiologists, orthopaedic surgeons
(both consultant and SPR), accident and emergency
consultants and senior radiographers. Results
The observers were asked to assess the clarity of
visualization of the femoral neck (sub-capital, mid neck 46 observers were randomly shown the two sets of films;
and intratrochanteric) and the sharpness of the fracture. 10 radiology consultants, 8 specialist registrars, 10 ortho-
paedic consultants, 9 middle graders, 8 senior radio-
graphers and 1 consultant emergency physician.
Analysis showed no statistically significant differences
Table 1. Angle of incident beam for each view, modified
(Kappa) between the two occasions of observation
and traditional
(p,0.001) or between grade/speciality of observer; i.e.
Radiograph Angle of incidence Anatomical Fracture excellent intraobserver and interobserver correlation.
(tube to femoral neck film neck 71% of observers ranked A and C as best.
angulation)
The questionnaire tried to differentiate between the
AP Pelvis (0 ˚) 70 ˚ D Y sub-capital, mid-cervical and intratrochanteric areas to
AP Hip (0 ˚) 75 ˚ X X see if any particular film out-performed in any one area.
Hip (15 ˚) 90 ˚ A W Pairwise comparisons of the means, using the
Hip (30 ˚) 105 ˚ C Z
Bonferroni correction for multiple comparisons, revealed
AP, anteroposterior. the following:

218 The British Journal of Radiology, March 2006

The Bristol Hip View
Sub-capital
A significantly out-performed B, C and D (p,0.001)
and B, C, D were not significantly different (p,0.05).
Mid-cervical
A performed similarly to C (p,0.13) and both were
significantly better than B and D (p,0.001).
Intratrochanteric
All films performed similarly with no statistical
variation (Figure 4). The fracture sharpness was better
demonstrated on W and Z than X and Y (p,0.001). Y
represented the traditional AP pelvis performed most
commonly (Figure 5). Z out-performed W by a similar
statistical difference (p,0.001).
Figure 4. Graphic representation of the radiographs A, B, C,
D showing the observers’ results by region. This shows
increased performance of A and C.
Figure 5. Graphic representation of the fracture sharpness.
This shows increased clarity of the fracture in film Z, out
performing the traditional views X and Y (p,0.001).
The British Journal of Radiology, March 2006
The CT data for femoral neck angles relative to the
AIIS baseline showed wide variation and overlap with
no statistical relationship for foot external rotation. A
random sample of scans was re-measured showing
agreement in the measurements.
Discussion
The reported incidence of occult femoral neck fractures
on plain radiographs is approximately 4% [1]. There is
very little in the recent literature regarding optimizing
plain radiography to decrease the incidence of occult
femoral neck fractures. The authors hope that by including
this further radiographic view it will decrease the numbers
of patients requiring further investigation.
When plain radiographs are negative, and there is a
high index of suspicion, MRI has been shown to be
sensitive and specific in diagnosis of occult femoral
fractures.
Studies have shown that in radiographic negative
cases, where clinical concern is high, MRI showed
femoral neck fractures in 23–50% [1–3, 6]. Fractures
other than those of the femoral neck were demonstrated
in 11–33% of cases [2, 3, 6]. Most commonly these were
insufficiency fractures of the pubic rami or sacrum.
A further modality for diagnosis is radionuclide bone
scans. The sensitivity has been reported as 93–100% [4,
5], the specificity as 96%, and the positive predictive
value as 97%. This was regardless of patient age, and
time from presentation to scanning [4]. However, there
have been reported cases of a negative bone scan in a
fractured neck of femur [7], and false positive results due
to ligamentous avulsion and periosteal injury [8].
Fluoroscopy has been used with success. By gently
internally rotating the femur and obtaining high quality
exposures the diagnosis of femoral fractures was made
in 8 out of 16 patients in whom the initial radiographs
were negative [9]. Internal rotation elongates the femoral
neck and hence improves visualization of fractures. Our
study used the same principle, having the X-ray beam
closer to 90 ˚ to the femoral neck making the fracture line
more obvious.
The observers ranked A and C as the preferred choice
for anatomy in 71%. C (Z), however, was significantly
better for fracture clarity than A (W), and both were
superior to the standard views. C (Z) tended to elongate
the femoral neck, for the same reasons as to the 40 ˚
angled scaphoid view now widely used routinely for
trauma, i.e. the central ray is no longer at right angles to
the bone, creating geometric distortion. The observers
were not used to looking at the femoral neck with this
elongated appearance and so this may explain why A
was ranked higher than C for the anatomical demonstra-
tion. The 40 ˚ angled scaphoid view is a good corollary as
to why the authors expect the angled hip view to out
perform the normal view for a fracture at right angles to
the femoral neck. It is recognized, however, that not all
femoral neck fractures will lie at 90 ˚, but it is proposed
that it is these fractures that are difficult to see on
standard views and therefore may be better demon-
strated on the new view.
Specialist investigations are both expensive and time
consuming, and if there is a quick and cheap method of
219

obtaining the diagnosis when the initial radiographs are
negative, then this will undoubtedly benefit both the
patient and institution.
Conclusion
Suspected fractured femoral necks are common clin-
ical problems. We have demonstrated that radiographs
angled at 15 ˚ and 30 ˚ towards the femoral head show
greater clarity of both the subcapital and the midcervical
areas than the standard views used in current clinical
practice. The femoral neck fracture was also better
demonstrated using these two views, but best on C (30 ˚
angulation).
The authors now intend to conduct a prospective trial
to evaluate this in clinical practice, to evaluate whether in
equivocal cases a radiograph angled 30 ˚ to the femoral
head (the Bristol view) should be considered to aid the
diagnosis of fracture.
References
1. Lee YP, Griffith JF, Antonio GE, Tang N, Leung KS. Early
magnetic resonance imaging of radiographically occult
220
M Bradley, M Shaw and D Fox
osteoporotic fractures of the femoral neck. Hong Kong
Med J 2004;10:271–5.
2. Pandey R, McNally E, Ali A, Bulstrode C. The role of MRI in
the diagnosis of occult hip fractures. Injury 1998;29:61–3.
3. Lim KB, Eng AK, Chng SM, Tan AG, Thoo FL, Low CO.
Limited magnetic resonance imaging (MRI) and the occult
hip fracture. Ann Acad Med Singapore 2002;31:607–10.
4. Holder LE, Schwarz C, Wernicke PG, Michael RH.
Radionuclide bone imaging in the early detection of fractures
of the proximal femur (hip): multifactorial analysis.
Radiology 1990;174:509–15.
5. Fairclough J, Colhoun E, Johnston D, Williams LA. Bone
scanning for suspected hip fractures. A prospective study in
elderly patients. J Bone Joint Surg Br 1987;69:251–3.
6. Galloway HR, Meikle GR, Despois M. Patterns of injury in
patients with radiographic occult fracture of neck of femur as
determined by magnetic resonance imaging. Australas
Radiol 2004;48:21–4.
7. Mulcahy D, O’Malley M. Negative radioisotope bone scan in
a patient with a fractured neck of femur. Ir J Med Sci
1995;164:42–4.
8. Schmidt C, Deininger HK. [The occult fracture in the
roentgen picture and its detection using bone scintigraphy].
Radiologe 1985;25:104–7 [In German].
9. Helland EB, Tollefsen I, Reksten G. Radiographic diagnosis
of the occult hip fracture: experience in 16 patients. Acta
Orthop Scand 2000;71:639–41.
The British Journal of Radiology, March 2006
The British Journal of Radiology, 79 (2006), 221–225

Visceral and testicular calcifications as part of the phenotype in

pseudoxanthoma elasticum: ultrasound findings in Belgian
patients and healthy carriers
1
O M VANAKKER, MD, 2D VOET, MD, PhD, 2M PETROVIC, MD, PhD,
3
F VAN ROBAEYS, MD,
1
B P LEROY,
1 1
MD, P COUCKE, PhD and A DE PAEPE, MD, PhD

1
Center for Medical Genetics , 2Department of Sonography and 3Department of Radiology and
Medical Imaging, Ghent University, Hospital, De Pintelaan 185, 9000 Ghent, Belgium

ABSTRACT. Occasionally calcifications in abdominal organs, breasts and testicles have

been reported in pseudoxanthoma elasticum (PXE) patients. In the present study, an
ultrasound evaluation was performed of the abdomen and – in male patients – of the
testicles in 17 PXE patients and 17 heterozygous carriers. Blood samples were taken to
evaluate calcium load, liver and kidney function. Calcifications in liver, kidneys and
spleen were detected in 59% of the patients and in 23.5% of healthy carriers.
Parameters of kidney and liver function were normal in both groups, suggesting that
the calcifications have no direct effect on organ function. Testicular ultrasound
revealed parenchymous calcifications in all males investigated. Widespread, small
hyperechogenic foci resembling testicular microlithiasis were seen. In some carriers,
focal calcifications were identified. The current data suggest that visceral and testicular Received 1 February 2005
calcifications are part of the phenotype of PXE patients. Their presence in some of the Revised 4 May 2005
healthy carriers are suggestive of subclinical manifestations in these relatives. The Accepted 15 July 2005
natural history and long-term effects of the parenchymal calcifications remain to be
DOI: 10.1259/bjr/20801330
elucidated. As testicular microlithiasis may be associated with a higher risk for
malignancy, regular clinical and ultrasound follow-up seems indicated in these ’ 2006 The British Institute of
patients. Radiology

Pseudoxanthoma elasticum (PXE – OMIM [Online
Mendelian Inheritance in Man]# 264800) is an autosomal
recessive connective tissue disorder with multiple
systemic manifestations. The phenotype consists of a
triad of papular lesions and increased skin laxity in the
flexural areas of the body, angioid streaks in the ocular
fundus with eventually retinal haemorrhages and loss of
central vision, and accelerated atherosclerosis leading to
cardiovascular complications [1–5]. The incidence of this
rare disease has recently been estimated to be 1:75 000
[6], although this may be an underestimation due to the
high variability of the phenotype. Clinical manifestations
of the disease are attributed to alterations of elastic fibres
within the extracellular matrix of the affected organs.
These fibres undergo progressive fragmentation and
mineralization, which is the histopathological hallmark
of the disease [2]. Nevertheless, other components of the
extracellular matrix, such as collagen, fibrillins and
proteoglycans have either an abnormal morphology or
distribution [7, 8].
The gene responsible for PXE (ABCC6 - MIM# 603234)
is located on chromosome 16p13.1. It encodes an
ATP-dependent transporter the substrate of which is
as yet unknown. The relationship between this protein
and the phenotype also remains to be elucidated
[9–11].
It has been shown that healthy carriers of PXE have
similar cutaneous abnormalities at the ultrastructural
level, suggesting that a mild phenotype may be seen in
these individuals [12]. Although a higher incidence of
cardiovascular disease has been reported, carriers do not
develop other manifestations of PXE such as cutaneous
and/or retinal disease [12–14].
Occasionally, PXE patients have been reported in which
calcifications in several organs, including kidney, pancreas,
spleen and breasts have been observed [15–21].
Additionally, one case report has described the presence
of multiple calcifications in the testicles of a
14-year-old PXE-patient [22]. These reports suggest a
possible association of organ calcifications and PXE. To our
knowledge, no systematic screening of patients nor of
healthy carriers has been performed. We present a
comprehensive ultrasound study of 17 PXE patients in
whom the association between visceral and/or testicular
calcifications and PXE was established. Furthermore, 17
heterozygous relatives were screened to detect whether
similar lesions could be found.

Patients and methods

Address correspondence to: Anne De Paepe, Center for Medical
Genetics, Ghent University Hospital, De Pintelaan 185, B-9000 Sixteen patients with clinical, molecular and biopsy-
Ghent, Belgium. proven PXE were examined. Informed consent was

The British Journal of Radiology, March 2006 221


obtained from all patients and the study was approved
by the Ethical Committee of the Faculty of Medicine of
the Ghent University Hospital. Our patient population
consisted of 5 men and 12 women. Ages ranged from
18 years to 80 years with an average of 54 years.
The group of 17 heterozygous carriers included
offspring as well as parents of patients (obligate carriers).
Additionally, siblings of patients proven to be hetero-
zygous carriers of an ABCC6 mutation were included.
The carrier group consisted of 11 men and 6 women.
Ages ranged from 16 years to 76 years with an average
age of 39 years.
All index-patients and carriers were personally exam-
ined at the PXE clinic of the Center for Medical Genetics
at the Ghent University Hospital. Thorough patient
histories were recorded in all individuals with special
consideration for signs and symptoms indicating hepatic,
renal or splenic dysfunction.
The full clinical protocol used at the PXE clinic of the
Center for Medical Genetics at the Ghent University
Hospital, including careful dermatological, ophthalmo-
logical and cardiovascular examinations and ultrasound
of the abdomen and testicles, was applied in both
groups. Ultrasound examinations were performed at
the Department of Sonography using a HDI 5000 system
(Philips, Brussels, Belgium) with a C5-2 and a L12-5
scanhead for the examination of the abdomen and
scrotum, respectively. To minimize interobserver varia-
tion three ultrasonographers performed the examina-
tions were blinded to patient information. Serum
analysis was performed to evaluate calcium load, liver
and kidney function in order to exclude other aetiologies
of parenchymal calcifications and to assess the possible
functional effect of the lesions. Parameters measured in
all individuals included serum concentrations of aspar-
tate amino transferase (AST), amino alanine transferase
(ALT), alkaline phosphatase (AF), gamma-glutamyl
transpeptidase (cGT), creatinine, urea, calcium and
phosphorus.
Skin biopsies were taken either in an affected skin area
or at the back of the neck when no lesion was
macroscopically apparent. Histological confirmation of
PXE was obtained with haematoxylin and eosin, van
Giesson and Von Kossa stains to detect the typical
anomalies of the elastic fibre.
Molecular screening of the ABCC6 gene was per-
formed using dHPLC (denaturing high performance
liquid chromatography) (Transgenomics, Cheshire, UK)
and subsequent sequencing of all ABCC6 exons in those
that showed abnormal dHPLC-patterns.
Results
Abdominal ultrasounds
Abdominal ultrasound revealed calcifications scat-
tered throughout the parenchyma of the kidneys (8
patients), liver (4 patients) or spleen (3 patients) in 10/17
(59%) of PXE patients (Figure 1a–d). In those with
visceral calcifications, kidneys were most frequently
affected (80%). In 3 out of 10 (30%) patients, two or
more organs were involved. The number of calcified
lesions ranged from a few in the spleen to widely
222
O M Vanakker, D Voet, M Petrovic et al
disseminated in the liver parenchyma. Calcifications
were seen as hyperechogenic foci with acoustic shadow-
ing, measuring 2–3 mm in diameter. Renal calcifications
were localized in the corticomedullary junction, but also
within the cortical tissue. Similar lesions could be
observed in 4 out of 17 healthy carriers. Two of those
had kidney calcifications while the others had lesions in
the liver. Other ultrasound findings included hepatic
haemangiomas and steatosis.
Serum tests to evaluate kidney and liver function were
performed in all patients and carriers examined. No
abnormalities of either liver enzymes nor serum creati-
nine and urea were observed. Calcium levels were
always within normal limits. None of the individuals
in this study had signs or symptoms indicative of
abnormal function of the liver, spleen or kidneys.
Testicular ultrasounds
Ultrasound of the scrotum was performed in four PXE
patients. In three multiple widespread, small hyperecho-
genic foci resembling a ‘‘heaven full of stars’’ were
identified throughout the parenchyma of both testicles
(Figure 2). This appearance matches the criteria of
classical testicular microlithiasis as described by
Dell’Acqua et al [23]. One patient had only few of these
lesions, compatible with limited testicular microlithiasis.
However, no histological confirmation of this diag-
nosis was obtained since none of the patients had any
complaint warranting a biopsy. No testicular tumours
were detected during the examination.
In two out of 11 healthy carriers examined, focal
calcifications of the testicular capsule or parenchyma
were observed. The parenchymatous calcification was a
small unilateral focus without acoustic shadowing. These
individuals were asymptomatic. Two carriers were
found to have a hyperechogenic mediastinum testis,
which can be considered a normal variant.
Discussion
PXE is a rare autosomal recessive disease character-
ized by fragmentation and calcification of the elastic
fibres. Clinical manifestations mainly consist of cuta-
neous, ophthalmological and cardiovascular lesions.
Case reports have mentioned the occurrence of calcifica-
tions in the visceral organs, breasts and testicles in some
individuals [15–21]. In this study, a standardized
examination protocol comprising abdominal and testi-
cular ultrasounds was used in 17 PXE patients to observe
whether calcified lesions in these organs could be
detected.
Due to the autosomal recessive inheritance of PXE,
parents and children of probands are obligate carriers of
one mutation in the ABCC6 gene. Previous ultrastruc-
tural studies in relatives of PXE patients have revealed
cutaneous morphologic alterations similar to those seen
in patients, although less severe in nature [12]. Trip et al
described a higher risk of coronary artery disease in
carriers of the frequent R1141X nonsense mutation [13].
These observations indicate that heterozygous carriers
may have mild PXE manifestations, albeit without
The British Journal of Radiology, March 2006
Visceral and testicular calcifications in PXE

Figure 1. Ultrasound images of calcified foci in several abdominal organs: (a) frontal cross-section through the abdomen with
multiple calcifications in the liver of a pseudoxanthoma elasticum (PXE) patient; (b) subcostal transverse cross-section of the liver
of a heterozygous carrier in which two calcifications with acoustic shadowing are seen; (c,d) frontal cross-section through the
abdomen with view of multiple hyperechogenic foci in (c) the right kidney and (d) spleen of PXE patients.

obvious cutaneous or ophthalmological symptoms.

Therefore, ultrasound evidence of subclinical manifesta-
tions was sought in mutation carriers.

Figure 2. Longitudinal cross-section of the testicle with
scattered parenchymatous calcifications in the right testicle
of a pseudoxanthoma elasticum (PXE) patient as a typical
example of testicular microlithiasis.
Abdominal ultrasound
The data presented suggest that visceral calcifications
in the kidneys, liver and spleen are indeed part of the
phenotype of PXE patients. Interestingly, similar lesions
were found to be present in some of the healthy carriers,
although less frequently and to a lesser extent.
All ages were represented in patients and carriers with
visceral calcifications, making our findings unlikely to be
attributed solely to the age of the individuals. Calcium
and phosphorus load were normal in all individuals,
excluding other aetiologies of visceral calcifications such
as chronic granulomatous diseases (e.g. sarcoidosis),
renal failure, hyper(para)thyroidism, pheochromocy-
toma, adrenal insufficiency or malignancy.

The British Journal of Radiology, March 2006 223


As serum tests for liver and kidney function revealed
no abnormalities and none of the individuals examined
suffered from any disturbances of renal, hepatic or
splenic function, the calcified foci probably do not
interfere with liver, kidney or splenic function.
However, their natural history and long-term effects
remain to be elucidated.
Therefore an abdominal ultrasound at the time of
diagnosis may be indicated. Furthermore, regular
re-evaluation with serum tests and ultrasound are
advisable.
It has been previously reported that both abdominal
plain radiographs and CT are unable to visualize these
lesions [18]. In the only patient with renal foci in whom
abdominal radiographs were performed in this study, no
calcifications were visible. We did not perform CT
imaging in our population and can therefore not rule
out that, due to technical improvements and new
developments, these lesions can now be visualized.
However, since ultrasound proved to give sufficient
data and comparing the costs and radiation load of both
examinations, we feel that at present CT is not an added
value in the work-up of a PXE patient in a clinical setting.
In a research setting, however, it would be interesting to
find out if these lesions are indeed visible with modern
CT techniques and to evaluate their extent and char-
acteristics in comparison with ultrasound findings. Thus,
in view of all known aspects, ultrasound should be
considered the investigation of first choice for detection
of these calcifications on a routine basis.
Testicular ultrasounds
Testicular parenchymal calcifications were identified
in all male patients so far examined. These lesions,
described as bilateral, small, hyperechogenic foci, meet
the ultrasound criteria of testicular microlithiasis (TM).
The TM pattern is defined as usually bilateral hyper-
echogenic multiple small foci without acoustic shadow
and with complete or partial extension to the paren-
chyma. Cases in which five or more foci can be
demonstrated are defined as classical TM [24–27].
Cases that do not meet this criterion are designated as
limited TM. The imaging diagnosis can be confirmed by
showing intratubular microliths on biopsy [24–27]. Since
none of our patients had either complaints or fertility
problems testicular biopsies were considered unethical.
TM is of special interest due to its reported association
with testicular malignancy [29–36]. Nevertheless, it
remains unclear whether primary testicular tumours
actually occur more frequently in patients with pre-
existing TM. Large prospective studies are needed to
further clarify this issue. Until further data are available,
it seems cautious to consider patients with a TM-like
ultrasound image as having a potentially increased risk
of developing a testicular malignancy and to offer
regular ultrasound screening [28–34, 36].
The findings in healthy carriers were different from
those in patients in their extent and/or location within
the testicle. Multiple hyperechogenic foci confined to
the capsule or the mediastinum testis were detected,
the latter probably being a normal variant. Although
anatomically this could also be compatible with
224
O M Vanakker, D Voet, M Petrovic et al
calcifications in the rete testis [37], we cannot be sure
of this without a biopsy which is unjustifiable in these
patients.
In another carrier, we observed one parenchymatous
calcification which could be considered as limited TM.
The remaining parenchyma, however, was completely
normal and we cannot exclude that these findings are
fortuitous. Since they have, to our knowledge, not
previously been described in PXE, further study on a
larger group of carriers would be of interest.
Acknowledgments
The authors are very grateful to all PXE patients and
families for their kind collaboration. This work was
supported by a grant from the Ghent University (GOA-
12051203). O Vanakker is a research assistant supported
by the Fund for Scientific Research – Flanders (Belgium).
References
1. Bercovitch L, Terry P. Pseudoxanthoma Elasticum 2004. J
Am Acad Dermatol 2004;51(1 Suppl.):S13–4.
2. Hu X, Plomp AS, van Soest A, Wijnholds J, de Jong PTVM,
Bergen A, et al. Pseudoxanthoma Elasticum: a clinical,
histopathological and molecular update. Surv Ophthalmol
2003;48:424–38.
3. Neldner KH. Pseudoxanthoma Elasticum. Clin Dermatol
1988;6:83–92.
4. Scherer DW, Sapadin AN, Lebwohl MG. Pseudoxanthoma
Elasticum: an update. Dermatology 1999;199:3.
5. Ringpfeil F. Selected disorders of connective tissue: pseudo-
xanthoma elasticum, cutis laxa and lipoid proteinosis. Clin
Dermatol 2005;23:41–6.
6. Bergen A, Plomp AS, Gorgels T, De Jong P. Van gen naar
ziekte; pseudoxanthoma elasticum en het ABCC6-gen. Ned
Tijdschr Geneeskd 2004;148:1586–9.
7. Baccarani-Contri M, Vincenzi D, Cicchetti F, Mori G,
Pasquali-Ronchetti I. Immunochemical identification of
abnormal constituents in the dermis of pseudoxanthoma
elasticum patients. Eur J Histochem 1994;38:111–23.
8. Longas MO, Wisch P, Lebwohl MG, Fleischmajer R.
Glycosaminoglycans of skin and urine in pseudoxanthoma
elasticum: evidence for chondroitin 6-sulfate alteration. Clin
Chem Acta 1986;155:227–36.
9. Ilias A, Urban Z, Seidl T, Le Saux O, Sinko E, Boyd C, et al.
Loss of ATP-dependent transport activity in Pseudo-
xanthoma Elasticum-associated mutants of human ABCC6
(MRP6). J Biol Chem 2002;277:16860–7.
10. Le Saux O, Beck K, Sachsinger C, Silvestri C, Treiber C,
Göring HHH, et al. A spectrum of ABCC6 mutations is
responsible for pseudoxanthoma elasticum. Am J Hum
Genet 2001;69:749–64.
11. Le Saux O, Urban Z, Tschuch C, Csiszar K, Bacchelli B,
Quaglino D, et al. Mutations in a gene encoding an ABC
transporter cause pseudoxanthoma elasticum. Nat Genet
2000;25:223–7.
12. Bacchelli B, Quaglino D, Gheduzzi D, Taparelli F, Boraldi F,
Trolli B, et al. Identification of heterozygote carriers in
families with a recessive form of pseudoxanthoma elasti-
cum (PXE). Mod Pathol 1999;12:1112–23.
13. Trip MD, Smulders YM, Wegman JJ, Hu X, Boer JMA,
ten Brink JB, et al. Frequent mutation in the ABCC6 gene
(R1141X) is associated with a strong increase in the preva-
lence of coronary artery disease. Circulation 2002;160:773–5.
14. van Soest S, Swart J, Tijmes N, Sandkuijl LA, Rommers J,
Bergen AAB. A locus for autosomal recessive
The British Journal of Radiology, March 2006
Visceral and testicular calcifications in PXE
Pseudoxanthoma Elasticum, with penetrance of vascular
symptoms in carriers, maps to chromosome 16p13.1.
Genome Res 1997;7:830–4.
15. Albertyn LE, Drew AC. Mammographically detected
microcalcifications due to Pseudoxanthoma Elasticum.
Australas Radiol 1990;35:81.
16. Crespi G, Derichi LE, Saffioti S. Sonographic detection of
renal changes in Pseudoxanthoma Elasticum. Urol Radiol
1992;13:223–5.
17. Ekim M, Tümer N, Atmaca L, Anadolu R, Salih M, Dönmez
O, et al. Pseudoxanthoma elasticum: a rare cause of
hypertension in children. Pediatr Nephrol 1998;12:183–5.
18. Erden A, Dumanli H, Yardimci C, Erden I, Aytaç S.
Ultrasound and color flow Doppler manifestations of
Pseudoxanthoma Elasticum. J Clin Ultrasound 1993;21:396–8.
19. Korn S, Scilnacht J, Huth C, Feller AM. Cardiovascular
manifestations of pseudoxanthoma elasticum (Grönblad-
Strandberg Syndrome). Thorac Cardiovasc Surg
1987;35:191–4.
20. Ortiz Gorraiz MA, Casares Aria A, Tallada Bunuel M,
Vicente Prados FJ, Honrubia Vilchez B, Fernandez Sanchez
A. Urologic findings in pseudoxanthoma elasticum: report
one case. Actas Urol Esp 2005;29:96–9.
21. Suarez MJ, Garcia JB, Orense M, Raimunde E, Lopez MV,
Fernandez O. Sonographic aspects of Pseudoxanthoma
Elasticum. Pediatr Radiol 1991;21:538.
22. Pinto K, Deogracias Pena R. Testicular calcifications in
Pseudoxanthoma Elasticum. J Urol 2004;171:1898.
23. Dell’Acqua A, Toma P, Oddone M, Ciccone MA, Marsili E,
Derci LE. Testicular microlithiasis: US findings in six
pediatric patients and literature review. Eur Radiol
1999;9:940–5.
24. Bushby LH, Miller FNAC, Rosairo S, Clarke JL, Sidhu PS.
Scrotal calcification: ultrasound appearances, distribution
and aetiology. Br J Radiol 2002;75:283–8.
25. Ganem JP. Testicular microlithiasis. Curr Opin Urol
2000;10:99.
The British Journal of Radiology, March 2006
26. Lawrentschuk N, Brough SJS, De Ryke RJ. Testicular
microlithiasis: a case report and review of the literature.
Aus NZ J Surg 2003;73:364–6.
27. Miller FNAC, Sidhu PS. Does testicular microlithiasis
matter? A review. Clin Radiol 2002;57:883–90.
28. Akin EA, Khati NJ, Hill MC. Ultrasound of the scrotum.
Ultrasound Q 2004;20:181–200.
29. Ganem JP, Workman KR, Shaban SF. Testicular micro-
lithiasis is associated with testicular pathology. Urology
1999;53:209–13.
30. Kocaoğlu M, Bozlar U, Bulakbaşi N, Sağlam M, Üçöz T,
Somuncu I. Testicular microlithiasis in pediatric age group:
ultrasonography findings and literature review. Diagn
Interv Radiol 2005;11:60–5.
31. Miller RL, Wissman R, White S, Ragosin R. Testicular
microlithiasis: a benign condition with a malignant associa-
tion. J Clin Ultrasound 1996;24:197–202.
32. Parra BL, Venable DD, Gonzalez LL, Eastham JA. Testicular
microlithiasis is a predictor of intratubular germ cell tumor.
AJR Am J Roentgenol 1996;167:889–90.
33. Pourbagher MA, Kilinc F, Guvel S, Pourbagher A, Egilmez
T, Ozkardes H. Follow-up of testicular microlithiasis for
subsequent testicular cancer development. Urol Int
2005;74:108–12.
34. Rashid HH, Cos LR, Weinberg E, Messing EM. Testicular
microlithiasis: a review and its association with testicular
cancer. Urol Oncol 2004;22:285–9.
35. Ringdahl E, Claybrook K, Teague JL, Northrup M.
Testicular microlithiasis and its relation to testicular cancer
on ultrasound findings of symptomatic men. J Urol
2004;172:1904–6.
36. Winter TC, Zunkel DE, Mack LA. Testicular carcinoma in a
patient with previously demonstrated testicular micro-
lithiasis. J Urol 1996;166:889–96.
37. Nistal M, Garcia-Cabezas MA, Regadera J, Castillo MC.
Microlithiasis of the epididymis and the rete testis. Am J
Surg Pathol 2004;28:514–22.
225
 The British Journal of Radiology, 79 (2006), 226–231

Life-threatening common carotid artery blowout: rescue

treatment with a newly designed self-expanding covered nitinol
stent
1
H S KIM, MD, 1D H LEE, MD, PhD,
4
H J KIM, MD,
1
S J KIM, MD, PhD,
2
W KIM, MD,
3
S Y KIM, MD, PhD
and 1D C SUH, MD, PhD

1
Department of Radiology, 2Emergency Medicine and 3Otorhinolaryngology, Asan Medical Center,
University of Ulsan, College of Medicine, 388-1 Poongnap Dong, Songpa-Gu, Seoul, 138-736 and
4
Department of Radiology, DaeJeon Catholic Hospital, Republic of Korea

ABSTRACT. Carotid blowout is a devastating complication in patients with head and

neck malignancy. A covered stent offers an alternative to treatment of a carotid
blowout patient thought to be at high risk for surgery or carotid occlusion. Stent
placement in the common carotid artery or carotid bulb is a technical challenge Received 1 June 2005
because of large luminal diameter and luminal calibre discrepancy between internal Accepted 15 July 2005
carotid artery and common carotid artery. We present four patients with common
DOI: 10.1259/bjr/66917189
carotid rupture and massive bleeding who were treated with self-expanding covered
stents, among them, two cases were treated with newly designed self-expanding ’ 2006 The British Institute of
polytetrafluoroethylene (PTFE)-covered nitinol stents. Radiology

Endovascular management of acute bleeding in the
head and neck by occlusion of the offending vessel with
coils or detachable balloons has been the alternative to
surgical exploration [1]. However, these procedures have
the potential for producing delayed cerebral ischaemic
complications in 15–20% of patients [2]. Covered stent
deployment has been developed as an effective treat-
ment option in carotid blowout patients thought to be at
high risk for surgery or carotid occlusion [2]. However,
stent placement for the management of carotid blowout
is not always effective in cases of head and neck
malignancy involving extensive segment of the common
carotid artery (CCA) with relatively large calibre or
carotid bulb with luminal calibre discrepancy between
internal carotid artery (ICA) and CCA.
We report four cases of CCA rupture with massive
bleeding in patients with head and neck malignancies
and a history of long-term radiation treatment who were
treated using self-expanding covered stents. Among
these patients, two cases were treated by a newly
designed covered stents which have a bare area in both
their proximal and distal portions.
Patients and methods
During a 5-year period between May 1999 and June
2004, we treated four patients (four males, aged 57–
68 years) with common carotid rupture, who had head
Address correspondence to: Dae Chul Suh, Department of
Radiology, Asan Medical Center, University of Ulsan College of
Medicine, 388-1 Poongnap-dong, Songpa-gu, Seoul 138-736, Korea.
This study was supported by a grant of the Korea Health 21 R&D
Project, Ministry of Health & Welfare, Republic of Korea (03-PJ1-
PG1-CH06-0001).
and neck malignancies, and who had histories of
radiation therapy alone or combined with chemotherapy.
The patients’ characteristics are listed in Table 1. All of
these patients presented with life-threatening massive
neck or oral bleedings, unstable vital signs and altered
mental changes. The procedures were performed under
local anaesthesia with 1% lidocaine and conscious
sedation with intravenously administered midazolam
hydrochloride (Versed; Roche Laboratories, Nutley, NJ).
A 9-F introducer sheath was positioned in the right
common femoral artery. The patients did not have
systemic heparinization as they were having massive
bleeding. Using digital roadmap guidance, a 0.0350
hydrophilic guidewire (Terumo, Tokyo, Japan) was
carefully manoeuvred into the ICA. A 4-F catheter
(Terumo, Tokyo, Japan) was then advanced over the
wire. After obtaining an angiogram with a 4 F catheter, a
steep 0.0350 exchange length wire (Terumo, Tokyo,
Japan) was introduced beyond the diseased segment
into ICA. A self-expanding covered nitinol stent (NITI-S
Stent; Taewoong Medical, Seoul, Korea) was then passed
over the exchange guidewire and carefully positioned at
the level of the bleeding site including pseudoaneurysm.
The stent was then deployed across the corresponding
segment of the pseudoaneurysm. 300 mg of oral clopi-
dogrel (Plavix; Bristol Myers-Squibb, New York, NY)
were given after deploying the stents to minimize
the risk of stent thrombosis resulting from platelet
aggregation.
Profile of PTFE covered nitinol stent
NITI-S stent is based on the longitudinal wire mesh
design (Figure 1). Nitinol wire with 0.0070 in diameter

226 The British Journal of Radiology, March 2006

Life-threatening common carotid artery blowout

Table 1. Summary of patients with covered stent placements in the common carotid artery
No. of Age/Gender Presentation Underlying Bleeding Treatment device Clinical course
cases disease location (diameter6length)

Case 1 62/M Massive bleeding Oesophageal Mid-CCA 10670 mm CS Re-bleeding after 11 days,
at neck open wound carcinoma 10650 mm CS stable for 2 months

Case 2 57/M Massive oral bleeding Nasopharyngeal Carotid bulb 10670 mm CS Re-bleeding after 6
carcinoma 9.0640 mm BS weeks, stable for 5
10650 mm CS months, died because
of massive infarcts due
to contralateral ICA
invasion

Case 3 68/M Massive oral bleeding Laryngeal Distal CCA 10670 mm CS Discharged in stable con-
carcinoma dition 1 day later and
lost follow-up
Case 4 61/M Massive oral bleeding Hypopharyngeal Carotid bulb 10670 mm CS Discharged in stable con-
carcinoma dition 1 day later and
transferred to other
hospital
CCA, common carotid artery; CS, covered stent; BS, bare stent; ICA, internal carotid artery.

was used for the single-wire woven stent making the
stent cells in the both ends closed. Three gold tip markers
were attached at each end of the stent margin to enhance
fluoroscopic visibility. Because the stent is adhered onto
the PTFE graft by polyurethane, the stent is disposed
between polyurethane at its outer surface and the PTFE
sheet at its inner surface. Polyurethane would also give
PTFE more durability.
The diameters and lengths of the stents were 10 mm
and 70 mm in all cases, and additional 10 mm 6 50 mm
stents placements were performed in cases 1 (Figure 2)
and 2 (Figure 3). The 10 mm 6 70 mm stent which was
designed for transjugular intrahepatic portosystemic
shunt (TIPS) consisted of a proximal covered area
(50 mm in length) and a distal bare area (20 mm in
length). The newly designed 10 mm 6 50 mm stent
consisted of a middle PTFE covered segment (40 mm)
and a 5 mm bare segment at both proximal and distal
ends (Figure 1).
Results
The stent delivery and placement were all success-
ful. Immediately after the procedure, vital sign and
neurological status of the patients became normalized
and follow-up angiogram showed occlusion of pseudoa-
neurysm and preservation of the parent arterial flow in
all cases. Re-bleeding at the proximal margin of the
stented segment, suggesting extension of the disease
beyond the stent margin, required another covered stent
deployment in two patients (case 1 and 2). In these two
patients, an additional newly designed 10 mm 6 50 mm
stent was placed in the lower portion of the previous
stent and no recurrent haemorrhage was found on
follow-up for 2 months and 5 months, respectively.
Discrepancy of vessel lumen size between the ICA and
the CCA required bare stenting within the covered stent
in a patient with carotid bulb blowout (case 2). Treatment
device, clinical course, and follow-up results for study
patients were listed in Table 1.
Discussion
The reported incidence of carotid rupture in patients
who have had a neck dissection with or without tumour
resection is 3–4% [3]. Carotid blowout is associated with
approximately 60% neurological morbidity and 40%
mortality in patients with associated conditions such as

(a) (b)

Figure 1. Photographs of the covered stent (Taewoong Medical, Seoul, Korea) composed of a self-expanding nitinol wire
covered with PTFE. (a) The 10 mm 6 50 mm stent used secondarily in cases 1 and 2 consists of proximal and distal bare segments
of 5 mm and a middle covered area of 40 mm. (b) Note the structural relationship of the stent wire and PTFE graft. Outer
polyurethane layer connects the stent wire and PTFE graft. Stent wire thickness is 0.0070 in size.

The British Journal of Radiology, March 2006 227

 H S Kim, D H Lee, H J Kim et al

(a) (b)

Figure 2. A 62-year-old male with unresectable oesophageal carcinoma presented with massive bleeding at the neck wound
site associated with deep neck infection after radiation therapy. (a) Conventional angiogram shows a large pseudoaneurysm in
the mid-portion of the left common carotid artery (CCA). (b) 11 days after the first stent placement. Conventional angiography
shows extension of the previous pseudoaneurysm at the lower margin of the stent. (Continued)

228 The British Journal of Radiology, March 2006

Life-threatening common carotid artery blowout
(c)
Figure 2. (Cont.) (c) After placement of an additional stent,
angiography shows no recurrent haemorrhage.
pharyngocutaneous fistula, recurrent tumour, or radia-
tion necrosis [4]. The history of irradiation therapy adds
a 7.6-fold increased risk of developing carotid blowout in
patients with head and neck malignancy [5].
Treatment of extracranial carotid artery pseudoaneu-
rysm has been open surgery with resection and
reconstruction or carotid artery ligation. However, this
condition often makes patient haemodynamically
unstable and causes significantly decreased cerebral
perfusion. Pseudoaneurysm is composed only of fibrous
tissue and contains no normal vessel wall elements:
neither do these aneurysms have a real neck. Therefore,
the dissection and preparation of the aneurysmal sac for
clipping involves an extremely high risk of perioperative
rupture [6].
The advent of various endovascular treatments includ-
ing permanent balloon occlusion or coil embolisation has
The British Journal of Radiology, March 2006
expanded the therapeutic options for patient with
rupture and pseudoaneurysm of ICA or CCA [3].
However, as many as 15–20% of patients whose carotid
blowout is managed with permanent balloon occlusion
may develop immediate or delayed cerebral ischaemia
[2]. In all of our patients, as their vital signs and mental
status were unstable and the examinations for the
cerebral perfusion before the procedure such as the
balloon occlusion test were impossible, it was not known
whether the occlusion procedure of the CCA would have
further compromised cerebral perfusion. Thus, endovas-
cular sacrifice of the CCA was not reliable. The choice of
endovascular carotid stent placement combined with
Guglielmi detachable coils (GDC) has been reported [7].
However, long-term radiation therapy in patients with
head and neck cancer could injure normal head and neck
structures, thus, the surrounding radiation induced soft
tissue changes cannot offer enough support to the parent
artery and to the pseudoaneurysmal sac. In case 2, the
patient was initially treated with coil embolisation
because stent placement was considered difficult due
to a discrepancy in the size of vessel lumen between the
ICA and the CCA. However, as coil embolisation alone
did not occlude the pseudoaneurysm of the CCA,
additional stent placement was then performed.
Kiyosue et al, reported dispersion and migration of coils
in carotid blowout patient treated by parent-artery
occlusion with coils [8].
Covered stents are already in clinical use for treating
occlusive, aneurysmal, and traumatic peripheral arterial
disease, for repairing aortic aneurysm, and in transju-
gular portosystemic shunting [9]. In several previous
reports, covered stents had been used in the treatment of
ICA pseudoaneurysm [10–12]. However, the stent place-
ment for the management of carotid blowout is not
always effective in case of head and neck malignancy
involving extensive segment of the CCA with relatively
large calibre or carotid bulb with luminal calibre
discrepancy between the ICA and the CCA. In general
treatment of ICA and CCA pseudoaneurysm, either
5 mm or 7 mm diameter stents consistent with the CCA
or ICA lumen were usually used as well as 9-F arterial
sheaths to accommodate the outer diameter as its
delivery system. In our cases, we used large-bore
10 mm diameter self-expanding stents via exchange
guidewire through a 9-F arterial sheath as its delivery
system. With these methods on the realtime roadmap
fluoroscopy, there was no difficulty in exact positioning
and deployment.
Additional pseudoaneurysm formation at the lower
end of the covered stented margin of the CCA can be due
to the radiation-induced surrounding soft tissue weak-
ness or due to the rigid lower end of the covered stent
structure because the covered stent used at the initial
attempt in our patients did not have a bare portion at the
lower end. The newly designed 10 mm 6 50 mm stents
were used in cases 1 and 2. It consisted of a covered
segment (40 mm) and 5 mm bare segments at proximal
and distal ends. Although the long-term patency rates of
these stents and their risk of thromboembolic or other
complications in the treatment of the CCA rupture and
pseudoaneurysm formation are unknown, the vital signs
and neurological status of these patients became stable
after these procedures and there were no complications
229
 H S Kim, D H Lee, H J Kim et al

(a) (b)

Figure 3. A 57-year-old man with inoperable nasopharyngeal carcinoma treated with radical neck dissection and radiation
therapy presented with massive oral bleeding. (a) The right common carotid arteriogram shows a pseudoaneurysm formation
and contrast leakage into pharynx and oral cavity near the carotid bulb. (b) Immediate angiography after stent placement and
coil embolisation revealed a small contrast leakage out of the distal portion of the covered stent due to luminal diameter
discrepancy between the proximal internal carotid artery (ICA) and the covered stent caused by the transitional lumen size of
the carotid bulb. (Continued)

230 The British Journal of Radiology, March 2006

Life-threatening common carotid artery blowout
(c)
Figure 3. (Cont.) (c) Final angiography shows no further
leakage of the contrast agent after deployment of another
self-expanding stent crossing the distal end of the
covered stent. The patient became stable immediately after
procedure.
leading to any neurological deficits during the short-
term follow-up periods.
Because of the limited follow-up periods in this series,
the long-term patency rates of these stents and their risk
of thromboembolic or other complications in the treat-
ment of the CCA pseudoaneurysm are unknown.
However, in all of our cases, there were no complications
during the short-term follow-up periods, and the causes
of death were not associated with stent complications or
bleeding pseudoaneurysms. The vital signs and neuro-
logical status of these patients were dramatically
improved after these procedures.
In summary, the newly designed self-expanding
covered nitinol stent may be a safe and useful tool for
the endovascular occlusion of the CCA pseudoaneur-
ysms. Delivering this 10 mm diameter stent via a 9-F
arterial sheath is easy. Although long-term follow-up
and larger series are required in order to evaluate the
The British Journal of Radiology, March 2006
stent efficacy, these four cases highlight the usefulness
and versatility of this covered stent for rescue treatment
of life-threatening bleeding pseudoaneurysm of the
CCA.)
Acknowledgments
We acknowledge the assistance of Eun Ja Yoon in
manuscript preparation, Sun Moon Whang, BS, in the
patients data collection and we also thank Bonie Hami,
MA, Department of Radiology, University Hospitals of
Cleveland, Cleveland, OH, for editorial assistance in
manuscript preparation. This study was supported by a
grant of the Korea Health 21 R&D Project, Ministry of
Health & Welfare, Republic of Korea (03-PJ1-PG1-CH06-
0001).
References
1. Scavee WJ, Mormont E, Coulier B, Trigaux JP, Schoevaerdts
JC. ‘‘Pseudoaneurysm of the internal carotid artery: treat-
ment with a covered stent.’’ Cardiovasc Intervent Radiol
2001;24:283–85.
2. Lesley WS, Chaloupka JC, Weigele JB, Mangla S, Dogar
MA. Preliminary experience with endovascular reconstruc-
tion for the management of carotid blowout syndrome.
AJNR Am J Neuroradiol 2004;24:975–81.
3. Morrissey DD, Andersen PE, Nesbit GM, Barnwell SL,
Everts EC, Cohen JI. Endovascular management of hemor-
rhage in patients with head neck cancer. Arch Otolaryngol
Head Neck Surg 1997;123:15–9.
4. Chaloupka JC, Roth TC, Putman CM, Mitra S, Ross DA,
Lowlicht RA, et al. Recurrent carotid blowout syndrome:
diagnostic and therapeutic challenges in a newly recog-
nized subgroup of patients. AJNR Am J Neuroradiol
1999;20:1069–77.
5. Maran AG, Amin M, Wilson JA. Radical neck dissection: a
19-year experience. J Laryngol Otol 1989;103:760–76.
6. Charbel FT, Gonzales-Portillo G, Hoffman W, Cochran E.
Distal internal carotid artery pseudoaneurysms: technique
and pitfalls of surgical management: two technical case
reports. Neurosurgery 1999;45:643–8.
7. Mericle RA, Lanzino G, Wakhioo AK, Guterman LR,
Hopkins LN. Stenting and secondary coiling of intracranial
internal carotid artery aneurysm: technical case report.
Neurosurgery 1999;43:1229–34.
8. Kiyosue H, Okahara M, Tanoue S, Sagara Y, Matsumoto S,
Mori H, et al. Dispersion of coils after parent-artery
occlusion of radiation-induced internal carotid artery
pseudoaneurysm. AJNR Am J Neuroradiol 2004;25:1080–2.
9. Razavi MK, Dake MD, Semba CP, Nyman UR, Liddell RP.
Percutaneous endoluminal placement of stent-grafts for the
treatment of isolated iliac artery aneurysms. Radiology
1995;197:801–4.
10. Satler LF, Promises MG. The covered stent. Catheter
Cardiovasc Interv 2000;50:89.
11. Simionato F, Righi C, Melissano G, Rolli A, Chiesa R, Scotti
G. Stent-graft treatment of a common carotid artery
pseudoaneurysm. J Endovasc Ther 2000;7:136–40.
12. Saket RR, Razavi MK, Sze DY, Frisoli JK, Kee ST, Dake MD.
Stent-graft treatment of extracranial carotid and vertebral
arterial lesions. J Vasc Interv Radiol 2004;15:1151–6.
231
 The British Journal of Radiology, 79 (2006), 232–238

Quantitative assessment of hip osteoarthritis based on image

texture analysis
1
I S BONIATIS, MSc, 1L I COSTARIDOU, PhD,
2
D A CAVOURAS, PhD,
3
E C PANAGIOTOPOULOS, MD, PhD
and 1G S PANAYIOTAKIS, PhD

1
University of Patras, School of Medicine, Department of Medical Physics, 265 00 Patras,
2
Technological Educational Institute of Athens, Department of Medical Instrumentation
Technology, 122 10 Athens and 3University of Patras, School of Medicine, Department of
Orthopaedics, 265 00 Patras, Greece

ABSTRACT. A non-invasive method was developed to investigate the potential capacity

of digital image texture analysis in evaluating the severity of hip osteoarthritis (OA)
and in monitoring its progression. 19 textural features evaluating patterns of pixel
intensity fluctuations were extracted from 64 images of radiographic hip joint spaces
(HJS), corresponding to 32 patients with verified unilateral or bilateral OA. Images were
enhanced employing custom developed software for the delineation of the articular
margins on digitized pelvic radiographs. The severity of OA for each patient was
assessed by expert orthopaedists employing the Kellgren and Lawrence (KL) scale.
Additionally, an index expressing HJS-narrowing was computed considering patients
from the unilateral OA-group. A textural feature that quantified pixel distribution non-
uniformity (grey level non-uniformity, GLNU) demonstrated the strongest correlation
with the HJS-narrowing index among all extracted features and utilized in further
analysis. Classification rules employing GLNU feature were introduced to characterize a
hip as normal or osteoarthritic and to assign it to one of three severity categories,
formed in accordance with the KL scale. Application of the proposed rules resulted in
Received 13 December
relatively high classification accuracies in characterizing a hip as normal or 2004
osteoarthritic (90.6%) and in assigning it to the correct KL scale category (88.9%). Revised 8 June 2005
Furthermore, the strong correlation between the HJS-narrowing index and the Accepted 1 July 2005
pathological GLNU (r520.9, p,0.001) was utilized to provide percentages quantifying
DOI: 10.1259/bjr/87956832
hip OA-severity. Texture analysis may contribute in the quantitative assessment of OA-
severity, in the monitoring of OA-progression and in the evaluation of a ’ 2006 The British Institute of
chondroprotective therapy. Radiology

Osteoarthritis (OA) is a common joint disease that
causes degenerative alterations in the hip as well as other
joints [1]. Characteristic radiological manifestation of hip
OA includes irregular superolateral, superior or super-
omedial hip joint space (HJS) narrowing, femoral and
acetabular subchondral bone sclerosis, development of
marginal osteophytes, as well as femoral and acetabular
subchondral cysts formation [2].
Plain film radiography remains the most prevalent
imaging modality for diagnosis of hip OA in clinical
routine, despite its limited sensitivity compared with
innovative imaging techniques, such as CT and MRI [3].
Radiographic assessment of hip OA comprises diagnosis,
evaluation of severity, and monitoring of progression of
structural alterations related to the disease [4]. A number of
qualitative or semi-quantitative grading systems have been
proposed for assessing hip OA [5–8], with the Kellgren and
Lawrence (KL) grading scale [5] being considered the gold
standard despite its deficiencies [9]. A reliable index for
Address correspondence to: George S Panayiotakis.
The first author was supported by a grant from the State Scholarship
Foundation (SSF), Greece.
monitoring hip OA progression on pelvic radiographs is
the progression of HJS-narrowing [6, 10], which may be
estimated either manually [11, 12], or by computerized
methods [13–15]. The latter are more sensitive, accurate,
reproducible, and thus more reliable [16].
Texture analysis refers to algorithms developed to
quantify image texture information that may, or may not,
be perceived visually [17]. Although texture analysis has
been previously employed in examining knee OA by
computer processing of radiographic images [18, 19], hip
OA has only been investigated in one study by computer
analysis (fractal geometry) of digitized histological
sections from the femoral head [20]. So far, the
quantitative assessment of hip OA has mainly relied on
measurements of HJS-width or HJS-area performed on
pelvic radiographs [11–15]. To our knowledge, the
textural properties of radiographic HJS in OA hips, as
well as the capability of computer based radiographic
texture analysis in evaluating the severity of hip OA
have not been previously investigated.
In the present study, a non-invasive method was
developed for analysing the structure of HJS from pelvic
radiographs and for evaluating the severity of hip OA,

232 The British Journal of Radiology, March 2006

Assessment of hip osteoarthritis using texture analysis
employing computerized texture analysis. In particular,
(i) textural features were extracted from the outlined
region of each radiographic HJS, (ii) textural-feature
thresholds, bearing good correlation to KL scale severity
grades, were established for grading OA automatically,
and (iii) an index was introduced for evaluating OA-
severity.
Methods and materials
Radiographs and patients
32 anteroposterior pelvic radiographs of standing
weight-bearing osteoarthritic patients were collected,
giving in total 64 hip joint images. All radiographs were
retrieved from the medical records of individuals who
were candidates for total hip arthroplasty at the
Department of Orthopaedics in our Hospital. From the
total number of patients, 18 were verified for unilateral
and 14 for bilateral hip OA. Patients’ ages ranged
between 49 years and 83 years with a mean age of 66.7
years. The American College of Rheumatology criteria
[21] were used for OA diagnosis.
All pelvic radiographs were obtained using a Siemens
X-ray unit (Polydoros 50; Siemens, Erlangen, Germany).
Radiographic protocol comprised alignment of the X-ray
beam 2 cm above the pubic symphysis, a focus–film
distance of 100 cm, tube voltage between 70 kVp and
80 kVp, and use of a fast screen and film cassette (30 cm
6 40 cm). Digitization of radiographs was performed at
12 bits (4096 grey levels) and 146 ppi (5.8 pixels mm21)
spatial resolution, using a laser digitizer for medical
applications (Lumiscan 75; Lumisys, Sunnyvale, CA)
[22]. Digitizer performance was evaluated employing a
quality control protocol [23]. All radiographs fulfilled a
specific criterion concerning safeguard against variations
in hip rotation, introduced by the experienced orthopae-
dists. According to this criterion, the difference between
the widths of projected lesser trochanters on each
(a)
Figure 1. Example of (a) an original and (b) the corresponding processed digitized radiograph with the contrast-limited
adaptive histogram equalization enhancement algorithm.
The British Journal of Radiology, March 2006
radiograph should not exceed 8 mm. Measurements on
radiographs were performed by custom developed
software [24–26].
Three experienced orthopaedists assessed the severity
of OA employing the KL grading scale. The KL scale
defines five categories of OA-severity (0–4), with KL
grades ¢2 corresponding to osteoarthritic pathology [5].
Based on the KL scale, patients were grouped into three
major OA-severity categories: Normal/Doubtful (KL50,
1), Mild/Moderate (KL52, 3), and Severe (KL54).
Accordingly, 18 unilateral-OA patients were assigned
to Normal/Doubtful category, 9 to Mild/Moderate and 9
to Severe. The corresponding numbers for the bilateral
patients were 0/7/21.
Radiograph enhancement
Pelvic radiographs were first processed by means of
custom developed software based on the contrast-limited
adaptive histogram equalization (CLAHE) method [27], in
order to emphasise the articular margins of the hip joint.
The CLAHE method partitioned the images into con-
textual non-overlapping regions. Within each region the
local histogram was obtained, clipped to a specific limit
and then histogram equalization was performed within
the region. Figure 1 shows a digitized radiograph
enhanced by the implementation of CLAHE algorithm.
On each enhanced radiograph two regions of interest
(ROIs), one from the osteoarthritic HJS and one from the
contralateral normal HJS, were manually outlined by three
experienced orthopaedists, in accordance with the method
proposed by Conrozier et al [13]. As shown in Figure 2,
each ROI was defined within an acute angle determined
by the patient’s standard anatomical landmarks.
Texture analysis of radiographic hip joint space
A total of 19 textural features were extracted from each
segmented HJS-ROI (see Figure 3), utilizing custom
(b)
233

Figure 2. Hip joint space-region of interest (HJS-ROI)
delineation within AOB. A: highest point of the homolateral
sacral wing, O: centre of the femoral head, and B: lateral rim
of the acetabulum.
Figure 3. Grey scale image of hip joint space region of
interest (ROI) delineated in Figure 2.
developed algorithms. (i) Four textural features were
computed from the ROI’s grey level histogram [28], (ii)
10 from the ROI’s grey level co-occurrence matrix [29]
and (iii) five using the ROI’s grey level run-length matrix
[30].
234
I S Boniatis, L I Costaridou, D A Cavouras et al
Textural feature selection
From the extracted 19 textural features, selection was
based on their correlation to an index evaluated for each
of the unilateral OA-group patient, employing Equation
(1):
HJSAnormal {HJSApath
HJS{narrowing~ |100 (1)
HJSAnormal
where HJSAnormal and HJSApath express the number of
pixels corresponding to the manually segmented con-
tralateral normal and osteoarthritic HJS-ROIs, respec-
tively. We have introduced this index, since it quantifies
OA-severity by expressing the HJS-narrowing as a
percentage of HJS-area difference between the osteoar-
thritic and contralateral normal HJS. This index is
expected to give zero value in case of normal joints,
since differences in hip joint spaces have been shown to
be negligible in normal individuals [13, 31].
The validity of the proposed HJS-narrowing index was
evaluated by examining its correlation with the KL scale,
since the latter is considered to be the gold standard for
OA-severity assessment. Analysis of HJS-narrowing
index performance compared with KL scale is provided
in a corresponding paragraph of the Results and
Discussion section.
Statistical analysis
The existence of statistically significant differences
between osteoarthritic and contralateral normal hips was
investigated in the patients of the unilateral OA-group.
Differences in HJS-area or in textural features were
examined by means of the two-tailed student’s paired t-
test. To assess the relationship between the HJS-narrow-
ing index and each textural feature extracted from
osteoarthritic HJS-ROIs, the Pearson’s correlation coeffi-
cient was used. To evaluate intraobserver and inter-
observer reproducibility concerning HJS-area
measurements and GLNU calculated values, the coeffi-
cient of variation (CV) was used [32]. Accordingly, all
radiographs were separately evaluated by each one of
the experienced orthopaedists twice, with about a
1 month interval between evaluations. Evaluation scores
were used to calculate the CV, which provides (e.g. see
Conrozier et al [13]) an assessment of interobserver or
intraobserver reproducibility; low coefficient values
correspond to high degree of reproducibility. Referring
to measured quantities, normality of their distributions
was assessed by means of the Lilliefors test [33]. For non-
gaussian distributions, a logarithmic transformation
(log10) was applied to corresponding data. Matlab
Statistics Toolbox and Matlab Curve Fitting toolbox
(The MathWorks Inc., Natick, USA) were used for
statistical and regression analysis.
Results and discussion
In a digital image, texture is characterized by intensity
properties (tone) and spatial inter-relationships (struc-
ture) of image pixels, depicting spatial distribution of
The British Journal of Radiology, March 2006
Assessment of hip osteoarthritis using texture analysis
grey level variations in the image [29, 34]. In a digitized
plain radiograph, a two-dimensional spatial distribution
of grey-level variation is formed by projecting on a two-
dimensional level the three-dimensional spatial distribu-
tion of the X-ray attenuation coefficients [17]. In the
present paper, textural properties of each analysed
radiographic HJS-ROI were attributed to X-ray attenua-
tion, due to superimposed three-dimensional anatomical
structures of articular cartilage, posterior acetabular wall
and iliac bone. Therefore, the analysed radiographic ROI
comprises of either osteoarthritic and/or normal super-
imposed anatomical components. Consequently, digital
image texture analysis attempts to assess the existence
and/or severity of structural alterations related to OA.
In patients with unilateral hip OA, statistical analysis
revealed the existence of statistically significant differ-
ences in 11 (out of 19) textural features values between
osteoarthritic and contralateral normal HJS-ROIs. Mean
values (¡ standard deviation (SD)) of significantly
differing textural features are presented in Table 1.
These differences demonstrate textural alterations in
radiographic HJS due to OA, which can be attributed to
cartilage and subchondral bone tissue alterations asso-
ciated to the disease. Articular cartilage performs
mechanical functions providing transmission and dis-
tribution of high loads to underlying bone, maintenance
of contact stresses at low levels, reduced frictional
resistance to movement and shock absorption with these
biomechanical properties being related to cartilage
molecular–biochemical composition [35, 36]. Typical
OA manifestations concern softening, ulceration, focal
disintegration and the final loss of articular cartilage [37].
Alterations in chemical composition of articular cartilage
have been associated with remodelling (increased den-
sity and stiffness) of subchondral bone in the form of
subchondral sclerosis [38, 39]. Taking into account that
structural alterations concern only osteoarthritic hips,
differentiation of textural properties between normal and
osteoarthritic HJS of unilateral OA-patients seems
reasonable.
Previous studies on quantitative assessment of hip OA
rely on measurements of the width or area of the
radiographic HJS [11–15]. In the present study, hip OA-
Table 1. Mean values (¡SD) of statistically significantly
differing textural features of contralateral normal and
osteoarthritic HJS-ROIs
Textural feature Normal Osteoarthritic R The strongest correlation was found between the HJS-
Grey level co-occurrence matrices-mean values
Entropy 0.7 (¡0.2) 0.8 (¡0.2) 0.2
Contrasta 20.7 (¡0.2) 20.6 (¡0.2) 0.3
Inverse difference moment 0.9 (¡0.1) 0.9 (¡0.1) 20.4
Sum of squaresa 20.3 (¡0.3) 20.2 (¡0.3) 0.2
Difference entropy 0.2 (¡0.1) 0.2 (¡0.1) 0.3
Difference variance 0.2 (¡0.1) 0.2 (¡0.1) 0.3
Grey level run length matrices-mean values
Short runs emphasisa 20.5 (¡0.1) 20.4 (¡0.1) 0.5
Long runs emphasis 12.9 (¡3.0) 10.1 (¡2.9) 20.6
Grey level non-uniformity 415.2 (¡92.8) 139.6 (¡73.8)20.9
Run length non-uniformity 305.0 (¡81.2) 110.0 (¡59.8)20.7
Runs percentagea 20.5 (¡0.1) 20.4 (¡0.1) 0.6
SD, standard deviation; HJS-ROIs, hip joint space regions of
interest.
a
Values after logarithmic transformation (log10).
The British Journal of Radiology, March 2006
severity was estimated by the introduction of the HJS-
narrowing index that evaluates the percentage of HJS-
area difference between the osteoarthritic and the
contralateral normal hip (Equation (1)). Repeated mea-
surements of the HJS-area concerning the same observer
were not found to differ significantly. Intraobserver
reproducibility was found on average high for both the
HJS area measurements (CV53.4%) and the correspond-
ing GLNU values (CV53.9%). Similarly, interobserver
reproducibility was also found high, 4.2% and 4.4% for
HJS-area measurements and GLNU values, respectively.
Mean values (¡SD) of HJS-area for osteoarthritic and
contralateral normal hips were found equal to 33.7
(¡20.3) mm2 and 105.0 (¡23.8) mm2, respectively. HJS-
area values were statistically smaller (p,0.001) in
osteoarthritic than in the contralateral normal hips, while
previous studies on normal individuals have found no
statistical differences between the two hips [13, 31]. HJS-
narrowing index was evaluated for each one of the 18
unilateral patients and the mean and standard deviation
of the HJS-narrowing index were calculated for the
Mild/Moderate and Severe OA categories. Based on
these means and standard deviations, classification rules
(see Table 2) regarding the assessment of OA-severity
were introduced (HJS-narrowing index Mean value ¡
2SD). Referring to Table 2, an osteoarthritic hip was
characterized as Severe if its HJS-narrowing index was
greater than 77.9, as Mild/Moderate for index values
within the interval [11.6, 77.9], and as Normal/Doubtful
for OA if index values were lower than 11.6.
The introduced classification rules were tested against
the KL classification of the unilateral OA patients
(Table 3). Referring to Table 3, all hips corresponding
to Mild/Moderate OA-severity category were classified
correctly, while classification accuracy of hips with
Severe OA was 77.8%, resulting in a relatively high
overall classification precision of 88.9%. Taking into
consideration that our method relies solely on the
assessment of HJS-narrowing, deviations of our results
from the KL scale may be attributed to the fact that the
KL scale evaluates, besides HJS-narrowing, the presence
of osteophytes, subchondral sclerosis, and subchondral
cysts.
Feature selection on the basis of Pearson’s correlation
coefficients between each of the textural features
extracted from osteoarthritic HJS-ROIs and the HJS-
narrowing index are summarized in column 4 of Table 1.
narrowing index and the pathological GLNU textural
feature (r520.9, p,0.001). This relationship is presented
graphically in Figure 4. As it can be observed, a
Table 2. Classification rules for assessment of osteoarthritis
severity concerning HJS-narrowing index
Osteoarthritis HJS-narrowing Classification rule
severity according index,
to KL grading scale mean ¡2SD
Severe 82.8 (¡2?8.2) HJS-narrowing
index .77.9
Mild/Moderate 50.5 (¡2?19.5) 11.6¡HJS-narrowing
Index¡77.9
KL, Kellgren and Lawrence; SD, standard deviation; HJS, hip
joint space.
235
 I S Boniatis, L I Costaridou, D A Cavouras et al

Table 3. Comparison of results obtained by the KL scale and the proposed classification rules concerning HJS-narrowing index
Osteoarthritis severity HJS-narrowing index.77.9 11.6¡HJS-narrowing Sum(s) Success percentage
according to KL scale (Severe) index¡77.9
(Mild/Moderate)

Severe 7 2 9 77.8%
Mild/Moderate 0 9 9 100%
Sum(s) 7 11 18 88.9%
KL, Kellgren and Lawrence; HJS, hip joint space.

Figure 4. Hip joint space (HJS)-

narrowing index versus pathologi-
cal grey level non-uniformity
(GLNU) textural feature. Solid line
is the best line fitted to data points
(N). Horizontal solid lines define the
regions of osteoarthritis severity.
Prediction bounds are denoted by
dotted lines.

regression line, described by Equation (2):
HJS{narrowing~{0:275|GLNUpath z105
fitted data adequately. The negative slope of the
regression line reflects the fact that in advanced stages
of the disease, characterized by greater HJS-narrowing
and thus higher HJS-narrowing index values, grey level
intensities are more uniformly distributed (see Appendix
1) [30] within the region of radiographic HJS.
Subsequently, the selected GLNU textural feature was
utilized in classification rules concerning the assessment
of hip OA-severity. Based on the mean and standard
deviation of the GLNU values, which were computed
from the normal hips of the unilateral OA-group, a
reference threshold value for GLNU, equal to 322.5 was
employed (see Table 4) for characterizing a hip as either
normal (GLNU.322.5) or osteoarthritic (GLNU¡322.5).
Using the contralateral normal hip for establishing
Table 4. Classification rules for assessment of osteoarthritis
severity concerning GLNU textural feature
Osteoarthritis GLNU (Mean¡SD) Classification rule
severity according
to KL grading scale
Severe (88.0¡49.1) GLNU,137.0
Mild/Moderate (191.1¡56.4) 137.0¡GLNU¡322.5
Normal/Doubtful (415.2¡92.8) GLNU.322.5
KL, Kellgren and Lawrence; SD, standard deviation; GLNU,
grey level non-uniformity
thresholds for hip osteoarthritis has been also employed
in previous studies. Conrozier et al [13] suggested the
(2) establishment of reference values by measuring the HJS-
width and HJS-area of the normal hips in patients with
unilateral OA, while other studies used the minimum
joint space width for classifying a hip as osteoarthritic [7,
40–44]. In the present study, however, a textural-feature
based classification rule was employed instead.
For verification purposes, GLNU values extracted
from the HJS-ROIs of the unilateral osteoarthritic
patients were subjected to the specific rule. It was found
that 17/18 (94.4%) of the contralateral normal hips and a
similar number (17/18) of the osteoarthritic hips were
characterized correctly. When the same classification rule
was applied to the bilateral OA-group, 24/28 (85.7%)
hips were correctly characterized as osteoarthritic. In
total, an overall classification accuracy of 90.6% (58/64),
for assigning normal and osteoarthritic hips to the
correct category, was achieved.
Besides characterizing a hip as normal or osteoarthitic,
the capacity of GLNU textural feature was tested in
establishing rules for differentiating hip OA-severity.
Accordingly, rules for distinguishing hip OA-severity
were formed, as shown in Table 4, that were defined on
the basis of the mean and standard deviation GLNU
values, obtained for hips assigned by the experienced
orthopaedists to the same KL scale severity category.
Referring to Table 4, a hip was characterized as Severe if
its corresponding GLNU value was lower than 137.0,
Mild/Moderate if its GLNU value was within the
interval [137.0, 322.5], and as Normal/Doubtful for OA
if its GLNU value was greater than 322.5. For verification

236 The British Journal of Radiology, March 2006

Assessment of hip osteoarthritis using texture analysis
Table 5. Comparison of results obtained by the KL scale and the proposed classification rules concerning GLNU textural feature
Osteoarthritis severity GLNU ,137.0 137.0 ¡GLNU¡322.5
according to KL scale (Severe) (Mild/Moderate)
Severe 7 2
Mild/Moderate 1 8
Normal/Doubtful 0 1
Sum(s) 8 11
KL, Kellgren and Lawrence; GLNU, grey level non-uniformity.
purposes, these classification rules were applied to the
unilateral OA-group and results were compared with the
KL scale classification (Table 5).
Referring to Table 5, the highest accuracy (94.4%) was
achieved for normal hips. For hips with Severe OA the
corresponding value was 77.8%, while for Mild/
Moderate hips the accuracy was 88.9%. Finally, an
overall classification accuracy of 88.9% (32/36) was
achieved. To our knowledge, textural-feature based
classification rules have not been proposed in previous
hip OA studies.
Finally, the strong correlation of GLNU textural feature
to the HJS-narrowing index was utilized to establish
means of quantification of hip OA from textural properties
of radiographic HJS (via GLNUpath feature) employing
Equations (1) and (2). This is important for monitoring the
progression of the disease and for assessing the effective-
ness of a treatment. Referring to the classification rules of
Table 4, percentages corresponding to OA-severity cate-
gories could be established using Equation (2). Thus, index
values for Severe OA were greater than 67.3%, for Mild/
Moderate OA within the interval [16.3%, 67.3%] and for
Normal/Doubtful OA less than 16.3%.
In this way, a hip may be assigned to an OA-severity
scale and its osteoarthritis, if it exists, can be evaluated
from its radiographic texture. This is of value because the
OA of patients suffering from bilateral-OA, which is often
the case, can be now quantified employing Equation (2),
whereas OA quantification in a manner similar to
Equation (1) applies only to patients with unilateral OA.
Conclusions
Alterations in the radiographic depiction of hip joint
space texture, due to osteoarthritis, were evaluated and
related to the severity of osteoarthritis, as defined by the
KL scale. Specifically, the GLNU textural feature, which
was selected considering its strong correlation to the
HJS-narrowing index, demonstrated high classification
accuracy in distinguishing hip OA-severity categories. In
addition, considering the high reproducibility derived
for the GLNU, the proposed method may have a
contribution in monitoring of OA-progression, as well
as in the evaluation of a chondroprotective therapy.
Acknowledgments
The authors thank the staff of the Departments of
Orthopaedics and Radiology for their contribution to this
work.
The British Journal of Radiology, March 2006
GLNU.322.5 Sum(s) Success percentage
(Normal)
0 9 77.8%
0 9 88.9%
17 18 94.4%
17 36 88.9%
References
1. Stoker DJ. Radiology now. Osteoarthrosis of the hip: one
disease or many? Br J Radiol 1977;50:81–3.
2. Altman RD, Bloch DA, Dougados M, Hochberg M,
Lohmander S, Pavelka K, et al. Measurement of structural
progression in osteoarthritis of the hip: the Barcelona
consensus group. Osteoarthritis Cartilage 2004;12:515–24.
3. Peterfy CG. Imaging of the disease process. Curr Opin
Rheumatol 2002;14:590–6.
4. Ory PA. Radiography in the assessment of musculoskeletal
conditions. Best Pract Res Clin Rheumatol 2003;17:495–512.
5. Kellgren JH, Lawrence JS. Radiological assessment of
osteoarthritis. Ann Rheum Dis 1957;16:494–501.
6. Altman RD, Fries JF, Bloch DA, Carstens J, Cooke DT,
Genant H, et al. Radiographic assessment of progression in
osteoarthritis. Arthritis Rheum 1987;30:1214–25.
7. Croft P, Cooper C, Wickham C, Coggon D. Defining
osteoarthritis of the hip for epidemiologic studies. Am J
Epidemiol 1990;132:514–22.
8. Lane NE, Nevitt MC, Genant HK, Hochberg MC. Reliability
of new indices of radiographic osteoarthritis of the hand
and hip and lumbar disc degeneration. J Rheumatol
1993;20:1911–8.
9. Spector TD, Cooper C. Radiographic assessment of osteoar-
thritis in population studies: whither Kellgren and
Lawrence? Osteoarthritis Cartilage 1993;1:203–6.
10. Altman R, Brandt K, Hochberg M, Moskowitz R. Design
and conduct of clinical trials in patients with osteoarthritis:
recommendations from a task force of the Osteoarthritis
Research Society. Osteoarthritis Cartilage 1996;4:217–43.
11. Lequesne M. Chondrometry: quantitative evaluation of
joint space width and rate of joint space loss in osteoar-
thritis of the hip. Rev Rhum Engl Ed 1995;62:155–8.
12. Hilliquin P, Pessis E, Coste J, Mauget D, Azria A, Chevrot
A, et al. Quantitative assessment of joint space width with
an electronic caliper. Osteoarthritis Cartilage 2002;10:542–6.
13. Conrozier T, Tron AM, Balblanc JC, Mathieu P, Piperno M,
Fitoussi G, et al. Measurement of the hip joint space using
computerized image analysis. Rev Rhum Engl Ed
1993;60:105–11.
14. Conrozier T, Vignon E. Quantitative radiography in
osteoarthritis: computerized measurement of radiographic
knee and hip joint space. Best Pract Res Clin Rheumatol
1996;10:429–33.
15. Gordon CL, Wu C, Peterfy CG, Duryea J, Klifa C, Genant
HK. Automated measurement of radiographic hip joint
space width. Med Phys 2001;28:267–77.
16. Conrozier T, Favret H, Mathieu P, Piperno M, Provvedini
D, Taccoen A, et al. Influence of the quality of tibial plateau
alignment on the reproducibility of computer joint space
measurement from Lyon schuss radiographic views of the
knee in patients with knee osteoarthritis. Osteoarthritis
Cartilage 2004;12:765–70.
17. Bocchi L, Coppini G, De Dominicis R, Valli G. Tissue
characterization from X-ray images. Med Eng Phys
1997;19:336–42.
237

18. Lynch JA, Hawkes DJ, Buckland-Wright JC. Analysis of
texture in macroradiographs of osteoarthritic knees using
the fractal signature. Phys Med Biol 1991;36:709–22.
19. Lynch JA, Hawkes DJ, Buckland-Wright JC. A robust and
accurate method for calculating the fractal signature of
texture in macroradiographs of osteoarthritic knees. Med
Inform 1991;16:241–51.
20. Fazzalari NL, Parkinson IH. Fractal properties of subchon-
dral cancellous bone in severe osteoarthritis of the hip. J
Bone Miner Res 1997;12:632–40.
21. Altman R, Alarcón G, Appelrouth D, Bloch D, Borenstein D,
Brandt K, et al. The American college of rheumatology
criteria for the classification and reporting of osteoarthritis
of the hip. Arthritis Rheum 1991;34:505–14.
22. Lumiscan 75, system specifications. Lumisys Inc. 1998;
http://www.lumisys.com/support/manuals.html [Accessed
18 October 2005].
23. Efstathopoulos EP, Costaridou L, Kocsis O, Panayiotakis G.
A protocol-based evaluation of medical image digitizers. Br
J Radiol 2001;74:841–6.
24. Sakellaropoulos P, Costaridou L, Panayiotakis G. An image
visualisation tool in mammography. Med Inform Internet
Med 1999;24:53–73.
25. Sakellaropoulos P, Costaridou L, Panayiotakis G. Using
component technologies for web-based wavelet enhanced
mammographic image visualization. Med Inform Internet
Med 2000;25:171–81.
26. Sakellaropoulos P, Costaridou L, Panayiotakis G. A wavelet
based spatially adaptive method for mammographic con-
trast enhancement. Phys Med Biol 2003;43:787–803.
27. Pizer SM, Amburn EOP, Austin JD, Cromartie R,
Geselowitz A, Greer T. Adaptive histogram equalization
and its variations. CVGIP (Computer Vision, Graphics and
Image Processing) 1987;39:355–68.
28. Gonzalez RC, Woods RE, editors. Digital image processing.
New Jersey, USA: Prentice-Hall Inc., 2002.
29. Haralick RM, Shanmugam K, Dinstein I. Textural features
for image classification. IEEE Trans Syst Man Cybern
1973;SMC-3:610–21.
30. Galloway MM. Texture analysis using gray level run
lengths. Comput Graph Image Proc 1975;4:172–9.
31. Goker B, Sancak A, Arac M, Shott S, Block JA. The
radiographic joint space width in clinically normal hips:
effects of age, gender, and physical parameters.
Osteoarthritis Cartilage 2003;11:328–34.
32. Altman DG, editor. Practical statistics for medical research.
London, UK: Chapman & Hall, 1991.
33. Lilliefors HW. On the Kolmogorov-Smirnov test for normal-
ity with mean and variance unknown. J Am Stat Assoc
1967;62:399–402.
34. Amadasun M, King R. Textural features corresponding to
textural properties. IEEE Trans Syst Man Cybern
1989;19:1264–74.
35. Armstrong CG, Mow VC. Variations in the intrinsic
mechanical properties of human articular cartilage with
age, degeneration and water content. J Bone Joint Surg Am
1982;64:88–94.
36. Jeffrey DR, Watt I. Imaging hyaline cartilage. Br J Radiol
2003;76:777–87.
37. Mankin HJ, Mow VC, Buckwalter JA, Ianotti JP, Ratcliffe A.
Form and function of articular cartilage. In: Simons S,
238
I S Boniatis, L I Costaridou, D A Cavouras et al
editor. Orthopaedic basic science. Chicago, USA: American
Academy of Orthopaedic Surgeons, 1994:1–44.
38. Radin EL, Rose RM. Role of subchondral bone in the
initiation and progression of cartilage damage. Clin Orthop
1986;213:34–40.
39. Burr DB, Schaffler MB. The involvement of subchondral
mineralized tissues in osteoarthrosis: quantitative micro-
scopic evidence. Microsc Res Tech 1997;37:343–57.
40. Jacobsen S, Sonne-Holm S, Soballe K, Gebuhr P, Lund B.
The distribution and inter-relationships of radiologic
features of osteoarthrosis of the hip. A survey of 4151
subjects of the Copenhagen City Heart Study: The
Osteoarthrosis Substudy. Osteoarthritis Cartilage
2004;12:704–10.
41. Bierma-Zeinstra SMA, Oster JD, Bernsen RMD, Verhaar
JAN, Ginal AZ, Bohnen AM. Joint space narrowing and
relationship with symptoms and signs in adults consulting
for hip pain in primary care. J Rheumatol 2002;29:1713–8.
42. Boegard T, Jonsson K. Hip and knee osteoarthritis.
Conventional X-ray best and cheapest method.
Lakartidningen 2002;44:4358–60.
43. Ingvarsson T, Hägglund G, Lindberg H, Lohmander SL.
Assessment of primary hip osteoarthritis: comparison of
radiographic methods using colon radiographs. Ann
Rheum Dis 2000;59:650–3.
44. Forsberg K, Nilsson BE. Coxarthrosis on the island of
Gotland. Increased prevalence in a rural population. Acta
Orthop Scand 1992;63:1–3.
Appendix 1
Description of the grey level non-uniformity (GLNU)
textural feature defined by Galloway [30]
The grey-level run is a set of consecutive linearly
adjacent pixels having the same grey level value. As
length of the run is considered the number of pixels
consisting the run. Each element p(i, j) of a grey-level run
length matrix represents the number of times an image
contains a run of length j for grey level i, in a specific
direction.
The mathematical definition of the GLNU textural
feature is:
!2
Ng
P P
Nr
pði, j Þ
i~1 j~1
GLNU~ Ng P
P Nr
pði, j Þ
i~1 j~1
where: p(i, j) is the (i, j)th element of grey level run length
matrix, Ng is the number of grey levels in the image and
Nr is the number of run lengths in the image.
Equally distributed runs throughout the grey levels,
correspond to low values for GLNU and vice versa.
The British Journal of Radiology, March 2006
The British Journal of Radiology, 79 (2006), 239–243

Trends in image quality in high magnification digital specimen

cabinet radiography
I P BIRCH, MSci, MPhys, C J KOTRE, PhD and R PADGETT, PhD

Regional Medical Physics Department, Newcastle General Hospital, Westgate Road, Newcastle NE4
6BE, UK

ABSTRACT. Advances in microfocus X-ray tube design together with the availability of
high resolution charge coupled device (CCD) detectors have led to the introduction of
high magnification digital specimen cabinets for the examination of tissue samples.
This paper explores the effect that the high magnification geometry permitted by such
units has upon image quality in terms of phase contrast edge enhancement, spatial
resolution and the appearance of test phantom images. Phase contrast effects and
spatial resolution were studied using a previously established method (using edge
profiles) and by computing the system spatial frequency response at various
geometries. It was demonstrated that the magnitude of the phase contrast Received 23 March 2005
enhancement effect reaches a stable maximum at a magnification of 6 4. It has also Revised 8 June 2005
been shown that a continual increase in both the spatial resolution together with an Accepted 4 July 2005
improved signal to noise ratio occurs up to the maximum permissible magnification
DOI: 10.1259/bjr/24723806
geometry, with effects of focal spot blur being negligible. In practice, the limited size
of the digital detector and the difficulty of object alignment can constrain the use of ’ 2006 The British Institute of
the very high magnification option. Radiology

Introduction small, but are most pronounced in regions of the object

Radiography of excised tissue samples is usually
carried out in specialized specimen cabinets. These units
commonly feature a focal spot size of approximately
0.05 mm, a film–focus distance of 50 cm and contain
movable shelves so that the distance between the sample
and the image receptor can be varied to provide a
geometric magnification up to 6 1.8. Low tube currents
are used, and low tube voltages in the region of 20 kVp
maximize contrast. A recently introduced model, the
MX20 (Faxitron, Wheeling, USA) features a nominal
focal spot size of only 0.02 mm, a receptor–focus distance
of 58 cm, magnification geometry of up to 6 5 and a
digital receptor consisting of a 5 cm 6 5 cm charge
coupled device (CCD) array with 1024 6 1024 pixels.
The aim of this paper is to investigate the image quality
trends with varying geometrical magnification on this
unit in terms of spatial resolution and signal to noise
ratio (SNR). In particular, the contribution of phase-
contrast information is assessed.
Phase contrast
Phase-contrast enhancement occurs at interfaces
between materials of differing X-ray refractive index.
As a spatially coherent X-ray beam propagates through
an X-ray transparent medium, the phase of the incident
wavefront becomes modified in a manner related to the
electron density of the medium. The resulting phase
gradient across the wavefront is equivalent to a change
in direction of the propagation of the wave. The angular
deflections from the initial direction of propagation are
where the X-ray refractive index is varying rapidly, such
as the interface between two different materials. The
direction of the deflection will vary from point to point
within a general object, depending on the structures
present, but produces a net effect of edge enhancement
between structures of differing X-ray refractive index
when imaged using an appropriate geometry. Smoothly
curved structures such as spheres and cylinders show
the effect particularly strongly, as they act in a manner
analogous to an optical lens [1].
Although phase-contrast imaging is frequently asso-
ciated with the use of monochromatic synchrotron
radiation [2], a simplified scheme based on conventional
microfocus X-ray tubes, with high spatial (lateral)
coherence, has been demonstrated [3, 4]. The lateral
coherence is enhanced by the use of low energy photons,
a small focal spot size and/or a large source–object
distance; many of these conditions are met by the
geometry used in specimen cabinet radiography.
The visual appearance of phase contrast enhancement
in the final image is edge enhancement at interfaces
between materials with differing X-ray refractive indices.
As there is also a change of X-ray attenuation across
these interfaces, the effect of the phase contrast is to
provide a subtle enhancement of the conventional
attenuation image.
Parameters under investigation and
experimental techniques
The investigation of the image quality trends in
magnification radiography took place on a Faxitron

The British Journal of Radiology, March 2006 239


MX20 microfocus specimen cabinet utilizing a 5 cm 6
5 cm CCD detector with a 50 mm pixel pitch. The
recorded pixel values from the detector were initially
verified to be linear with dose using an aluminium step
wedge. The focal spot was measured by the slit method
as 0.02 mm 6 0.02 mm and all experiments were
performed at a nominal 20 kV and 0.3 mA. Some
comparative measurements also took place on a film–
screen Micro50 specimen cabinet (measured focal spot of
0.08 mm 6 0.11 mm) using Kodak MinR2000 film and
screens.
Phase contrast detection
Phase contrast enhancement was demonstrated from
the imaged profiles of a low attenuation edge test object
where the magnitude of the phase signal is comparable
with that of the attenuation signal. A simple phase
contrast test object was constructed from the edge of a
standard radiography film (approximate thickness
180 mm). Thin aluminium foil (50 mm) was used to create
a ‘‘non-phase contrast’’ edge of similar linear attenuation
properties. Thin aluminium edges have been shown not
to produce measurable phase enhancement effects due to
the small phase signal being swamped by the larger
attenuation signal [4].
Both edges were imaged at all available magnifications
( 6 1, 6 1.5, 6 2, 6 3, 6 4 and 6 5). In each case, the
test edges were rotated by approximately 30 ˚ to the
coordinate system of the CCD pixel array to allow
oversampling of the edge profiles.
The phase contrast enhancement effects were further
analysed using the pre-sampled modulation transform
function (MTF) calculated using data from the edge
profiles. By comparing the frequency response of the
phase contrast edge with that of the non-phase contrast
edge (which yields the conventional MTF), the effect of the
contrast enhancement was quantified in frequency space.
Spatial resolution/geometric blurring
An inherent limitation of all forms of magnification
radiography is the finite size of the X-ray focus, causing
geometric blurring of an imaged object edge. When using
the 50 mm focus, this blurring limits specimen cabinet
radiography to approximately magnification 6 2, after
which blurring becomes unacceptable.
For digital radiography systems the spatial resolution
is also limited by the Nyquist frequency of the detector
defined by (2p)21 where p is the pixel size. The
theoretical maximum spatial resolution in the image
plane for the MX20 system using a 50 mm pixel detector
therefore is 10 cycles mm21. As this value is low
compared with that for film/screen, where over 20 line
pairs mm21 is more typical, the performance of the
digital detector, in terms of limiting spatial resolution for
specimen assessment, was also investigated.
The limiting spatial resolution for each magnification
geometry ( 6 1 to 6 5) was assessed by two methods; by
the 5% MTF cut-off frequency (cycles per mm), and with
a Huttner line-pairs test object (Type 25a) orientated at
45 ˚ to the pixel coordinate system (line pairs per mm).
240
I P Birch, C J Kotre and R Padgett
Visual appearance
The perceived SNR was visually evaluated using the
Leeds TOR(MAM) phantom which is usually associated
with the performance testing of mammography equip-
ment. The phantom contains three groups of test objects:
fibres, simulated microcalcification clusters and low
contrast plastic discs [5].
Specimen cabinets are often used for evaluation of
mammography core samples that may contain small
calcification clusters associated with developing cancers.
For this reason the microcalcification clusters in the
TOR(MAM) phantom were used to assess the overall
image quality.
The simulated microcalcification clusters in the
phantom were imaged at all magnification geometries.
The digital images were then rescaled (with no
pixel interpolation) and windowed so that the
features in each image appeared at the same size and
grey level. The visual appearance of the microcalcifica-
tion clusters was assessed on a standard computer
monitor.
Results
Phase contrast enhancement
Figure 1 demonstrates the averaged edge profiles
(pixel values) for the film and aluminium edge test
objects. The distance across each edge (the x-axis) has
been rescaled to account for the oversampling angle of
the edge profiles. The profiles of the film edge in Figure
1a demonstrate ‘‘overshoots’’ that become more appar-
ent with increasing image magnification. This is the
characteristic appearance of phase contrast for this type
of object [3]. The gradient of each profile also appears to
increase slightly with magnification. This occurrence
suggests that the phase contrast enhancement serves to
counteract geometrical blurring effects.
Figure 1b shows that the aluminium edge profiles have
no phase contrast overshoots. In addition, all these
profiles are comparable for each of the magnification
geometries used. This suggests that effects from geome-
trical blurring are small, meaning that maximum
magnifications can be used for all object types, with
little detectable image degradation in image spatial
resolution.
The frequency response curves of Figure 2a show that
phase contrast effects preferentially enhance the mid
spatial frequency range for magnification geometries
whilst the overall calculated limiting spatial resolution
(taken as the 5% level) is left relatively unchanged at
9 cycles mm21. For the non-phase contrast object it is
seen from Figure 2b that the shape of the frequency
response curve is consistent for all magnification
geometries from 6 1 to 6 4 with magnification 6 5
slightly lower, especially at the high frequency end. We
draw two conclusions from this observation. First, it
further demonstrates that the mid-frequency enhance-
ment shown in Figure 2a is a true phase contrast effect
and does not occur as a consequence of the magnification
geometry or changes in signal to noise ratio. Second,
geometrical unsharpness due to focal spot blurring is
The British Journal of Radiology, March 2006
Image quality trends in digital specimen cabinet radiography
minimal up to magnification 6 4 but there is some
degree of blurring at magnification 6 5.
In addition, Figure 2a shows that there is little
difference in mid-frequency response between the
geometries of 6 4 and 6 5 magnification. Therefore,
between these two geometries there is little to be gained
from additional phase contrast signal. This is likely to be
caused by the drop in lateral coherence of the poly-
chromic X-ray beam; a consequence of shortened focus to
object distance. At magnification 6 4 the geometry
appears optimized between the amount of phase contrast
created (focus to object distance) and the capability of the
detector to record the small angular phase contrast
deflections (object to detector distance).
Spatial resolution/geometrical blurring
To calculate the 5% MTF cut-off frequency in the object
plane (consistent with the Huttner test object) the spatial
frequency axis of the frequency response curves in
The British Journal of Radiology, March 2006
Figure 1. (a) Pixel value profiles
across image edge acquired at
various magnification geometries
using a phase contrast test object.
(b) Pixel value profiles across image
edge acquired at various magnifica-
tion geometries using a non-phase
contrast test object.
Figure 2b was rescaled to correct for the magnification
effect. These results together with those from the Huttner
test object are given in Table 1. This table shows that
despite the spatial resolution of the digital detector being
constrained to 10 lp mm21, much higher object plane
resolution is possible through image magnification. Note
that the maximum spatial frequency measurable with the
Huttner test object is 20 lp mm21 and this was reached
by magnification 6 3 geometry.
Figure 3 shows rescaled images of the 14th Huttner
group (16.6 lp mm21) for each of the 6 magnification
geometries. These images demonstrate a continual
improvement of image overall sharpness from magnifi-
cation 6 1 to magnification 6 5, consistent with the
estimated spatial resolution from the MTF assessment.
Visual appearance
Images of the 5th group (group ‘‘E’’, size 90–141 mm
[5]) of the simulated microcalcification clusters for the
241
 I P Birch, C J Kotre and R Padgett

Figure 2. (a) Image plane system

frequency response curves (MTFs)
calculated from edge profiles
acquired at various magnification
geometries using a phase contrast
test object. (b) Image plane system
frequency response curves (MTFs)
calculated from edge profiles
acquired at various magnification
geometries using a non-phase
contrast test object.

TOR(MAM) phantom are presented in Figure 4. The

images have been rescaled for magnification to represent
them as if in the object plane (i.e. to display all features at
the same size). An improvement of overall detail
detectability is seen with increasing magnification.

Table 1. Limited spatial resolution measurements using the

Huttner test object and frequency response method for
digital MX20 and conventional Micro50 units
Nominal lp mm21 from Cycles mm21 lp mm21 from
magnification Huttner at 5% Huttner
(MX20unit) MTF (MX20unit) (Micro50unit)

1.0 9 9 5
1.5 13.4 13.5 7.5
2 18.3 16 10
3 .20 24 n/a
4 .20 32 n/a Figure 3. 14th Group of Huttner (Type 25a) spatial
5 .20 40 n/a resolution test object. Images acquired at various magnifica-
tions and rescaled. (a) mag 6 1, (b) mag 6 1.5, (c) mag 6 2,
MTF, modulation transform function. (3) mag 6 3, (e) mag 6 4, (f) mag 6 5.

242 The British Journal of Radiology, March 2006

Image quality trends in digital specimen cabinet radiography
Figure 4. Microcalcification cluster number 5 of Leeds
TOR(MAM) mammography test object. Images acquired at
various magnifications and rescaled. (a) mag 6 1, (b) mag 6
1.5, (c) mag 6 2, (3) mag 6 3, (e) mag 6 4, (f) mag 6 5.
The standard deviation in pixel value in the uniform
background region was measured to be approximately
equal for each image. This result is expected as the
photon flux at the detector for a fixed exposure is
independent of the magnification geometry selected.
Increasing magnification with constant exposure factors
would, however, be expected to increase the amplitude
of large area signals in the plane of the object, and
therefore improve the SNR in the image, due to the
increased number of photons per unit area in the object
plane. At high magnification more photons will interact
with any given object feature, resulting in a larger
difference in total number of photons recorded due to
the presence of that feature. By simple geometry, the
photon flux will increase as the square of the magnifica-
tion factor. If quantum noise is considered to be the
dominant noise source, then the SNR for large area
objects would be expected to increase approximately in
proportion to the magnification factor. In addition, for
small objects comparable in size with the system point
spread function such as the microcalcifications in
Figure 4, increasing magnification will increase the size
of the object projected at the plane of the detector,
shifting the spatial frequencies down the system MTF
(Figure 2b) so that they are imaged at a larger signal
amplitude.
Discussion
The results above demonstrate that significant
improvements in the overall image quality of specimen
cabinet radiography can be achieved when using the
high magnification geometry available with the digital
Faxitron MX20 unit. We have shown that the 20 mm
focus permits up to 6 5 magnification with no
demonstrable loss in spatial resolution in the image
plane. This is compounded with the result of continual
increases in resolution in the object plane with increasing
The British Journal of Radiology, March 2006
magnification and a predicted maximum object resolu-
tion of 40 lp mm21 at magnification 6 5.
An improvement of SNR and spatial resolution has
also been shown to occur at high geometric magnifica-
tion, with the increased visibility of small, low contrast
objects in the TOR(MAM) phantom. However, improve-
ment occurs at the expense of a much reduced field size
(in the object plane).
An interesting result from this study is the fact the
Faxitron MX20 unit produces phase contrast image
enhancement at mid to high magnification geometries.
These enhancement effects produced an improvement in
the mid-frequency response. The overall effect of phase
contrast enhancement on clinical images will depend
entirely on the object being imaged, but the most
noticeable effects should be seen in the visibility of
filamentous and spherical objects and at interfaces
between materials of similar attenuation contrast [3].
Although we have demonstrated that a phase contrast
contribution is present, the improvement in image
quality is mainly governed by the increase in SNR and
object plane resolution produced at high magnification
geometries.
Conclusions
The results above suggest that, for a modern digital
specimen cabinet, with focal spot sizes in the order of
0.02 mm, image quality in terms of spatial resolution and
SNR in the object plane can be maximized by the use of
the highest magnification factor ( 6 5 in this case). Phase
contrast is also produced at high magnification geome-
tries with 6 4 magnification producing the optimum
results. However, it is appreciated that in practice, the
limited size of the digital detector and the difficulty of
object alignment may constrain the use of these very high
magnification options.
Acknowledgments
We would like to thank staff at the Breast Screening
Unit at Queen Elizabeth Hospital, Gateshead for their
help with this study.
References
1. Ingal V, Beliaevskaya E, Brianskaya A, Merkurieva R. Phase
mammography – a new technique for breast investigation.
Phys Med Biol 1998;43:2555–67.
2. Lewis R. Medical applications of synchrotron radiation x-
rays. Phys Med Biol 1997;42:1213–43.
3. Wilkins SW, Gureyev TE, Gao D, Pogany A, Stevenson AW.
Phase-contrast imaging using polychromatic hard x-rays.
Nature 1996;384:335–8.
4. Kotre CJ, Birch IP. Phase contrast enhancement of x-ray
mammography: a design study. Phys Med Biol
1999;44:2853–66.
5. Cowen AR, Brettle DS, Coleman NJ, Parkin GJS. A
preliminary investigation of the imaging performance of
photostimuable phosphor computed radiography using a
new mammographic quality control test object. Br J Radiol
1992;62:528–35.
243
 The British Journal of Radiology, 79 (2006), 244–247

Margins between clinical target volume and planning target

volume for electron beam therapy
S J THOMAS, MA, MSc, PhD

Medical Physics, Addenbrooke’s Hospital, Cambridge CB2 2QQ, UK

ABSTRACT. When growing a clinical target volume (CTV) to a planning target volume
(PTV), it is necessary to determine suitable margins, based on the systematic and Received 21 April 2005
random uncertainties. For electron therapy, where treatments are usually given with Revised 23 June 2005
single fields, the factors affecting the margin are very different in the direction of the Accepted 4 July 2005
incident beam from those in the perpendicular directions, since set-up errors do not
DOI: 10.1259/bjr/70202978
affect the depth of the 90% isodose. For a typical case, the perpendicular margins are
three times the margin in the direction of the incident beam. This gives rise to problems ’ 2006 The British Institute of
with volume growing algorithms if the beam axis is not aligned with a cardinal axis. Radiology

The International Commission of Radiation Units and
Measurements (ICRU), in reports 50 and 62 [1, 2], defines
the gross tumour volume (GTV), the clinical target
volume (CTV) and the planning target volume (PTV).
Both reports discuss factors contributing to the CTV-PTV
margin, but do not give any recipes for its calculation.
The British Institute of Radiology (BIR) has recently
published a report on Geometric Uncertainties in
Radiotherapy [3], which reviews sources of uncertainty
and describes methods of calculating CTV-PTV margins.
All the specific advice relates to photon beams rather
than electron beams.
ICRU report 71 [4] extends the work of ICRU 62 to
electron beam therapy. This report gives a recipe for
calculating the CTV-PTV margin, based on work by
Stroom et al [5]. A CTV-PTV margin which ensures at
least 95% of the dose to 99% of the CTV is given by:
CTV-PTV margin~2Sz0:7s
is the unblurred beam penumbra width, and b is a value
that depends on the beam configuration, being always 1.64
in the superior–inferior (sup-inf) direction for coplanar
beams, and taking lower values in transverse planes
depending on the number and arrangement of beams.
When b51.64 and sp53.2 mm, the last term of Equation (2)
approximates to 0.7s, as in Equation (1).
The derivation is based on the assumption of a CTV that is
approximately spherical, with an arrangement of beams
designed to conform the dose distribution to it in three
dimensions. However, the more usual situation in electron
therapy is as shown in Figure 1. A single beam, shaped by a
metal cut-out, is chosen with an energy appropriate to the
depth required to ensure that the 90% isodose covers the
PTV. It is apparent that the effect of geometrical uncertain-
ties in the x and y directions in Figure 1 is very different from
the effect of geometrical uncertainties in the z direction.
(1)

where S is the standard deviation for the systematic Systematic errors

(preparation) error, and s is the standard deviation for
the random (execution) error. However, this margin
recipe is based on photon beam therapy, making a
number of assumptions that do not hold for electron
beams.
The aim of the work described below is to develop a
method of calculating margins that is valid for the
conditions applying in electron beam therapy.
Theory
The methods described by the BIR report on Geometric
Uncertainties in Radiotherapy [3], for photon therapy with
multiple beams give the following margin, to ensure a
minimum dose to the CTV of 95% for 90% of patients:
CTV-PTV margin~2:5Szazbzb((s2 zsp 2 )0:5 {sp ) (2)
where S and s are as in Equation (1), a and b are corrections
for planning algorithm error and breathing, respectively, sp
The BIR report [3] describes four gaussian sources of
systematic error, the standard deviations of which may
be combined in quadrature; the doctor’s delineation
error Sdoctor, the organ position and shape (excluding
breathing) error Smotion, the set up error Ssetup and the
phantom transfer error Stransfer.
Sdoctor is the systematic error resulting from inter-
clinician and intraclinician variation in volume delinea-
tion. The issue of whether Sdoctor can be combined in
quadrature with other errors is still a matter of debate;
recent work by McKenzie [6] suggests that it cannot be
handled in the same manner as the other gaussian errors,
but requires an alternative theoretical basis. In the
example below I have omitted it, and assumed that it
has been included in the CTV.
Smotion is the systematic error in position and shape
(excluding breathing). It will not be affected by modality,
so can be treated in the same way as for photons. Ssetup (the
standard deviation of the systematic set-up error) can be
treated in the same way as for photons in the x and y
direction. However, in the z direction, most errors have no

244 The British Journal of Radiology, March 2006

Margins between CTV and PTV for electron beam therapy
Figure 1. A typical electron treatment. The planning target volume (PTV) is shown in dark grey, the collimator (cut-out) is
shown in light grey. The 90% isodose conforms to the PTV in the xy plane, at the depth of maximum PTV width.
effect on the position of the isodoses. A systematic shift of a
few millimetres in the z position of the patient relative to
the end of the applicator may have a small effect on
delivered dose (generally less than 1%), but will not affect
the depth of the isodoses.
The phantom transfer error Stransfer, is the error
accumulated in transferring image data through the
treatment planning system to the linear accelerator,
including errors in imaging, planning, and linear
accelerator geometry. Since the component from accel-
erator geometry will have no effect on the depth of
isodoses, Stransfer will be less in z than in x and y.
An additional systematic uncertainty, which affects
only the z direction, is uncertainty in electron density
derived from CT. For low atomic number materials,
published data for eight CT scanners showed a max-
imum error in electron density of 2.5%, with a standard
deviation below 1% [7]. Most electron treatments are
given through soft tissue. The Sdensity, in the depth of the
90% depth dose, varies with energy from 0.2 mm at
6 MeV to 0.6 mm at 21 MeV. For bone, if a standard
curve is used for all scanners, errors of up to 6% can be
observed, with a standard deviation below 2.5%. If
10 mm of the depth is bone of density 1.5, Sdensity
becomes 0.4 mm at 6 MeV, 0.8 mm at 21 MeV.
In all the clinical examples in the BIR report [3], the
dominant systematic gaussian errors are Ssetup and
Stransfer. Since these are insensitive to systematic errors
in the z direction, the problem will change from a 3D case
to a 2D case. As shown by Van Herk et al [8], this reduces
the systematic margin from 2.5S to 2.15S.
Linear errors
Geometric Uncertainties in Radiotherapy [3] defines two
linear errors, the ‘‘breathing positional error’’ b and the
The British Journal of Radiology, March 2006
‘‘treatment planning system photon-beam algorithm
error’’ a.
The breathing error b can be treated in the same
manner as for photons, since the derivation of the margin
is not dependent on modality.
Electron treatment planning algorithms do not usually
give as good agreement with measurement as do photon
algorithms. The errors are very dependent on the shape
of the patient surface, and the size and shape of
inhomogeneities. For photon planning, simple measure-
ments can determine whether the planning system over-
corrects or under-corrects the field sizes, and corrections
can be made. For electrons, this error is very plan
dependent, and is probably best not included in the PTV
margin. Hence a has been taken as zero.
Treatment execution errors
There are two random gaussian errors considered by
the BIR report [3], the daily set-up error sset-up and the
organ position and shape execution error, smotion. Both of
these combine in quadrature to give the s of Equation (2).
sset-up can be treated in the same way as for photons in
the x and y direction. However, in the z direction, most
errors have no effect on the position of the isodoses, for
the same reasons as given for systematic set-up errors.
smotion will be unaffected by modality, so can be
treated in the same way as for photons.
Unblurred penumbra width
The unblurred beam penumbra width sp requires a
very different treatment for electrons than for photons. In
the x and y directions, although the penumbra of an
electron field can still be defined by a gaussian, the width
245
 S J Thomas

of the gaussian varies very rapidly with depth and The depth dose fall-off can also be approximated with
energy. A penumbra can be described by an error a gaussian. Values of spdd can be chosen such that the
function with a parameter sp; an approximation for sp shape of the measured percentage depth dose curve
(in mm) can be derived from the data of Lax et al [9]: (centred on the depth of the 50%) is matched by values of
d50ð{d


 rffiffiffiffiffi 100  
Z E pffiffiffiffiffiffi exp {x2 =2spdd 2 dx (5)
sp ~ 21:3 {3:9 (3)
R 10 2pspdd
{?

Figure 2 gives an example of this fit; the shape of the

where Z is depth in mm, E is the electron energy at the
depth dose is well modelled from the depth of dose
surface in MeV, and R is the range in mm, which can
maximum to the depth of 5% dose. Table 1 shows values of
itself be approximated by:
spdd derived by this method, which vary from 7 mm at
6 MeV to 23 mm at 21 MeV. spdd is used in place of sp in
R~5:21E{3:76 (4)
Equation (2).
Fitting to a 90% dose level with a single beam gives a b
The depth to be used depends on the exact shape of the of 1.28, as derived in Van Herk et al [8]. If a simplified
systematic target volume (STV), which is the volume version of Equation (2) is required, to enable comparison
resulting after a margin is added to the CTV to account with Equation (1), the final term of Equation (2) can be
for systematic errors [3]. The STV-PTV margin accounts approximated to 0.3s, for values of s up to 5 mm, using
for the random (execution) errors. Table 1 calculates the the linear approximation of Van Herk et al [8], for x and
sp for two different depths. In the first case, the STV is y, if sp 55 mm,. For z, it approximates to 0.2s at 6 MeV,
assumed to be ellipsoidal, and symmetrically situated reducing to 0.1s at 18 MeV.
between 10 mm deep and the depth of the 90% isodose
(D90). In this case the widest point of the target volume
will be depth D15(10 mm + D90)/2. In the second case, Example
the calculation has been done at D90; this has been chosen
to deal with the extreme cases where the target volumes Let us assume we are treating a target volume in the
are widest at their deepest position. All penumbral head or neck. We will assume an anterior beam, so that the
widths are within 1 mm of 5 mm at D1, and within z direction of Figure 1 corresponds to the posteroanterior
2 mm of 6 mm at D90. (PA) direction. We will use values for the systematic and
random errors based on those used in Chapter 7 of
Geometric Uncertainties in Radiotherapy [3]. For the example
chosen, where the patient is immobilized, breathing errors
Table 1. Data used to characterize the beam profiles and are taken to be negligible, so b is omitted.
depth doses. D1, halfway between 10 mm and the depth of
Table 2 shows the resulting margins. For the values
the 90% isodose (D90), is the depth at the widest part of the
target volume in Figure 1. sp and spdd describe the shape of shown, the anteroposterior (AP) margin is about 3 mm,
the penumbra and depth-dose fall-off, respectively the superior-inferior (Sup-Inf) and lateral margins about
10 mm. This means that geometrical uncertainties will
Energy Practical D1 D90 sp at D1 sp at D90 spdd have a larger effect on the size of cut-out required than
(MeV) range (mm) (mm) (mm) (mm) (mm) (mm)
they do on the electron energy.
6 28 13 16 5 5 7
9 43 19 27 5 6 8
12 59 23 36 5 7 10
15 74 28 45 5 8 12
Discussion
18 90 31 52 5 8 16 For electron treatments, a much smaller CTV-PTV
21 106 33 56 4 8 23 margin is required in the direction of the incident beam

Figure 2. This illustrates the use of a

gaussian to model the shape of the
depth dose curve. The line is
measured beam data for a 15 MeV
beam, the points are calculated
from Equation (5), using a spdd of
12 mm. A close fit is observed from
the depth of maximum dose down
to the depth of 5%.

246 The British Journal of Radiology, March 2006

Margins between CTV and PTV for electron beam therapy

Table 2. Example of typical clinical target volume-planning of the incident beam, the effect of set up errors has no
target volume (CTV-PTV) margins for electron therapy. The effect on the margin, so margins are smaller.
beam is assumed to be an anterior beam. All distances are in In the direction perpendicular to the incident electron
millimetres. Values of sp and spdd for 12 MeV have been
beam, the margin required is approximately
used; changing the energy between 5 MeV and 21 MeV will
change the margin by a maximum of 0.1 mm in anterior-
2.15Sz+b+0.3s, where S and s are the standard devia-
posterior (AP), and a maximum of 0.2 mm right-left (R-L) and tions for systematic (preparation) errors and random
superior-inferior (S-I) (execution) errors, respectively, and b is the linear
breathing margin. In the direction of the incident beam,
Systematic errors AP R-L S-I this reduces to 2.15Sz +b+0.15sz, where Sz and sz
Smotion 1.0 1.0 1.0 exclude any set-up errors.
Stransfer 1.0 2.9 3.8
Sset-up 0.0 2.5 2.5 References
Sdensity 0.2 0 0
S(combined) 1.4 4.0 4.7 1. International Commission on Radiation Units and
Systematic52.15 S 3.0 8.5 10.0 Measurements. ICRU Report 50. Prescribing, recording and
reporting photon beam therapy. Bethesda MD: ICRU, 1993.
Treatment execution errors
2. International Commission on Radiation Units and
sset-up 0.0 2.5 2.5 Measurements. ICRU Report 62 (Supplement to ICRU report
smotion (target shape) 1.0 1.0 1.0 50). Prescribing, recording and reporting photon beam
s 1.0 2.7 2.7 therapy. Bethesda MD: ICRU, 1999.
sp (or spdd for AP ) 10.0 5.0 5.0 3. British Institute of Radiology Working Party. Geometric
Planning parameter (b) 1.28 1.28 1.28 uncertainties in radiotherapy. London, UK: British Institute
 0:5 
0.1 0.9 0.9 of Radiology, 2003.
Execution~b s2 zsp 2 {sp
4. International Commission on Radiation Units and
Total CTV-PTV margin 3.1 9.4 10.9 Measurements. ICRU Report 71. Prescribing, recording and
reporting electron beam therapy. Oxford University Press,
2004.
than perpendicular to it. With most volume growing 5. Stroom JC, deBoer HC, Huizenga H, Visser AG. Inclusion of
software, this is only straightforward in cases where the geometrical uncertainties in radiotherapy planning by means
beam direction is along one of the cardinal axes, e.g. an of coverage probability. Int J Radiat Oncol Biol Phys
anterior, a posterior or a true lateral beam. In cases where 1999;43:905–19.
the beam is being applied obliquely, the geometry is an 6. McKenzie AL. A novel way to allow for uncertainties in
expansion ellipsoid whose principal axes are not aligned delineation and changes in shape of target volumes in
with the cardinal axes. This cannot be dealt with by most radiotherapy. In: Chambers LA, Chambers IR, editors.
planning systems. One method of dealing with this Proceedings of the 11th Annual Scientific Meeting; 2004
would be to avoid making a CTV-PTV expansion, but September 6–8; York, UK. York, UK: Institute of Physics in
Engineering and Medicine, 2004.
instead to use the values directly in the field shaping and
7. Thomas SJ. Relative electron density calibration of CT
choice of energy.
scanners for radiotherapy treatment planning. Br J Radiol
1999;72:781–6.
8. Van Herk M, Remeijer P, Rasch C, Lebesque JV. The
Conclusions probability of correct target dosage: dose-population histo-
grams for deriving treatment margins in radiotherapy. Int J
The methodology of the BIR report on geometrical Radiat Oncol Biol Phys 2000;47:1121–35.
uncertainties [3] can be followed for electrons in the 9. Lax I, Brahme A, Andreo P. Electron beam dose planning
direction perpendicular to the incident beam, but using a using Gaussian beams. Improved radial dose profiles. Acta
2.15 multiplier for the systematic errors. In the direction Radiol Suppl 1983;364:49–59.

The British Journal of Radiology, March 2006 247

 The British Journal of Radiology, 79 (2006), 248–253

SHORT COMMUNICATION

Gold nanoparticles: a new X-ray contrast agent

1 1 1
J F HAINFELD, PhD, D N SLATKIN, MD, T M FOCELLA, BS and 2H M SMILOWITZ, PhD

1
Nanoprobes, Inc., 95 Horse Block Road, Yaphank, NY 11980 and 2University of Connecticut Health
Center, Farmington, CT 06030, USA

ABSTRACT. There have been few fundamental improvements in clinical X-ray contrast
agents in more than 25 years, and the chemical platform of tri-iodobenzene has not
changed. Current agents impose serious limitations on medical imaging: short imaging
times, the need for catheterization in many cases, occasional renal toxicity, and poor
contrast in large patients. This report is the first demonstration that gold nanoparticles
may overcome these limitations. Gold has higher absorption than iodine with less bone
and tissue interference achieving better contrast with lower X-ray dose. Nanoparticles
clear the blood more slowly than iodine agents, permitting longer imaging times. Gold
nanoparticles, 1.9 nm in diameter, were injected intravenously into mice and images
recorded over time with a standard mammography unit. Gold biodistribution was
measured by atomic absorption. Retention in liver and spleen was low with elimination
by the kidneys. Organs such as kidneys and tumours were seen with unusual clarity and
Received 4 February 2005
high spatial resolution. Blood vessels less than 100 mm in diameter were delineated, Revised 24 May 2005
thus enabling in vivo vascular casting. Regions of increased vascularization and Accepted 1 September
angiogenesis could be distinguished. With 10 mg Au ml21 initially in the blood, mouse 2005
behaviour was unremarkable and neither blood plasma analytes nor organ histology
DOI: 10.1259/bjr/13169882
revealed any evidence of toxicity 11 days and 30 days after injection. Gold nanoparticles
can be used as X-ray contrast agents with properties that overcome some significant ’ 2006 The British Institute of
limitations of iodine-based agents. Radiology

Contrast agents for X-rays are based on tri-iodoben-
zene with substituents added for water solubility.
Diatrizoate, an ionic form, was introduced in 1954, but
the high osmolality of this compound (1.57 osm kg21 for
a 300 mg I ml21 solution) was found to be the source of
chemotoxicity [1]. In the 1970s, a non-ionic form, iohexol,
lowered osmolality (0.67 osm kg21), and is still widely
used today under the names OmnipaqueH and
ExypaqueH, Amersham Health, Amersham, UK (now
GE Healthcare). Because osmolality was still excessive, a
dimeric form was introduced, iodixanol (AcupaqueH and
VisipaqueH, Amersham Health, Amersham, UK (now GE
Healthcare); 0.29 osm kg21). Intravascular agents based
on other mid-Z to high-Z elements have not been
successful due to toxicity, performance, or cost. The
low molecular weights of the iodine agents (diatrizoate,
613; iohexol, 821; iodixanol, 1550) effect rapid renal
clearance and vascular permeation, necessitating short
imaging times. Intra-arterial catheterization is therefore
commonly needed, but carries the risks of arterial
puncture, dislodgement of plaque, stroke, myocardial
infarction, anaphylactic shock and renal failure. A
further shortcoming of the current agents is in molecular
imaging, since their conjugates with antibodies or other
This study was supported in part by a National Cancer Institute
Small Business Innovative Research Phase 1 Grant 1R43CA83576-
01. JFH is part owner of Nanoprobes, Inc. Other authors do not have
any financial interest.
targeting moieties fail to deliver iodine to desired sites at
detectable concentrations.
Several other experimental X-ray contrast materials
show promise as blood pool agents, including standard
iodine agents encapsulated in liposomes [2, 3], a
dysprosium-DTPA-dextran polymer [4], polymeric
iodine-containing PEG-based micelles [5], perfluoroctyl
bromide [6], dervatized polylysine linked to iodine [7],
and iodine linked to a polycarboxylate core (P743,
MW512.9 kDa) [8]. Iron nanoparticles have been used
successfully as MRI contrast agents [9], but our report is
the first, to our knowledge, to use gold as an X-ray
contrast agent in vivo.
With a higher atomic number (Au, 79 vs I, 53), and a higher
absorption coefficient (at 100 keV: gold: 5.16 cm2 g21;
iodine: 1.94 cm2 g21; soft tissue: 0.169 cm2 g21; and bone:
0.186 cm2 g21), gold provides about 2.7 times greater
contrast per unit weight than iodine [10]. Imaging gold at
80–100 keV reduces interference from bone absorption
and takes advantage of lower soft tissue absorption which
would reduce patient radiation dose. Gadolinium has been
used instead of iodine to image the chest with half the X-
ray dose [11]. The higher molecular weight of nanoparti-
cles (here ,50 kDa) permits much longer blood retention,
so that useful imaging may be obtained after intravenous
injection, possibly obviating invasive catheterization for
diagnostic triage. Molecular imaging may also be possible
as each nanoparticle bound to a targeting agent would
deliver a ‘‘truckload’’ of ,250 gold atoms to a cognate

248 The British Journal of Radiology, March 2006

Short communication: Gold nanoparticles
receptor thereby increasing the signal. Although gold is
more costly than iodine, low detectable amounts and
significant benefits should enable feasible gold-mediated
clinical radiography.
Materials and methods
Animals and injections
Balb/C mice were injected subcutaneously in the thigh
with 106 EMT-6 syngeneic mammary carcinoma cells [12]
suspended in 0.05 ml of equal volumes of medium and
Matrigel. 10 days after tumour initiation, gold nanopar-
ticles were injected via a tail vein. Experimental protocols
using animals were approved by the University of
Connecticut Health Center animal care committee.
Gold nanoparticles
1.9¡0.1 nm gold nanoparticles were obtained from
Nanoprobes, Inc. (preparation # 1101, Yaphank, New
York, USA). The size of the nanoparticles was deter-
mined by electron microscopy. The concentration of
injected gold was 270 mg Au cm23, and volume injected
was 0.01 ml g21 mouse weight. Nanoparticles were
suspended in phosphate-buffered saline at pH 7.4.
Gold analysis
Tissues were excised, placed in tared vials, and
analysed for gold by graphite furnace atomic absorption
spectrometry using a Perkin Elmer 4100Z instrument
(Wellesley, Massachusetts, USA).
Radiographs
A Lorad Medical Systems mammography unit
(Hologic, Inc., Danbury, CT; model XDA101827) was
used with 8 mAs exposures (0.4 s at 22 kVp). Kodak
Min-R2000 mammography film, 18 cm 6 24 cm
(Eastman Kodak, Rochester, NY) was used.
Toxicity tests
60 outbred CD1 mice (male and female) were rando-
mized into four groups of 15 animals per group receiving
700 mg Au kg21, 70 mg Au kg21, or 7 mg Au kg21, or
sham-injected with phosphate buffered saline. Animals
were weighed and observed regularly for clinical signs.
Animals were euthanized by CO2 narcosis 1 day, 11 days,
and 30 days after intravenous gold injections and ,0.4 ml
blood was removed from the right ventricle immediately
after the cessation of breathing. Haematology analytes
included haematocrit, haemoglobin, total white [WBC]
and red [RBC] blood cell counts, neutrophil, lymphocyte,
monocyte, and eosinophil counts, mean corpuscular
volume, mean corpuscular haemoglobin, mean corpus-
cular haemoglobin concentration, WBC differential (per-
cent neutrophils, bands, lymphocytes, monocytes, and
The British Journal of Radiology, March 2006
eosinophils), and blood smear microscopy. Blood chem-
istry analytes included glucose, blood urea nitrogen (BUN),
creatine, calcium, phosphate, total protein, albumin, globin,
albumin:globulin ratio, alanine aminotransferase (ALT),
aspartate aminotransferase (AST), AST/ALT ratio, alkaline
phosphatase, total bilirubin, and direct bilirubin. Livers
and kidneys were weighed and slices of the following 24
tissues were prepared for microscopic study by formalin
fixation, paraffin embedding, and haematoxylin/eosin
staining: kidneys, liver, testes, epididymis, lungs, heart,
adrenals, bone, bone marrow, spinal cord, sciatic nerves,
oesophagus, stomach, duodenum, ileum, jejunum, colon,
cecum, lymph nodes, spleen, thymus, trachea, ovaries, and
uterus. Histopathology was evaluated by a board-certified
veterinarian (A G Richter, DVM, DACVP) and assessed
independently by a physician certified in anatomic
pathology (D N Slatkin, MD, DABP).
Results
Gold nanoparticles, 1.9 nm in diameter, were sus-
pended in phosphate-buffered saline and injected via a
tail vein into Balb/C mice bearing EMT-6 subcutaneous
mammary tumours. The vascular system was imaged in
planar projection using a clinical mammography unit.
Blood vessels as fine as 100 mm in diameter could be
distinguished (Figure 1). A 5 mm tumour growing in
one thigh was clearly evident from its increased
vascularity and resultant higher gold content
(Figure 1). These nanoparticles thus enable direct ima-
ging, detection, and measurement of angiogenic and
hypervascularized regions. Images taken at various
times after intravenous injection show that the small
nanoparticles do not concentrate in the liver and spleen,
but clear through the kidneys (Figure 2). A closer
examination of the kidneys revealed a remarkably
detailed anatomical and functional display (Figure 3).
Toxicity and clearance are critical issues for clinical
imaging. Mice intravenously injected with the gold
nanoparticles at 2.7 g Au kg21 survived over 1 year
without signs of illness. The LD50 for this material is
approximately 3.2 g Au kg21. In a 30-day toxicity study
using 60 mice, intravenous injection of the gold
nanoparticles (initially, 10 mg Au ml21 blood) showed
normal haematology (Table 1) and blood chemistry
(Table 2). Histological examination of 24 vital organs
and tissues from each mouse, assayed 11 days or 30 days
after injection of the nanoparticles, showed no evidence
of toxicity in any animals.
Quantitative pharmacokinetics using graphite furnace
atomic absorption spectroscopy (Figure 4) showed that
blood gold concentration decreased in a biphasic
manner, with a 50% drop between 2 min and 10 min
followed by a slower decrement of another 50% between
15 min and 1.4 h. The highest tissue gold concentration
15 min after injection was in the kidney (10.60¡0.2
percent of the injected dose per gram of measured tissue
[%id/g]), followed by tumour (4.2¡0.4%id/g), liver
(3.6¡0.3%id/g) and muscle (1.2¡0.1%id/g). Whole
body gold clearance was 77.5¡0.4% of the total injected
gold after 5 h. Muscles and blood were almost gold-free
24 h after injection (0.28¡0.07%id/g and 0.10¡0.01%id/g,
respectively), whereas tumour at 24 h retained 64% of
249

Figure 1. X-ray images mouse hind legs in vivo. (a) Before
injection; (b) 2 min post tail vein injection of gold
nanoparticles; (c) 2 min after equal weight of iodine contrast
agent (OmnipaqueH). Arrow points to leg with tumour and
increased vascularity. Arrowhead points to 100 mm diameter
vessel. Bar55 mm. (Figures 1a and 1b are reproduced with
permission from Phys Med Biol [12]).
its value reached at 15 min. The tumour:muscle gold
ratio was 3.4 at 15 min post injection, improving to 9.6
at 24 h, enabling clear delineation of the tumour. In
addition to imaging, higher tumour X-ray absorption
due to the gold has been shown to greatly improve the
efficacy of radiotherapy [13].
Even when concentrated, gold nanoparticle solutions
were similar to water in viscosity, in sharp contrast to the
high viscosity of iodine contrast media. The gold
nanoparticles may be completely dried and later re-
suspend easily in water or aqueous buffers, such as
phosphate buffered saline, pH 7.4. Solubility was found
to be at least 1.5 g Au ml21. The gold nanoparticles are
stable, showing no change in spectra or aggregation after
6 months storage at 4 ˚C or 220 ˚C.
Discussion
Gold’s K-edge at 80.7 keV compared with iodine’s at
33.2 keV confers higher absorptivity and ,3-fold better
250
J F Hainfeld, D N Slatkin, T M Focella and H M Smilowitz
contrast at ,100 keV, a useful range for clinical CTs and
fluoroscopes. Absorption is also higher at low energies
(,30 keV), where mammography machines operate,
again providing gold with an approximately 3-fold
absorption advantage over iodine.
Some iodine agents’ side effects are due to high
osmolality. Iodine agents contain 3 (monomer) or 6
(dimer) iodine atoms per molecule. In contrast, the
nanoparticles used here each contain about 250 gold
atoms per molecule and at the same elemental concen-
tration as iodine agents (350 mg Au ml21), therefore
have a negligible osmolality of 0.0072 M. Saline could, of
course, be added to provide iso-osmolality. The low
viscosity of gold nanoparticle solutions would also
facilitate injections.
Deliberately high amounts of gold were used to clarify
printed images. CT is much more sensitive than planar
imaging, and studies of high-Z agents indicate that good
contrast-to-noise images can be obtained at gold con-
centrations of 100 mg ml21 [14]. This level is ,100 times
lower than a dose of gold nanoparticles at which we
found no evidence of toxicity. Use of these lower
amounts of gold clinically would not only improve the
safety margin, but also lower the cost.
The extended imaging time and high contrast pro-
vided by gold nanoparticles after a non-toxic intravenous
injection might enable such applications as: non-invasive
imaging of coronary and cerebral arteries, assessment of
atherosclerotic plaque and stenoses, delineation of
stroke, arteriovenous malformations, aneurysms,
renal angiography, determination of vascularity, and
enhancement of mammography and virtual colono-
scopy. Improved contrast might enable non-invasive
detection of small tumours (e.g. , 1 cm) that are
currently missed, yielding better prognoses. Tumour
vascularity is correlated with invasiveness [15], so
indices of vascularity make non-invasive staging
possible. These gold nanoparticles might be useful to
distinguish vulnerable plaque since it is more highly
vascularized than stable plaque [16, 17]. With the
advent of faster CT machines that lessen motion
artefacts, gold-enhanced imaging of coronary arteries,
especially those in obese patients or those with mural
calcifications, might prove feasible via transvenous
injection without resorting to arterial catheterization.
Contrasting during transarterial catheterization might
also benefit from the use of gold nanoparticles, especially
for large patients where additional contrast is needed
at present, since the concentration of gold can be
made ,5 times higher than that of iodine agents.
With the absorbance of gold 3 times higher at 100 keV,
and the concentration 5 times higher (1.5 g Au cm23 vs
0.3 g I cm23), the overall contrast gain could be greater
than 10-fold.
The size of gold nanoparticles used here facilitates
specific extravasation through angiogenic endothelium
such as in tumours. This is evident from the extensive
retention of gold in tumours after the blood has cleared,
compared with the much lower accumulation in muscle
(Figure 4). This gold retention may be useful for
enhanced tumour detection and possibly for therapy
due to the improved tumour:non-tumour ratio reached
hours after injection.
The British Journal of Radiology, March 2006
Short communication: Gold nanoparticles

Figure 2. Pharmacokinetics of gold nanoparticles (a–d) and iodine contrast agent (e–h, OmnipaqueH) in mice. (a,e) Before
injection. (b,f) 2 min after injection; (c,g) 10 min after injection; (d,h) 60 min after injection. The gold nanoparticles show low
liver and spleen uptake and clearance via kidneys and bladder (b–d). At 60 min (d), the contrast in the gold-injected mouse is
similar to the uninjected mouse (a), indicating efficient clearance.

Figure 3. Planar X-ray image of kidneys in live mouse 60 min after intravenous injection of gold nanoparticles (a) or iodine
contrast medium (b, OmnipaqueH). Arrow points to 100 mm ureter. Bar51 mm.

Conclusion
These animal studies demonstrate that gold nanopar-
ticles are useful X-ray contrast agents that offer novel
physical and pharmacokinetic advantages over current
agents. They appear to be non-toxic and enable higher
contrast and longer imaging times than currently
possible using standard iodine-based agents.
Acknowledgments
We are indebted to Lahmer Lynds, PhD, Alexander
Richter, MS, DVM, DACVP, Terry M Button, PhD, Bipin
Jagjivan, MD, and Harold Atkins, MD for their expert
advice and assistance. This study was supported in part
by a National Cancer Institute Small Business Innovative
Research Phase 1 Grant 1R43CA83576-01.

The British Journal of Radiology, March 2006 251

252

Table 1. Blood haematology of mice intravenously injected (0.2 ml) with an initial gold dose producing 10 mg Au ml21 in the blood, sampled at day 1, 11, and 28, compared with
mice sham injected with 0.2 ml phosphate buffered saline (control). Five animals were used per measurement, and all values were within normal limits
Hct % Hgb g dl21 WBC 6 103 RBC 6 106 RBC indices % differential

MCV MCH MCHC Bands Neutro Lymphs Monos Eosins

Day 1 Gold 44.3¡4.9 15.7¡1.2 5.4¡3.5 9.0¡0.7 49.3¡1.9 17.5¡0.5 35.6¡1.3 0.2¡0.4 45.0¡13.0 54.0¡13.0 0.4¡0.5 0.4¡0.5
Control 44.3¡2.8 14.4¡0.8 5.6¡3.1 8.4¡0.4 52.6¡1.7 17.2¡0.8 32.8¡1.2 0.2¡0.4 36.2¡4.5 60.8¡5.0 2.4¡1.5 0.4¡0.9
Day 11 Gold 42.2¡2.9 14.9¡0.9 9.2¡2.6 9.2¡0.6 45.6¡1.9 16.1¡0.6 35.3¡0.6 0.0¡0.0 11.2¡7.0 87.6¡7.3 0.4¡0.5 0.8¡1.3
Control 44.6¡4.5 15.6¡1.3 7.2¡1.7 9.6¡1.2 46.6¡3.0 16.3¡0.7 35.0¡1.2 0.0¡0.0 15.8¡12.4 81.0¡11.9 0.6¡0.5 2.6¡2.1
Day 28 Gold 39.8¡1.5 14.8¡0.4 8.0¡4.0 8.7¡0.5 45.4¡0.6 16.9¡0.6 37.1¡0.8 0.0¡0.0 36.8¡8.5 60.6¡10.7 60.6¡10.7 2.4¡3.0
Control 43.2¡0.8 15.8¡0.1 5.8¡2.0 9.3¡0.2 46.6¡1.5 17.1¡0.4 36.7¡0.8 0.0¡0.0 33.4¡5.9 58.2¡3.1 58.2¡3.1 7.2¡6.6
Hct, haematocrit; Hgb, haemoglobin; WBC, white blood cells; RBC, red blood cells; MCV, mean corpuscular volume; MCH, mean corpuscular haemoglobin; MCHC, mean corpuscular
haemoglobin concentration; Neutro, neutrophils; Lymphs, lymphocytes; Monos, monocytes; Eosins, eosinophils.

J F Hainfeld, D N Slatkin, T M Focella and H M Smilowitz

Table 2. Serum clinical chemistry of mice intravenously injected (0.2 ml) with an initial gold dose producing 10 mg Au ml21 in the blood, sampled at day 1, 11, and 28, compared
The British Journal of Radiology, March 2006

with mice sham injected with 0.2 ml phosphate buffered saline (control). Five animals were used per measurement, and all values were within normal limits
BUN CR GLU CA PO4 TP ALB GLB A/G AST ALT AST/ALT ALP TBILI
mg dl21 mg dl21 mg dl21 mg dl21 mg dl21 g dl21 g dl21 g dl21 ratio Units Units ratio units mg dl21

Day 1 Gold 30.8¡6.7 0.2¡0.1 341.6¡83.7 11.7¡0.7 13.9¡3.7 6.0¡0.8 4.3¡0.6 1.8¡0.3 2.5¡0.5 368¡698 362¡274 4.2¡4.2 174.4¡39.3 0.2¡0.1
Control 28.0¡7.3 0.2¡0.2 390.8¡105.8 11.3¡1.2 15.3¡3.5 6.8¡0.7 4.7¡0.6 2.1¡0.2 2.3¡0.3 38.8¡9.9 226¡105 6.0¡2.5 141.5¡55.0 0.3¡0.1
Day 11 Gold 28.4¡6.7 0.4¡0.5 319.6¡47.7 11.6¡0.6 12.3¡2.0 6.3¡0.3 4.5¡0.5 1.8¡0.3 2.6¡0.7 58.8¡8.9 96.0¡10.6 1.7¡0.2 157.6¡25.7 0.2¡0.0
Control 26.4¡5.9 0.4¡0.2 321.2¡115.9 12.0¡0.9 15.6¡2.2 6.6¡0.3 4.6¡0.4 2.0¡0.2 2.3¡0.5 61.6¡6.7 109.2¡37.2 1.8¡0.5 180.0¡60.4 0.2¡0.0
Day 28 Gold 28.4¡4.3 0.0¡0.1 389.5¡77.8 10.7¡1.4 13.0¡1.4 6.0¡0.3 4.0¡0.1 1.9¡0.3 2.1¡0.3 60.4¡17.1 33.6¡7.3 2.0¡1.2 139.6¡32.5 0.2¡0.0
Control 33.6¡6.2 0.0¡0.1 309.2¡66.7 11.1¡0.9 11.7¡0.9 6.1¡0.3 4.2¡0.4 1.9¡0.3 2.3¡0.5 123.2¡52.0 40.8¡6.7 3.0¡1.1 100.4¡28.7 0.2¡0.0
BUN, blood urea nitrogen; CR, creatine; GLU, glucose; CA, calcium; PO4, phosphate; TP, total protein; ALB, albumin; GLB, globin; A/G, albumin:globulin ratio; AST, aspartate
aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; TBLI, total bilirubin.
Short communication: Gold nanoparticles
Figure 4. Pharmacokinetic data showing biodistribution of gold in mice following a 0.2 ml intravenous injection of gold
nanoparticles showing percent injected dose per gram in the blood, kidneys, liver, muscle, and tumour over a 24 h period.
References
1. Yu S-B, Watson AD. Metal-based x-ray contrast media.
Chem Rev 1999;99:2353–77.
2. Kao CY, Hoffman EA, Beck KC, Bellamkonda RV,
Annapragada AV. Long-residence-time nano-scale liposo-
mal iohexol for X-ray-based blood pool imaging. Acad
Radiol 2003;10:475–83.
3. Schmiedl UP, Krause W, Leike J, Sachse A. CT blood pool
enhancement in primates with -lopromide-carrying lipo-
somes containing soy phosphatidyl glycerol. Acad Radiol
1999;6:164–9.
4. Vera DR, Mattrey RF. A molecular CT blood pool contrast
agent. Acad Radiol 2002;9:784–92.
5. Torchilin VP. PEG-based micelles as carriers of contrast
agents for different imaging modalities. Adv Drug Deliv
Rev 2002;54:235–52.
6. Fruman SA, Harned RK 2nd, Marcus D, Kaufman S,
Swenson RB, Bernardino ME. Perfluoroctyl bromide as a
blood pool contrast agent for computed tomographic
angiography. Acad Radiol 1994;1:151–3.
7. Torchilin VP, Frank-Kamenetsky MD, Wolf GL. CT visua-
lization of blood pool in rats by using long-circulating,
iodine-containing micelles. Acad Radiol 1999;6:61–5.
8. Idee JM, Port M, Robert P, Raynal I, Prigent P, Dencausse A,
et al. Preclinical profile of the monodisperse iodinated macro-
molecular blood pool agent P743. Invest Radiol 2001;36:41–9.
9. Bulte JW, Kraitchman DL. Iron oxide MR contrast agents for
molecular and cellular imaging. NMR Biomed 2004;17:484–99.
The British Journal of Radiology, March 2006
10. http://physics.nist.gov/PhysRefData/XrayMassCoef
[Accessed 22 November 2005].
11. Atkins HL, Fairchild RG, Robertson JS, Greenberg D. Effect
of absorption edge filters on diagnostic x-ray spectra.
Radiology 1975;115:431–7.
12. Rockwell SC, Kallman RF, Fajardo LF. Characteristics of a
serially transplanted mouse mammary tumor and its tissue-
culture-adapted derivative. J Natl Cancer Inst 1972;49:
735–49.
13. Hainfeld JF, Slatkin DN, Smilowitz HM. The use of gold
nanoparticles to enhance radiotherapy in mice. Phys Med
Biol 2004;49:N309–15.
14. Dilmanian FA, Wu XY, Parsons EC, Ren B, Kress J,
Button TM, et al. Single-and dual-energy CT with mono-
chromatic synchrotron x-rays. Phys Med Biol 1997;42:
371–87.
15. Weidner N, Carroll PR, Flax J, Blumenfeld W, Folkman J.
Tumor angiogenesis correlates with metastasis in invasive
prostate carcinoma. Am J Pathol 1993;143:401–9.
16. Higuchi ML, Gutierrez PS, Bezerra HG, Palomino SA,
Aiello VD, Silvestre JM, et al. Comparison between
adventitial and intimal inflammation of ruptured and
nonruptured atherosclerotic plaques in human coronary
arteries. Arq Bras Cardiol 2002;79:20–4.
17. Tenaglia AN, Peters KG, Sketch MH Jr, Annex BH.
Neovascularization in atherectomy specimens from
patients with unstable angina: implications for pathogen-
esis of unstable angina. Am Heart J 1998;135:10–4.
253
 The British Journal of Radiology, 79 (2006), 254–257

SHORT COMMUNICATION

Calculation of high-LET radiotherapy dose required for

compensation of overall treatment time extensions
1 2
B JONES, MD, FRCR, MedFIPEM, A CARABE-FERNANDEZ, MSc, MPhys and 2R G DALE, PhD, FinstP, FIPEM

1
Birmingham Cancer Centre, Queen Elizabeth University Hospital, Birmingham B15 2TH and
2
Department of Radiotherapy Physics & Radiobiology, Charing Cross Hospital, London W6
?8RF, UK

ABSTRACT. A method is presented that allows biological effective dose (BED) equations
to be used to calculate compensatory doses for treatment time extensions when high-
LET (linear energy transfer) radiotherapy schedules are used. The principles involved are
Received 6 July 2005
the same as those for low-LET radiations, but incorporate two relative biological Revised 15 August 2005
effectiveness (RBE) factors, RBEmax and RBEmin, which represent the RBE at very low and Accepted 5 September
very high fraction doses, respectively, with the actual RBE changing between these 2005
extremes. The method has the advantage that low-LET a/b ratios and low-LET daily
DOI: 10.1259/bjr/49977661
dose-equivalent repopulation factors are used in the calculations. The daily dose
repopulation equivalents and increments in dose per fraction in the case of high LET ’ 2006 The British Institute of
radiotherapy are smaller than those for low LET. Radiology

The loss of tumour control following an extension in aH

RBEmax ~ (2)
treatment duration can, in principle, be overcome by aL
increasing the total dose after the extension. This can be
achieved – as well as our assumptions allow – by where aH and aL are the respective radiosensitivities for
calculating the dose per day (dc) that should compensate high and low LET radiations. Similarly, from Appendix
for the additional tumour cell repopulation. The magni- A, at very high dose RBE approaches an asymptotic
tude of dc can be estimated from a logical extension of the minimum value (RBEmin), where:
linear-quadratic (LQ) model of radiation effect through sffiffiffiffiffiffi
consideration of the biologically effective dose (BED) bH
RBEmin ~ (3)
concept [1–4]. Dale et al [5] have stressed that the bL
compensatory dose per day is normally expressed in
BED Gy units, which are those that pertain for a specific Consequently, where RBEmin is significantly greater than
tumour a/b ratio and from this the actual physical dose
unity and in contradiction with the conventional
(dc) can be separately calculated. There are some
assumption first put forward in the Theory of Dual
additional subtleties in the case of high-LET radio-
therapy because additional RBE (relative biological Radiation Action by Kellerer and Rossi [7], it follows that
effectiveness) correction factors need to be included. bH.bL. Also, the high-LET b parameter can be expressed
The general mathematical approach is that originally in terms of the low-LET b value and the RBEmin, at high
followed by Bewley [6] for high LET radiations, but is dose per fraction, i.e.:
now adapted to the LQ model and the BED concept in
particular, such that the use of the Cobalt equivalent Gy bH ~ðRBEmin Þ2 bL (4)
concept is avoided. These Equations (2–4) are used to form BED equations
for high LET radiations as shown in Appendix B.
Next we consider how to compensate for a practical
Methods problem using the standard BED equations for low LET
radiations and also the newly derived equation for the
RBE is conventionally defined as the ratio of the dose
high LET BED (BEDH):
per fraction of the low LET radiation (dL) to the dose per 2 3
fraction of the high LET radiation (dH), so that it is
usually greater than 1, as in: 6 RBE2min dH 7
BEDH ~NH dH 6
4RBEmax z aL  5
7 (5)
dL
RBE~ (1)
dH bL
As shown in Appendix A, RBE is maximized at low dose where N is the number of fractions, d the dose per
to a value given as RBEmax, where: fraction and (aL/bL) is the low LET a/b ratio. The

254 The British Journal of Radiology, March 2006

Short communication: Calculation of high-LET radiotherpay dose
repopulation equivalent equations are also given in
Appendix B: essentially, the repopulation terms in the
high LET equations are the same as for low LET.
Worked example
A schedule of 45 Gy in 25 fractions using megavoltage
X-rays is to be followed by a highly localized ‘‘boost’’ of
6 Gy in 2 fractions of 3 Gy each, using a high-LET
radiation for which RBEmin51.3 and RBEmax58; these
values are assumed to apply for both cancer and normal
tissues. There is a delay of 7 days in the provision of the
boost, due to patient illness. The tumour type is assumed
to have a daily repopulation equivalent of 0.6 Gy per day
after a lag interval of 25 days during megavoltage X-ray
treatment. The normal tissue BED is assumed to be
governed by a/b52 Gy and the tumour a/b510 Gy.
The intended BED to normal tissue from X-rays545 6
(1+1.8/2)585.5 Gy2.
The intended BED to any normal tissue that receives
the added high-LET boost of 2 fractions of 3 Gy56 6
(8+1.32 6 3/2)563.2 Gy2, so that the total intended
maximum BED to same volume of normal
tissue585.5+63.25148.7 Gy2.
The intended BED to tumour by X-rays, BEDL545 6
(1+1.8/10)553.1 Gy10, plus the intended BED to tumour
by high LET, BEDH56 6 (8+1.32 3/10)551.04 Gy10, so
that the total tumour BED is 104.14 Gy10 before allowing
for repopulation.
The additional 7 days of repopulation must be allowed
for because of the treatment interruption in providing
the boost, which is equivalent to 0.6 6 754.2 Gy10.
The boost must therefore accommodate the original
high-LET BED plus 4.2 Gy, i.e. 51.04+4.2555.24 Gy10.
As this is to be given in two fractions, then: 2 6 d 6
(8+1.32d/10)555.24, and the solution for d is 3.23 Gy/
fraction instead of the originally prescribed 3 Gy per
fraction, prescribed before the treatment gap had
occurred.
The normal tissue BED will then be: 2 6 3.23 6 (8+1.32
6 3.23/2)569.31 Gy2.
Thus the total (low- plus high-LET) normal tissue
maximum BED will have increased by
69.31263.256.11 Gy2, an increase of 4.1% on an already
high BED in the localized boost volume, in order to
maintain the same tumour BED. This could cause
enhanced tissue side effects.
In practice a compromise solution such as a dose per
fraction of 3.15 Gy instead of 3.23 Gy might be used. This
would lead to 67.17 Gy2 maximum high-LET BED to the
normal tissues and 53.75 Gy10 to the tumour.
A plot of the calculated compensatory dose per day,
given as a single fraction, for high-LET radiations is
shown in Figure 1. It can be seen that the a/b ratio has
little influence on the value of the compensatory dose,
with marginal differences only if a/b is very low and
when K is large. This result is dominated by the RBEmax
value, since small fractional doses are operational – e.g.
for a K value of 1 Gy per day the equivalent high LET
dose per day is 0.12 Gy per day which is close to the ratio
1/RBEmax51/8 (51.25). It needs to be stressed that these
calculated doses, which are single fraction equivalents,
must not be used as additional doses added to the
The British Journal of Radiology, March 2006
Figure 1. Relationship between low linear energy transfer
(LET) K value and the calculated single fraction
compensatory dose per day for a high LET radiation with
relative biological effectiveness (RBE)max58 and RBEmin51.3,
with variation of the a/b ratio, obtained using Equation (B10)
(given in Appendix B).
prescribed dose per fraction; to be correctly incorporated
the method used in the worked example should be
followed to adjust the dose per fraction.
Discussion
As in the case of low-LET radiations, extension of
treatment time can cause significant dilemmas to the
radiation oncologist. Gaps occurring later in the treat-
ment are particularly difficult to compensate for since an
extension to the prescribed treatment time becomes
unavoidable [4]. Compensations for overall treatment
time extensions typically involve a compromise between
delivery of a higher BED to normal tissues and a reduced
BED to tumour, or the same BED to tumour and an even
higher BED to normal tissues. The incorporation of RBE
within BED equations has enabled the calculation of
estimated compensatory doses for high LET radiations.
The use of a fixed RBE weighting factors is an
alternative approach. Although approved in interna-
tional definitions, this method will under estimate RBE at
low doses per fraction and over estimate RBE at high
doses per fraction. In contrast, the use of RBEmax and
RBEmin overrides this potential problem. The RBEmax
dominates the effective RBE at low dose per fraction,
whereas RBEmin dominates the RBE at large doses per
fraction. We have used these two RBE parameters
because of the increasing use of large fraction sizes in
high LET radiotherapy [8].
The actual high LET daily dose correction factors are
much lower than the K value doses, as are the increments
in dose per fraction required when compared with those
required in low LET radiotherapy.
These aspects should be additionally considered in
situations where protraction of relatively high LET
radiotherapy occurs, e.g. radioiodine seed implants and
in the case of hadrontherapy, for instance using ion
beam or neutron radiotherapy schedules. For high-
energy proton beams with significant spread out
255
 B Jones, A Carabe-Fernandez and R G Dale

Bragg peaks, the correction factors will be much smaller, aL dL zbL dL2 ~aH dH zbH dH
2 (A1)
since the average RBE for protons is only around 1.1,
although the refinement of a variable RBE with dose per
fraction might lead to better results [9, 10]. In the case of At near-zero doses the quadratic dose terms become
ions beams, treatments are hypofractionated [8, 11], negligible and:
varying from 1 fraction to 20 fractions with treatment
times of up to 1 month, but these times – apart from the aL dL ~aH dH (A2)
single fraction case – could be extended for a variety of
Leading to:
reasons such as patient illness, synchrotron breakdown
etc. dL aH
RBE?RBEmax ~ ~ (A3)
dH aL

Acknowledgments and declarations Similarly, at exceedingly high doses, the quadratic terms
dominate, i.e.:
Author AC-F is a Cyclotron Trust Research Fellow. All
authors are members of the EPSRC-funded medical bL dL2 ~bH dH
2 (A4)
applications of ion beams network. BJ and RGD are Hon.
Professors at the University of Caen as part of the and:
ASCLEPIOS project, part of the French National sffiffiffiffiffiffi
Hadrontherapy Project. dL bH
RBE?RBEmin ~ ~ (A5)
dH bL
References
1. Barendsen GW. Dose fractionation, dose rate and iso-effect
relationships for normal tissue responses. Int J Radiat Oncol
Biol Phys 1982;8:1981–97.
2. Fowler JF. The linear quadratic formula and progress in Appendix B
fractionated radiotherapy. Br J Radiol 1989;62:679–94.
3. Jones B, Dale RG, Deehan C, Hopkins KI, Morgan DAL. The
role of biologically effective dose (BED) in clinical oncology.
To obtain high LET BED equations
Clin Oncol 2001;13:71–81. The standard linear quadratic equation for the low
4. Dale RG, Hendry JH, Jones B, Deehan C, Sinclair J, LET surviving fraction (SF) with compensation for
Robertson G. Practical methods for compensating for
cellular repopulation is:
missed treatment days in radiotherapy, with particular
reference to head & neck schedules. Clin Oncol {aL NL d L {bL NL d 2L z
0:693(T{T DEL )

2002;14:382–93. SF~e T EFF

5. Dale RG, Jones B, Sinclair JA. Dose equivalents of tumour
repopulation during radiotherapy: the potential for confu-
sion. Br J Radiol 2000;73:892–4.
6. Bewley DK. The physics and radiobiology of fast neutron
beams. Bristol & New York: Adam Hilger, 1989:113.
7. Kellerer AM, Rossi HH. RBE and the primary mechanism of
radiation action. Radiat Res 1971;47:15–34.
8. Tsujii H, Mizoe J, Kamada T, Baba M, Kato S, Kato H, et al.
Overview of clinical experiences on carbon ion radio-
therapy at NIRS. Radiother Oncol 2004;73 (Suppl. 2):S41–9.
9. Paginetti H, Niemierko A, Ancukiewicz M, Gerweck LE,
Goitein M, Loeffler JS, et al. Relative biological effectiveness
(RBE) values for proton beam therapy. Int J Radiat Oncol
Biol Phys 2002;53:407–21.
10. Jones B, Dale RG. Estimation of optimum dose per fraction
for high LET radiations: implications for proton radio-
therapy. Int J Radiat Oncol Biol Phys 2000;48:1549–57.
11. Schulz-Ertner D, Nikoghosyan A, Thilmann C, Haberer T,
Jakel O, Karger C, et al. Results of carbon ion radiotherapy
in 152 patients. Int J Radiat Oncol Biol Phys 2004;58:631–40.
(B1)
where aL and bL are the low LET radiosensitivity
parameters, NL is the number of low LET fractions of
dose dL, T is the overall treatment time in days, TDEL is
the delay time in days before which repopulation
becomes significant and is optional according to tumour
type and TEFF is the effective cellular doubling time in
days.
For high LET radiations the surviving fraction expres-
sion is changed to be:
0:693(T{T DEL )
{aH NH d L {bH NH d 2H z
SF~e T EFF
(B2)
where the subscripts are changed to H in order to refer to
high LET radiations. Next, taking the natural logarithm
of each side and multiply by 21 to, respectively, obtain
from Equations (B1) and (B2) the log cell kills designated
by EL and EH:

0:693(T L {T DEL )
Appendix A EL ~aL NL dL zbL NL dL2 { (B3)
T EFF
To obtain RBE terms
RBEs are derived by intercomparing the single doses 0:693(T L {T DEL )
EH ~aH NH dH zbH NH dH2 { (B4)
(low- and high-LET) required to obtain a given iso-effect. T EFF
Since only single doses are involved there is no need to
consider the repopulation effect, i.e. the iso-effect Assume that Equations (B3) and (B4) refer to the same
equation is simply: biological effect.

256 The British Journal of Radiology, March 2006

Short communication: Calculation of high-LET radiotherpay dose

To obtain the BED, divide each equation by aL, the low offset the effect of TH days-worth of repopulation, as
LET radiosensitivity parameter, then, for Low LET, the quantified by the second (subtractive) factor. Equating
BED is: these two components we obtain:
2 3
2 3
EL 6 dL 7
BED~ ~NL dL 6 7
41z aL 5{K ðTL {TDEL Þ (B5) 6 RBE2min dH 7
aL NH dH 6 7
4RBEmax z aL  5~K ðTH {TDEL Þ (B8)
bL
  bL
0:693
and where K ~ is the daily dose equivalent
aL Teff Once the time-point (TDEL) is passed, the BED-equivalent
for repopulation in units of BED Gy per day. of repopulation for each additional day (for which
The High LET BED is also obtained by dividing by the TH2TDEL51) is found by setting NH51 in Equation
low LET aL parameter to give (B8). On rearrangement, this leads to:
  2 3
EH aH bH dH
BED~ ~NH dH z {K ðTL {TDEL Þ (B6) 6
aL aL aL RBE2min dH 7
K~dH 6
4RBEmax z   7 5 (B9)
aL
Replacement of the high LET radiosensitivity parameters bL
in Equation (B6) with RBEmax and RBEmin – as given in
Equations (2–4) in the main text – results in: In Equation (B9) the repopulation term K is expressed as
an equivalent BED dose per day. The actual (physical)
2 3
dose per day (dc) – given as a single fraction – required to
6 RBE2min dH 7 compensate for repopulation is the solution for dH in
BEDH ~NH dH 4RBEmax z aL 5{K ðTH {TDEL Þ (B7)
Equation (B9), i.e.:
bL
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
u 4KRBE2
For calculations involving overall treatment time varia- {RBEmax zuRBE2max z   min
t a
tions and compensation of unintended treatment inter-
b L
ruptions, the same low-LET daily BED dose equivalent dc ~ (B10)
values (K) are used in both cases. RBE2
2  min
The first component of the right hand side of Equation a
(B7) represents the BED which must be delivered to b L

The British Journal of Radiology, March 2006 257

 The British Journal of Radiology, 79 (2006), 258–260

SHORT COMMUNICATION

CT fluoroscopic guided insertion of inferior vena cava filters

P IGNOTUS, FRCS Ed, FRCR, C WETTON, MRCP, FRCR and J BERRY, FRCR

Kent and Sussex Hospital, Mount Ephraim, Tunbridge Wells, Kent TN4 8AT, UK

ABSTRACT. The value and use of inferior vena cava (IVC) filters is well documented and has
been growing since the first reported filter placement in 1973 and the first percutaneous Received 25 July 2005
insertion in 1982. Access routes now include both jugular veins, both ante-cubital veins and Accepted 12 September
2005
both femoral veins. However, all insertions require some form of imaging, usually
fluoroscopy, to identify the location of the filter with respect to the IVC and the renal veins. DOI: 10.1259/bjr/43028256
We describe two cases where the patients’ weight was significantly greater than the weight
limit of the angiography table, necessitating insertion under CT fluoroscopic guidance. ’ 2006 The British Institute of
Radiology

The first non-occlusive inferior vena cava (IVC) filter (Terumo, Tokyo, Japan) was inserted. The patient was
was described in 1973 [1], and in 1982 the first then transferred to the CT scanner (Toshiba X-press GX;
percutaneous device became available [2]. Since then Toshiba Medical Systems, Japan).
the use of these devices has become widely accepted for An Amplatz 0.0350 guidewire (William Cook Europe,
the treatment and prevention of pulmonary thromboem- Bjaeveskov, Denmark) and modified 5 Fr left vertebral
bolism. Many different filter designs are available, but all catheter (William Cook Europe) were then advanced
usually require fluoroscopy to assist insertion. until CT fluoroscopy confirmed that they lay at the level
We describe two cases where fluoroscopy was not pos- of the renal veins. The diameter of the vena cava was
sible as the weight of the patient significantly exceeded confirmed to be less than 28 mm.
the manufacturer’s limits for the angiography table in our The vascular sheath and catheter were removed,
department. We therefore elected to insert the filters using leaving the wire in situ, and the femoral vein access site
CT assisted by CT fluoroscopy as an imaging modality. dilated to 8 Fr. The sheath of a Gunther tulip filter
(William Cook Europe) was inserted using CT fluoro-
scopy at the level of the renal veins until the introducer
Case reports sheath appeared. The guidewire was then removed, and
the filter loaded into the sheath and advanced. By
Case 1 maintaining CT fluoroscopy at the level of the renal
veins, it was easy to see when the hook of the filter
A 57-year-old woman was diagnosed with Stage I appeared (Figure 1). With this level fixed it was easy to
carcinoma of the left kidney. This was an incidental ascertain that the level of the feet of the filter would be
finding during a CT scan of the abdomen as part of the
investigation of upper abdominal pain, which was not
thought to be related to her carcinoma.
The patient had a past history of lower limb deep
venous thrombosis (DVT) on two occasions, and on one
occasion this had led to a proven diagnosis of pulmonary
embolism (PE). The patient weighed 115 kg (the weight
limit for our angiography table is 110 kg) and during a
pre-operative attempt to lose weight to allow filter
insertion, her weight actually rose to 122 kg.
As she was deemed a very high risk of further DVTs and
pulmonary emboli, it was decided to attempt placement of
a filter using fluoroscopic guidance from an operating
theatre ‘‘C-arm’’. Unfortunately the quality of imaging was
such that filter placement was not possible. CT fluoroscopic
guidance was therefore employed to perform the procedure.

Technical details
The right femoral vein was punctured using ultra- Figure 1. Unenhanced CT fluoroscopic image showing the
sound guidance, and a 7 French vascular sheath hook of the filter at the level of the renal veins.

258 The British Journal of Radiology, March 2006

Short communication: CT guided insertion of IVC filter
comfortably within the infrarenal IVC, above the
confluence of the two iliac veins, since a previous CT
scan had confirmed that the infrarenal IVC was greater
than 45 mm in length (the length of the device) and that
the IVC and iliac veins were free of thrombus.
The introducer sheath was then withdrawn to the hub
of the release mechanism, so that the filter was in its
‘‘pre-release’’ configuration, and CT fluoroscopy again
confirmed its adequate location and the characteristic
‘‘flaring’’ of the unconstrained body of the filter above
the still constrained feet (Figure 2). The filter was then
released in the usual manner, and its position and
full opening confirmed (Figure 3). The introducer was
then removed and pressure applied to the groin until
haemostasis was achieved.
The patient underwent surgery 2 days later and made
an uneventful recovery.
Case 2
A 47-year-old lady was admitted acutely from long
term care with right lower abdominal/back pain. She
was on long term anticoagulation with Warfarin for two
proven episodes of pulmonary emboli. She weighed
118 kg.
Clinically, a retroperitoneal bleed was suspected. CT
scanning confirmed the presence of a large haematoma
within the abdominal wall and the right iliacus muscle. It
was therefore felt that anticoagulation should be dis-
continued, yet she remained at high risk of pulmonary
emboli due to her past history of venous thromboem-
bolic disease and her obesity. An IVC filter was thus
indicated, but, once again, she exceeded the weight limit
of all our fluoroscopy facilities.
An IVC filter was easily and successfully inserted
using the same CT technique described above.
Figure 2. Unenhanced CT fluoroscopic image showing the
body of the filter in the ‘‘pre-release’’ configuration.
The British Journal of Radiology, March 2006
Figure 3. Unenhanced CT fluoroscopic image showing the
deployed filter. The feet are seen embedded in wall of the
infrarenal inferior vena cava.
Discussion
The use of IVC filters is now well accepted in current
medical practice. Indications include massive pulmonary
emboli, recurrent pulmonary emboli in spite of adequate
anticoagulation, proven pulmonary emboli with a con-
traindication to anticoagulation and a failure to tolerate
anticoagulation in a patient who has had recent
pulmonary emboli. Relative indications include patients
at very high risk of pulmonary emboli, particularly those
who have significant medical co-morbidity, past histories
of DVT and those about to undergo surgery for
malignant disease or pelvic surgery [3, 4].
The indications for filtration for Case 1 were her
medical co-morbidity (obesity), a past history of pul-
monary emboli and the fact that she was about to
undergo surgery for malignancy.
In Case 2 the indication was a proven history of
pulmonary emboli (recurrent) for which the patient
would normally be on anticoagulation for life, but she
had failed to tolerate this and she had now developed a
retroperitoneal bleed while her international normali-
zation ratio (INR) was within the therapeutic range.
Insertion is usually performed in an angiography
suite. Fluoroscopy (with or without digital subtraction) is
the main imaging tool. This was not possible in these
cases as the patients significantly exceeded the weight
limit for our angiography table and also the barium table
which could otherwise have provided fluoroscopic
guidance. Ultrasound guided insertion has been
described [5, 6] but was not technically possible in our
patients. Open surgical implantation together with the
use of intraoperative ultrasound has also been success-
fully performed in morbidly obese patients [7], but
this is significantly more invasive than a percutaneous
approach. Use of an electromagnetic position sensor
(Biosense, Setauket, NY) to superimpose positional data
has also been used in animal experiments [8]. However,
we felt that CT with CT fluoroscopy as an adjunct would
provide us with the simplest guidance system for
insertion.
259

The technique of IVC filter insertion is relatively
simple. In terms of imaging, there are relatively few
pieces of information that are required, namely:
(1) the level of the renal veins;
(2) the diameter of the vena cava (the Gunther tulip
filter can only be used if the caval diameter is
28 mm or less);
(3) an adequate length of infrarenal IVC;
(4) the position of the vena cava filter during device
implantation;
(5) patency of the IVC.
All of this information can be obtained using CT
scanning. It would be possible to do this with a ‘‘scan
and view’’ approach, but the use of CT fluoroscopy
makes it much easier and significantly faster.
The length of the infrarenal IVC had previously been
determined from the CT scans, and was well over
45 mm, the length of the deployed Gunther tulip filter.
This meant that we could be certain that if the hook was
visualized at the level of the renal veins the filter could
be deployed with confidence knowing that the feet and
barbs would be securely in the infrarenal cava rather
than in the iliac veins.
One possible difficulty would have been if we had had
trouble passing the catheter and wire up through the
iliac veins. Occasionally some manipulation is required
under fluoroscopy to achieve this, particularly if the wire
passes into small branch veins. We had anticipated the
need for possible manipulations under CT fluoroscopy,
but in fact the wire passed straight up into the IVC with
no manipulation in both patients.
Had patient 1 not exceeded the weight limit for the
angiography table, she could have been considered for
later filter removal, i.e. the filter would only have been
used temporarily to cover the perioperative period.
However, we do not feel that the filter could be removed
easily under CT guidance and so the filter has been left in
situ permanently.
260
P Ignotus, C Wetton and J Berry
The CT technique of insertion could also be extended
to include patients with a history of significant contrast
allergy. In thinner patients in whom contrast and
fluoroscopy were contraindicated we would usually
favour the use of ultrasound, since this can also
accurately pinpoint the filter within the vena cava and
provide all the other necessary information. However,
due to the obesity of these patients this was not a
practical solution in these cases.
Conclusion
CT scanning, particularly if combined with CT fluoro-
scopy, provides a simple, safe and practical alternative to
conventional fluoroscopic insertion of IVC filters in
selected patients.
References
1. Greenfield LJ, McCrudy JR, Brown PP, Elkins RC. A new
intracaval filter permitting continued flow and resolution of
emboli. Surgery 1973;73:599–606.
2. Tadavarthy SM, Castaneda-Zuniga W, Salomonowitz E, et al.
Kimray Greenfield vena cava filter: percutaneous introduc-
tion. Radiology 1984;151:525–6.
3. Menon K, Insall RL, Ignotus PI. Inferior vena cava filters: an
overview of current use. Hosp Med 1998;59:224–5.
4. Kinney TB. Update on inferior vena cava filters. J Vasc Interv
Radiol 2003;14:425–40.
5. Neuzil DF, Garrard CL, Berkman RA, Pierce R, Naslund TC.
Duplex-directed vena caval filter placement: report of initial
experience. Surgery 1998;128:470–4.
6. Matsumura JS, Morasch MD. Filter placement by ultrasound
technique at the bedside. Semin Vasc Surg 2000;13:199–203.
7. Snyder JM, Arita A, Inampudi C, Gwin-Stewart JA. Intra-
operative ultrasound guidance of vena caval umbrella
placement in a super obese patient. Obes Surg
2002;12:679–81.
8. Solomon SB, Magee CA, Acker DE, Venbrux AC.
Experimental nonfluoroscopic placement of inferior vena
cava filters: use of an electromagnetic navigation system
with previous CT data. J Vasc Interv Radiol 1999;10:92–5.
The British Journal of Radiology, March 2006
The British Journal of Radiology, 79 (2006), 261–263

CASE REPORT

Gastric carcinoma presenting with extensive bone metastases

and marrow infiltration causing extradural spinal haemorrhage
S HUSSAIN, MRCP and S CHUI, FRCR

Alexandra Hospital, Woodrow Drive, Redditch B98 7UB, UK

ABSTRACT. Gastric carcinoma is the third most common gastrointestinal (GI)

malignancy after colon and pancreatic carcinoma. A Japanese study showed that the
incidence of bone metastases of gastric cancer was 13.4% among autopsies. It is very
rare for the primary presentation of a gastric malignancy to be with bone metastases.
This case report is of a 46-year-old female patient, who presented with a thoracic
vertebral wedge fracture and was subsequently found to have widespread vertebral
metastatic deposits with marrow infiltration. The infiltration and suppression of
marrow function was complicated by an acute bleed into the extradural space causing Received 14 January 2005
spinal cord compression. This case demonstrates two important features. First, that Revised 26 April 2005
gastric cancer, although far less common than breast, kidney, thyroid, prostate and Accepted 12 May 2005
bronchial cancer; is a cause of metastases to bone. Second, it highlights the
DOI: 10.1259/bjr/64677209
complications of bone metastases, marrow suppression, leukoerythroblastic anaemia,
spinal canal haematoma and cord compression. The case is illustrated by axial and ’ 2006 The British Institute of
sagittal MRI slices. Radiology

Metastases to the bony skeleton from gastric carcinoma
are uncommon, with an incidence of 14% in autopsy
specimens [1]. It is rare for the primary presentation of a
gastric carcinoma to be metastases to the bone [2, 3]. This
case is notable for that reason and also because the
widespread bone deposits and marrow infiltration were
complicated by a leukoerythroblastic anaemia and con-
sumptive coagulopathy, which resulted in haemorrhage
into the spinal canal and cord compression.
Case report
A 46-year-old woman presented to the medical
admissions unit with a 4–5 week history of lower back
pain of increasing severity. The pain had become
significantly worse over the previous 10 days and, no
longer being relieved by non-steroidal analgesia, the
patient visited her GP. The GP requested blood tests
which showed a normocytic anaemia – haemoglobin
8.9 g dl21 mean corpuscular volume (MCV) 80.6 fl mean
corpuscular haemoglobin (MCH) 27.0 pg; thrombocyto-
penia with platelet count of 141 6 109 l21; raised liver
function tests with alkaline phosphatase 955 u l21,
alanine aminotransferase (ALT) 66 u l21, gamma-gluta-
myl transpeptidase (GGT) 326 u l21 and raised inflam-
matory markers with erythrocyte sedimentation rate
(ESR) 86 mm h21. A radiograph of the thoracic spine
showed a partial compression fracture of the T10
vertebral body. With these initial investigations com-
pleted and the pain uncontrolled by diazepam, the
patient was referred to hospital for further investigation
and treatment.
Her past medical history was unremarkable. She had
been started on diazepam and codeine/paracetamol based
analgesics and iron tablets by her GP following his consul-
tation. There was no acute history of trauma to trigger her
back pain symptoms. The patient was married with no
children and worked as a teacher. She was a non-smoker and
only drank alcohol occasionally. She was still having regular
periods. Physical examination was unremarkable with no
chest, abdominal or central nervous system signs and only
features of anaemia were found. Blood tests on admission
were similar to those reported by the GP. In addition,
haematinics – B12, folate and ferritin – were normal.
Initial management was targeted towards obtaining
control of her severe back pain and treating her anaemia
with a blood transfusion of three units of packed red
cells. Upper gastrointestinal (GI) endoscopy was delayed
until her back pain was controlled, but was eventually
performed and showed a small amount of fresh blood in
the oesophagus, and a thickened gastric mucosa in the
pre-pyloric area was found with a small amount of blood
oozing from this area. The duodenum was normal.
Biopsies of this gastric lesion were taken.
A repeat blood film done after admission showed
continuing anaemia and thrombocytopenia and also a
raised neutrophil 7.79 6 109 l21, eosinophil 0.47 6 109 l21
and basophil 0.12 6 109 l21 count. The blood film showed a
leukoerythroblastic blood picture containing immature
granulocytes and some nucleated red blood cells. This
picture was highly suggestive of marrow infiltration by
tumour. At this point, 1 week after admission, the patient
developed a nosebleed, which was very hard to control
despite nasal packing. Platelet count was re-checked and
was found to show a significant drop to 21 6 109 l21.

The British Journal of Radiology, March 2006 261


Coagulation screen showed elevated prothrombin time 16 s,
international normalization rate (INR) 1.4, activated partial
thromboplastin time 37 s, and D-Dimer (fibrinogen degra-
dation fragment) 7047 ng ml21. This coagulation picture
was indicative of disseminated intravascular coagulation
(DIC). The following day, the on-call doctor was called to
urgently review the patient since she had lost the use of her
lower limbs. On examination there was paraparesis with
reduced tone and areflexia of the lower limbs. Muscle power
was determined to be 1/5 in both legs. An urgent MR
examination of the lumbar spine was requested.
MR showed two main abnormalities. First, in the cervical,
thoracic and lumbar vertebrae there were numerous
abnormal vertebrae with marrow replacement in several
vertebral bodies with one or two completely infiltrated.
Several vertebrae showed bi-concave collapse, particularly
D2, D10, D12, L3 and L5. Second, there was a large
cylindrical soft tissue mass lying in the spinal cord canal
from D11/12 down to the level of L3. The axial scans
showed it to be lying posterior in the extradural position
behind the dural sac. The dural sac was displaced anteriorly
with maximal compression at L2 level. The soft tissue mass
was of mixed signal and typical for a haematoma in the
extradural space, and was thought to have arisen from a
bleed from a lumbar dural vein (Figures 1 and 2).
Histology results from the biopsy taken at endoscopy
showed antral type gastric mucosa infiltrated by poorly
differentiated adenocarcinoma of diffuse type. Trephine
core bone marrow biopsy from the iliac crest showed
clumps of non-haemopoietic malignant cells. The marrow
architecture was largely replaced by a carcinomatous
infiltrate. Special stains showed mucin positivity and
positive immunohistology with cytokeratins 7 and 20, EP4
and carcinoembryonic antigen (CEA). Oestrogen and
progesterone receptors were negative, as was CA125.
These features were entirely consistent with gastric
primary site of tumour and against a breast primary.
CT of thorax, abdomen and pelvis was performed for
staging of the tumour. CT showed normal lungs with
(a)
Figure 1. (a) T1 and (b) T2 weighted MRI axial slices at the level of D12 vertebra show a cylindrical mass in the extradural space displacing
the dural sac anteriorly and causing compression. The haematoma is isointense to neural tissue on the axial T1 and shows central
hypointensity and peripheral hyperintensity on T2. This pattern is seen in hyperacute to early subacute haematomas (first few days).
262
S Hussain and S Chui
bilateral pleural effusions but no lymphadenopathy.
Abdominal CT showed normal liver, spleen, kidneys,
adrenals and pancreas. Para-aortic lymphadenopathy was
seen with nodes around coeliac axis and superior mesenteric
vessels. No pelvic masses or lymphadenopathy were seen.
Following oncology review it was felt that the patient
was unlikely to benefit from cytotoxic chemotherapy and
prognosis was very poor. Surgical decompression of the
haematoma was considered, but felt to be of too great a
risk for causing further complications in view of the
haematological difficulty in controlling the consumptive
coagulopathy. Referral to palliative care was made.
Discussion
Gastric carcinoma is the third most common GI
malignancy after colon and pancreatic carcinoma. It shows
a significant geographic variation with a five times higher
incidence in Japan and Finland than in the UK. 95% of
gastric carcinomas are adenocarcinomas. A Japanese study
showed that the incidence of bone metastases of gastric
cancer was 13.4% among autopsies [1]. Bone metastases
tend to occur in invasive cancers – Bormann types III and
IV. Examination of the histology has shown that more than
80% of bone metastasis from gastric cancer was poorly
differentiated adenocarcinoma. Thoracic and lumbar
vertebrae are the most frequent sites of bone metastasis
[2] although there are reported cases of deposits in
calcaneum [4], pelvis [5] and skull base. The prognosis of
patients with bone metastases is poor with mean survival
times of less than 5 months with the longest survival
period in the literature of being 3K years [6].
Leukoerythroblastic anaemia arises when the bone
marrow is replaced. Metastases to bone usually occur in
areas that contain red marrow. Metastatic deposits destroy
and replace bone, partly by their own expansion and
partly by stimulating active bone resorption. The pre-
sentation on the peripheral blood film is of normocytic,
(b)
The British Journal of Radiology, March 2006
Case report: Gastric cancer presenting with bone metastaes and spinal haemorrhage
(a)
Figure 2. (a) T1 and (b) T2 weighted MRI sagittal slices showing extension of the haematoma from D11/12 disc level down to the
body of L3. Widespread metastatic foci are seen in the lumbar and lower thoracic vertebral bodies, appearing as low intensity on
the T1 weighted image.
normochromic anaemia with numerous poikilocytes,
normoblasts, reticulocytes, immature white cells and
thrombocytopaenia. The presence of leukoerythroblastic
anaemia is most commonly seen in association with
carcinoma of the breast, stomach, prostate and lung [7].
The low haemoglobin in this patient prompted
investigation for a cause of bleeding resulting in the
discovery of the primary tumour. The additional
presence of thrombocytopenia should raise the suspicion
for DIC and/or bone marrow involvement by malignant
disease and coagulation status and a blood film should
be examined. DIC is rare and can have a varied
presentation with severe bleeding and thrombosis in
large and small vessels due to activation of the
coagulation cascade resulting in a consumption coagulo-
pathy. Although it is reported to be associated with
every type of malignancy, the strongest association is
with GI adenocarcinomas and leukaemia. Treatment
directed at decreasing the risk of spontaneous bleeding
should be considered. In this case, the posterior position
of the haematoma in the spinal canal suggests that it was
The British Journal of Radiology, March 2006
(b)
not related to the vertebral fractures and more likely to
be a spontaneous haemorrhage.
References
1. Nisidoi H, Koga S. Clinicopathological study of gastric
cancer with bone metastases. Gan To Kagaku Ryoho
1987;14:1717–22.
2. Yoshikawa K, Kitaoka H. Bone metastasis of gastric cancer.
Jpn J Surg 1983;13:173–6.
3. Carstens SA, Resnick D. Diffuse sclerotic skeletal metastasis
as an initial feature of gastric carcinoma. Arch Intern Med
1980;140:1666–8.
4. Bergqvist D, Mattson J. Solitary calcaneal metastases as the
first sign of gastric cancer. Ups J Med Sci 1978;83:115–8.
5. Chung YS, Choi TY, Ha CY, et al. An unusual case of
osteoblastic metastasis from gastric carcinoma. Yonsei Med J
2002;43:377–80.
6. Kobayashi M, Araki K, Matsuura K, et al. Early gastric cancer
giving rise to bone and brain metastases: a review of the
Japanese literature. Hepatogastroenterology 2002;49:1751–4.
7. Rubins JR. The role of myelofibrosis in malignant myelo-
sclerosis. Cancer 1983;51:308–11.
263
 The British Journal of Radiology, 79 (2006), 264–265

CASE REPORT

Diagnosis of myocardial contusion after blunt chest trauma using

18
F-FDG positron emission tomography
M PAI, MD

Department of Radiology, College of Medicine, Ewha Womans University, Ewha Mokdong

Hospital, 911-1 Mok-dong, Yangchun-gu, Seoul 158-710, Republic of Korea

ABSTRACT. Cardiac contusion is an infrequent complication of blunt chest trauma. The

definite diagnosis of myocardial contusion is complex and needs a number of
examinations such as electrocardiography, echocardiography, cardiac enzyme and Received 6 March 2005
radionuclide perfusion scan. We present a patient who had a blunt chest trauma from a Revised 10 May 2005
car accident resulting in an acute myocardial infarction without injury to coronary Accepted 18 May 2005
arteries. The non-viable myocardium was diagnosed with 18F-fluorodeoxyglucose
DOI: 10.1259/bjr/17221499
positron emission tomography (FDG PET) combined with 201Tl perfusion single photon
emission tomography (SPECT). This is the first report of FDG PET for the diagnosis of ’ 2006 The British Institute of
myocardial contusion in blunt myocardial trauma. Radiology

Cardiac contusion is usually caused by blunt chest
trauma arising from car accidents or falling injuries. The
reported incidence of cardiac contusion in patients with
blunt chest trauma ranges between 3% and 56%
depending on the diagnostic methods [1–4]. Acute
myocardial infarction is a rare but serious complication
after blunt chest trauma and is usually caused by injuries
to coronary arteries such as coronary thrombosis,
laceration, fistula and pseudoaneurysm [1, 5, 6]. A
number of methods have been employed to diagnose
cardiac contusion and to predict the associated compli-
cations. Electrocardiography, echocardiography and car-
diac enzymes are routine and additional coronary
angiography or 201Tl myocardial perfusion single photon
emission tomography (SPECT) have been used to
localize the injured myocardial walls [7].
18
F-fluorodeoxyglucose positron emission tomography
(FDG PET) is known to have great advantages in the
assessment of myocardial viability [8]. There has been no
report citing findings of myocardial FDG PET in patients
with blunt chest trauma [1]. We report a case of acute
myocardial infarction without coronary artery injuries
after a blunt chest trauma, in which myocardial viability
status was evaluated by FDG PET combined with 201Tl
perfusion SPECT.
Case report
A 45-year-old previously healthy, ‘‘10 packs a year
cigarette smoking man’’ with no other risk factors for
ischaemic heart disease was admitted to hospital after a car
accident, in which he had been hit by the steering wheel on
his anterior chest. On arrival, he complained of chest pain
and dyspnoea. On physical examination, no evidence of
hypotension or arrhythmia was found. He was diagnosed
with fractures of sternum and bilateral costochondral
junctions with bone scan. The electrocardiography showed
sinus rhythm and non-specific T-wave inversion in leads
V3–6. The serum creatinine kinase was 221 U l21 with a
slight elevation of MB fraction (6.7 ng ml21).
Echocardiography revealed normal cardiac dimensions
and hypokinesia of the anterior wall with ejection fraction
of 51%. However, coronary angiography showed no
evidence of coronary artery disease. For the assessment
of myocardial viability, FDG PET scan (Allegro; Phillips
Medical Systems, Best, The Netherlands) was performed
after intravenous administration of 370 MBq (10 mCi) of
18
F-FDG with usual oral glucose loading method [9]. A
metabolic defect was noted in the anteroseptal wall on
FDG PET corresponding to the perfusion defect on 201Tl
SPECT (ECAM; Siemens, Knoxville, TN) (Figure 1). This
matching defect is consistent with myocardial infarction
arising as a consequence of blunt chest trauma in the
context of normal coronary arteries. On 8 month follow-up
studies of 201Tl SPECT and echocardiography, there was
no improvement of perfusion and left ventricular wall
motion abnormality.
Discussion
Early diagnosis of cardiac lesions following blunt
trauma is essential but there remains no sensitive test for
myocardial contusion. Although the true diagnosis of
contusion can only be established by direct inspection of
myocardium [10], in the majority of cases, the diagnosis
of cardiac contusion with myocardial injury is made on
the basis of elevated creatinine kinase-MB fraction,
which is not predictive of the complications of cardiac
contusion [11, 12]. Myocardial regulatory contractile
proteins such as troponin I and T are considered to be
sensitive and specific for myocardial injury [2], despite
problems with timing of blood sampling after trauma.

264 The British Journal of Radiology, March 2006

M Pai
Figure 1. Short axis slice of (a) myocardial 18F-fluorodeoxyglucose positron emission tomography (FDG PET) and (b) 201Tl single
photon emission tomography (SPECT) demonstrated perfusion-metabolic matching defect in anteroseptal wall (arrows),
indicating non-viable myocardium.
Echocardiography can be used as a screening tool for
ventricular dysfunction which is one of the most serious
complications of cardiac contusion but is often hampered
by a suboptimal ultrasound window [13]. There are a
few reports that blunt chest trauma may lead to acute
myocardial infarction caused by coronary artery dissec-
tion, thrombosis or fistula [6, 14, 15]. Contused myocar-
dial tissue is histologically characterized by
intramyocardial haemorrhage, oedema and necrosis of
myocardial muscle cells, which are similar to the
findings of acute myocardial infarction [10]. In the
assessment of myocardial viability, FDG PET has been
used as the method of choice for the diagnosis of non-
viable, infarcted myocardium. In this study, FDG PET
identified the exact extent and location of the injured
tissue by describing non-viable myocardium despite the
presence of normal coronary arteries. The presence of
infarction was confirmed by the lack of improvement in
myocardial perfusion and wall motion at 8 month
follow-up studies by 201Tl SPECT and echocardiography.
This is the first report to use FDG PET for diagnosis of
myocardial contusion after blunt chest trauma in a
patient with normal coronary artery.
In conclusion, FDG PET helped to identify the
contused myocardium as a result of perfusion-metabo-
lism matching defect suggesting non-viable infarcted
myocardium. FDG PET may be useful in early decision
making for patients with blunt chest trauma in a case
with indistinct laboratory and imaging findings.
References
1. Sybrandy KC, Cramer MJ, Burgersdijk C. Diagnosing
cardiac contusion: old wisdom and new insights. Heart
2003;89:485–9.
2. Collins JN, Cole FJ, Weireter LJ, Riblet JL, Britt LD. The
usefulness of serum troponin levels in evaluating cardiac
injury. Am Surg 2001;67:821–5.
3. Karalis DG, Victor MF, Davis GA, McAllister MP,
Covalesky VA, Ross JJ Jr, et al. The role of echocardio-
graphy in blunt chest trauma: a transthoracic and
The British Journal of Radiology, March 2006
transesophageal echocardiographic study. J Trauma
1994;36:53–8.
4. Bertinchant JP, Polge A, Mohty D, Nguyen-Ngoc-Lam R,
Estorc J, Cohendy R, et al. Evaluation of incidence, clinical
significance, and prognostic value of circulating cardiac
troponin I and T elevation in hemodynamically stable
patients with suspected myocardial contusion after blunt
chest trauma. J Trauma 2000;48:924–31.
5. Salmi A, Blank M, Slomski C. Left anterior descending
artery occlusion after blunt chest trauma. J Trauma
1996;40:832–4.
6. Marcum JL, Booth DC, Sapin PM. Acute myocardial
infarction caused by blunt chest trauma: successful treat-
ment by direct coronary angioplasty. Am Heart J
1996;132:1275–7.
7. Maenza RL, Seaberg D, D’Amico F. A meta-analysis of
blunt cardiac trauma: ending myocardial confusion. Am J
Emerg Med 1996;14:237–41.
8. Ghesani M, Depuey EG, Rozanski A. Role of 18F FDG
positron emission tomography (PET) in the assessment of
myocardial viability. Echocardiography 2005;22:165–77.
9. Tamaki N, Yonekura Y, Konishi J. Myocardial FDG PET
studies with the fasting, oral glucose-loading or insulin
clamp methods. J Nucl Med 1992;33:1266–8.
10. Tenzer ML. The spectrum of myocardial contusion: a
review. J Trauma 1985;25:620–7.
11. Keller KD, Shatney CH. Creatine phosphokinase-MB assays
in patients with suspected myocardial contusion: diagnostic
test or test of diagnosis? J Trauma 1988;28:58–63.
12. Biffl WL, Moore FA, Moore EE, Sauaia A, Read RA, Burch
JM. Cardiac enzymes are irrelevant in the patient with
suspected myocardial contusion. Am J Surg 1994;168:
523–7.
13. Miller FAJ, Seward JB, Gersh BJ, Tajik AJ, Mucha P Jr. Two-
dimensional echocardiographic findings in cardiac trauma.
Am J Cardiol 1982;50:1022–7.
14. Yoon SJ, Kwon HM, Kim DS, Hong BK, Kim DY, Cho YH,
et al. Acute myocardial infarction caused by coronary artery
dissection following blunt chest trauma. Yonsei Med J
2003;44:736–9.
15. Kahn JK, Buda AJ. Long-term follow-up of coronary artery
occlusion secondary to blunt chest trauma. Am Heart J
1987;113:207–10.
265

Correspondence
(The Editors do not hold themselves respondible for opinions expressed by correspondents)
Source to image receptor distance
The Editor—Sir,
With regard to the paper ‘‘The effect of source to image-
receptor distance on effective dose for some common X-
ray projections’’ by Poletti and McLean [1], published in
the September 2005 issue of this journal, the authors have
performed the work competently and have presented
some valuable results. There are, however, some points
we would like to raise.
We believe that the description of the dose reductions
demonstrated in the paper as ‘‘modest’’ undervalues to
an extent the contribution that increasing the focus to
image receptor distance can have on optimization of
radiation doses; especially in light of their findings. In
their paper, reductions of up to 46% and 68% were
shown for thoracic exposures following consistent
collimation to a central region of interest and the
entrance surface, respectively, with increased distance.
Whilst we accept that the latter measurement is clinically
unimportant, since the structure of interest is rarely
positioned at the entrance surface, we believe that the
reductions shown at the central region of interest are
impressive and compare very favourably with other,
more expensive methods of dose-reduction. Admittedly,
when collimation was fixed to the image receptor, little
or no reduction was shown at increased distances.
However, under optimized radiographic protocols, this
is of no clinical relevance since a region of interest can
never be coincident with the image receptor. The data
provided therefore demonstrate that good radiographic
practice is critical in that the level of collimation and
entrance surface markings employed for a specific body
part should change depending on the position of the
structure along the anteroposterior axis and the distance
between the focus and the image receptor (unless the
region of interest is at the entrance surface, which as
stated above is highly unlikely). In practice this appears
to be rarely done.
On a separate note, the authors highlight the impor-
tance of theoretical studies to examine a topic of this
nature and describe some of the problems encountered
in experimental verification of theory. We acknowledge
that computer simulation can generate an array of
valuable data and presents answers to difficult scientific
questions in a very precise way, but rarely do patients
demonstrate such a high level of precision. If dose-
reducing measures are going to be useful clinically
(which presumably is the main reason for doing these
types of investigations), a thorough understanding of the
effect of specific measures on patients of varying size,
state of health, gender and so on must be established
before widespread implementation. Mathematical mod-
elling alone can rarely achieve this.
The data presented by Poletti and McLean are
important and of much use to the radiographic and
radiological communities. We believe that their results,
266
The British Journal of Radiology, 79 (2006), 266–267
like those of some previous workers [2], have shown that
increasing source to image receptor distance can reduce
stochastic radiation risk for common X-ray examinations,
and further studies investigating more widespread usage
should be encouraged.
Yours etc.,
P C BRENNAN and
D O’LEARY
School of Medicine and Medical Sciences
University College Dublin
Belfield
Dublin 4
Ireland
(Received 24 October 2005 and accepted 25 October 2005)
DOI: 10.1259/bjr/60546865
References
1. Poletti JL, McLean D. The effect of source to image-receptor
distance on effective dose for some common X-ray projec-
tions. Br J Radiol 2005;78:810–5.
2. O’Leary D, McDonnell S, Brennan PC. Increasing film focus
distance (FFD) reduces radiation dose for X-ray examina-
tions. Radiat Prot Dosim 2004;108:263–8.
Authors’ reply
The Editor—Sir,
In response to the comments by Brennan and O’Leary on
our paper in the September 2005 issue of this journal [1],
we are pleased that they agree that collimation to a fixed
field size at the entrance surface, as the source to image
distance (SID) is changed, is clinically irrelevant. They
further claim that collimation to the image receptor as
the SID is changed is also clinically irrelevant. Whilst we
agree that this should be the case we would like to point
out, however, that collimation to the image receptor is
the default requirement in many codes of practice.
Furthermore, where autocollimation systems or digital
image systems that permit collimation of the image post-
exposure are used, there is at least anecdotal evidence
that collimation to the image receptor only, becomes all
too common.
As for our claim that the dose reduction to be obtained
by an increase in SID is modest, we stand by this. The
case quoted by Brennan and O’Leary from our results is
obtained by changing from 75 cm to 200 cm SID, an
extreme case not likely to occur in practice. The effective
dose reduction to be obtained by increasing from the
commonly used 100 cm to 125 cm is typically 10% or so
according to our results. This is not to say that such dose
reductions are not worth obtaining.
The British Journal of Radiology, March 2006
Correspondence

Finally, we accept that patient variability is an issue. Auckland, 1003

This leads to obvious difficulties in experimental work New Zealand
with current theoretical approaches also deficient. We
2
look forward to developments in voxel patient modelling Adjunct A/Prof University of Sydney
for Monte Carlo simulations that can more realistically School of Medical Radiation Sciences
give insight into changes in organ doses and hence Faculty of Health Sciences, University of Sydney
effective doses that occur due to patient variability, and PO Box 170 Lidcombe NSW Australia 2141
consequently the impact of this on changes to the SID.
Yours etc.,
1 (Received 14 November 2005 and accepted 18 November 2005)
J POLETTI and
2 DOI: 10.1259/bjr/71660401
D MCLEAN

1
Health and Community Studies References
Unitec 1. Poletti JL, McLean D. The effect of source to image-receptor
Private Bag 92025 distance on effective dose for some common X-ray projec-
Carrington Road tions. Br J Radiol 2005;78:810–5.

The British Journal of Radiology, March 2006 267


Book review
Cross-sectional imaging made easy. By SA Jackson,
RM Thomas. pp. 147, 2004 (Churchill Livingstone,
Edinburgh) £14.99
ISBN 0-443-07187-X
Cross sectional Imaging Made Easy is a smart 147-page
pocket book from the successful medical book series
‘‘Made Easy’’. A word of warning: an exhaustive
reference book for the initiated this is certainly not!
Instead, this book is written to offer a simplified, brief
and easily digestible approach to cross sectional imaging
for the uninitiated – medical students, junior doctors and
clinicians with an interest in radiology. This is a
commendable ambition and one that I believe the
authors have achieved quite successfully. Only essential
information is included in the text, which is presented in
a bullet point fashion. This enables the book to remain
concise and easy to read while permitting twice the
amount of information to be included when compared
with a standard prose format. The sections dedicated to
the physics of image generation for CT, MRI and
ultrasound are particularly well written and easy to
understand. This is a splendid achievement, as radi-
ological physics is often perceived as alien and difficult
to grasp, as many radiology trainees studying physics for
the Part 1 FRCR examination will testify! Another good
point of the book is that it is well illustrated with a
268
The British Journal of Radiology, 79 (2006), 268
combination of schematic diagrams and numerous high
quality radiological images. The radiological images are
presented not just in standard axial, coronal and sagittal
views but also using 3D reconstructions with surface
shaded display and maximal intensity projection for-
mats. There is on average at least one radiological image
per page, usually more. In general, the authors have
strived to focus on the imaging of clinical cases that
represent the most appropriate use of the imaging
modalities that are available. For CT, cases are chosen
to highlight the usefulness of CT in the clinical setting of
an acutely unwell patient. For MRI, cases are chosen to
demonstrate the usefulness of MRI in detecting an
abnormality on the basis of its superior soft tissue
contrast resolution. For ultrasound, cases are chosen to
demonstrate common clinical applications where ultra-
sound is utilized to provide an easy, reliable and non-
radiation means of dynamic assessment.
Thus in summary, I believe that this book achieves its
aim of providing non-radiologists with an insight into
cross sectional imaging using a simple, no ‘‘frills’’ and
‘‘cringe free’’ approach. Despite the book’s small size
and the fact that it can be read ‘‘cover-to-cover’’ in a
couple of days, it packs a surprisingly ‘‘large punch’’
with a sizeable amount of useful information, and
represents good value for money.
G LOW
The British Journal of Radiology, March 2006
The British Journal of Radiology, 79 (2006), 269–270

CASE OF THE MONTH

An intranasal mass
K GOWDA, MD, M FARRUGIA, MD, FRCR and C PADMANATHAN, FRCR

Department of Radiology, Newham University Hospital, Glen Road, Plaistow, London E13 8SL, UK

Received 31 May 2005

Accepted 27 June 2005

DOI: 10.1259/bjr/50814804

’ 2006 The British Institute of

Radiology

A 10-year-old male child with suspected nasal poly- nose and throat (ENT) surgeon (Figure 1). What is the
posis was referred for an MRI examination by the ear, diagnosis?

(a) (b) (c)

Figure 1. (a,b,c) Coronal T1 weighted,

T2 weighted, (d) axial T2 weighted spin
echo sequences and (e) sagittal T 1
(d) (e) weighted images.

The British Journal of Radiology, March 2006 269


Intranasal encephalomeningocele
Anterior encephalocele is a rare condition and only a
few large series have been published in the literature.
Surprisingly, the incidence is much higher in Southeast
Asian countries, including some parts of India.
Frontoethmoidal encephaloceles are the most common
type, followed by the nasopharyngeal and orbital types.
Among the frontoethmoidal encephaloceles, the nasoe-
thmoid variation is the most common type, and these
patients present with swelling over the bridge of the nose
with significant hypertelorism and orbital deformities.
The nasopharyngeal type remains occult and presents
with nasal obstruction or cerebrospinal fluid (CSF)
rhinorrhoea. Rarely, the patient may present with
meningitis [1]. The successful correction of frontoeth-
moidal encephaloceles has been shown to depend on
the detailed understanding of the pathological anatomy
[2].
MRI is the best modality for imaging encephalome-
ningocoeles as accurate information regarding the origin
270
K Gowda, M Farrugia and C Padmanathan
and site of herniation can be obtained with minimal
invasiveness.
In MRI, nasal encephaloceles are always identified as
complex masses of mixed soft tissue and CSF intensity that
are contiguous with intracranial cerebral matter.
Sometimes, differentiation of actual cortical grey matter
and subcortical white matter can also be made if a large
part of the gyrus has herniated. The differential diagnosis
includes a dermal sinus extending into the nasal cavity,
also known as an intranasal dermoid and nasal cerebral
heterotopias, more commonly known as nasal gliomas [3].
References
1. Mahapatra AK, Suri A. Anterior encephaloceles: a study of
92 cases. Pediatr Neurosurg 2002;36:113–8.
2. Holmes AD, Meara JG, Kolker AR, Rosenfeld JV, Klug GL.
Frontoethmoidal encephaloceles: reconstruction and refine-
ments. J Craniofac Surg 2001;12:6–18.
3. Barkovich AJ, Vandermarck P, Edwards MS, Cogen PH.
Congenital nasal masses: CT and MR imaging features in 16
cases. AJNR Am J Neuroradiol 1991;12:105–16.
The British Journal of Radiology, March 2006
The British Journal of Radiology, 79 (2006), 271
Book review
Get through nuclear medicine for the FRCR and MRCP.
By J Frank, T Nunan. pp. xi+179, 2004 (Royal Society
of Medicine, London, UK) £19.95
ISBN 1-85315-591-8
It seemed helpful to have two (registrar and consultant)
review this book – given its title. This 180 page book is
set out in a random question and answer format of 80
cases as one might expect to face in the FRCR or MRCP
examinations. Each case is introduced with an image
followed by short questions sometimes with clinical
information. The questions often start with radiographs
or non-radionuclide scans allowing the candidate to
suggest appropriate isotope imaging. The answers and
teaching points are found on the following pages. A wide
range of nuclear medicine techniques are discussed
covering all parts of the body. Emphasis is given to
more commonly encountered investigations such as
bone scans, V/Q scans and renograms. Less frequent
techniques are also covered including cardiac, neuro-
logical, biliary, GI and imaging techniques for infection.
The use of PET for oncological imaging is also included
with CT-PET having a brief mention. References for
further reading and an index of the cases are given at the
end for easy reference.
Knowledge required for the FRCR 2A exam is covered
with discussion of agents and techniques used for
various investigations. Illustrations of normal appear-
ances and uptake for specific agents are given. Helpful
tips on potential pitfalls in image interpretation are given
such as artefacts mimicking pathology. Less commonly
encountered techniques such as HMPOA brain scanning
and imaging of neuroendocrine studies are also briefly
covered. Cases also introduce newer applications such as
PET scanning in lung cancer staging and follow-up.
There is discussion of the role nuclear medicine plays
in the management of a variety of conditions. A major
plus point for the radiologist is the integration of nuclear
medicine with radiology with discussion of the relative
advantages of each, a feature often missing from
dedicated nuclear medicine texts. Understanding is
aided by cases that reflect real-life clinical situations,
making the reading much less dry than trawling through
a textbook. As always there are some minor points of
The British Journal of Radiology, March 2006
contention, e.g. recommending spiral CT following high
probability V/Q scan and suggesting a post treatment
baseline scan – a counsel of perfection rarely practiced
and contentious in view of the additional radiation
exposure.
One criticism of the book is that some of the images are
very small and the quality of the images (particularly
radiographs) does not always allow the abnormality to
be spotted which can be frustrating. This is partly a
consequence of a laudable attempt to keep the cost of the
book low. However, the majority of images are of high
quality and illustrate the teaching points being made
adequately. The teaching points provided are very
practical and one can imagine having to trawl through
many pages of deeper textbook to find such information.
Overall, this is an enjoyable book suitable for
candidates about to sit their FRCR or MRCP exams.
Though short and concise, it covers a wide range of
nuclear medicine techniques and clinical scenarios. It can
be used as a starting point to stimulate further reading of
investigations not frequently seen on a daily basis.
Though not an essential read, it is a good starting point
for radiology candidates who feel a bit rusty with their
nuclear medicine interpretation prior to the exam. With
the emphasis on plain films and cross sectional imaging,
and the modular training nuclear medicine is one of the
subjects that (sometimes) falls to the wayside during
exam preparation and registrars may not have had much
exposure to nuclear medicine except during film viewing
tutorials. It did help this registrar prepare for the exam,
having not encountered many inexpensive and concise
books that cover nuclear medicine in an exam type
format. One would be unlucky to meet a nuclear
medicine technique in the long case section of the
FRCR 2B exam not covered by this book.
For those who do pass the exam (with or without the
help of this book) there are two short, clear and helpful
sections at the end describing the role of ARSAC, the
IRMER regulations and the current routes by which
training in nuclear medicine/radionuclide radiology can
be achieved.
P LI and
P GUEST
271
BJR
The British Journal
of Radiology

April 2006
Volume 79
Issue 940
April 2006, Volume 79, Issue 940
● BJR Advance – A new service of Publish Ahead of Print

● Particle matters

● HCPBM [Heavy Charged Particles in Biology & Medicine] and

ENLIGHT [European Network for Light Ions] Meetings, Oropa
(Italy), June 2005

● What are the risks from medical X-rays and other low dose
radiation?

● Diagnosing a parotid lump: fine needle aspiration cytology or

core biopsy?

● Intra-arterial MR angiography in the iliac system: initial clinical

experience with 25 patients

● Prevalence and distribution of adnexal findings suggesting

endometriosis in patients with MR diagnosis of adenomyosis

● Influence of post-treatment delay on the evaluation of the

response to focused ultrasound surgery of breast cancer by
dynamic contrast enhanced MRI

● Introduction of grids to mobile ICU radiography in a teaching

hospital

● Colour Doppler ultrasound patterns and clinical follow-up of

incidentally found hypoechoic, vascular tumours of the spleen:
evidence for a benign tumour

● A review of current local dose–area product levels for paediatric

fluoroscopy in a tertiary referral centre compared with national
standards. Why are they so different?

● Peripheral dose from uniform dynamic multileaf collimation

fields: implications for sliding window intensity-modulated
radiotherapy

● The DXL Calscan heel densitometer: evaluation and diagnostic

thresholds
 ● Time-dependent observer errors in pulmonary nodule detection

● Study of scattered radiation for in-air calibration by a multipledistance

method using ionization chambers and an HDR 192Ir
brachytherapy source

● Implications of quality adjusted survival for clinical trials in

radiation oncology

● Correspondence

● The angel in the marble

The British Journal of Radiology, 79 (2006), 275
GUEST EDITORIAL
BJR Advance – A new service of Publish Ahead of Print
Introduction by the Honorary Editors
It is increasingly important for medical journals and
more specifically for those focusing on best practice
using health care technology to remain at the cutting
edge of publishing and to provide such facilities for its
authors and readers. In particular they must take
advantage of all the electronic developments available
in modern-day publishing. In this context we are very
pleased to announce our new online publication service.
In this Guest Editorial our Publisher, Ms Sherry Dixon,
explains how BJR Advance – A new service of Publish Ahead
of Print will work.
BJR Advance – A new service of Publish Ahead
of Print
From April 2006, readers of The British Journal of
Radiology (BJR) will find even more content online. BJR
Advance, the journal’s website for publishing papers
electronically ahead of print will be launched at http://
www.bjr.birjournals.org.
Since the BJR launched its online submission and peer
review system in June 2004, the number of submissions
has doubled, and the number of papers accepted for
publication has increased. With the introduction of BJR
Advance, the majority of these papers will be published
online earlier than the printed version. Notwithstanding,
we are also committed to ensuring that those articles
accepted for publication in print will appear within a
reasonable time scale.
BJR Advance will bring readers full papers, case
reports, reviews, pictorial reviews and commentaries
on Diagnostic Radiology, Radiotherapy and Oncology,
Medical Physics, and Radiobiology several weeks in
advance of the printed edition of BJR. The site will be
updated bi-monthly with the latest research.
All BJR Advance manuscripts will undergo the com-
plete peer-review process and are essentially the same
version of the article that will appear in print. They will
have been peer reviewed, revised by the authors,
copyedited, typeset, and approved by the authors and
BJR editors. There may be minor typographical errors,
which will be corrected in the final print and online
versions. However, there will be no substantive factual
changes.
The British Journal of Radiology, April 2006
DOI: 10.1259/bjr/22015456
’ 2006 The British Institute of
Radiology
Digital Object Identifiers (DOIs) will be used as the
Article IDs for all pre-press articles. The DOI is a
specification and a system to support unique and
permanent identifiers for content in an electronic
environment, developed by the non-profit International
DOI Foundation (http://www.doi.org). The DOIs will
serve as ‘‘serial numbers’’ for articles, unambiguously
identifying each one.
Like regular journal content, all BJR Advance articles will
be deposited with the National Library of Medicine for
inclusion in Medline and PubMed to ensure early and
wide dissemination. In addition all pre-press articles will
be fully searchable and references will be hyperlinked.
Authors and researchers should be assured that for
citation purposes the DOI can replace the year, volume,
and page numbers for a given article. For example,
N For a BJR Advance article just posted online:
Published May 6, 2006, Br J Radiol, 10.1259/bjr/
83624598
N Then, once the article appears in print or is accessible
on BJR Online:
Br J Radiol 2006;75:356–361.
N If the exact publication date is important (e.g. for an
author)
Published May 6, 2006, 10.1259/bjr/83624598
Br J Radiol, 2006;75:356–361.
This new section of BJR online will have a link from
the journal homepage. A reader who clicks on this link
will be able to view, download and cite articles
published in advance of print – just as they do now for
current and archive journal articles. Once a printed issue
of BJR is mailed, articles in the printed issue that were
published on BJR Advance will move to the regular BJR
Online.
We hope that this addition to the BJR will significantly
improve our ability to make our authors work more
speedily available to the medical community, and
enhance the process of submission, publication and
using our journal.
S DIXON
Publisher
275

EDITORIAL
Particle matters
A J MUNRO
The commentary in this issue [1], along with several
other recent publications [2–8], draws our attention to
important developments that should be of particular
interest to readers of the BJR. Physicists, engineers,
radiobiologists and clinicians should be aware that in
Europe, the Far East and in the USA there has recently
been considerable activity in the field of particle therapy.
For worse, or perhaps for better, there has been little
progress in the UK. The outstanding exception being the
meticulous work on proton therapy for ocular tumours
carried out using the cyclotron at Clatterbridge, near
Liverpool [9, 10]. There are two ways of looking at the
British inertia when it comes to particle therapy: one is the
perspective of Professor Jones, that it is reprehensible sloth
which puts us in danger of being left behind scientifically
and which denies a significant number of patients with
cancer the treatment they need and deserve; the other view
is that the recent renaissance of particle therapy is based on
hope, rather than evidence, and that, as a nation, we do
ourselves no great disservice by failing enthusiastically to
adopt unproven therapeutic technologies. Should we be
prudent, or should we be (relatively) early adopters?
Perhaps some of the explanation for our apparent
reluctance to rush out and embrace particle therapy lies
with our recent history. We were, after all, enthusiastic
early adopters of neutron therapy and spent the best part
of 25 years finding out that, in the long run, we might have
been better putting our efforts elsewhere. The lessons
learnt were: that an apparently compelling biological
rationale does not always translate into improved clinical
results; and that even ‘‘randomized controlled’’ trials may
be no bulwark against misplaced enthusiasm. These things
reverberate. It is not just about the wasted money, or even
the futile intellectual effort. It is about the false hopes that
we raised in the public; we ended up expending patients’
goodwill and altruism on proving a negative. There was
also an opportunity cost to all of this: because we were
doing things we need have not done, we could not do the
things we should have done.
This journal published the inquest into the MRC trials of
neutron therapy for Head & Neck cancer [11], and in 1987
the BIR awarded the Röntgen Prize to Professor William
Duncan and team for their meticulous clinical studies
demonstrating that fast neutron therapy offered no
significant advantage over conventional therapy with
photons. It is entirely appropriate, therefore, that the BJR
should take the lead in encouraging an open debate
concerning particle therapy in the 21st century. We cannot
276
The British Journal of Radiology, 79 (2006), 276–277
DOI: 10.1259/bjr/14849445
’ 2006 The British Institute of
Radiology
assume that, because particle therapy ought to be better
than conventional therapy, it necessarily will be better.
Nor can we assume that our melancholy experience with
neutrons will not be repeated. For proton therapy, at least,
we only have to consider potential physical advantages;
with a relative biological effectiveness (RBE) of 1.1 we are
unlikely to encounter unpleasant biological surprises. The
position of ion therapy is, however, more complex: we
have to consider both the physical advantages associated
with the Bragg peak as well as the potential biological
advantages associated with high linear energy transfer. If
particle therapy does have a role in cancer management,
then that role needs to be carefully defined, on the basis of
evidence, not opinion.
The peer-reviewed literature contains little evidence
upon which we might base decisions concerning the
efficacy and cost-effectiveness of particle therapy. Most
of the information is hidden in meeting reports and
discussion documents. For the moment at least, the
benefits from particle therapy appear to be limited to
certain uncommon tumours: ocular tumours; tumours of
the base of skull; chordomas. Good evidence for cost-
effectiveness is non-existent: such studies as have been
performed have been methodologically unsound and
poorly reported. The onus is now firmly on the
advocates of particle therapy to expose their results to
peer review rather than simply presenting them to
meetings of like-minded people.
As Professor Jones points out, the main benefit from
particle therapy may be for patients with rare tumours. If
this is the case, then randomized controlled trials will be
neither feasible nor appropriate. We need to think of other
methods to construct an evidence-base for particle
therapy. The main criticism of the literature on particle
therapy at present is that it represents little more than a
series of anecdotal reports on the results of treating
patients who may, or may not, be representative of
patients with a given type of tumour. It is interesting to
find that although over 8000 patients have been referred
for particle therapy in Italy, it would seem that only 54
have been treated [1]. I would argue strongly that, as soon
as possible, we need to construct a central database
including information on all patients treated with particle
therapy. In the first instance, a Europe-wide initiative
would be sufficient. The database would function pro-
spectively as a particle therapy register and should, for
each patient, include demographic data as well as
information on tumour type, stage and grade, treatment
The British Journal of Radiology, April 2006
Editorial
details and outcome. The register could then be linked to
cancer registries and we would therefore be able to bring a
sense of perspective to the potential role of particle therapy
as well as sensibly plan future developments. The idea of
centres springing up simply because some EU develop-
ment funds are available, or there happens to be a basic
physics research facility, or because a private health care
company thinks there might be a market to exploit, is
unappealing. Particle therapy for cancer is an unproven
intervention and its development needs to be part of a co-
ordinated scientific research programme. Professor Jones’s
commentary is an extremely useful reminder of where we
might be investing – but it is not just about where: it is also
about why, how and when.
Conflict of interest and acknowledgments
The author worked at both the Hammersmith Hospital,
London and the Western General Hospital, Edinburgh at
the time of the clinical trials of fast neutron therapy. He
would like to thank the following colleagues for helpful
discussions: Mark Lodge; Dirk De Ruysscher, Tom
Jefferson, Madelon Pijls-Johannesma, Lisa Stirk and
Nancy Owens. The opinions expressed above are entirely
his own.
References
1. Jones B. HCPBM [Heavy Charged Particles in Biology &
Medicine] and ENLIGHT [European Network for Light
The British Journal of Radiology, April 2006
Ions] Meetings, Oropa (Italy), June 2005. Br J Radiol
2006;79:277–83.
2. Jones B. The case for particle therapy. Br J Radiol
2006;79:24–31.
3. Jones B, Burnet N. Radiotherapy for the future. BMJ
2005;330:979–80.
4. Jones B, Dale R. The clinical radiobiology of high LET
radiotherapy with particular reference to proton radio-
therapy. Clin Oncol (R Coll Radiol) 2003;15:S16–22.
5. Jones B, Price P. Proton therapy: expanding clinical
indications. Clin Oncol (R Coll Radiol) 2004;16:324–5.
6. Jones B, Price P, Burnet NG, Roberts JT. Modelling the
expected increase in demand for particle radiotherapy:
implications for the UK. Br J Radiol 2005;78:832–5.
7. Jones B, Rosenberg I. Particle Therapy Co-operative
Oncology Group (PTCOG 40) meeting, Institute Curie
2004. Br J Radiol 2005;78:99–102.
8. Price P, Errington RD, Jones B. Report on the U.K. meeting
September 2001 to discuss the clinical and scientific case for
a high-energy proton therapy facility in the U.K. Clin Oncol
(R Coll Radiol) 2003;15:S1–9.
9. Damato B, Kacperek A, Chopra M, Campbell IR, Errington
RD. Proton beam radiotherapy of choroidal melanoma: the
Liverpool-Clatterbridge experience. Int J Radiat Oncol Biol
Phys 2005;62:1405–11.
10. Damato B, Kacperek A, Chopra M, Sheen MA, Campbell IR,
Errington RD. Proton beam radiotherapy of iris melanoma.
Int J Radiat Oncol Biol Phys 2005;63:109–15.
11. A comparative review of the Hammersmith (1971–75) and
Edinburgh (1977–82) neutron therapy trials of certain
cancers of the oral cavity, oropharynx, larynx and hypo-
pharynx. Medical Research Council Neutron Therapy
Working Group. Br J Radiol 1986;59:429–40.
277

COMMENTARY
HCPBM [Heavy Charged Particles in Biology & Medicine] and
ENLIGHT [European Network for Light Ions] Meetings, Oropa
(Italy), June 2005
B JONES, MSc, MD, FRCR
Birmingham Cancer Centre, University Hospital Birmingham, Birmingham B15 2TH, UK
Some 13 years ago, the European Organisation for
Research and Treatment of Cancer (EORTC) organized a
large meeting devoted entirely to neutron therapy, its
original promise and, of course, its clinical disappoint-
ments. The term ‘‘charged particles’’ has changed the
emphasis, with realistic expectations of better clinical
outcomes; the best available radiation dose distributions
can now be achieved due to the Bragg peak effect that
results from the deceleration of charged particles (CP) in
tissues; neutrons and X-rays carry no charge and do not
exhibit such selectivity in dose deposition. Consequently,
the application of curative radiotherapy in the future is
expected to rely more and more on these physical
advantages [1]. The reader who may be unfamiliar with
terms and some of the acronyms used should refer to
Tables 1 and 2. This article is written to inform those
working in the various disciplines associated with radio-
therapy in the UK about the present and intended
developments in Europe, Japan and North America and
how these are organized and funded. It is not intended to
be an account that proves that these new forms of radiation
therapy, which are being tested in phase I/II studies in an
increasing number of cancer sites, are better than the
techniques presently available in the UK; considerable
prospective research and development will be necessary to
show not only that this is unequivocally so, but also to
investigate the actual cost-benefits compared with other X-
ray based techniques such as intensity-modulated radio-
therapy (IMRT)/tomotherapy/conformal radiotherapy.
Other articles have been written elsewhere regarding this
increasingly important topic [2–4].
Under the leadership of Prof. Ugo Amaldi and Prof.
Andre Wambersie for Heavy Charged Particles in
Biology & Medicine (HCPBM) and European Network
for Light Ions (ENLIGHT), respectively, this combined
meeting – held over 4 days and attended by about 150
delegates – discussed the merits of various CP ranging
278
The British Journal of Radiology, 79 (2006), 278–284
Received 23 June 2005
Revised 21 July 2005
Accepted 21 July 2005
DOI: 10.1259/bjr/23066263
’ 2006 The British Institute of
Radiology
from protons to helium and carbon ions in clinical
practice. The potential for routine use of CP has
increased for a variety of reasons: impressive initial
clinical results; improved imaging that allows safe
application of the better dose deposition; industrial
production of turnkey centres that can be situated in a
large hospital, and also increasing critique of the
limitations of X-ray based radiotherapy based on three-
dimensional dose distributions made available by faster
computers. Although CP have been used for a many
years (50 years at Boston, USA) the treatments were
disadvantaged by several factors: relatively poor imag-
ing; location in what were essentially physics labora-
tories; limited beam energies and beam availability,
resulting in treatments of a very small range of cancers.
Japan, followed by Germany, has pioneered the use of
light ions in modern medicine, after some promising
results were initially obtained at Berkeley, USA and PSI in
Switzerland (initially with pions – or pi-mesons – using
more traditional radiation field arrangements and treat-
ment planning methods). The ions, in contrast to protons,
have substantially increased linear energy transfer (LET)
properties that cause additional biological effects, which
can counteract tumour intrinsic and hypoxic radio-
resistance; the latter properties were those from which
clinical advantages were predicted for neutrons, but
neutral particles cannot be deposited with selective
advantage as in the case of CP and, largely as a
consequence of this, they failed to improve the therapeutic
index [5, 6]. The initial claims of the efficacy of neutron
therapy at the phase I/II study level – and which were not
substantiated in subsequent phase III randomized studies
– cannot be compared fairly with the results presently
being claimed for CP therapy because tissue side effects
were classified in a rudimentary manner during the initial
neutron studies, whereas CP studies are presently
assessed with the most modern criteria.
The British Journal of Radiology, April 2006
Commentary: HCPBM and ENLIGHT meetings

Table 1. Nuclear physics, radiobiological and radiotherapy terms used in main text
Term Explanation

RBE The relative biological effect of different types of radiations;

applicable to normal tissue and tumours. Cobalt c-rays have
RBE51; higher values for other radiations indicate more biological
effect for the same dose exposure
X-ray High energy electromagnetic radiation (photons) that react with
electrons in matter to cause ionization
Conformal radiotherapy X-ray based radiotherapy which uses better shaped beams that match
tumour geometry
IMRT Intensity-modulated radiotherapy – X-ray based radiotherapy which
uses changes in beam intensity to match tumour and normal tissue
geometry
Hadronsa Collective term for nuclear particles that are not fundamental, i.e. they
are combinations of fundamental particles (known as leptons) held
together by strong nuclear forces
Neutron Neutral subatomic particle (a hadron); reacts with atomic nuclei to
cause dense ionization; attenuation in tissue is similar to
X-rays (i.e. there is no Bragg peak). RBE is increased
Electron A fundamental charged particle with very little mass such that the
complete Bragg peak effect does not occur in tissue. Clinically useful
for superficial cancer targets
Proton Positive charged nuclear particle which has mass sufficient to cause
Bragg peak effect, main interactions are with electrons, but some
nuclear collisions; small increase in RBE
Ions Charged atoms; these have greater mass than subatomic particles and
have more exaggerated Bragg peaks depending on their mass and
energy. Require higher energies than protons to reach deeper
tissues. They react with electrons and atomic nuclei and have
elevated RBE values as in case of neutrons
Mesons (pions) Subatomic particles that decay into leptons and c-rays
Cyclotron/synchrotron Complex equipment required to accelerate particles or ions to high
energies
BED Biological effective dose – a concept used to compare different
radiation schedules in terms of dose fractionation and duration of
exposure
a
Hadrons as a collective term in radiotherapy can be confusing since neutrons are included in this classification; the author
prefers the term charged particles.

The present international position
There is considerable interest in the formation of a
network of active CP Centres situated in as many
European countries as possible. Germany and Italy have
already commenced construction work; the French
Government decided in June 2005 to proceed with a
CP Centre in Lyon, while also encouraging a second
centre in Caen. The existing GANIL (Grand Accélérateur
National d’Ions Lourds) facility at Caen can be adapted
at relatively low cost for therapy and there is financial
support from the Conseil Régional de Basse-Normandie.
Austria is well advanced in its facility planning and
Sweden, as part of the Nordic Light Ion Project, will soon
finalize their options. These centres are separate from
existing projects to develop high-energy proton therapy,
which requires a lower energy facility than for light ions,
e.g. at Munich, Essen, Cologne and an expansion of the
service at Orsay, Paris. ENLIGHT is an ambitious 3-year
network project that reports back to the EC this year (see
estroweb.org); membership has included the scientifi-
cally advanced countries in Western Europe with the
rather enigmatic absence of the UK. CERN, the European
Organization for Nuclear Research based in Switzerland
and to which all advanced countries including the UK
contribute substantial sums of research monies, is also
highly supportive of the now attainable goal of provid-
ing CP treatment to a large number of patients so that the
promised benefits of Nuclear Research can be spread as
widely as possible.
The USA, by comparison, is extending only proton
therapy, but at a rapid rate with eight centres either in
operation or under construction; more contracts are
being considered. The throughput of proton treatments
is increasing by 40% per year at the relatively new
Massachusetts General Hospital facility. It seems likely
that full treatment capacity can be obtained within 3–
4 years of starting a new centre. Whether the USA will
invest in Centres that can deliver ions remains to be seen;
it appeared that there was some interest from the Boston
group who attended the meeting.
There are presently no plans to extend proton or ion
therapy in the UK beyond continuing the relatively low
energy proton work at CCO, Clatterbridge, where targets
are restricted to the eye.
Funding and organization of future CP therapy
The major problem appears to be reimbursement,
rather than the initial purchase costs of around J100–140
million. In Japan, one third of the cost is paid by the

The British Journal of Radiology, April 2006 279


Table 2. Acronyms used in text and websites for further information
Acronym Expansion
ENLIGHT European Network for Light Ions
HCPMB Heavy Charged Particles in Biology
& Medicine
NIRS National Institute Radiological Sciences,
Japan
HIMAC The NIRS funded Heavy Ion Medical
Accelerator, sited in Chiba, Japan
CERN Advanced internationally funded facility
for nuclear research; high profile
hadron collider project underway
PSI The Paul Scherrer Institute, Villigen,
Switzerland – centre for nuclear
research; proton therapy facilities with
isocentric gantry
GSI Gesellschaft für Schwerionenforschung
(the German Nuclear Research facility,
Darmstadt); the German pilot studies
using carbon ions performed at GSI
DKFZ/HICAT Refers to the German Cancer Research
Centre [Mannheim/Heidelberg], where
the new clinical ion facility is being built in
the Heidelberg University Hospital. This is
termed the HICAT (heavy ion cancer
therapy) Facility
ETOILE Espace de Traitment Oncologique par ions
legers dans le cadre European; to be the
first French ion beam centre
ASCLEPIOS Aire européenne de soins du cancer
par les protons et les ions. Situated in
Caen – there is an existing accelerator
MED-AUSTRON The Austrian facility based in Vienna
CPO Centre for Proton Therapy, Orsay
(south-west Paris)
CCO Clatterbridge Centre for Oncology, UK
PTCOG Particle therapy co-operative oncology group
(holds international meetings and
facilitates research)
patient, the remainder by health insurance companies,
while the facilities receive separate subsidies. The
Austrian state has agreed to fund treatments in a
privately owned but university managed facility; some
40% of the beam time – at night and weekends – would
be used to generate income from other research projects
in physics, astronomy and medicine in order to cover the
initial operating costs of approximately J18 million per
year (www.medaustron.at). The funding pattern differs
in Italy, where 90% of the funding will be from its state
health service, with plans to treat 3000 patients per year
at Pavia (see www.tera.it), where building has now
commenced. This large facility is to have advanced
electronic links to the existing nine large Italian Cancer
Centres.
The French Etoile and ASCLEPIOS projects are a
consequence of the French Plan against Cancer (2002).
Their national programme for research in hadrontherapy
(therapy using particle irradiation) has now officially
commenced: funding will be via their Ministry of Health
and Ministry for Research. The Royal College of
Radiologists have recently nominated a team of three
to initially represent the British health service in
280
B Jones
Further information (URL)
estroweb.org
www.tera.it
www.nirs.go.jp
www.nirs.go.jp/ENG/particl3.htm (for physics)
and www.nirs.go.jp/ENG/rd/ikagaku_e.html (for
medical applications)
cern.ch
www.psi.ch, then forschung/forschung_mensch_e.shtml
for health applications
www.gsi.de
www.dkfz.de and www.gsi.de
etoile.univ-lyon1.fr
www.asclepios.org
www.medaustron.at
www.protontherapie-orsay.fr
www.ccotrust.nhs.uk/patient/treatment/
cyclotron.asp
ptcog.web.psi.ch and ptcog.mgh.harvard.edu
collaboration with ASCLEPIOS: these are Professors R
G Dale, Dr I Rosenberg and the present author. Formal
agreements are presently being processed. It is envisaged
that a British team will need to be employed for selection
and care of patients treated in France. The UK govern-
ment will be obliged to fund treatments approved in
other EC countries that are not available on our islands.
Construction of a new German facility at The
University Hospital of Heidelberg is proceeding well
and randomized clinical trials (RCT) are being planned:
this is in marked contrast to neutron therapy where no
RCT were performed in a plethora of German depart-
ments, quite unlike the exemplary efforts made in the
UK under the auspices of the MRC.
There is widespread concern that the present
European work directives will adversely influence salary
costs and might limit staff recruitment. In the UK, the
state would have to reimburse the treatment costs under
present legislation. Somehow, one cannot see the present
system – with its multiplicity of health purchasers –
allowing this to occur efficiently. It is perhaps time for
the UK to re-think the arrangements for funding
radiation cancer services: a central (i.e. single) purchasing
The British Journal of Radiology, April 2006
Commentary: HCPBM and ENLIGHT meetings
system would solve at least some of the problems
encountered over the last 15 years.
Governments will need to decide not only the
reimbursement scales and the mixture of private and
public partnership, but also the ratio of proton to ion
facilities; the proportion of research and development;
how to ensure equity of access and consequently the
geographic spread of Centres per 10 million population.
It is feared that the newest privately managed proton
centres in the USA and Europe will be of most benefit to
rich patients with cancer, who may have relatively poor
indications for such sophisticated therapy. This is
potentially a serious political issue. In the UK, the
excellent progress made in development of Cancer
Networks would allow CP therapy to be mapped onto
these existing structures, even if patients may require
referral abroad for therapy.
Several epidemiological and cost effectiveness studies
– to estimate the likely demand for CP – were presented.
It is clear that the indications will extend to previously
unconsidered conditions such as lymphoma where dose
can be confined better to target structures in order to
reduce second malignancy, e.g. of the breast. Also, for
example, in cancer of the cervix, ions can be used to treat
a U-shaped nodal volume in the pelvis and the para-
aortic nodes without the delivery of substantial bowel
and kidney dose, which can be a problem in X-ray based
IMRT.
For costs, a detailed German study from the University
of Mannheim, in conjunction with Heidelberg, showed
that in 10 chordoma patients the overall management
was more cost effective by ions with respect to conven-
tional radiotherapy combined with multiple surgical
procedures (these are based on the available tumour
control rates of 74% at 4 years in Germany [7] and 80% in
5 years in Japan [8]), and the best published result for X-
ray based therapy (50% with stereotactic guided radio-
therapy) [9]. The permitted reimbursement for chordoma
treatment in Germany is J20 000, which is consistent
with charges of around J1000 per fraction proposed at
the new Heidelberg facility and for the Italian, French
and Austrian projects. This cost must be balanced against
the overall average cost of J81 470 for multiple surgical
procedures, mostly for recurrences. The break-even cost
point for carbon ions given in 20 fractions is reached if
the control rate is 70%, whereas for 16 fractions a control
rate of 65% would produce the same degree of economy.
Interestingly, the reimbursement costs for radiotherapy
appear to be lower in Europe than the USA, almost by an
order of magnitude.
A Belgian consortium has proposed a pan-European
solution based on Airbus Industries, where technical
specifications and standards would be agreed. This may
have been feasible a few years ago, but with up to seven
industrial companies capable of providing the necessary
equipment this seems unlikely – at least in the short term
– unless there is significant industrial collaboration or
consolidation.
The organization of the medical service for
hadrontherapy is quite advanced in Italy: the TERA
foundation team (Novara) consists of five radiation
oncologists and a secretary, who have dealt with
approximately 1000 requests per year resulting in 54
patients being accepted for therapy since 1997; this is
The British Journal of Radiology, April 2006
because referrals were limited mostly to base of skull
tumours at that time and also there were large numbers
with metastatic cancer. The UK will soon need to
establish some form of selection system and could learn
from this experience.
Education of existing staff is clearly a major problem;
few specialists in conventional radiotherapy are fully
aware of the potential of CP therapy. It is noteworthy
that Germany and Italy will soon be offering Masters
degree-courses in this subject for medical physicists;
Marie Curie Fellowships are also being promised.
Contracts are already operational at some European
centres for training given by experts from PSI,
Switzerland.
There are proposals for e-health related developments
with specific health networks for fast transfer of imaging
data sets and grid-computing for rapid analysis and
treatment planning. Progress is already being made
between some Italian cancer hospitals in this respect.
Experts at CERN consider CP departments to be an
excellent test bed for e-health care; a similar system was
proposed in the UK as part of the unsuccessful CASIM
project. Existing projects such as EGEE (egee-intranet.
web.cern.ch) and MIAS, developed for mammography in
the UK (www.wiau.man.ac.uk/services/MIAS) should
be considered for further development in cancer care.
The future scope for CP therapy is being actively
assessed using evidence based medicine methods. As
part of ETOILE, and linked to ENLIGHT, the following
steps are being followed: (1) definition of a rationale
based on physical and biological properties; (2) collegial
screening of potential indications; (3) assessment of
current therapy results using evidence based processes;
(4) incidence and prevalence studies; (5) a second round
of collegiate work as advisory committees to validate
previous steps, ascertain indications and estimate poten-
tial recruitment; (6) proposition of a set of clinical trials in
a European framework. Currently these multiple projects
dealing with multiple anatomical sites of interest
(cancers of the head and neck, thoracic, pelvic, central
nervous system (CNS), and sarcomas) have reached
stage 5.
New technologies
Work in CERN, China, and at the TERA foundation in
Italy explores new technologies for the construction of
smaller cyclotrons; the cyclinac concept, for example,
involves use of existing small cyclotrons coupled to a
linac booster with multiple klystrons capable of deliver-
ing carbon ions up to 400 MeV. Clatterbridge (UK) had
requested funding for a similar research project pre-
viously, only to be repeatedly turned down because of
cost – which could not be born by an NHS Trust – and, of
course, demands for evidence based medical proof.
Technical developments at Boston include the setting
up of a dedicated radiosurgery system, with novel
developments including miniature multileaf collimation,
beam steering and focusing. The use of PET/CT imaging
to verify dose distributions due to nuclear activation
after initial 1–2 Gy exposures is being investigated in
several centres worldwide. The use of robotic treatment
tables is also increasing, as well as more sophisticated
281

immobilization devices and methods for the correction of
organ movement, with continuous modifications of
tissue density with time, as in the lung, leading to more
realistic four-dimensional treatment planning. Various
Monte Carlo computational codes, containing funda-
mental physics interactions, are now capable of accurate
dose predictions. Better methods for beam monitoring,
dosimetry and the use of proton radiography show
considerable promise. These developments should
change the conduct of proton and ion therapy.
The topic of intensity-modulated proton therapy
(IMPT) – using the ‘‘spot scanning’’ technique – was
described in detail by Tony Lomax of PSI (Switzerland),
who has pioneered this technique in proton therapy [10,
11]. In ‘‘passive scattering’’ – the usual method for
creating a broad particle beam – all fields are homo-
geneous over the target volume, whereas in IMPT the
fields are individually non-homogeneous. IMPT is used
in patients where single optimized fields cannot provide
similar target coverage or critical organ sparing cannot
be achieved with single optimized fields. Patch fields are
often required; these have greater dose uncertainties in
the case of IMPT. The possibility of an enhanced risk of
delivery error and additional physical effects such as
secondary particle production exist when using IMPT.
They have used IMPT in only 43 of 209 patients between
1999 and 2004, sometimes as boosts or in the later part of
treatment. Considerable further research is required with
this powerful newly developed method for radiotherapy
delivery in the most difficult clinical situations to
determine if outcomes are significantly improved.
Radiobiology
One interesting development is that of micro-ion
beams, which can study selective radiation down to
individual cells or even to sub-cellular structures. The
bystander effect, where adjacent non-irradiated cells die,
saturates at a relatively low dose and so cannot dominate
the tumour cure probability achieved by radiotherapy,
particularly at high dose per fraction [12]. Nevertheless,
a fuller understanding of the mechanisms might usefully
inform new pathways or mechanisms that might be
exploited within normal tissues. Of particular relevance
to CP beams is that bystander effects have been found
with respect to malignant induction, so that the
possibility of reducing the yield of second malignancies
might enhance the potential benefits predicted for proton
radiotherapy.
The thorny question of how best to use relative
biological effect (RBE) within treatment planning arose
repeatedly in the meeting. The longstanding interna-
tional recommendation is to use constant weighting
factors to compensate for RBE; the dose used in
treatment is described as the cobalt equivalent Gray
and this factor should be varied with dose per fraction in
the case of ion beams; a similar correction although
theoretically required for protons will only produce
small differences in the equivalent dose. However, more
recent British modelling developments based on biolo-
gical effective dose have revised the method of dose
conversion [13, 14], with full continual allowance for
changes in RBE with dose per fraction, which the
282
B Jones
previous method did not fully allow. Some Centres, as
at Darmstadt/Heidelberg and CHIBA, Japan, have used
sophisticated methods to derive an RBE at various
depths of the beam. In Germany a/b ratios are also used
to determine the actual dose given to patients.
Calculations presented from the UK showed that the
cobalt equivalent Gray method may incorrectly estimate
the actual BED. Consequently, it was not surprising to
learn that the Japanese authorities have further increased
their single fraction lung dose from 28 Co-Gy Eq to
34 Co-Gy Eq; our estimated equivalent single fraction
doses for X-rays using the newer method was 18 Gy
instead of 28 Gy, hence the scope for dose escalation was
entirely predictable. There were several pleas for more
radiobiological funding in a clinically applied direction,
with scope for improved model based understanding
and outcome assessment.
Further presentations included alternative models for
the estimation of second cancer risk, e.g. using the
concept of organ equivalent dose with epidemiological
data from A-bomb exposures and the risk of second
malignancy developing in tissues after radiotherapy for
Hodgkin’s disease [15]. In the example of prostate
cancer, risks were compared with conventional confor-
mal radiotherapy: the use of IMRT appears to increase
the risk by 15%, and by 20% if the X-ray beam energy is
15 MeV, whereas spot scanned protons should reduce
the incidence by up to 50%.
The Japanese clinical experience
Carbon ion therapy is considered as ‘‘Highly
Advanced Medical Technology’’ in Japan, where over
2000 patients have now received carbon ion radiotherapy
within 40 different protocols. The dose distributions
appear better to those of protons, although there remains
some concern over the straggling effect beyond the Bragg
peak and enhanced auto activation when compared with
protons; this might result in more second malignancies
than with protons. Prof. Tsujii and others gave detailed
presentations of their experience in large pelvic and
other inoperable sarcomas, prostate cancer, in recurrent
rectal cancers and in T1 and T2 lung cancer. Their most
recent publication is from the previous ENLIGHT and
HCPBM meeting [8], where the specific patient numbers
treated until 2003 can be seen. The following summary is
of unpublished work; numbers of patients are small in
some categories as these are essentially phase I/II dose
escalation studies with more limited follow up of only a
few years in the highest dose categories. Interested
readers should seek out their future publications and
judge the results for themselves; it is possible that higher
rates of late toxicity might occur.
In rectal cancer recurrences, three dose levels have been
tested in 16 fractions: 67.2 Co-Gy Eq, 70.4 Co-Gy Eq and
73.6 Co-Gy Eq; the overall local control is 73% at 2 years
(surgery provides 10–20% control in various series). The
control rate at the highest dose level is so far 100% and side
effects are extremely low if Gore-Tex sheet displacement of
bowel is also used. For chordomas the 5 year survival is
80% with low morbidity. In large mucosal malignant
melanomas, concomitant carbon ion and chemotherapy
pilot schedules in small numbers of patients appear to be
The British Journal of Radiology, April 2006
Commentary: HCPBM and ENLIGHT meetings
giving better survival at 3 years than the use of both
treatments given sequentially.
Their impressive experience in liver cancer [8] was not
fully presented, although there was a statement that the
only serious side effect appears to be grade 3 acute liver
failure in 3% of cases – all of whom had pre-existing
cirrhosis – and which tends to slowly recover. Perhaps it
was thought that this was less of a problem in Europe,
although the prevalence of biliary and hepatic cancers is
increasing. It is clear that the results are extremely good
and that the incidence of serious complications is very
low in the pilot studies that have been completed since
1994. Urethral dose appears to be a problem in prostatic
ion-therapy, where uniform dose has been given to the
gland in 272 patients (stages T1–T3), the overall survival
being 79.8% at 7 years with grade 2 or above rectal
toxicity rate of only 0.4% and 2.85% for haematuria
originating from the urethra or bladder; because of the
latter, future protocols will use the ‘‘dose painting’’
technique – using spot scanning of targets – to reduce the
urethral dose in the same way as selective peripheral
loading of radioactive sources in prostate brachytherapy.
Recently, research has started on pancreatic cancers and
primary brain tumours in Japan. Techniques continue to
be refined and fractionation protocols reduced to levels
that cannot be used in the case of X-ray based therapy.
In lung cancer, 125 patients with 131 peripheral
primary lung cancers were treated with increasingly
hypofractionated carbon ions (see above) between 1999
and 2003 with 91.9% local control (100% for T1 stage and
78% for T2 stages) with crude survival rates of 54.7% and
46.1%, respectively. The 28 Co-Gy Eq single fraction
results (66.7% local control) have been improved to
90% with 32 Gy; so far 34 Gy equivalent single fractions
are without detectable side effects but these schedules
have only been used over the past 12–18 months, so
follow-up is short. The progressive changes in dose
fractionation have allowed them to construct dose
response curves so that any changes in technique can
be rapidly compared with the most recent cohorts.
There is a clear concept that a Japanese fisherman can
be rapidly screened for lung cancer, and flown at short
notice to the National Institute for Radiological Sciences
for treatment in a single day with high probability of
success and almost no side effects; this is not a viable
prospect for the UK population (e.g. for a Scottish
crofter), until perhaps the French CP Facilities are
complete, and even then our treatment slots may be
limited to a few hundred patients per year. In contrast, a
London stockbroker – if well informed – might negotiate
privately funded CP treatment in Europe or elsewhere.
Clearly, we face very difficult decisions in the UK.
The UK position
Our present commitments, and future dilemmas about
how best to improve research in radiation therapy, can
be summarized by the following points:
(1) Further development of micro-ion beam cell and
sub-cellular based studies at the Gray Laboratory
and the University of Surrey should be encouraged.
The British Journal of Radiology, April 2006
(2) As a result of our fast neutron experience we have
naturally been sceptical about cyclotron based
treatments, although this fear is to some extent
irrational because of the improved dose distribu-
tions with CP therapy. From the neutron experi-
ence, improved radiobiological models for high
LET radiotherapy have emerged, which should
further improve its safety and efficacy [13, 14].
(3) Our National Physical Laboratory already has
considerable expertise in proton dosimetry and
has worked closely with CCO (Clatterbridge).
(4) The CCO team has made several important
international contributions to proton therapy in
the eye [16].
(5) Preliminary UK collaboration has commenced with
ASCLEPIOS (Caen) and the ‘‘National French
Hadrontherapy Project’’. The first centre will be
operational in Lyon, probably in 2011.
(6) It remains to be seen whether ions will be better
than protons; although different results are claimed
for skull base tumours, the fractionation schedules
used are different: the least expensive short term
solution would be to purchase a UK high energy
proton therapy facility in the first instance; other
options include the use of ‘‘cyclolinac’’ technology
at Clatterbridge – by means of cooperation with
CERN – which would effectively upgrade that
facility.
(7) The UK Government needs to decide whether it
wishes our NHS and Research Councils to be ‘‘at
the heart of Europe’’ with respect to cancer
therapy, or provide a peripheral supportive role.
The additional question of an ion beam facility
might form part of new funding reforms proposed
within the EU as a partnership in the fight against
cancer; at least some of the resources used to
bolster the common agricultural policy (CAP)
could be used to provide the best obtainable
treatment for cancer. Perhaps the political maxim
should be more CP and less CAP.
(8) The clinical physics and oncology community in
the UK needs to speak with a stronger collective
voice. Previous prescient bids for CP facilities have
been unsuccessful for a variety of reasons [4] and
there is a risk that further pusillanimous and
parsimonious reactions to newer bids would
prejudice the already low morale and technological
capability of British radiotherapy. A very positive
response is essential if we are to compete with the
‘‘best available’’ elsewhere. There is a danger that
we shall remain obsessed with ‘‘evidence based
medicine’’ and ‘‘cost-effectiveness’’ and yet not be
capable of making contributions towards improved
care because we will simply be unable to do the
necessary research.
(9) British politicians and those that administer
research funding [17] may continue not to invest
in expensive radiation facilities, with possible
deleterious consequences for British patients [18].
There will certainly be other research projects in
cancer medicine that will demand resources with
good justification, although there are presently few
examples of benefits from large investments
already made in so many molecular based projects.
283

(10) Current plans to increase research in radiotherapy
and radiobiology should be coupled with therapy
using CP; molecular biology applications will
inevitably be helpful alongside such therapy and
should also increase the diagnostic yield of appro-
priately small cancers that might be safely treated
by this approach [2–4, 17].
Conclusions
Just as we seek better early diagnosis of cancer, so we
should also seek a safe alternative to radical cancer
surgery through the development of safer radiation
techniques; the fear of cancer will then be much
alleviated by the benefits promised by nuclear physics.
At a time like this, with the availability of such advanced
medical technology, priority should be given to research
and development aimed to find the best curative cancer
therapies along these lines. As an absolute minimum, the
UK needs a large high energy proton therapy centre as
soon as possible and to continue its present cooperation
with France in ion beams until longer follow up data is
available in larger numbers of patients. There must be no
complacency about this matter – short term expediency
and prolonged prevarication may prove to be politically
embarrassing in the long term, with additional costs and
consequences to British cancer patients. The UK is
capable of making a substantial contribution to this area
of medicine, providing a positive decision is made quite
soon.
Acknowledgments and declarations
The author was funded to attend this meeting entirely
by ASCLEPIOS & Conseil Regional de Basse Normandie
(France) and is a member of the UK EPSRC Medical
Applications of Ion Beams Network and a Trustee of The
Cyclotron Trust (UK). The author regrets not being able
to mention the names of so many excellent speakers.
References
1. Suit H, Goldberg S, Niemerko A, Trofimov A, Adams J,
et al. Proton beams to replace photon beams in radical dose
treatments. Acta Oncologica 2003;42:800–8.
284
B Jones
2. Jones B, Burnett NG. The future for radiotherapy: protons
and ions hold much promise. Br Med J 2005;330:979–80.
3. Jones B, Rosenberg I. Particle Therapy Co-operative
Oncology Group (PTCOG 40) Meeting, Institute Curie
2004. Br J Radiol 2005;78:99–102.
4. Jones B. The case for particle therapy. Br J Radiol 2006;79:
24–31.
5. Errington RD, Ashby D, Gore SM, et al. High energy
neutron treatment for pelvic cancers: study stopped
because of increased mortality. Br Med J 1991;302:1045–51.
6. Maor MH, Errington RD, Caplan RJ, et al. Fast neutron
therapy in advanced head & neck cancer: a collaborative
international randomised trial. Int J Radiat Oncol Biol Phys
1995;32:99–604.
7. Schulz-Ertner D, Nikoghosyan A, Thilmann C, Haberer T,
Jakel O, Karger C, et al. Results of carbon ion radiotherapy
in 152 patients. Int J Radiat Oncol Biol Phys 2004;58:631–40.
8. Tsujii H, Mizoe J, Kamada T, Baba M, Kato S, Kato H, et al.
Overview of clinical experiences on carbon ion radio-
therapy at NIRS. Radiother Oncol 2004;73 (Suppl. 2):S41–9.
9. Debus J, Schulz-Ertner D, Schad L, Essig M, Rhein B,
Thillmann CO, al. Stereotactic fractionated radiotherapy for
chordomas and chondrosarcomas. Int J Radiat Oncol Biol
Phys 2000;47:591–6.
10. Lomax AJ, Cella L, Weber D, Kurtz JM, Miralbell R.
Potential role of intensity-modulated photons and protons
in the treatment of the breast and regional nodes. Int J
Radiat Oncol Biol Phys 2003;55:785–92.
11. Lomax AJ, Goitein M, Adams J. Intensity modulation in
radiotherapy: photons versus protons in the paranasal
sinus. Radiother Oncol 2003;66:11–8.
12. Prise KM, Schettino G, Folkard M, Held KD. New insights on
cell death from radiation exposure. Lancet Oncol 2005;6:520–8.
13. Dale RG, Jones B. The assessment of RBE effects using the
concept of Biologically Effective Dose. Int J Radiat Oncol
Biol Phys 1998;43:639–45.
14. Jones B, Dale RG. Estimation of optimum dose per fraction
for high LET radiations: implications for proton radio-
therapy. Int J Radiat Oncol Biol Phys 2000;48:1549–57.
15. Schneider U, Zwahlen D, Ross D, Kaser-Hotz B. Estimation
of radiation cancer from 3-D dose distributions: concept of
organ equivalent dose. Int J Radiat Oncol Biol Phys
2005;61:1510–5.
16. Damato B, Lecuona K. Conservation of eyes with choroidal
melanoma by a multimodality approach to treatment: an
audit of 1632 patients. Opthalmology 2004;111:977–83.
17. Jones B. Science and Innovation Summit 2004. Br J Radiol
2005;78:17–9.
18. Jones B, Price P, Burnet NG, Roberts JT. Modelling the
expected increase in demand for particle radiotherapy:
implications for the UK. Br J Radiol 2005;78:832–5.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 285–294

REVIEW ARTICLE

What are the risks from medical X-rays and other low dose
radiation?
B F WALL, BSc, G M KENDALL, PhD, A A EDWARDS, MSc, S BOUFFLER, PhD, C R MUIRHEAD, PhD and
J R MEARA, FFPH

Health Protection Agency, Radiation Protection Division, Centre for Radiation, Chemical and
Environmental Hazards, Chilton, Didcot, Oxon. OX11 0RQ, UK

ABSTRACT. The magnitude of the risks from low doses of radiation is one of the central
questions in radiological protection. It is particularly relevant when discussing the
justification and optimization of diagnostic medical exposures. Medical X-rays can
undoubtedly confer substantial benefits in the healthcare of patients, but not without
exposing them to effective doses ranging from a few microsieverts to a few tens of
millisieverts. Do we have any evidence that these levels of exposure result in significant
health risks to patients? The current consensus held by national and international
radiological protection organizations is that, for these comparatively low doses, the
most appropriate risk model is one in which the risk of radiation-induced cancer and
hereditary disease is assumed to increase linearly with increasing radiation dose, with
no threshold (the so-called linear no threshold (LNT) model). However, the LNT
hypothesis has been challenged both by those who believe that low doses of radiation
are more damaging than the hypothesis predicts and by those who believe that they
Received 29 April 2005
are less harmful, and possibly even beneficial (often referred to as hormesis). This article Revised 30 September
reviews the evidence for and against both the LNT hypothesis and hormesis, and 2005
explains why the general scientific consensus is currently in favour of the LNT model as Accepted 11 October 2005
the most appropriate dose–response relationship for radiation protection purposes at
DOI: 10.1259/bjr/55733882
low doses. Finally, the impact of the LNT model on the assessment of the risks from
medical X-rays and how this affects the justification and optimization of such exposures ’ 2006 The British Institute of
is discussed. Radiology

What evidence do we have that low or moderate doses
of radiation, such as those involved in diagnostic
imaging, are harmful? Some medical X-rays, for example
extremity, chest and dental radiographs, involve effec-
tive doses of only a few microsieverts. However, organ
doses and effective doses can be tens of millisieverts for
extensive fluoroscopic or CT examinations [1] and can
easily rise to 100 mSv or more when such examinations
are repeated through an episode of disease or trauma. Do
these levels of exposure result in significant health risks?
The most direct evidence on radiation risks comes
from epidemiological studies of increased levels of
cancer in exposed human populations. However, these
epidemiological studies inevitably suffer from problems
of insufficient statistical power at low doses. When these
limitations are fully recognized, epidemiological studies
are generally unable to provide clear evidence of the
effects of protracted low doses of radiation of less than
about 50–100 mSv. Judgements about extrapolation to
lower doses are made in the light of information from
Current address for G M Kendall: Childhood Cancer Research
Group, University of Oxford, 57 Woodstock Road, Oxford OX2 6HJ,
UK.
cellular studies and animal experiments that provide
radiobiological insights into the basic underlying
mechanisms of radiation interaction with living cells
and organisms. Radiobiological studies also have some
significant limitations, such as the fact that the radiation-
induced biological endpoints observed in cells or
laboratory animals are not always reproducible and that
they are not necessarily directly indicative of radiation-
induced carcinogenesis in humans.
Radiation risks are reviewed by international and
national organizations, such as the International
Commission on Radiological Protection (ICRP), the
United Nations Scientific Committee on the Effects of
Atomic Radiation (UNSCEAR), the UK’s Radiation
Protection Division of the Health Protection Agency
(formerly the National Radiological Protection Board,
NRPB) and the National Council on Radiation Protection
and Measurement (NCRP) in the USA. It is an important
function of these bodies to continually assess and review
publications from all over the world on the effects of
exposure to ionizing radiation on human health and to
reach a balanced view of the risks involved. The current
consensus of these bodies is that for radiation protection
purposes the most appropriate risk model at low doses is

The British Journal of Radiology, April 2006 285


one in which the risk of radiation-induced cancer and
hereditary disease is assumed to increase with increasing
radiation dose, with no threshold. Any increment of
exposure above natural background levels will produce
a linear increment of risk (the so-called linear no
threshold (LNT) model).
The LNT hypothesis is not regarded as immutable law,
proven in every circumstance, but rather as a robust
working rule. However, the LNT hypothesis has been
attacked both by those who believe that low doses of
radiation are more damaging than the hypothesis
predicts and by those who believe that they are less
harmful, and possibly even beneficial (this latter hypo-
thesis is often referred to as hormesis). This paper
reviews the evidence for and against the LNT hypo-
thesis, taking particular note of the debates at the UK
Radiological Congress in 2004 in which the case for
hormesis was strongly argued [2–4]. The LNT hypothesis
was, of course, also ably defended in these debates [5, 6].
Finally, we discuss the impact of the LNT model on
the assessment of the risks from medical X-rays and how
this affects the justification and control of medical
exposures to ensure adequate protection of patients
without denying them the undoubted clinical benefits
that modern diagnostic radiology has to offer.
National and international assessment of the
risks at low doses
Reviews of the available information on the effects of
low dose radiation are undertaken at regular intervals by
national and international organizations. Recent reviews
include those conducted by NRPB in 1995 [7], UNSCEAR
in 2000 [8] and NCRP in 2001 [9]. The evidence
considered by these bodies comprises epidemiological
studies of human populations, and radiobiological
studies involving laboratory experiments with animals
and living cells.
Epidemiological studies
Epidemiological studies of human populations pro-
vide the most direct and easily interpretable evidence.
Some of the major studies are summarized in Table 1,
which is based on data in the UNSCEAR 2000 Report [8].
The epidemiological studies summarized here are of two
designs: cohort and case/control. A cohort study
involves a large group of individuals who have been
exposed to different levels of the agent of interest (here,
radiation); the investigators see whether disease levels
correlate with the different exposures. A case/control
study on the other hand involves a group who have
developed the disease being studied and a group of
matched controls who did not; the investigators then
retrospectively examine whether the cases had been
exposed to higher levels of the agent under study than
the controls.
Perhaps the most important of the epidemiological
studies is the Life Span Study of the Japanese atomic
bomb survivors in Hiroshima and Nagasaki. This study
has high statistical power because it has followed, for
over 50 years, a large cohort of survivors of all ages and
286
B F Wall, G M Kendall. A A Edwards et al
both sexes, exposed to a wide range of reasonably well
established doses. It is significant that important infor-
mation also comes from a number of studies of medical
exposures, particularly those of children or pregnant
mothers, where the low natural prevalence of childhood
cancers improved the ability to detect a small increase in
cancer rates following paediatric or in utero exposures.
Studies of occupationally exposed nuclear workers are
also beginning to provide information [10, 11]. All of the
above reviews [7–9] came to essentially the same
conclusions regarding the scientific evidence for the
effects of radiation at low doses. They agreed that human
epidemiological data demonstrate significantly increased
cancer risks from radiation doses above about 100 mGy.
Moreover, all the reviews were satisfied that there was
good evidence for a significant increase in childhood
cancers following fetal doses of 10–20 mGy.
A recent paper by a large group of eminent epide-
miologists and radiation scientists from both sides of the
Atlantic [12] carefully assessed the latest data from the
most recent epidemiological studies including some that
were not available to the review bodies listed above.
Brenner et al [12] concluded that there is now good
epidemiological evidence for an increased cancer risk in
humans for acute doses of X-rays down to about 10–
50 mSv and down to 50–100 mSv for protracted expo-
sures. In particular, a detailed analysis of the many
studies of childhood cancer risk after fetal exposure from
diagnostic X-rays [13] demonstrated that a dose of
around 10 mSv to the fetus does cause a statistically
significant and quantifiable increase in the risk of
childhood cancer.
Extrapolation to low doses and dose rates
Despite the fact that epidemiological studies provide
direct evidence of the effects of radiation exposure on
human populations, there are problems in interpreting
the evidence that they provide in the context of radiation
protection in general. One of the most important of these
problems arises because many of the studies involved
populations which received relatively high radiation
exposures in a relatively short time. At high doses and
dose rates, there is evidence that the effects of radiation
exposure are proportionately greater than at the low
doses and dose rates which are usually more relevant in
radiological protection. This reduction in effects at low
doses and dose rates is quantified using a Dose and Dose
Rate Effectiveness Factor, DDREF. Animal studies
suggest a DDREF in the range 2–10 [8]. The limited
human data and cellular studies support values at the
lower end of this range. NRPB [7], UNSCEAR [8] and
NCRP [9] all suggest a DDREF of two as a reasonable
judgement to be used for cancer risks for radiological
protection purposes. Consequently the quantitative
estimates of the risks per unit dose derived from the
epidemiological studies at high doses and high dose
rates are halved in order to estimate the probabilities of
radiation-induced cancer following diagnostic medical
exposures.
While future epidemiological studies in humans will
remain of great importance for quantitative risk assess-
ment, it is accepted that they are unlikely to have the
The British Journal of Radiology, April 2006
The British Journal of Radiology, April 2006 Table 1. Epidemiological studies of the effects of exposures to external low-LET radiation

Review article: What are the risks of low dose radiation?

Study Type of study Follow up, years (mean) Type of exposure Cancers studied

External high dose rate exposures

Exposure to atomic bombings
Life Span Study [53] Cohort mortality 50113 5–45 (32.5) Gamma and neutron radiation various*
persons . 5 mSv Japan from nuclear explosions
Life Span Study [54, 55] Cohort incidence 37270 13–42 (24.4) Gamma and neutron radiation various*
persons .10 mSv Japan from nuclear explosions
Treatment of malignant disease
Cervical cancer cohort [56] Cohort incidence 82616 0–.30 (7.0) Radiotherapy various*
exposed women 8 countries
Leukaemia following cancer Case-control 218 cases 775 1–50 Radiotherapy Leukaemia*
of the uterine corpus [57] controls 5 countries
Lung cancer following Case-control 98 cases 259 1–.10 Radiotherapy Lung cancer
Hodgkin’s disease controls 7 countries
(international) [58]
Childhood cancers Case-control within 5–48 (5.5) Adjuvant radiotherapy Thyroid*, leukaemia, bone
(international) [59–61] 9170-member cohort 6 sarcoma*
countries
Retinoblastoma [62] Cohort incidence 962 exposed 1–.60 (median 20) Radiotherapy Various*
persons 642 unexposed
persons USA
Treatment of benign disease
Childhood skin haemangioma: Cohort incidence/mortality 1–67 (39) Radiotherapy Thyroid*, breast*, leukaemia,
Stockholm [63–66] 14351 exposed persons all other sites
Sweden
Childhood skin haemangioma: Cohort incidence 11914 exposed 0–69 (31.1) Radiotherapy Various*
Gothenburg [67, 68] persons Sweden
Ankylosing spondylitis [69, 70] Cohort mortality 13914 exposed 1–57 (17.6) X-ray therapy Leukaemia*, other neoplasms*
persons UK (except colon)
Israel tinea capitis [71–74] Cohort incidence/mortality 26–38 (25.3) X-ray induced epilation Various*
10834 exposed persons Israel
New York tinea capitis [75] Cohort incidence 2226 exposed 20–39 (25.4) X-ray induced epilation Various*
persons USA
New York acute post-partum Cohort incidence 571 exposed women 20–35 (25.1) X-ray therapy Breast*
mastitis [76] USA
Rochester thymic irradiation Cohort incidence 2652 exposed 23–.50 (29.5) X-ray therapy Thyroid*, breast*, skin
[77–79] persons USA
Metropathia haemorrhagica [80] Cohort mortality 2067 exposed women >5–.30 (,26) X-ray therapy Various*
UK
226
Benign gynaecological disease Cohort mortality 4153 exposed women 0–60 (26.5) Intrauterine Ra various*
[81, 82] USA
Massachusetts TB fluoroscopy Cohort incidence 2367 exposed women 0–.50 (11.4) Multiple X-ray chest fluoroscopies Breast*, skin
[79, 83] USA
287

Canadian TB fluoroscopy [84, 85] Cohort mortality 25007 exposed 0–57 (30) Multiple X-ray chest fluoroscopies Lung, breast*
persons Canada

(Continued )
 B F Wall, G M Kendall. A A Edwards et al

*Sites for which statistically significant excesses were reported in the exposed group (cohort studies), or for which a higher proportion of the cases were exposed to radiation (case/
statistical power to provide direct evidence on radiation
effects in humans for doses much below 10–50 mGy.

Maternal X-rays during pregnancy Leukaemia*, all solid tumours*

This is because of the difficulty of observing a small

Leukaemia, all other cancers

Leukaemia, all other cancers

Maternal X-rays during pregnancy Leukaemia*, solid tumours
number of additional cancers against very high back-
ground incidence rates; ever larger study populations
would be required to detect ever smaller effects as doses
decrease, and this is simply impractical. Moreover,
epidemiology is an observational and not an experi-
Cancers studied

mental science. Epidemiologists make strenuous efforts

to optimize study design, but it is not possible to select
the exposed and reference populations on strict statis-
tical grounds and there are likely to be residual effects of
confounding factors and possible biases (e.g. selection
bias), as well as other practical problems such as
uncertainties in the dose estimates. UNSCEAR, in
Exposures in nuclear industry

Annex I of its 2000 report [8], gives a discussion of the

Exposures mainly in nuclear

potential problems of bias and confounding, particularly

in the low dose region where attempts are being made to
resolve very small effects.
The low dose region, where epidemiology is unable to
Type of exposure

produce clear evidence of risk, provides a fertile area for

those who wish to argue that radiation risks have been
industry

overestimated or underestimated. The former can con-

clude, quite correctly, that there is ‘‘no significant
evidence for an effect’’; this must, however, not be
confused with there being ‘‘significant evidence for no
effect’’. Furthermore, as discussed later, selection effects
Table 1 Epidemiological studies of the effects of exposures to external low-LET radiation (Cont.)

Follow up, years (mean)

in epidemiological studies, in particular the ‘‘Healthy

Worker Effect’’ must be allowed for. Those who believe
that radiation risks are greater than the LNT extrapola-
Up to 43 (22.2)

Up to 47 (16.5)

tion suggests can point to selected studies where the

play of chance has resulted in apparently elevated risks
16 (max)

20 (max)

at low doses, while ignoring studies which contradict

this view. It is also true that publication bias will result
in studies with significant findings reaching the litera-
ture more readily than those which are inconclusive. All
this assumes that studies are well designed. Badly
designed epidemiological studies, for example, with
serious bias in the selection of the study populations, are
available in the ‘‘grey’’ literature without peer review
Cohort mortality 96673 workers
Case-control 14491 cases 14491

Case-control 1342 cases 14292

and are cited as evidence that radiation risks are

underestimated or overestimated [14].
monitored workers UK
Cohort mortality 124743
Canada, UK, USA

Radiobiological studies
controls USA
controls UK

The national and international review bodies men-

Type of study

tioned above also agreed that increased understanding

of biological mechanisms will increasingly underpin
judgements about the shape of the dose–response
relationship in the low dose region. This increased
Low-dose or low-dose-rate exposures

understanding will come from qualitative and quantita-

tive data from cellular and molecular studies of the
National Registry for Radiation
NE USA childhood cancers [88]

biological mechanisms underlying the health effects of

Oxford Survey of Childhood

radiation. A comprehensive review of recent develop-

ments in this area has been undertaken by a Task Group
After UNSCEAR (2000).
Occupational exposure

of the International Commission on Radiological

Nuclear workers [10]

Workers, UK [11]
Pre-natal exposure

Protection [15]. Currently these studies indicate that

control studies).
Cancers [86, 87]

the carcinogenic effects of radiation are caused largely

by double strand breaks and complex lesions in stem cell
DNA. Mechanistic modelling of radiation-induced car-
cinogenesis based on these radiobiological studies is still
Study

at an early stage of development. However, data on the

role of gene mutations and DNA damage and repair

288 The British Journal of Radiology, April 2006

Review article: What are the risks of low dose radiation?
mechanisms are now sufficiently well established to
support the thesis that the risk of radiation-induced
cancer at low doses rises as a simple function of dose
without threshold for most types of cancer. While some
experiments suggest a curve that is concave downwards
and others one that is concave upwards, depending on
the biological endpoint, there are sound biophysical
arguments supporting the LNT model as the most
appropriate general model for cancer induction. It
should be appreciated that this may not be the most
conservative approach and it might result in an under-
estimation of some radiation-induced cancer risks and an
overestimation of others. Those involved in protection
should also be alert to the fact that truly low dose
experiments are difficult in cellular systems. Also, it
remains the case that relatively few studies address
directly the effects of low doses. Until the uncertainties
are resolved, all the reviews concluded that the current
weight of evidence on fundamental cellular processes
supports the view that an increase in risk proportionate
to the radiation dose is the most scientifically defensible
approximation of the low-dose response.
Could radiation stimulate beneficial adaptive
responses and hormesis?
It has been suggested [16] that low or moderate doses
of radiation might stimulate responses, for example to
DNA repair processes, which might counteract the
harmful effects of the radiation damage. At its most
extreme it has been suggested that these effects are so
great as to confer a net benefit, at least in certain dose
ranges. The general hypothesis of radiation stimulated
beneficial changes is known as the Adaptive Response
and the idea of net benefit is called Hormesis. The
Adaptive Response was considered by the United
Nations Scientific Committee on the Effects of Atomic
Radiation (UNSCEAR) in its 1994 report [17] and by
NCRP [9]. Explanations for an Adaptive Response tend
to involve stimulation of DNA repair processes, although
other mechanisms have also been suggested. However,
the arguments put forward for adaptive responses and
hormesis claim to have supporting evidence from
epidemiological, ecological and radiobiological studies.
Epidemiological studies (the Healthy Worker Effect
and unintentional bias)
Various epidemiological studies have been conducted
of populations exposed to low and/or protracted doses.
In its 2000 report UNSCEAR [8] has undertaken a
comprehensive review of these studies. Here we will
focus on studies of people exposed to radiation either in
the workplace or the home, because such studies have
often been quoted to support claims of hormesis.
The essence of the epidemiological method is to
compare disease rates in two or more populations which
differ qualitatively or quantitatively in their exposure to
the agent under investigation. However, a common
complicating factor is the existence of selection effects,
typically factors which mean that those entering the
exposed group tend to be more healthy than those in the
The British Journal of Radiology, April 2006
comparison group. It is a very common finding in
epidemiological studies of working populations that
death rates are lower than in the general population.
Consequently the Standardized Mortality Ratio (SMR) –
the ratio (expressed as a percentage) of the number of
deaths in the exposed group relative to the number
expected based on rates for the general population – is
often below 100. This is known as the Healthy Worker
Effect [18, 19] and it is generally accepted that this is a
reflection of the selection of the fit and healthy into
employment and their retention in work. It would be a
mistake to interpret low SMRs in such epidemiological
studies as evidence for hormesis.
For example, recent issues of an epidemiological
journal described three studies of working populations.
More examples could be cited, but these are sufficient to
make the point. The cohorts were, respectively, workers
exposed to ethylene oxide [20], workers in the petroleum
industry [21] and workers exposed to formaldehyde in
the garment industry [22]. The SMRs were:
N Ethylene oxide exposed workers SMR 90 (95% CI 88–
93)
N Workers in the Petroleum Industry SMR 68 (95% CI
63–73)
N Garment workers exposed to formaldehyde SMR 92
(85% CI 88–96).
All three populations thus have significantly lower
mortality than the general population. Are we to
conclude that exposure to ethylene oxide, to hydrocar-
bons and to textile fibres or dust (in the presence of
formaldehyde) are all beneficial to human health? The
investigators who conducted these three studies did not
draw such a conclusion, nor would epidemiologists
generally. It is clear that what is being seen is the effect of
selection factors which mean that working populations
are healthier than the population as a whole, which
includes the chronically sick and unemployed. It would
be perverse to try to impose a different explanation in the
case of exposures to radiation.
One of the studies sometimes cited as demonstrating a
beneficial effect of radiation exposure is that of mortality
amongst British radiologists by Berrington et al [23].
Berrington et al observed significantly lower SMRs for
radiologists employed after 1920, when occupational
doses had fallen to moderately low levels, compared
with other contemporary (unexposed) doctors. It has
been suggested that this demonstrates a highly signifi-
cant beneficial effect of radiation [24]. However, great
care has to taken when attributing SMRs for different
groups of people (even if they are all doctors) to a
particular cause like radiation exposure, when no
account can be taken of other possible confounding
factors, like smoking. There is evidence in a study of
mortality among doctors by Doll and Peto [25] that,
during the 1950s and 1960s, GPs smoked more than other
doctors and radiologists smoked less. This alone could be
the cause of the lower SMRs for radiologists compared
with other doctors (a large proportion of whom would be
GPs), but without specific smoking habit information for
the cohort of British radiologists and the control group of
other doctors, it is impossible to be sure. It is because of
these problems with external comparisons with other
289

populations that epidemiologists generally pay more
regard to comparisons or trends in mortality (or disease
incidence) within cohorts as evidence for radiation risks.
Berrington et al focused on internal trends in mortality
risks with time since entry into the radiology profession
(as an indicator of cumulative dose), rather than relying
solely on external comparisons of SMRs with other
groups of doctors or the general population. Thus they
found evidence for an increasing trend in risk of cancer
mortality with time since first registration with a
radiological society; for example, such that for those
registered for more than 40 years there was a 41% excess
risk. This was mostly due to those who registered
between 1921 and 1954 when exposures were higher
than in more recent years.
As well as the problem of confounding, Brenner and
Hall [26, 27] point out that a similar but much larger and
more detailed study of male North American radiolo-
gists [28] showed a higher SMR for the radiologists
compared with other doctors. Moreover, a study in the
UK of the mortality of 20 000 NHS consultants employed
between 1962 and 1979 [29] showed that SMRs for
radiologists and radiotherapists were not significantly
different from those for all consultants taken together,
either for all causes of death or for cancer. Brenner and
Hall suggest that such inconsistent results are entirely to
be expected when the doses are so low (a few mSv or less
per year after the 1950s), since the radiation effects are
likely to be below the limit of detectability for epide-
miological methods. In such a situation, most studies
would be expected to show no statistically significant
effects but there will be occasional ones showing slightly
positive or negative results. Such results provide no
evidence for health effects one way or the other, they
merely rule out large risks or large benefits.
A second study which has been cited [2] as supporting
the hypothesis that a moderate dose rate of radiation is
beneficial to health is that by Matanoski et al [30] of
workers at US nuclear shipyards. This is on the basis of
significantly lower SMRs seen in those workers with
cumulative effective doses greater than 5 mSv than in
those with lower doses, and in the latter compared with
non-radiation shipyard workers. However, the authors
of the study do not suggest that these results provide
evidence for a beneficial effect of radiation, but instead
regard selection bias as a more likely cause. Those
selected to work on nuclear powered ships were given a
physical examination prior to assignment and so are
likely to be healthier than those working elsewhere in the
shipyard. In addition, those with cumulative doses
.5 mSv, mainly through being employed for longer,
are further self-selected for enduring good health.
Matanoski et al go on to point out that although lower
SMRs are seen for all causes of death and for lung cancer,
they are not seen for leukaemia where the SMR is 2.17
times higher for the radiation workers with doses
.5 mSv than for those with doses ,5 mSv.
Surprisingly, only an abstract of this study has appeared
in the peer reviewed literature [31], but in it the authors
recommended an extension of the study population and
the application of more powerful methods of analysis.
Report 136 of the US National Council on Radiation
Protection and Measurements (NCRP) [9], which pro-
vides a detailed evaluation of the LNT model, dismisses
290
B F Wall, G M Kendall. A A Edwards et al
the suggestion that the nuclear shipyard worker study
provides support for the beneficial effects of radiation.
There have been various other studies of nuclear
industry workers, reviewed by UNSCEAR [8], where the
interpretation of low dose risks is not always clear, for
the reasons described previously, but which do not
provide any strong support for hormesis. Indeed, such
studies provide some evidence of raised leukaemia risks
associated with radiation exposure [10, 11]. In the UK’s
National Registry for Radiation Workers [11, 32], radia-
tion workers had lower SMRs than the general popula-
tion, not because of any beneficial effect of their
exposures, but simply due to the Healthy Worker Effect.
Another study that has been cited as demonstrating
radiation hormesis, claimed to show very low cancer
rates for residents in Taiwanese office blocks built with
Co60 contaminated steel [33]. However, it now seems
overwhelmingly likely that this study was seriously
flawed and grossly underestimated the true number of
cancers in the study population [34].
Ecological studies (confounding factors)
Reports of a strong negative correlation between mean
natural background radiation levels and cancer mortality
in different states of the USA [35] and between mean
radon levels and mean lung cancer rates in different US
counties [36, 37] have also been cited as further evidence
for radiation hormesis [2]. However, it is well-recognized
that ecological studies based on aggregated data for large
geographical areas, rather than on the individual data
used in case-control or cohort studies, have the potential
for serious statistical problems [9]. The author of the first
study [35] agreed that confounding factors such as
smoking, poverty or environmental pollution could be
affecting the mortality rates rather than background
radiation levels [38].
However, the exceptional strength of the negative
correlations between lung cancer and radon levels in the
ecological studies reported by Cohen [36, 37] appears to
contradict the LNT model. The fact that smoking is
responsible for a large majority of lung cancers
suggested that different smoking habits between the
populations in the US counties might be influencing the
results, but no direct data on smoking were available.
Indirect evidence finally emerged when Puskin exam-
ined correlations between radon levels and a variety of
cancers other than lung cancer [39]. Some of these other
cancers are related to smoking and some are definitely
not. Radon, of course, gives virtually all its dose to the
lung so any effect of radiation, beneficial or harmful,
could not be expected in cancers of any other organs. For
the other smoking-related cancers, Puskin also found
negative correlations with radon levels but there was no
association between radon and those cancers that are not
linked to smoking. He concluded that the negative
correlation seen between radon levels and lung cancer
in the earlier studies is largely a consequence of a
negative correlation between smoking and radon levels
across the US counties. Thus the results could be
explained in terms of confounding by smoking without
invoking any kind of beneficial effect of low level
radiation exposure.
The British Journal of Radiology, April 2006
Review article: What are the risks of low dose radiation?
Moreover, there have been numerous more recent
analyses of cohort and case-control studies of residential
radon concentrations and lung cancer incidence through-
out the world, where smoking habit information was
also available [40]. Most of them observe an excess lung
cancer risk from residential radon, though the risks may
not always achieve statistical significance. However,
recent pooled analyses of case-control studies of radon
and lung cancer from Europe [41] and from North
America [42] have much more statistical power than the
individual studies, and provided unequivocal evidence
of the risks of domestic exposure to radon. In particular
the European pooling demonstrated a risk of exposure
to radon in homes down to concentrations less than
200 Bq m23 which is equivalent to an effective dose of
about 5 mSv using the dose conversion convention
recommended by ICRP [43].
Radiobiological studies (differences between
metabolic and radiation-induced DNA damage and
repair)
There is general agreement that DNA lesions are
continuously being produced in the body, for example
by reactive oxygen species (ROS) generated by normal
oxidative metabolism. The vast majority of these lesions
are repaired by normal cellular processes. The total
number of these endogenous DNA lesions exceeds the
number produced by normal background radiation
levels by several orders of magnitude.
Proponents of hormesis suggest that acute low doses
of radiation induce a temporary protective response
against DNA damage that could counteract the ever-
present endogenous DNA damage from the ROS. Some
cellular and animal experiments suggest that this
adaptive response to ionizing radiation appears to
increase initially with dose but starts to decrease when
the dose exceeds 100 mGy and disappears completely at
higher acute doses. By combining such a non-linear
model for a protective response with a linear, no
threshold (LNT) model for radiation induced cancer, it
can be argued that the net risk–dose relationship is more
likely to exhibit a threshold than to be linear down to
zero dose. It may even result in lower than spontaneous
cancer incidence (i.e. a beneficial effect) at doses below
200 mGy [3]. However, it should be appreciated that the
validity of the predictions of any mechanistic model is
critically dependent on the appropriateness of the
underlying assumptions.
A fundamental objection that many radiobiologists
have to this model is that the DNA damage caused by
ROS is believed to consist mostly of base damage and
single strand breaks. Ionizing radiation induces these
simple DNA lesions but it also induces double strand
breaks and more complex lesions [9]. It is only these
latter two types of damage that are considered to be the
initiating lesions in radiation-induced carcinogenesis
[15].
There are also serious doubts about the validity of
some of the radiobiological studies that are quoted to
provide evidence for a protective effect at low doses. In
experiments on chromosome aberrations in human
lymphocytes, cells pre-treated with a low (‘‘adapting’’)
The British Journal of Radiology, April 2006
dose of radiation apparently acquired increased resis-
tance to a second high (‘‘challenging’’) dose. These
studies have been interpreted as giving support for
adaptive responses, but serious limitations have been
recognized in two of them [44, 45], leading to a
subsequent publication by some of the original authors
and other scientists who found it very difficult to repeat
the results [46]. Another study purporting to show a
decreased chromosome aberration score in human
lymphocytes in vitro after a single acute low dose of
about 10 mGy followed by a dose-dependent increase
above 50 mGy [47], could also not be repeated by other
laboratories. The apparent dip in aberration yield at
10 mGy was thought to be due to an erroneous and
unusually high control yield at zero dose [48]. Criticism
can also be directed at two other sets of experiments [49,
50], which studied thymidine kinase activity in mouse
bone marrow cells and apoptosis in mouse thymocytes,
respectively. Both thymidine uptake and the percentage
of apoptotic cells observed after different doses of whole
body irradiation could be affected by radiation-induced
perturbations in the cycling characteristics of the cells as
much as by any assumed adaptive response or protective
effect.
The 1994 report of the UN Scientific Committee on the
Effects of Atomic Radiation, UNSCEAR [17], included an
annex which specifically considered evidence for adaptive
responses to radiation. However, UNSCEAR decided that
‘‘it would be premature to conclude that cellular adaptive
responses could convey possible beneficial effects to the
organism that would outweigh the detrimental effects of
exposures to low doses of radiation’’. Broadly similar
conclusions were reached by NCRP [9]. In a recent draft
report, an ICRP Task Group has concluded that current
understanding of mechanisms and quantitative data on
dose and time–dose relationships support a linear dose
response at low doses with no compelling evidence for the
existence of a threshold dose below which there would be
no effect [15].
The impact of the LNT model on the
justification and control of medical exposures
One obvious implication of the LNT model is that
when the doses are very low, so are the risks. There
comes a point when the risks are so low they can be
considered negligible, i.e. they are so small in relation to
the other everyday risks that surround us that they do
not need to be considered in any rational decisions about
lifestyle choices. There is in a sense a dose corresponding
to a threshold risk that, although not zero, is safe enough
for the risks to be ignored. Above this ‘‘safe enough’’
dose there will be a range of doses where the risks are
very small, but are sufficient to require some justification
for allowing people to be exposed to them, in terms of an
overriding benefit. As the doses and the risks increase so
should the concomitant benefits, for the exposures to
remain justified. This principle of justification is one of
the main planks of ICRP’s recommendations regarding
radiation protection for medical exposures. In recogni-
tion of the substantial potential health benefits to patients
from medical exposures, ICRP does not place any
restrictions on the levels of exposure that can be used
291
 B F Wall, G M Kendall. A A Edwards et al

in diagnostic radiology. It only requires that they be Table 2. X-ray examinations divided into four broad risk
justified in terms of an expected improvement in the bands
clinical management of the patient and that all reason-
Risk banda Risk rangeb Typical type of X-ray
able steps are taken to keep the exposures as low as examination
possible without compromising their diagnostic efficacy
(i.e. the exposures should be justified and optimized). Negligible , 1 in a million Radiography of
Thus although the LNT model implies that no dose is chest, limbs and
without risk, in practice there is a dose below which the teeth
Minimal 1 in a million to Radiography of
risks are considered negligible and are of no conse-
1 in 100000 head, neck and
quence in decisions regarding the radiation protection of joints
patients. Moreover, precise quantified risk estimates are Very low 1 in 100000 to Radiography of
not critical for the control of medical exposures accord- 1 in 10000 spine, abdomen
ing to the justification and optimization principles. It is and pelvis
very difficult to quantify the benefits of diagnostic X-ray Low 1 in 10000 to CT, angiography,
examinations in any way that is comparable with the 1 in 1000 contrast studies of
radiation risks, so an accurate quantitative weighing of the alimentary,
benefits against risks is usually impossible. Justification biliary and urinary
generally consists of the diagnostic radiology practi- tracts, and
interventional
tioner confirming that the exposure is clinically indicated
radiology
for the patient and making a mostly subjective judge-
a
ment that the expected benefits will outweigh the likely Nomenclature according to reference [89].
b
radiation risks. For such subjective judgements an Lifetime risk of cancer per examination for patients aged
approximate estimate of the risks will usually be 30–60 years.
sufficient. However, in the advice given to practitioners
on radiation risks, sources of bias should still be reduced Acknowledgment
to a minimum, so that the estimated risk lies centrally
within the range of uncertainty, which will unavoidably We would like to thank Dr Will Atkinson for helpful
become wider as the doses become lower. Use of the discussions.
LNT model at the relatively low doses typical of all
diagnostic X-ray exposures (, ,100 mGy) and a DDREF References
of 2 to extrapolate from the effects seen in epidemiolo- 1. ICRP Publication 87. Managing patient dose in computed
gical studies, will provide sufficiently robust risk tomography. Annals of the ICRP, Vol 30, No 4. 2000.
estimates for justification purposes. 2. Cameron J. Moderate dose rate ionising radiation increases
As well as the uncertainties in the radiation risk longevity. Br J Radiol 2005;78:11–3.
estimates, there are also large variations in the doses 3. Feinendegen L. Evidence for beneficial low level radiation
delivered to individual patients by the same type of X- effects and radiation hormesis. Br J Radiol 2005;78:3–7.
ray examination. The risks also depend markedly on the 4. Dawson P. Patient dose in multislice CT: why is it
age and sex of the patient and might be quite different increasing and does it matter? Br J Radiol 2004;77:S10–3.
for a few individuals in the population who are 5. Chadwick K, Leenhouts H. Radiation risk is linear with
genetically predisposed to cancer. Consequently, it is dose at low doses. Br J Radiol 2005;78:8–10.
usually not appropriate to resolve X-ray examinations 6. Martin C. The LNT model provides the best approach for
practical implementation of radiation protection. Br J Radiol
into any more than a few broad risk categories, each
2005;78:11–3.
spanning quite a wide range of risks. This has been done
7. NRPB. Risk of Radiation-induced Cancer at Low Doses and
in an information leaflet on the safety of X-ray examina- Low Dose Rates for Radiation Protection Purposes. Doc of
tions published by NRPB [51], where all X-ray examina- NRPB 1995;6(1).
tions have been divided into just four risk bands each 8. UNSCEAR 2000. Sources and effects of ionising radiation,
spanning a factor of ten in risk. This broad classification Volume 2 Effects: Annex G. Biological effects at low
into negligible, minimal, very low and low risk bands is radiation doses. New York, NY: UN, 2000.
shown in Table 2 and should be sufficient for most 9. NCRP Report 136. Evaluation of the Linear Nonthreshold
justification purposes. Dose-Response Model for Ionizing Radiation. Bethesda,
Those types of examination falling into the highest risk MD: NCRP, 2001.
band (still , 1 in 1000 risk of delayed cancer) are those 10. Cardis E, Gilbert ES, Carpenter L, et al. Effects of low doses
where the potential benefits can be correspondingly and low dose rates of external ionizing radiation: cancer
high, since many of them are used in the investigation of mortality among nuclear industry workers in three coun-
tries. Radiat Res 1995;142:117–32.
symptoms that suggest life-threatening diseases and will
11. Muirhead CR, Goodill AA, Haylock RGE, Vokes J, Little
easily comply with ICRP’s justification requirement. It
MP, Jackson DA, et al. Occupational radiation exposure and
would therefore appear that beneficial medical X-ray mortality: second analysis of the National Registry for
exposures need not be unnecessarily restricted by Radiation Workers. J Radiol Prot 1999;19:3–26.
adoption of the ICRP recommendations for patient 12. Brenner DJ, Doll R, Goodhead DT, Hall EJ, Land CE, Little
protection [52]. At the same time, use of the LNT model JB, et al. Cancer risks attributable to low doses of ionising
at low doses provides sufficiently reliable risk estimates radiation: assessing what we really know. Proc Natl Acad
to ensure that patients are being adequately protected Sci USA 2003;100:13761–6.
from unnecessary medical exposures that are either 13. Doll R, Wakeford R. Risk of childhood cancer from fetal
unjustified or not fully optimized. irradiation. Br J Radiol 1997;70:130–9.

292 The British Journal of Radiology, April 2006

Review article: What are the risks of low dose radiation?
14. COMARE Statement on Green Audit Occasional Paper
2002/5. COMARE Web site 2004 http://www.comare.org.
uk/statements/comare_statement_burnham.htm [Accessed
16 December 2005].
15. ICRP. Committee 1 Task Group Report. Draft Foundation
Document on Biological and Epidemiological Information
on Health Risks Attributable to Ionising Radiation: A
Summary of Judgements for the Purposes of Radiological
Protection of Humans. ICRP website 2005. www.icrp.org
[Accessed 16 December 2005].
16. Feinendegen LE, Bond VP, Sondhaus CA, Altman KI.
Cellular signal adaptation with damage control at low
doses versus the predominance of DNA damage at high
doses. C.R. Acad. Sci. Paris, Sciences de la vie/Life Sciences
1999;322:245–51.
17. UNSCEAR 1994. Sources and effects of ionising radiation.
United Nations Scientific Committee on the Effects of
Atomic Radiation 1994 Report to the General Assembly.
Annex B. Adaptive responses to radiation in cells and
organisms. New York, NY: UN, 1994.
18. Fox AJ, Collier PF. Low mortality rates in industrial cohort
studies due to selection for work and survival in the
industry. Br J Prev Social Medicine 1976;30:225.
19. Monson RR. Occupational epidemiology. Boca Raton, FL:
CRC Press, 1980.
20. Steenland K, Stayner L, Deddens J. Mortality analyses in a
cohort of 18,235 ethylene oxide exposed workers: follow-up
extended from 1987 to 1998. Occup Environ Med
2004;61:2–7.
21. Gun RT, Pratt NL, Griffith EC, Adams GG, Bisby JA,
Robinson KL. Update of a prospective study of mortality
and cancer incidence in the Australian petroleum industry.
Occup Environ Med 2004;61:150–6.
22. Pinkerton LE, Hein MJ, Stayner LT. Mortality among a
cohort of garment workers exposed to formaldehyde: an
update. Occup Environ Med 2004;61:193–200.
23. Berrington A, Darby SC, Weiss HA, Doll R. 100 years of
observation on British radiologists: mortality from cancer
and other causes 1897–1997. Br J Radiol 2001;74:507–19.
24. Cameron J. Radiation increased the longevity of British
radiologists. Br J Radiol 2002;75:637–40.
25. Doll R, Peto R. Mortality among doctors in different
occupations. Br Med J 1977;1:1433–6.
26. Brenner DJ, Hall EJ. Mortality patterns in British and US
radiologists: what can we really conclude? Br J Radiol
2003;76:1–2.
27. Hall EJ, Brenner DJ. The weight of evidence does not
support the suggestion that exposure to low doses of x-rays
increases longevity. Radiology 2003;229:18–9.
28. Matanoski GM, Sternberg A, Elliott EA. Does radiation
exposure produce a protective effect among radiologists?
Health Phys 1987;52:637–43.
29. Carpenter LM, Swerdlow AJ, Fear NT. Mortality of doctors
in different specialities: findings from a cohort of 20,000
NHS hospital consultants. Occup Environ Med
1997;54:388–95.
30. Matanoski GM. Health effects of low-level radiation in
shipyard workers: final report. DOE contract no. DE-AC02-
79 EV10095. Baltimore, MD. 1991.
31. Matanoski GM. Nuclear shipyard workers study. Radiat
Res 1993;133:126–7.
32. Kendall GM, Muirhead CR, MacGibbon BH, O’Hagan JA,
Conquest AJ, Goodill AA, et al. Mortality and occupational
exposure to radiation: first analysis of the National Registry
for Radiation Workers. Br Med J 1992;304:220–5.
33. Chen WL, Luan YC, Shieh MC. Is chronic radiation an
effective prophylaxis against cancer? J Am Phys Surg
2004;9:6–10.
34. Wakeford R. Radiation exposure in Taiwan. J Am Phys
Surg 2004;9:33–5.
The British Journal of Radiology, April 2006
35. Jagger J. Natural background radiation and cancer death in
Rocky Mountain States and Gulf Coast States. Health Phys
1998;75:428–30.
36. Cohen BL. Test of the linear no-threshold theory of
radiation carcinogenesis for inhaled radon decay products.
Health Phys 1995;68:157–74.
37. Cohen BL. Lung cancer rate vs. mean radon level in US
counties of various characteristics. Health Phys
1997;72:114–9.
38. Jagger J. Response to Archer. Health Phys 1999;76:694–5.
39. Puskin JS. Smoking as a confounder in ecologic correlations
of cancer mortality rates with average county radon levels.
Health Phys 2003;84:526–32.
40. Lubin LH. Studies of radon and lung cancer in North
America and China. Radiat Prot Dosim 2002;104:315–9.
41. Darby S, Hill D, Auvinen A, Barros-Dios JM, Baysson H,
Bochicchio F, et al. Radon in homes and risk of lung cancer:
collaborative analysis of individual data from 13 European
case-control studies. Br Med J 2005;330:223–6.
42. Krewski D, Lubin JH, Zielinski JM, Alavanja M, Catalan VS,
Field RW, et al. Residential radon and risk of lung cancer: A
combined analysis of 7 North American case-control
studies. Epidemiology 2005;16:137–45.
43. ICRP. Protection against radon-222 at home and at work.
ICRP Publication 65. Ann ICRP 23:(2) 1993.
44. Wolff S, Afzal V, Wienke JK, Olivieri G, Michaeli A. Human
lymphocytes exposed to low doses of ionising radiations
become refractory to high doses of radiation as well as to
chemical mutagens that induce double strand breaks in
DNA. Int J Radiat Biol 1988;53:39–49.
45. Olivieri G, Bodycote J, Wolff S. Adaptive response of
human lymphocytes to low concentrations of radioactive
thymidine. Science 1984;223:594–7.
46. Wojcik A, Aghamohammadi S, Aillaud M, Bosi A, Dai G,
Oliviera G, et al. Adaptive response to ionising radiation in
human lymphocytes: the problem of scoring aberrations in
cells irradiated during asynchronous growth. Mutation
Research 1996;366:137–43.
47. Pohl-Rueling J, Fischer P, Haas O, et al. Effect of low-dose
acute x-irradiation on the frequencies of chromosomal
aberrations in human peripheral lymphocytes in vitro.
Mutation Research 1983;110:71–82.
48. Edwards AA, Lloyd DC, Prosser JS. Chromosome aberra-
tions in human lymphocytes – A radiobiological review. In:
Baverstock, Stather, editors. Proceedings of L H Gray
Conference, Oxford 1989, Low Dose Radiation: Biological
Bases of Risk Assessment. Taylor & Francis, 1989.
49. Feinendegen LE, Muehlensiepen H, Lindberg C, Marx J,
Porscen W, Booz J. Acute and temporary inhibition of
thymidine kinase in mouse bone marrow cells after low-
dose exposure. Int J Radiat Biol 1984;45:204–15.
50. Zambouglou N, Porschen W, Muehlensiepen H, Booz J,
Feinendegen LE. Low dose effect of ionising radiation on
incorporation of iodo-deoxyuridine into bone marrow cells.
Int J Radiat Biol 1981;39:83–93.
51. NRPB, College of Radiographers, Royal College of
Radiologists, Royal College of General Practitioners. X-rays
- How safe are they? An information leaflet for patients.
2001. www.hpa.org.uk/radiation [Accessed 16 December
2005].
52. ICRP. 1990 Recomendations of the International
Commission on Radiological Protection. ICRP Publication
60. Ann ICRP 21 (1–3) 1991.
53. Pierce DA. Shimizu Y, Preston DL, et al. Studies of the
mortality of A-bomb survivors. Report 12, Part 1. Cancer:
1950–1990. Radiat Res 1996;146:1–27.
54. Preston DL, Kusumi S, Tomonaga M, et al. Cancer
incidence in atomic bomb survivors. Part III: Leukemia,
lymphoma and multiple myeloma, 1950–87. Radiat Res
1994;137:S68–S97.
293

55. Thompson DEK, Mabuchi ER, et al. Cancer incidence in
atomic bomb survivors. Part II. Solid tumors, 1958–87.
Radiat Res 1994;137:S17–S67.
56. Boice JD Jr, Day NE, Andersen A, et al. Second cancers
following radiation treatment for cervical cancer. An
international collaboration among cancer registries. J Natl
Cancer Inst 1985;74:955–75.
57. Curtis RE, Boice JD Jr, Stovall M, et al. The relation of
leukemia risk to radiation dose after cancer of the uterine
corpus. J Natl Cancer Inst 1994;86:1315–24.
58. Kaldor JM, Day NE, Bell J, et al. Lung cancer following
Hodgkin’s disease: a case-control study. Int J Cancer
1992;52:677–81.
59. Tucker MA, Morris Jones PH, Boice JD Jr, et al. Therapeutic
radiation at a young age is linked to secondary thyroid
cancer. Cancer Res 1991;51:2885–8.
60. Tucker MA, Meadows AT, Boice JD Jr, et al. Leukemia after
therapy with alkylating agents for childhood cancer. J Natl
Cancer Inst 1987;78:459–64.
61. Tucker MA, D’Angio GJ, Boice JD Jr, et al. Bone sarcomas
linked to radiotherapy and chemotherapy in children. N
Engl J Med 1987;317:588–93.
62. Wong FL, Boice JD Jr, Abramson DH, et al. Cancer
incidence after retinoblastoma. Radiation dose and sarcoma
risk. J Am Med Assoc 1997;278:1262–7.
63. Lundell M, Hakulinen T, Holm L-E. Thyroid cancer after
radiotherapy for skin hemangioma in infancy. Radiat Res
1994;140:334–9.
64. Lundell M, Holm L-E. Risk of solid tumors after irradiation
in infancy. Acta Oncol 1995;34:727–34.
65. Lundell M, Mattsson A, Hakulinen T, et al. Breast cancer
after radiotherapy for skin hemangioma in infancy. Radiat
Res 1996;145:225–30.
66. Lundell M, Holm L-E. Mortality from leukemia after
irradiation in infancy for skin hemangioma. Radiat Res
1996;145:595–601.
67. Karlsson P, Holmberg E, Lundberg LM, et al. Intracranial
tumors after radium treatment for skin hemangioma during
infancy - a cohort and case-control study. Radiat Res
1997;148:161–7.
68. Lindberg S, Karlsson P, Arvidsson B, et al. Cancer incidence
after radiotherapy for skin haemangioma during infancy.
Acta Oncol 1995;34:735–40.
69. Weiss HA, Darby SC, Doll R. Cancer mortality following
x-ray treatment for ankylosing spondylitis. Int J Cancer
1994;59:327–38.
70. Weiss HA, Darby SC, Fearn T, et al. Leukemia mortality
after x-ray treatment for ankylosing spondylitis. Radiat Res
1995;142:1–11.
71. Ron E, Modan B, Boice JD Jr. Mortality after radiotherapy
for ringworm of the scalp. Am J Epidemiol 1998;127:713–25.
72. Ron E, Modan B, Preston D, et al. Thyroid neoplasia
following low-dose radiation in childhood. Radiat Res
1989;120:516–31.
73. Ron E, Modan B, Preston D, et al. Radiation-induced skin
carcinomas of the head and neck. Radiat Res
1991;125:318–25.
294
B F Wall, G M Kendall. A A Edwards et al
74. Ron E, Modan B, Boice JD Jr, et al. Tumors of the brain and
nervous system after radiotherapy in childhood. N Engl J
Med 1988;319:1033–9.
75. Shore RE, Albert RE, Reed M, et al. Skin cancer incidence
among children irradiated for ringworm of the scalp. Radiat
Res 1984;100:192–204.
76. Shore RE, Hildreth N, Woodard E, et al. Breast cancer
among women given X-ray therapy for acute postpartum
mastitis. J Natl Cancer Inst 1986;77:689–96.
77. Hildreth NG, Shore RE, Dvoretsky PM. The risk of breast
cancer after irradiation of the thymus in infancy. N Engl J
Med 1989;321:1281–4.
78. Hildreth N, Shore R, Hempelmann L, et al. Risk of extra-
thyroid tumors following radiation treatment in infancy for
thymic enlargement. Radiat Res 1985;102:378–91.
79. Shore RE. Overview of radiation-induced skin cancer in
humans. Int J Radiat Biol 1990;57:809–27.
80. Darby SC, Reeves G, Key T, et al. Mortality in a cohort of
women given X-ray therapy for metropathia haemorrha-
gica. Int J Cancer 1994;56:793–801.
81. Inskip PD, Kleinerman RA, Stovall M, et al. Leukemia,
lymphoma and multiple myeloma following pelvic
radio-therapy for benign disease. Radiat Res
1993;135:108–25.
82. Inskip PD, Monson RR, Wagoner JK, et al. Cancer mortality
following radium treatment for uterine bleeding. Radiat Res
1990;123:331–44.
83. Boice JD Jr, Preston DL, Davis FG, et al. Frequent chest x-
ray fluoroscopy and breast cancer incidence among
tuberculosis patients in Massachusetts. Radiat Res
1991;125:214–22.
84. Howe GR. Lung cancer mortality between 1950 and 1987
after exposure to fractionated moderate-dose-rate ionizing
radiation in the Canadian fluoroscopy cohort study and a
comparison with lung cancer mortality in the atomic bomb
survivors study. Radiat Res 1995;142:295–305.
85. Howe GR, McLaughlin J. Breast cancer mortality between
1950 and 1987 after exposure to fractionated moderate-
dose-rate ionizing radiation in the Canadian fluoroscopy
cohort study and a comparison with breast cancer mortality
in the atomic bomb survivors study. Radiat Res
1996;145:694–707.
86. Bithell JF, Stewart AM. Prenatal irradiation and childhood
malignancy: a review of British data from the Oxford
survey. Br J Cancer 1975;31:271–87.
87. Muirhead CR, Kneale GW. Prenatal irradiation and child-
hood cancer. J Radiol Prot 1989;9:209–12.
88. Monson RR, MacMahon B. Prenatal x-ray exposure and
cancer in children. In: Boice JD Jr, Fraumeni JF Jr, editors.
Radiation carcinogenesis: epidemiology and biologi-
cal significance. New York, NY: Raven Press, 1984:
97–105.
89. Chief Medical Officer of the Department of Health, On the
State of the Public Health 1995, the annual report of
the CMO of the DoH for the year 1995, HMSO, London,
1996.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 295–297

REVIEW ARTICLE

Diagnosing a parotid lump: fine needle aspiration cytology or

core biopsy?
D C HOWLETT, MRCP(UK), FRCR

Department of Radiology, Eastbourne District General Hospital, Kings Drive, Eastbourne, East
Sussex BN21 2UD, UK

ABSTRACT. Fine needle aspiration cytology (FNAC) has been widely adopted for the
cytological diagnosis of parotid lumps. FNAC does have drawbacks, even under
optimum conditions and may be associated with poor levels of diagnostic accuracy, Received 21 April 2005
particularly outside the specialized clinic environment. Ultrasound-guided core biopsy Revised 25 August 2005
(USCB) is a relatively recently described technique in the parotid gland which has been Accepted 11 October 2005
well tolerated and has demonstrated a high degree of diagnostic accuracy in several
DOI: 10.1259/bjr/74329476
studies. This article discusses the merits and pitfalls of FNAC, together with the
technique of USCB and also highlights the potential advantages benefit provided by ’ 2006 The British Institute of
USCB in parotid diagnosis. Radiology

A broad spectrum of pathologies that present with
parotid swelling and extraglandular masses can also
mimic parotid lesions clinically. It is frequently difficult
on clinical grounds alone to distinguish between neoplas-
tic and non-neoplastic causes for a parotid mass and also
to reliably differentiate between benign and malignant
neoplasms. If an accurate pre-operative diagnosis can be
achieved using a combination of imaging and cytology or
histology, then many non-neoplastic lesions will not
require excision. Surgery may also be avoided for certain
parotid neoplasms in the elderly or unfit, e.g. Warthin’s
tumour. To ensure that surgery is indicated and to allow
appropriate operative planning and patient consent an
accurate pre-operative diagnosis is essential.
Following initial demonstration and characterization
of a parotid lesion with imaging, usually ultrasound or
MRI, needle biopsy is used to confirm its nature if
required. Following the demise of open biopsy due to
high rates of tumour seeding [1] fine-needle aspiration
cytology (FNAC) has become an established technique.
FNAC is most commonly performed blindly in the
outpatient clinic and has a number of advantages – it is
quick, safe and accurate in the hands of a skilled
practitioner and high levels of diagnostic accuracy have
been quoted [2–4]. This process may occur within a
specialized clinic and diagnostic accuracy can be
improved using ultrasound-guidance and an on-site
cytologist [5]. Clinics may also be cytologist-led,
although this may necessitate multiple aspirations and
there is currently a severe shortage of cytology staff
available with the necessary expertise, making it difficult
to offer this facility outside larger centres.
In the absence of ultrasound-guidance, or an on-site
cytologist the diagnostic accuracy of FNAC often falls off
dramatically. A recent study reported a sensitivity of
only 38% in distinguishing benign from malignant
disease using blind FNAC [6].
There are also well-recognized pitfalls of FNAC even
in experienced hands – particular difficulties may occur
in the cytological diagnosis of pleomorphic adenoma [7,
8], Warthin’s tumour [9] and lymphoma [10]. Indeed the
diagnosis of lymphoma with FNAC is not generally
considered definitive, with FNAC often acting solely as a
guide for the need for surgical biopsy if lymphoid
proliferation is present cytologically [11]. There are
specialized ancillary cytological techniques, e.g. flow
cytometry and in situ hybridization, which are used to
improve the diagnosis of lymphoid proliferation but
these are not widely available outside larger centres. The
cytological diagnosis of parotid involvement by systemic
disease, e.g. sarcoidosis and Sjögren’s syndrome may be
difficult and it is also not possible to accurately grade or
type malignant tumours or lymphomas or to distinguish
in situ from invasive disease with FNAC. FNAC has a
low predictive value for benign non-neoplastic lesions
[3] and similarly there is a low negative predictive value
if a negative FNAC result is obtained [4]. As a
consequence of the diagnostic difficulties that may arise
with FNAC there is a high incidence of surgical biopsy
with subsequent delays in referral to the appropriate
clinical team.
Performing ultrasound-guided core biopsy (USCB)
using a spring-loaded biopsy device does provide a
recently described alternative to parotid FNAC. Core
biopsy possesses an inherent advantage over FNAC in
that it provides a sample of tissue for immunohisto-
chemical analysis. This allows typing and grading of
carcinomas and lymphomas and also improved differen-
tiation of reactive nodal hyperplasia from lymphoma. A
core of tissue can also be used to evaluate parotid
involvement by systemic disease. An on-site cytologist is
not required and USCB can be combined with an initial
diagnostic ultrasound, allowing lesion characterization
and compartmentalization into superficial or deep lobe.

The British Journal of Radiology, April 2006 295


In two series of patients with parotid masses USCB has
shown promising results with diagnostic samples
obtained in all patients and reported accuracies of
100% [12] and 97% [13] when comparing core biopsy
with final surgical histology. There were no reported
complications of USCB and 26 out of 54 [12] and 22 out of
53 [13] patients avoided surgery as a result of core
biopsy.
USCB utilizes a small bore needle (18 G or 20 G) which
is introduced through a small skin incision following
informed written consent and infiltration with 1%
lignocaine local anaesthesia. Using a high-resolution
linear-array transducer for guidance, the needle-tip is
positioned just adjacent to the lesion, such that following
discharge the needle traverses, but does not exit deep to
the lesion (Figures 1 and 2). The use of ultrasound is
important as it allows avoidance of adjacent structures
Figure 1. Ultrasound through the tail of the right parotid
gland demonstrates a 12 mm pleomorphic adenoma over-
lying the angle of the mandible (M). The biopsy needle has
been placed just anterior to the lesion, using a 15 mm setting
on a variable-throw biopsy device.
Figure 2. The biopsy device has been discharged and the
needle can be seen to traverse the mass, sampling periphery
and core, but has not exited deep to the lesion.
296
D C Howlett
during biopsy and also ensures that both the periphery
and core of the lesion are sampled, with bypassing of
necrotic areas, increasing diagnostic yield [12]. Some
biopsy devices can allow the needle throw to be varied
(Magnum gun; Bard, Covington, GA – 15–22 mm
variable throw), which is useful for smaller or deeper
lesions in proximity to the parotid vessels. There are
important potential complications of needle biopsy
including haemorrhage and facial nerve injury and also
tumour seeding in the needle tract. The identification of
the intraparotid vessels on ultrasound should allow the
main parotid vessels and thereby the adjacent facial
nerve to be avoided during biopsy. There are two reports
of tumour seeding post needle biopsy [14, 15] although
these occurred with large bore needles and this
phenomenon is not described using smaller needle sizes.
Some surgeons may choose to excise the biopsy tract at
the time of operation. USCB is a little more invasive than
FNAC in that it requires local anaesthesia and a skin
incision. USCB also does not lend itself to the ‘‘one-stop’’
clinic setting due to longer requirements for histological
reporting than a cytology aspirate.
In conclusion USCB represents a recently described
technique for diagnosing a parotid mass which has
potential benefits over FNAC. These advantages may be
particularly relevant in the district general hospital
setting where ancillary cytological facilities are not
routinely available. The use of USCB may help reduce
the need for diagnostic surgical biopsy and thereby
facilitate referral to the appropriate clinical team.
References
1. McGuirt WF, McCable BF. Significance of node biopsy
before definitive treatment of cervical metastatic carcinoma.
Laryngoscope 1978;88:594–7.
2. Das DK, Petkar MA, Al-Mane NM, et al. The role of fine
needle aspiration cytology in the diagnosis of swellings in
the salivary gland regions: a study of 712 cases. Med Princ
Prac 2004;13:95–106.
3. Cohen EG, Patel SG, Lin O, et al. Fine needle aspiration
biopsy of salivary gland lesions in a selected patient
population. Arch Otolaryngol Head Neck Surg
2004;130:773–8.
4. Zbären P, Schär C, Hotz MA, Loosli H. Value of fine needle
aspiration cytology of parotid gland masses. Laryngoscope
2001;111:1989–92.
5. Robinson IA, Cozens NJ. Does a joint ultrasound – guided
cytology clinic optimize the cytological evaluation of head
and neck masses? Clin Radiol 1999;54:312–6.
6. Balakrishnan K, Castling B, McMahan J, Imrie J, Feeley KM,
Parker AJ, et al. Fine needle aspiration cytology in the
management of parotid mass: a two centre retrospective
study. Surgeon 2005;2:67–72.
7. Brachtel EF, Pilch BZ, Khettry V, et al. Fine needle
aspiration biopsy of a cystic pleomorphic adenoma with
extensive adnexa-like differentiation: differential diagnostic
pitfall with mucoepidermoid carcinoma. Diagn Cytopathol
2003;28:100–3.
8. Verma K, Kapila K. Role of fine needle aspiration cytology
in the diagnosis of pleomorphic adenoma. Cytopathology
2002;13:121–7.
9. Parwarni AV, Ali-Sayed Z. Diagnostic accuracy and pitfalls
in the fine needle aspiration interpretation of Warthin’s
tumour. Cancer 2003;99:166–71.
10. Loggins JP, Urquhart A. Pre-operative distinction of parotid
lymphomas. J Am Coll Surg 2004;199:58–61.
The British Journal of Radiology, April 2006
Review article: Diagnosing a parotid lump
11. Lioe TF, Elliot H, Allen DC, Spence RA. The role of fine
needle aspiration cytology (FNAC) in the investigation of
superficial lymphadenopathy – uses and limitations of the
technique. Cytopathology 1999;10:291–7.
12. Kesse KW, Manjaly G, Violaris N, Howlett DC. Ultrasound-
guided biopsy in the evaluation of focal lesions and diffuse
swelling of the parotid gland. Br J Oral Maxillofac Surg
2002;40:384–9.
The British Journal of Radiology, April 2006
13. Yung-Liang W, Siu-Cheung C, Yao-Liang C, et al.
Ultrasonography-guided core-needle biopsy of parotid
gland masses. Am J Neuroradiol 2004;25:1608–12.
14. Yamaguchi KT, Strong MS, Shapsey SM, Soto E. Seeding of
parotid carcinoma along Vim-Silverman needle tract. J
Otolaryngol 1979;81:49–52.
15. Peacock EE, Byars LT. Management of tumours of the
parotid gland. N C Med J 1958;19:1–9.
297
 The British Journal of Radiology, 79 (2006), 298–302

Intra-arterial MR angiography in the iliac system: initial clinical

experience with 25 patients
C PAETZEL, MD, N ZORGER, MD, O W HAMER, MD, J SEITZ, MD, T SCHLEICHER, MD, S FEUERBACH, MD,
W R NITZ, PhD, M LENHART, MD and T HEROLD, MD

Department of Diagnostic Radiology, University of Regensburg, Hospital, 93042 Regensburg,

Germany

ABSTRACT. The aim of this study was to evaluate intra-arterial magnetic resonance
angiography (MRA) of the iliac arteries. Therefore, 25 consecutive patients (17 male, 8
female) suffering from symptomatic occlusive disease of the lower limbs were
investigated prospectively. Catheter angiography was performed before MRA and
served as the standard of reference. Contrast-enhanced intra-arterial MRA was
performed using a 1.5 Tesla MRI system. Contrast agent (gadodiamide) was injected by
a conventional pigtail-shaped angiography catheter placed in the abdominal aorta.
Vascular lesions were assessed by four investigators. The degree of stenosis was
compared with the findings of conventional catheter angiography. Additionally, the
diagnostic quality of the MR angiograms was assessed by the investigators using a semi Received 29 March 2005
quantitative five-point scale. All lesions shown by catheter angiography were detected Revised 10 June 2005
and correctly localized by intra-arterial MRA. MR angiograms exhibit a specificity of Accepted 18 July 2005
95% and a sensitivity of 96% for stenoses of 50% or more. The diagnostic quality of the
DOI: 10.1259/bjr/19491401
images was judged from good to excellent, on average. Intra-arterial MRA exhibits a
specificity and sensitivity comparable with intravenous angiography. The image quality ’ 2006 The British Institute of
appears to be adequate for supporting MR-guided vascular intervention. Radiology

Fluoroscopy-guided percutaneous interventional pro-
cedures like transluminal angioplasty are well estab-
lished methods for the treatment of peripheral arterial
occlusive disease of the lower limbs. Although MR
guidance of manoeuvres for vascular intervention
provides several advantages, such as the absence of
ionizing radiation, avoidance of potential hazards from
iodinated contrast material and superior soft tissue
contrast, it has not yet gained acceptance in the clinical
routine. Specific problems with MR-guided intervention
are the lack of availability of MR compatible materials
(e.g. stents, guidewires), their visibility and the asso-
ciated high costs. Furthermore, the precise localization of
stenosis and post-interventional workup may cause
difficulties. Manke et al [1] showed the feasibility of
MRI-guided stent placement in iliac artery stenoses in 14
cases using intravenous magnetic resonance angiogra-
phy (MRA) for pre- and post-interventional imaging.
Due to the restricted total amount of applicable MR
contrast agent, recurrent control examinations cannot be
performed. Manke reported one stent misplacement due
to patient motion during the procedure while relying on
a previously acquired MRA roadmap. A newer report
describes MR-guided angioplasty using intra-arterial
MRA [2]. Intra-arterial MRA with a catheter positioned
in the infrarenal abdominal aorta allows a reduction in
the amount of contrast material applied. Based on this
approach, the correct intraluminal site of the catheter can
Current address for Christian Paetzel: Department of Diagnostic
Radiology, Klinikum Weiden, Söllnerstraße 16, 92637 Weiden,
Germany.
be identified, and repeated control examinations due to
patient motion can be performed. There are reports
regarding this technique in the literature, but sensitivity
and specificity have not yet been systematically assessed
[3, 4].
The aim of this study was to investigate the sensitivity,
specificity and the image quality of intra-arterial MR
angiograms of the iliac arteries using conventional
catheter angiography as the standard of reference.
Materials and methods
Patient population
Over a time period of 5 months, 25 consecutive patients
(17 men, 8 women; mean age 61.4 years, range 33–
82 years, standard deviation ¡8.5) suffering from
peripheral arterial occlusive disease of the lower limbs
underwent intra-arterial conventional catheter angiogra-
phy and intra-arterial MRA of the iliac arteries. The
grading of the disease according to the classification of
Table 1. Patient population: Fontaine’s classification of
peripheral arterial occlusive disease
Grading (Fontaine’s classification) Number of patients
I 0
II a 3
II b 16
III 3
IV 3

298 The British Journal of Radiology, April 2006

Intra-arterial MRA in the iliac system

Figure 1. Examination of a 53-year-old man suffering from symptomatic arterial occlusive disease of the lower limbs with rest
pain. (a) Maximum intensity projection of the intra-arterial magnetic resonance angiography showing an occlusion of the left
common iliac artery. The artefact of the angiography catheter is indicated by an arrow. (b,c) Corresponding conventional
catheter angiography confirms the finding of the MR angiogram.

Fontaine is presented in Table 1. Written informed consent
was obtained from all patients in advance. The study
protocol was approved by the institutional review board.
Protocol of conventional angiography and intra-
arterial MRA
Digital subtraction angiography (DSA) was the stan-
dard of reference in all patients. Femoral access with a
5 F sheath was performed in the angiography suite. To
avoid thromembolic complications due to the extended
investigation time, all patients received 2500 IE heparin
intravenously and 40 mg butylscopolamin. A pigtail
catheter was placed in the infrarenal aorta at the level
of the second lumbar vertebral body. Non-ionic contrast
material with 320 mgI ml21 was injected intra-arterially
(volume 20 ml; injection rate 10 ml s21; frame rate 2 s21).
Imaging was performed using the 35 ˚ left and right
anterior oblique position. The degree of stenosis was

The British Journal of Radiology, April 2006 299


Figure 2. 62-year-old man with ischaemic ulcer of the right foot. (a) Intra-arterial MR angiography detects a high grade stenosis
of the right external and an arteriosclerotic dilation of the left common iliac artery as effectively (b) as conventional
angiography. The arrow marks the artefact of the angiography catheter.
assessed using the commercially available software on
the Polytron T.O.P. (Siemens, Erlangen, Germany). Using
this software to analyse the angiographic findings, the
degree of stenosis was calculated automatically by
comparing the geometry and density of a lesion.
After conventional DSA the patients were taken to the
MR-scanner with a permanently flushed pigtail-catheter.
MRA images were acquired on a 1.5 T scanner (Sonata;
Siemens Medical Solutions) with a gradient strength of
40 mT m21 and a slew rate of 200 T m21 s21 using a
phased array receiver coil and breath-hold acquisitions
(22 s). A fast-low-angle-shot (FLASH) sequence was
utilized with elliptical scanning encoding, 40 partitions
(75% partial Fourier), 280 mm 6 263 mm field of view,
1.5 mm partition thickness, 25 ˚ excitation angle, 266 6
512 matrix size (75% partial Fourier) with an echo time of
1.92 ms and a repetition time of 5.87 ms for a bandwidth
of 200 Hz/pixel. 10 ml of gadodiamide (Omniscan;
Amersham Buchler, Braunschweig, Germany) was
applied via the arterial catheter with a concentration of
0.5 mmol l21, which was diluted with 50 ml of 0.9% of
saline solution at an injection rate of 3.5 ml s21. The
injection was followed by 20 ml of 0.9% saline solution
with an injection rate of 3.5 ml s21 [4]. The application of
the contrast agent was performed with an MR injection
system (Spectris; Medrad, Indianola, Iowa).
A maximum intensity projection (MIP) algorithm was
applied to all contrast-enhanced MR angiography stu-
dies after subtraction of the unenhanced measurement
by using the commercially available software on the MRI
system (Magnetom Symphony, software Numaris 3.5,
version VA11A; Siemens). MIPs were reconstructed in
steps of 9 ˚ (range of 180 ˚).
300
C Paetzel, N Zorger, O W Hamer et al
Image analysis
Four radiologists experienced in cardiovascular ima-
ging assessed the MR angiograms independently in a
randomized order for stenoses of the common and
external iliac arteries and image quality of MRA.
Observers were not aware of DSA findings when
analysing contrast enhanced MRA, and vice versa.
Image analysis was based on original contrast-
enhanced data sets, MIPs and digital subtraction
angiograms. None of the observers were aware of the
clinical history or the interpretations of the other
observers. The degree of stenosis was defined as the
ratio of the narrowest diameter (A) within the stenosis
and the diameter of the nearest downstream uninvolved
segment of the artery (B): 100 6 (1–A/B). The stenoses
were classified into occlusions, high grade stenoses (75–
99%), moderate stenoses (50–74%), mild stenoses (25–
49%) and non-stenoses (0–24%).
Additionally, the four radiologists assessed MR
angiograms independently for the following: diagnostic
value of the infrarenal abdominal aorta; common iliac
arteries; internal and external iliac arteries, and the
overall impressions were rated on a subjective scale
(15excellent, no limitations; 25good, minor limitations;
35moderate, moderate limitations; 45poor, major
limitations but still of diagnostic quality; 55non-diag-
nostic). Intermediate scores at 0.5 intervals were not
allowed.
The catheter angiograms were assessed by a consensus
of two of the investigators. The degree of stenosis was
quantified and the lesions were classified in the same
manner as described for the MR angiograms. If there was
no vascular lesion detected by either catheter angiography
The British Journal of Radiology, April 2006
Intra-arterial MRA in the iliac system

Table 2. Classification of stenoses; results of intra-arterial MR angiography compared with conventional catheter angiograms
Degree of stenosis Underestimated Correct scores Overestimated
(number of lesions)

Occlusion (n54) 0% (0/16) 100% (16/16)

75–99% (n53) 0% (0/12) 100% (12/12) 0% (0/12)
50–74% (n512) 6.3% (3/48) 89.6% (43/48) 4.2% (2/48)
25–49% (n510) 7.5% (3/40) 82.5% (33/40) 10% (4/40)
0–24% (n511) 79.5% (35/44) 20.5% (9/44)
All patients (n540) 3.8% (6/160) 86.9% (139/160) 9.4% (15/160)

or intra-arterial MRA, it was scored as a patient with one Discussion

lesion with a stenosis of 0%. MR-guided vascular interventions provide many
advantages over fluoroscopy-guided techniques as men-
tioned above. Nevertheless, MR-guided procedures are
Results not yet well established in the clinical routine. The
reasons for this are the lack of availability of visible
In total, 40 vascular lesions were assessed by catheter materials, real time imaging during the intervention and
angiography. There were 10 patients with one lesion, the peri-interventional check-up. Manke et al used
nine patients with two, three patients with three and commercially available catheters and guidewires [1] for
three patients with none. the treatment of iliac artery stenoses by stent placement
The catheter angiography detected four occlusions based on the safety precautions as described by Nitz et al
(Figure 1), three high grade (Figure 2), 12 moderate and [4]. Pre-interventional imaging was performed using
10 mild stenoses (Table 2). All of the seven lesions with a intravenous MRA. Even though the usefulness of
degree of stenosis higher than 75% were correctly intravenous MRA for imaging the iliac arteries has been
assessed by intra-arterial MRA. The moderate stenoses demonstrated by several authors [6–16], the restricted
were overestimated by one of the investigators in two total amount of contrast medium of approximately
cases out of 48 (4.2%) and underestimated in three cases 0.3 mmol kg21 body weight prohibits the investigator
(6.3%). Mild stenoses were overestimated in 4 out of 40 from performing repeated angiograms, which can lead to
(10%), no stenoses in 9 out of 44 (20.5%) cases. Mild stent misplacement due to patient motion. To avoid this,
stenoses were underestimated three times (7.5%). The intra-arterial MRA was used in a recent study for peri-
lesions which were assessed incorrectly by at least one of interventional imaging [2]. Thus, the contrast agent dose
the observers are outlined in Table 3. With respect to is reduced 50% compared with MRA with intravenous
stenoses of more than 50%, the sensitivity and specificity injection, allowing the investigator to perform repeated
were calculated at 96% and 95%, respectively. angiograms up to four times. The contrast agent dose
The diagnostic value of all intra-arterial MR angio- chosen was based on a conservative level for a wide
grams was judged to be 1.54 (range 1–4, median 1) on range of interindividual conditions. A further reduction
average by all investigators. The absolute and relative in the total amount of contrast agent administered seems
frequency of the scores for diagnostic quality is listed in possible. Even if initial reports about intra-arterial MRA
Table 4. already exist in the literature [3, 4], there is no further
experience with this method concerning its sensitivity
and specificity.
In this study, all of the 40 lesions assessed by catheter
Table 3. Lesions scored incorrectly by at least one observer angiography were detected and correctly localized by
Degree of Observer 1 Observer 2 Observer 3 Observer 4
intra-arterial MRA. Thus, verification of lesions is
stenosis possible using intra-arterial angiography with a sensi-
(catheter tivity of 100% in this study.
angiography) The indication to treat a lesion depends on the degree
0.19 0 + + 0 of the stenosis and the clinical symptoms. The degree of
0.20 0 0 + 0 stenosis was correctly classified in 86.9% (139/160);
0.22 + 0 0 + stenoses were overestimated in 9.4% (15/160) and
0.22 0 + + 0 underestimated in 3.8% (6/160) of the cases. Focusing
0.24 + + 0 0 on obstructions or stenosis of more than 50%, the
0.29 0 0 0 2 sensitivity and specificity of approximately 95% is
0.33 0 2 0 2
0.42 + 0 0 0
0.48 + + 0 0 Table 4. Estimation of image quality by four investigators in
0.49 0 0 0 + 25 patients
0.51 2 0 2 0 Score 1 2 3 4 5
0.58 0 0 + 0
0.59 0 0 0 2 Absolute frequency 58 31 10 1 0
0.69 0 0 0 + Relative frequency 0.58 0.31 0.10 0.01 0
0, Correct score; 2, underestimated; +, overestimated. 1, excellent; 2, good; 3, moderate; 4, poor; 5, non-diagnostic.

The British Journal of Radiology, April 2006 301


comparable with intravenous MRA [6, 10–15]. The value
of contrast-enhanced intravenous MRA has been demon-
strated in a study by Lenhart et al, showing that MRA
can take the place of catheter angiography in the routine
work-up of patients with peripheral arterial occlusive
disease. Additional examinations are necessary in
approximately 5% of the patients, mainly due to
inadequate diagnostic quality of the vessels of the lower
leg. Thus, with respect to the clinical symptoms,
intravenous and intra-arterial MRA give us sufficient
information for the treatment of iliac artery lesions in
patients with symptomatic arterial occlusive disease of
the lower limbs. The outcome of this study is comparable
with the results of a recent study concerning the
accuracy of intra-arterial MRA of the femorocrural
arteries [16].
Using gadodiamide for intra-arterial injection repre-
sents an ‘‘off-label’’ use. There are some studies dealing
with the intra-arterial application of gadolinium-com-
plex in catheter angiography [17, 18]. The reasons for
such application are allergic reactions to iodide contrast
agents, renal insufficiency and hyperthyroidism. We
used gadodiamide due to the low osmolality of
780 mmol kg21 compared with the other available
gadolinium complexes. Similar to reports in the litera-
ture, no adverse reactions where observed during this
study.
The confidence of the radiologists in the diagnostic
quality of the intra-arterial MR angiograms is demon-
strated by their subjective assessment of the image
quality. They judged 89% of the investigations as good
or excellent using a five point scale. All MR angiograms
were assessed as having diagnostic quality. This result
indicates that no additional examinations would have
been required by the readers to make decisions about
further treatment.
Conclusion
Lesions of the iliac arteries can be detected and
localized by intra-arterial MRA. MR angiograms pre-
sented with a sensitivity and specificity comparable with
intravenous MRA. Thus, intra-arterial MRA seems to
support MR-guided interventional procedures.
References
1. Manke C, Nitz WR, Djavidani B, Strotzer M, Lenhart M,
Völk M, et al. MR imaging-guided stent placement in iliac
arterial stenoses: a feasibility study. Radiology
2001;219:527–34.
2. Paetzel C, Zorger N, Bachthaler M, Völk M, Seitz J, Herold
T, et al. Feasibility of MR guided angioplasty of femoral
artery stenoses using real-time imaging and intraarterial
contrast-enhanced MR-angiography. Rofo 2004;176:1232–6.
3. Tello R, Mitchel PJ, Melhem ER, Witte D, Thomson KR.
Interventional catheter magnetic resonance angiography
with a conventional 1.5-T magnet: work in progress.
Australas Radiol 1999;43:435–9.
302
C Paetzel, N Zorger, O W Hamer et al
4. Paetzel C, Zorger N, Seitz J, Völk M, Nitz WR, Herold T,
et al. Intraarterial contrast-enhanced magnetic resonance
angiography of the aortoiliac system. J Vasc Interv Radiol
2004;15:981–4.
5. Nitz WR, Oppelt A, Renz W, Manke C, Lenhart M, Link J.
On the heating of linear conducting structures as guide-
wires and catheter in interventional MR. J Magn Reson
Imaging 2001;13:105–14.
6. Kai Yiu JAM, Leiner T, de Haan MW, Kessels AG, Kitslaar
PJ, van Engelhoven J. Periphereal vascular tree stenoses:
evaluation with moving-bed infusion-tracking MR angio-
graphy. Radiology 1998;206:683–92.
7. Quinn SF, Sheley RC, Semonsen KG, Leonardo VJ, Kryss K,
Szumowski J. Aortic and lower-extremity arterial disease:
evaluation with MR angiography versus conventional
angiography. Radiology 1998;206:693–701.
8. Snidow JJ, Johnson MS, Harris VJ, Margosian PM, Aisen
AM, Lalka SG, et al. Three-dimensional gadolinium-
enhanced MR angiography for aortoiliac inflow assessment
plus renal artery screening in a single breath hold.
Radiology 1998;198:725–32.
9. Lenhart M, Finkenzeller T, Paetzel C, Strotzer M, Mann S,
Djavidani B, et al. Contrast-enhanced MR angiography in
the routine work-up of the lower extremity arteries. Rofo
2002;174:1289–95.
10. Sueyoshi E, Sakamoto I, Matsuoka Y, Ogawa Y, Hayashi H,
Hashmi R, et al. Aortoiliac and lower extremity arteries:
comparison of three-dimensional dynamic contrast-
enhanced subtraction MR angiography and conventional
angiography. Radiology 1999;210:683–8.
11. Lenhart M, Herold T, Volk M, Seitz J, Manke C, Zorger N,
et al. Contrast media-enhanced MR angiography of the
lower extremity arteries using a dedicated peripheral vascular
coil system. First clinical results. Rofo 2000;172:992–9.
12. Lenhart M, Djavidani B, Volk M, Strotzer M, Manke C,
Requardt M, et al. Contrast medium-enhanced MR angio-
graphy of the pelvic and leg vessels with an automated
table-feed technique. Rofo 1999;171:442–9.
13. Kreitner KF, Kalden P, Neufang A, Duber C, Krummenauer
F, Kustner E, et al. Diabetes and peripheral arterial
occlusive disease: prospective comparison of contrast-
enhanced three-dimensional MR angiography with con-
ventional digital subtraction angiography. AJR Am J
Roentgenol 2000;174:171–9.
14. Huber A, Scheidler J, Wintersperger B, Baur A, Schmidt M,
Requardt M, et al. Moving-table MR angiography of the
peripheral runoff vessels: comparison of body coil and
dedicated phased array coil systems. AJR Am J Roentgenol
2003;180:1365–73.
15. Douek PC, Revel D, Chazel S, Falise B, Villard J, Amiel M.
Fast MR angiography of the aortoiliac arteries and arteries
of the lower extremity: value of bolus-enhanced, whole-
volume subtraction technique. AJR Am J Roentgenol
1995;165:431–7.
16. Zorger N, Völk M, Hamer OW, Lenhart M, Seitz J, Butz B,
et al. Intraarterial gadolinium-enhanced MR angiography in
humans: feasibility and accuracy for detection of infra-
inguinal arterial stenoses before and after percutaneous
angioplasty. AJR Am J Roentgenol 2005;185:867–72.
17. Schild HH, Weber W, Boeck E, Mildenberger P, Strunk H,
Duber C, et al. Gadolinium-DTPA (Magnevist) as a contrast
medium for arterial DSA. Rofo 1994;160:218–21.
18. Hatrick AG, Jarosz JM, Irvine AT. Gadopetate dimeglumine
as an alternative contrast agent for use in interventional
procedures. Clin Radiol 1997;52:948–52.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 303–307

Prevalence and distribution of adnexal findings suggesting

endometriosis in patients with MR diagnosis of adenomyosis
T T ZACHARIA, MD and M J O’NEILL, MD

Abdominal Imaging and Interventional Radiology, Massachusetts General Hospital, Harvard

Medical School, USA

ABSTRACT. The purpose of this investigation was to establish the prevalence and
distribution of MR findings associated with pelvic endometriosis in patients with a MRI
diagnosis of adenomyosis. Retrospective review of the pelvic MRI in 59 patients (age
32–54 years, mean 42 years) who met strict MRI criteria for adenomyosis was
performed. T1 weighted fat saturated and T2 weighted images in these patients were
reviewed for the presence or absence of T1 bright signal suggesting endometriosis in
any of nine locations (uterine serosa, right and left ovary, right and left fallopian tube,
right and left broad ligament, and right and left pelvic side wall). 20 (20/59) patients
(34%), showed characteristic MRI features associated with endometriosis. A total of 54
sites of involvement were identified (uterine serosa n517, ovaries n514, broad
ligaments n510, fallopian tubes n58, pelvic side walls n55) in 20 patients with an
average of 2.7 sites per patient. Implants (n543) were more common than
endometriomas (n511). Endometriomas occurred most often in the ovaries (ovaries
n59, broad ligament n52) whereas implants were seen on all locations (uterine serosa Received 19 May 2005
n517, ovaries n55, broad ligaments n58, fallopian tubes n58, pelvic side walls n55). Revised 17 July 2005
One third of patients with adenomyosis diagnosed by MRI also had MRI findings of Accepted 12 August 2005
endometriosis, with serosal implants being the most common finding. Imaging
DOI: 10.1259/bjr/70121266
protocols should routinely include T1 weighted fat saturated imaging sequences in
order to detect coexistent endometriosis in patents undergoing pelvic MRI for the ’ 2006 The British Institute of
diagnosis of adenomyosis. Radiology

Adenomyosis and endometriosis are gynaecological
disorders with similar pathological appearance and
proposed aetiologies. We sought to define whether these
disease entities coexist.
Recent studies have shown high sensitivity and
specificity of MRI in the diagnosis of endometriosis [1,
2]. The detection and characterization of endometrioma
is relatively easier when compared with detection of
serosal endometriotic implants. A diagnosis of serosal
implants can be easily missed during routine pelvic MRI
evaluation.
Some investigations performed recently have shown
MRI to have very high sensitivity and specificity in the
diagnosis of adenomyosis [3, 4]. MRI has shown added
value in the evaluation of associated diseases like uterine
leiomyomas when compared with transvaginal ultra-
sound (TVUS).
Similar theories of abnormal migration and metaplasia
[5–7] are postulated in the aetiology of both adenomyosis
and endometriosis. The most commonly proposed
aetiology for adenomyosis is direct migration of endo-
metrial tissue into the myometrium. This mechanism of
direct extension is similar to the most widely accepted
Address correspondence to: T Thomas Zacharia, MD, #4225,
Elmhurst Ave, Ithaca Street, Queens, New York City, NY 11373,
USA.
Presented at a meeting: 104th ARRS (American Roentgen Ray
Society) meeting, Miami, Florida, May 2004.
theory for the aetiology of endometriosis, i.e. reflux and
retrograde extension of endometrial tissue into the
peritoneal cavity through the fallopian tube.
To date, no large series of patients with both disease
entities has been reported. The purpose of this investiga-
tion was to establish the prevalence and distribution of
MR findings associated with pelvic endometriosis in
patients with a MRI diagnosis of adenomyosis.
Materials and methods
Patients
Retrospective review of an MRI database from 1
January 2000 to 6 January 2003 identified 63 patients
(aged 32–54 years, mean 42 years) in whom the diag-
nosis of adenomyosis was included in the differential
diagnosis. Indications for MRI included menstrual
irregularities (like dysmenorrhoea, menorrhagia), pelvic
pain, infertility and clinically palpable uterus. IRB
(Institutional Review Board) approval and informed
consent was obtained in all patients. Retrospective
review by a subspecialty trained women’s imaging
radiologist of the pelvic MRI exams in these patients
found 59 patients who met strict MRI criteria (Reinhold’s
criteria) for adenomyosis (junctional zone (JZ) .12 mm
and sub endometrial T2 bright foci within abnormal
JZ) [8] (Figure 1a). These 59 patients served as the

The British Journal of Radiology, April 2006 303

 T T Zacharia and M J O’Neill

(a) (b)

(c)

Figure 1. 35-year old woman with posterior wall adenomyoma and coexisting right ovarian endometrioma. (a) Sagittal T2
weighted MR image (repetition time (TR) 5800/echo time (TE) 100) shows classical findings of adenomyosis involving the
posterior wall of the uterus. Ill defined adenomyoma with T2 bright focus within (arrow). (b) Coronal T2 weighted MR image (TR
5000/TE 125) shows right ovarian endometrioma with T2 shading (arrow). (c) Axial T1 weighted fat saturated MR image (TR 150/
TE 1.8) shows very hyperintense right ovarian endometrioma (arrow).

investigational study group. 20 (20/59) patients had should ideally have been a review of MR exams
severe symptoms. Hysterectomy was performed in all performed for the clinical indication of endometriosis,
these 20 patients. To serve as a control group there to evaluate co-existence of adenomyosis. However, a

304 The British Journal of Radiology, April 2006

Endometriosis coexisting with adenomyosis

converse study (detection of adenomyosis in patients
with MR diagnosis of endometriosis) was not performed
as part of our investigation.
MRI protocols
MRI was performed with 1.5 T signa scanners
(General Electric Medical Systems, Milwaukee, WI). T2
weighted images were acquired with 4 mm slices with
1 mm spacing in the sagittal, coronal, and axial planes
relative to the orientation of uterine cavity, using fast
(turbo) spin echo sequences in all three planes. T1
weighted spin echo and gradient echo images were
obtained both with and without fat saturation in all 59
patients. High resolution surface coils were used for data
acquisition and exams were completed in 30–45 min.
Image analysis
T1 weighted fat saturated and T2 weighted images in
these patients were then reviewed for the presence or
absence of T1 bright/T2 dark signal suggesting endome-
triosis in any of nine anatomical locations. Endometrial
implant has a varied appearance depending on the age of
associated blood products. Diagnosis of implant was
made when hyperintense lesions were detected on fat
saturated T1 weighted imaging. The diagnosis of
endometrial implant was made after ruling out artefacts
of motion, incomplete fat saturation and bowel contents.
Lesions larger that 10 mm were considered to represent
endometriomas (Figure 1b,c) and lesions less than
10 mm were classified as implants [9]. The sites (n59)
evaluated included uterine serosa, bilateral ovaries,
bilateral fallopian tubes, bilateral broad ligaments and
pelvic side walls. Fallopian tubes were distinguished
from broad ligament on coronal MR images as thin
tubular structures.
Data analysis
The reviewed data were collected on a Microsoft Excel
(Redmond, DC) spreadsheet. The data entered included
patient age, sex, MRI findings and sites of involvement.
The percentage of patients with coexistent endometriosis
and adenomyosis was calculated.
Results
Approximately one third of patients diagnosed to have
adenomyosis by MRI were found to have coexisting
endometriosis. 20 (20/59) patients with adenomyosis
(34%, 95% confidence interval (CI) 22–47%)) had coex-
isting endometriosis (Figure 2).

(a) (b)

Figure 2. 38-year-old woman with adenomyosis and serosal endometriosis. (a) Sagittal T2 weighted MR image (repetition time
(TR) 5800/echo time (TE) 100) shows junctional zone thickening (arrow) consistent with adenomyosis. (b) Sagittal gradient echo
T1 weighted fat saturated MR image (TR 150/TE 1.8) shows hyperintense focus (arrow) suggestive of serosal implant.

Table 1. Distribution of endometriotic lesions

Uterine serosa Ovary Broad Ligament Fallopian tubes Sidewall Total

Endometrioma 0 9 2 0 0 11
Implants 17 5 8 8 5 43
Total 17 14 10 8 5 54

The British Journal of Radiology, April 2006 305


Uterine serosa and ovaries were the most common
sites of involvement in our series (Table 1). Single site
involvement was noted in two patients and multiple sites
were involved in 18 patients. Extra ovarian involvement
was noted in all patients (n520).
Sites of endometriosis
The various sites of involvement studied were uterine
serosa, bilateral ovaries, broad ligaments, fallopian tubes
and the pelvic wall. A total of 54 sites of involvement
were identified (uterine serosa n517/54, ovaries n514/
54, broad ligaments n510/54, fallopian tubes n58/54,
pelvic side walls n55/54) in 20 patients with an average
of 2.7 sites per patient.
Implants
Implants (n543) were more common than endome-
triomas (n511) and formed 79% of the lesions. Implants
(lesions less than 10 mm) were most frequently seen over
the uterine surface (Figure 2b). Most of these lesions
involving uterine serosa were more than 2 mm thick
(Figure 1), but occasionally these lesions were subtle and
barely perceptible (Figure 2). Implants were seen on all
locations (uterine serosa n517, ovaries n55, broad
ligaments n58, fallopian tubes n58, pelvic side walls
n55).
Endometriomas
Endometriomas were less common than serosal
implants (n511) (20.4%). Endometriomas occurred most
often in the ovaries (ovaries n59, broad ligament n52).
‘‘T2 shading’’ was observed in some of these lesions
(Figure 1b). The diagnosis was obvious also on T1 fat
saturated sequence.
Histopathology
Histopathological confirmation was obtained in 20
patients (20/59) (34%) who underwent hysterectomy. All
of them had findings correlating with MRI. Seven (7/20)
patients had coexisting adenomyosis and endometriosis.
The presence of multiple patchy or diffuse areas of
endometrial glands and/or stroma associated with
surrounding muscle hypertrophy located more than
halfway from the endometrial-myometrial junction was
considered diagnostic of adenomyosis on histopathol-
ogy. Endometriosis was diagnosed when endometrial
tissue (endometrial gland and stroma) was found at
histopathological examination of resected specimen in
uterine serosa, ovaries, broad ligaments, fallopian tubes
or pelvic sidewalls.
Discussion
Common aetiological factors like abnormal migration,
metaplasia and catalase are implicated in the pathoge-
nesis of both endometriosis and adenomyosis [5–7].
Sampson’s theory of retrograde menstruation is the most
306
T T Zacharia and M J O’Neill
accepted theory postulated in the pathogenesis of
endometriosis and has been supported by convincing
circumstantial and experimental evidence [5]. Metaplasia
of serosal epithelium is another postulated theory in the
aetiology of endometriosis [7].
Adenomyosis is a fairly frequent disorder in adult
women characterized by infiltration of endometrial
glands and stroma within the myometrium of the uterus.
Recent experimental data have supported the hypothesis
that adenomyosis represents a condition of the uterine
body in which the stromal cells have a primary
pathogenetic role, although some contribution of accel-
erated epithelial down growth cannot be entirely
excluded [7]. This migration of endometrial epithelium
to the inner myometrium can act as an irritant focus and
cause smooth muscle hyperplasia.
There are a number of clinical and pathological
studies, which throw light on the coexistence of both
the diseases [1, 7, 10]. However, no imaging study exists
which has studied the coexistence of adenomyosis and
endometriosis. We sought to evaluate the patients with
adenomyosis who had coexisting serosal or ovarian
endometriosis on MRI.
MRI findings in endometriosis has been well docu-
mented in the literature. The endometriotic lesions are
typically hyperintense on T1 weighted fat saturated
sequence with ‘‘shading’’ on T2 weighted imaging [11–
13]. Identification of these lesions by MRI relies on
detection of haemorrhagic lesions.
Signal characteristics vary according to the age of the
haemorrhage. Typically, these lesions appear hyperin-
tense on T1 weighted spin echo images and hypointense
(shading) on T2 weighted fast spin echo images
(T2WFSE) due to the presence of deoxyhaemoglobin
and met haemoglobin. Acute haemorrhage occasionally
appears hypointense on T1WSE and T2WFSE sequences.
Old haemorrhage occasionally appears hyperintense on
T1WSE and T2WFSE images. T1 weighted fat saturation
sequence was performed in all our patients. The addition
of fat saturated T1 weighted imaging improved diag-
nostic accuracy in the evaluation of both endometriomas
and peritoneal disease by increasing lesion conspicuity
and differentiating lipid containing ovarian masses from
those containing blood. Recent studies have shown that
administration of gadolinium will enhance detection of
deep endometriotic implants.
The coexistence of endometriosis and adenomyosis
has got significant clinical implications. The purpose of
this investigation is to prove the association between
endometriosis and adenomyosis. The observation that
one third of patients with adenomyosis have coexisting
endometriosis should alert the radiologist to look for
endometriosis, which is quite often subtle, represented
by uterine serosal implants. The pelvic sidewall implants
can also be easily missed. In patients presenting with
infertility, the coexistence of these two conditions can
result in treatment failures if both the diseases are not
detected and treated. Hysterectomies can be avoided if
endometriosis can be diagnosed and treated by drugs
like progestin, GnRH (gonadotrophin releasing hor-
mone) analogues and Danazol [14, 15]. If symptomatic
relief is attained in this group of patients after medical
management of endometriosis, hysterectomy may not be
needed to treat adenomyosis.
The British Journal of Radiology, April 2006
Endometriosis coexisting with adenomyosis
Our findings have important implications in deciding
the MRI protocols too. Fat saturated T1 weighted
imaging should be routinely included in the pelvic
MRI protocol done for patients undergoing evaluation
for adenomyosis.
The most important drawback of our study is the lack
of pathological evidence in patients who did not have the
severe symptoms to undergo hysterectomy. Although
MRI has high specificity in the diagnosis of both these
disease entities, certain centres depend upon deep
myometrial biopsy for the diagnosis of adenomyosis
and laparoscopic biopsy for the diagnosis of endome-
triosis [16]. Another limitation is the absence of a control
group. There should have been a converse study (a
review of MR exams performed for endometriosis, to
evaluate associated adenomyosis) and this was not
performed as part of our investigation. Without rigorous
statistics, this association will be difficult to prove
absolutely. However, similar pathogenesis, histological
appearance and presentation in a similar age group
suggest that the association might be real.
Conclusion
MRI evidence of endometriosis was seen in one third
of patients diagnosed to have adenomyosis on MRI.
Serosal implants especially on the uterine surface were
the most common observation. Endometriomas were
also commonly noted. Imaging protocols should routi-
nely include T1 weighted fat saturated imaging
sequences in order to detect coexistent endometriosis in
patents undergoing pelvic MRI for the diagnosis of
adenomyosis.
References
1. Dumontier I, Roseau G, Vincent B, et al. Comparison of
endoscopic ultrasound and magnetic resonance imaging in
severe pelvic endometriosis. Gastroenterol Clin Biol
2000;24:1197–204.
2. Stratton P, Winkel C, Premkumar A, et al. Diagnostic
accuracy of laparoscopy, magnetic resonance imaging, and
The British Journal of Radiology, April 2006
histopathologic examination for the detection of endome-
triosis. Fertil Steril 2003;79:1078–85.
3. Bazot M, Cortez A, Darai E, Rouger J, Chopier J, Antoine
JM, et al. Ultrasonography compared with magnetic
resonance imaging for the diagnosis of adenomyosis:
correlation with histopathology. Hum Reprod
2001;16:2427–33.
4. Dueholm M, Lundorf E, Hansen ES, Sorensen JS, Ledertoug
S, Olesen F. Magnetic resonance imaging and transvaginal
ultrasonography for the diagnosis of adenomyosis. Fertil
Steril 2001;76:588–94.
5. Sampson JA. Benign and malignant endometrial implants
in the peritoneal cavity, and their relation to certain ovarian
tumors. Surg Gynecol Obstet 1924;38:287–311.
6. Ota H, Igarashi S, Sato N, Tanaka H, Tanaka T. Involvement
of catalase in the endometrium of patients with endome-
triosis and adenomyosis. Fertil Steril 2002;78:804–9.
7. Starzinski-Powitz A, Zeitvogel A, Schreiner A, Baumann R.
In search of pathogenic mechanisms in endometriosis: the
challenge for molecular cell biology. Curr Mol Med
2001;1:655–64.
8. Reinhold C, McCarthy S, Bret PM, Mehio A, Atri M,
Zakarian R, et al. Diffuse adenomyosis: comparison of
endovaginal US and MR imaging with histopathologic
correlation. Radiology 1996;199:151–8.
9. Manfredi R, Valentini AL. Magnetic resonance imaging of
pelvic endometriosis. Rays 1998;23:702–8.
10. Sharpe-Timms KL. Endometrial anomalies in women with
endometriosis. Ann NY Acad Sci 2001;943:131–47.
11. Gougoutas CA, Siegelman ES, Hunt J, Outwater EK. Pelvic
endometriosis: various manifestations and MR imaging
findings. AJR Am J Roentgenol 2000;175:353–8.
12. Umaria N, Olliff JF. Imaging features of pelvic endome-
triosis. Br J Radiol 2001;74:556–62.
13. Glastonbury CM. The shading sign. Radiology
2002;224:199–201.
14. Ylanen K, Laatikainen T, Lahteenmaki P, Moo-Young AJ.
Subdermal progestin implant (Nestorone) in the treatment
of endometriosis: clinical response to various doses. Acta
Obstet Gynecol Scand 2003;82:167–72.
15. Rotondi M, Labriola D, Rotondi M, Ammaturo FP, Amato
G, Carella C, et al. Depot leuprorelin acetate versus danazol
in the treatment of infertile women with symptomatic
endometriosis. Eur J Gynaecol Oncol 2002;23:523–6.
16. McLucas B, Perrella R, Adler L. Embolization for the
treatment of adenomyosis. AJR Am J Roentgenol
2002;178:1028–9.
307
 The British Journal of Radiology, 79 (2006), 308–314

Influence of post-treatment delay on the evaluation of the

response to focused ultrasound surgery of breast cancer by
dynamic contrast enhanced MRI
A KHIAT, PhD, D GIANFELICE, MD, M AMARA, MSc and Y BOULANGER, PhD

Département de Radiologie, Hôpital Saint-Luc du CHUM, 1058 St-Denis, Montreal, Quebec, H2X 3J4
Canada

ABSTRACT. The assessment of the effectiveness of MRI-guided focused ultrasound

surgery (MRIgFUS) of breast carcinomas can be performed by dynamic contrast-
enhanced magnetic resonance imaging (DCE-MRI) parameters which monitor the
presence of residual tumour. The aim of this study was to evaluate the effect of the
post-treatment delay on this assessment. DCE-MRI data were acquired immediately and
3–14 days after MRIgFUS treatment of 26 tumours (,7 days, n56; ¢ge;7 days, n520).
The percentage of residual tumour was determined histologically on the resected mass
and correlated with two DCE-MRI parameters: increase in signal intensity (ISI) and
Received 7 June 2005
positive enhancement integral (PEI). No correlation could be found between DCE-MRI Revised 15 August 2005
data acquired immediately after treatment and the percentage of residual tumour. Accepted 15 September
Good correlation coefficients were found for data acquired several days after 2005
treatment (ISI, r50.749; PEI, r50.778). However, they were higher when the post-
DOI: 10.1259/bjr/23046051
treatment time interval was 7 days or more (ISI, r50.962; PEI, r50.934). These results
suggest that a post-treatment delay of 7 days is necessary for the accurate assessment ’ 2006 The British Institute of
of the presence of residual tumour by DCE-MRI parameters. Radiology

Magnetic resonance imaging-guided focused ultra-
sound surgery (MRIgFUS) is a non-invasive thermother-
apeutic method allowing the elimination of different
types of lesions without incision, thus reducing the risk
of infection and scarring as well as diminishing the pain
considerably [1]. MRIgFUS studies have been reported
for patients with breast fibroadenomas [2], breast
carcinomas [3–6] and uterine fibroids [7, 8]. The ultra-
sound radiation produced by a transducer elevates the
temperature at a focal point and destroys the tissue by
protein denaturation and tissue necrosis. The repeated
application of the procedure leads to the ablation of
multiple points covering the entire lesion or the desired
part of the lesion. As the exact location of the focal points
may vary depending on the tissue composition through
which the focused ultrasound beam passes, monitoring
the effects of the focused ultrasound is done using a non-
invasive method such as MRI.
Once the MRIgFUS is performed, early detection of
residual disease is of crucial importance to evaluate the
response. Indeed, clinical examination is compounded
by necrosis and fibrosis which are often intermixed with
residual tumour, breast oedema, or inflammation.
Dynamic contrast-enhanced magnetic resonance imaging
(DCE-MRI) allows the monitoring of the presence of
tumour by demonstrating changes in the enhancement
Address correspondence to: Dr Y Boulanger, PhD, Département de
Radiologie, Hôpital Saint-Luc du CHUM, 1058 St-Denis, Montréal,
Québec, H2X 3J4 Canada.
Sources of support: InSightec, Inc. (Haifa, Israel and Dallas, TX) and
the Canadian Institutes of Health Research (Ottawa, ON, Canada).
curves following the injection of contrast agent [9, 10]. In
the breast, this technique has been applied to detect the
presence of tumours [11–13], to assess their grade [14, 15]
and to monitor the results of different types of treatment
[16–18]. Enhancing lesions on post-treatment DCE-MR
images are normally indicative of the presence of
significant residual tumour, but growing evidence
suggests that the situation may be more complicated
due to the fact that some benign processes such as
oedema, fibrosis, necrosis and inflammation can mimic
malignant contrast [19, 20]. Hence, the presence of
contrast enhancement alone is not exclusively specific
to residual cancer. An enhancing lesion indicates
abnormal tissue, but the shape of the enhancement curve
must be considered to identify the nature of this tissue.
Indeed, inspection of the lesion morphology and
enhancement profile shows that most malignancies have
an irregular border and display very rapid enhancement
and a distinct early washout phase in DCE-MRI. This
pattern can be used to distinguish malignant masses
from benign lesions or normal tissue which enhance and
washout slowly [21]. Since benign phenomena due to the
therapy such as necrosis, fibrosis, oedema and inflam-
mation can often occur and be responsible for abnormal
enhancement, the time interval between treatment and
post-treatment must be taken into account when analys-
ing DCE-MRI data to minimize the contribution of these
processes. In the case of conventional treatments such as
lumpectomy and conventional surgery, time intervals
ranging from 4 weeks to 12 months have been reported
to be necessary for a reliable assessment of residual
tumour by DCE-MRI [19, 20, 22, 23].

308 The British Journal of Radiology, April 2006

Effect of post-MRIgFUS delay on DCE-MRI evaluation of residual breast cancer

In our efforts to assess the value of the MRIgFUS
technique for the treatment of breast carcinomas, DCE-
MRI data were collected immediately and several days
after treatment, and then compared with the percentage of
residual tumour determined by histopathology of the
excised treated mass. In our previous report, data acquired
several days after treatment were presented for 17 patients
which demonstrated a strong correlation between DCE-
MRI parameters and the percentage of residual tumour [5].
However, for a few patients the correlation was inexplic-
ably poor. In the present study, DCE-MRI data acquired
both immediately and several days after treatment were
analysed for 25 patients (26 tumours) to determine the
effect of the time interval between MRIgFUS and post-
treatment MRI data acquisition.
Inclusion criteria were a life expectancy of at least
5 years and the possibility of receiving hormone replace-
ment or Tamoxifen therapy. Major exclusion criteria
were pregnancy or lactation, having previously under-
gone radiation or laser therapy, having received immu-
nosuppressive therapy and having been treated with
chemotherapy in the past 3 months. In addition, women
with a history of grand mal seizures, severe cerebrovas-
cular disease, haemolytic anaemia, or dialysis treatment;
women who were receiving anticoagulation therapy; and
women with breast implants were excluded. Patients
were also excluded if their lesion was outside the
margins of safety for ultrasound ablation, consisted only
of microcalcifications, or had extensive intraductal
components.

Patients and methods MRIgFUS

Patients
25 women aged between 45 years and 87 years (mean
age¡SD, 61.3 years¡11.0) fulfilling the inclusion/exclu-
sion criteria completed the protocol between March 2000
and June 2004. Each patient was diagnosed by core
breast biopsy with an invasive breast carcinoma of less
than 3.5 cm in diameter (Grade 1). The tumour volumes
determined by MRI ranged between 0.11 cm3 and
11.2 cm3 (Table 1). All tumour types were invasive
ductal carcinomas except for one. In one patient (patient
7), two tumours were treated. All patients signed an
informed consent form approved by the Scientific and
Ethics Committees of our institution prior to any exam or
MRIgFUS.
All MRI data were acquired with the GE 1.5 T MRI
unit (Signa; GE Medical Systems, Milwaukee, WI) using
a breast coil. The MRIgFUS treatments were performed
by different versions of the ExAblate system developed
by InSightec Inc. (Haifa, Israel and Dallas, TX). The
MRIgFUS system and procedure have been described in
detail previously [3]. Briefly, the patient is prone on the
focused ultrasound table inside the MRI magnet bore. An
acoustic ultrasound beam generated by a multielement
phased-array transducer focused on a point that was
thermally ablated. The location of the focal point was
verified by the thermally sensitive fast spoiled gradient-
echo MRI sequence [3]. Multiple adjacent focal points
covering the entire tumour were treated sequentially to
complete the FUS ablation.

Table 1. Lesion and exam parameters for breast tumours treated by MRI-guided focused ultrasound surgery (MRIgFUS)
Patient number Tumour type Original tumour Residual tumour Post-treatment
volume (cm3) volume (cm3) delay (days)

1 Invasive ductal carcinoma 9.36 8.42 7

2 Invasive ductal carcinoma 0.57 0.42 7
3 Invasive ductal carcinoma 5.18 2.59 7
4 Invasive ductal carcinoma 8.80 3.52 7
5 Invasive ductal carcinoma 7.80 3.12 7
6 Invasive ductal carcinoma 7.20 2.16 3
7 Invasive ductal carcinoma 0.88 0.20 7
8 Invasive ductal carcinoma 0.78 0.07 7
9 Invasive ductal carcinoma 1.85 0.18 3
10 Invasive ductal carcinoma 0.50 0.05 7
11 Invasive ductal carcinoma 3.42 0.27 3
12 Invasive ductal carcinoma 5.76 0.29 3
13 Invasive ductal carcinoma 9.60 0.48 7
14 Invasive ductal carcinoma 2.89 0.14 7
15 Invasive lobular carcinoma 0.70 0.035 14
16 Invasive ductal carcinoma 0.17 0.005 3
17 Invasive ductal carcinoma 1.30 0.03 7
18 Invasive ductal carcinoma 0.21 0.002 7
19 Invasive ductal carcinoma 4.80 0 7
20 Invasive ductal carcinoma 0.11 0 10
21 Invasive ductal carcinoma 0.75 0 3
22 Invasive ductal carcinoma 1.44 0 7
23 Invasive ductal carcinoma 0.28 0 7
24 Invasive ductal carcinoma 0.70 0 7
7 Invasive ductal carcinoma 11.2 0 7
25 Invasive ductal carcinoma 0.41 0 7

The British Journal of Radiology, April 2006 309


Post-treatment MRI
The MR images used in this study were recorded both
immediately and at a follow-up visit 3–14 days after the
MRIgFUS but before conventional surgery. Among the
25 treated patients, six had their post-treatment evalua-
tion 3 days after ultrasound ablation, 17 patients 7 days
after ultrasound ablation (18 tumours), one patient
10 days and one patient 14 days after ultrasound
ablation (Table 1). The timing of the post-treatment
examination depended on the availability of the MRI
scanner and on the scheduling of the surgery. DCE-MR
images were acquired before and after injection of the
MR contrast agent gadopentetate dimeglumine (Berlex
Canada Inc., Lachine, Canada; 0.1 mmol kg21 body
weight) using a three-dimensional fast spoiled gradient
echo (FSPGR) sequence with fat saturation (repetition
time (TR)56.4 ms; echo time (TE)52.4 ms; preparation
time522 ms; flip angle510 ˚, slice thickness54 mm; no
intersection gap; 256 6 128 points; number of
images532; acquisition time53.2 min).
Analysis of DCE-MRI data
To calculate the most enhancing pixel (1 mm2) within
the entire tumour in the DCE-MR images, a utility
program was written using MATLAB programming
(Mathworks, Natick, MA) operating on Linux software.
From a spherical ROI centred on the original tumour
centre and covering the entire tumour region in the
intensity vs phase image series, a Bézier curve was
constructed by picking the control points (phase) to
identify the most enhancing pixel. DCE-MRI parameters
at the most enhancing pixel were evaluated by the
FUNCTOOL version 2.5.36b software program (GE
Medical Systems, Fremont, CA) from time-signal inten-
sity curves. The dynamic contrast uptake at the most
enhancing point was evaluated by calculating two
parameters: percentage increase in signal intensity (ISI)
and the positive enhancement integral (PEI). ISI corre-
sponds to the percentage of the maximal increase in
signal intensity after the injection of contrast agent
(SIpost) relative to the pre-contrast signal intensity
(SIpre): ISI5(SIpost2SIpre)/SIpre 6 100. PEI is the
integral of the area under the enhancement curve
observed after the injection of contrast agent for the
time-signal intensity graph acquired for a time t:
Pt
PEI~ SIt . Data were analysed separately for all patients
0
taking into account the time interval between the
MRIgFUS and the DCE-MRI exam.
Surgery and histopathological evaluation
Patients underwent a routine segmental tumour
resection 3–21 days after MRIgFUS, which included
resection of wide margins around the area treated by
focused ultrasound. Three-dimensional macroscopic and
microscopic histopathological measurements were per-
formed on the resected mass by a pathologist immedi-
ately after surgery [3, 24]. The total tumour volume, the
volume of tumour in the treated zone and the volume of
310
A Khiat, D Gianfelice, M Amara and Y Boulanger
necrosed tumour were determined. Volumes (V) were
determined by measuring the largest tumour dimensions
in each axis (i.e. distances a, b, and c) and performing an
ellipsoid volume calculation: V5(p/6) abc.
Statistical analysis
Correlation coefficients between the DCE-MRI para-
meters ISI and PEI, and the percentage of residual
tumour volume determined by histopathology were
determined by a Pearson analysis. Correlations were
considered significant at p,0.05 with r values . 0.50.
Comparison of the correlation coefficients was per-
formed using a t-test [25]. A receiver operating char-
acteristic (ROC) curve analysis was performed for both
DCE-MRI parameters. Statistical analyses were per-
formed using the SPSS software version 9.0.1 (SPSS
Inc., Chicago, IL).
Results
Processing of DCE-MRI data
Our home written program allowed rapid and auto-
matic determination of the most enhancing point in the
DCE-MR images acquired at the post-treatment exam.
This method was found to be reliable and exactly
reproducible on repeated measurements.
Histopathological data
Histopathology results of the excised mass following
the MRIgFUS of breast carcinomas show a variable
degree of success, as summarized in Figure 1 and
Table 1. Among the 26 treated tumours, seven tumours
showed no detectable residual cancer at the site of
MRIgFUS, 11 tumours had residual cancer below 10%,
and seven tumours showed a larger percentage of
residual cancer, which ranged between 20% and 90%.
DCE-MRI results
In general, complete tumour necrosis should be
reflected by the absence of enhancement in the time-
signal intensity DCE-MRI curves. Figure 2 presents two
types of time-signal intensity curves before, immediately
after and several days after MRIgFUS observed for
patients 19 (Figure 2a–c) and 21 (Figure 2d–f). Before
treatment, a strong enhancement is observed in all cases
at the most enhancing point of the breast tumour region
(Figure 2a,d). In the case of patient 19, the tumour
necrosis determined by histopathology was complete
(Figure 1, Table 1) and no apparent enhancement was
observed 7 days after treatment (Figure 2c). However,
this assessment could not have been made immediately
after treatment since a slowly enhancing curve was
observed (Figure 2b). When residual cancer was present,
as was the case for patient 21, a significant enhancement
(Figure 2f) was observed with a steep washout pattern
similar to the patterns observed before treatment
The British Journal of Radiology, April 2006
Effect of post-MRIgFUS delay on DCE-MRI evaluation of residual breast cancer

(Figure 2d). In that case, the enhancement patterns were

very similar immediately and 3 days after MRIgFUS
(Figure 2e,f).

Correlation between DCE-MRI and histopathology

Figure 1. Graph showing the percentage of residual tumour
determined by histopathology following MRI-guided
focused ultrasound surgery (MRIgFUS) for 26 breast tumours
in 25 patients. A variable degree of success was achieved at
various stages of development of the technique.
data
Figure 3a,b presents the correlations between the ISI
parameter and the percentage of residual tumour
volume determined by histopathology for all tumours
and for the 20 tumours evaluated at least 7 days after the
MRIgFUS, respectively. Similar correlations are pre-
sented in Figure 3c,d between the PEI parameter and
the percentage of residual tumour volume determined
by histopathology for all patients and for the patients
having been evaluated at least 7 days post-treatment,
respectively. Strong correlations between the DCE-MRI
parameters ISI (r50.749) and PEI (r50.778), and the
percentage of residual tumour volume were found.
When six patients who had their evaluation only 3 days
after MRIgFUS were excluded from the analysis, the
correlations were much stronger, the correlation coeffi-
cients for ISI increasing to 0.962 (Figure 3b) and for PEI to
0.934 (Figure 3d). Using a t-test to compare the

Figure 2. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) signal intensity curves as a function of image
number (6.0 s/image) at the maximally enhancing pixel for patient 19 (a) before, (b) immediately after and (c) 7 days after MRI-
guided focused ultrasound surgery (MRIgFUS), and for patient 21 (d) before, (e) immediately after and (d) 3 days after MRIgFUS.
The absence of signal enhancement in (c) is consistent with the absence of residual tumour and the progressive signal
enhancement in (b), (e) and (f) is attributable to benign tissue phenomena occurring in the first few days following focused
ultrasound ablation.

The British Journal of Radiology, April 2006 311


correlation coefficients for the ISI and PEI parameters for
all values and for values obtained more than 7 days after
treatment [25], a statistically significant difference was
found for ISI (p,0.01) and an almost significant
difference for PEI (p50.055).
Due to the presence of enhancement in practically all
cases, no correlation could be found between the ISI and
PEI parameters and residual tumour volume for data
acquired immediately after treatment.
ROC curve analysis
Receiver operating characteristic (ROC) curves are
presented in Figure 4 for the ISI parameter. When the
data for all tumours were used, the area under the curve
was 0.853 with a p-value of 0.005 for asymptotic
significance (Figure 4a). The same analysis was made
for the 20 tumours whose DCE-MRI data were recorded
7 days or more after treatment, which yielded an area
under the curve of 0.986 with a p-value of 0.001 for
asymptotic significance (Figure 4b). The cut off ISI value
for both curves corresponded to 7% with a sensitivity of
100% and a specificity of 83%. The ROC curve analysis
for the PEI parameter showed an area under the curve of
0.732 without asymptotic significance (p50.067) for all
tumours, and an area under the curve of 0.785 without
asymptotic significance (p50.055) for patients whose
DCE-MRI exam was performed 7 days or more after
treatment. The cut off value for PEI was 4% with a
sensitivity of 67% and a specificity of 83%.
Discussion
The calculation of the ISI and PEI parameters from
DCE-MR images requires an accurate determination of
312
A Khiat, D Gianfelice, M Amara and Y Boulanger
Figure 3. Graphs showing the correla-
tion between parameters calculated
from the post-treatment dynamic con-
trast-enhanced magnetic resonance ima-
ging (DCE-MRI) data and the percentage
of residual viable cancer determined by
histopathology on the excised mass for
breast tumours treated by MRI-guided
focused ultrasound surgery (MRIgFUS):
increase in signal intensity (ISI) for (a) all
patients (r50.749) and (b) patients who
had their post-treatment evaluation at
least 7 days after treatment (r50.962);
positive enhancement integral (PEI) for
(c) all patients (r50.778), and (d) patients
who had their post-treatment evaluation
at least 7 days after treatment (r50.934).
Points corresponding to patients whose
post-treatment was performed 3 days
after MRIgFUS are shown as asterisks (*)
in (a) and (c).
the most enhancing point in the tumour area. The
strategy of using the automatically identified most
enhancing pixel was based on the conclusions of
previous work clearly demonstrating improved results
with this strategy compared with using large regions of
interest [26]. This was achieved by using a home written
program which allowed rapid identification of this point,
thus significantly reducing the time for analysis com-
pared with the manual location of this point used in our
previous report [5]. This study has also the advantage of
using volumes for histological quantitation of tissue. In
previous studies, diameter values were utilized. As
tumour and non-tumour masses are generally not
spherical, the volume calculation leads to increased
accuracy.
The analysis of DCE-MRI data recorded immediately
after performing MRIgFUS showed enhancement pat-
terns in practically all cases and no correlation could be
found between the ISI or PEI parameters and the
percentage of residual tumour determined by histo-
pathology. These enhancements are attributable to
benign phenomena such as haemorrhage, fibrosis,
inflammation or oedema caused by the treatment and
which prevent accurate assessment of the percentage of
residual tumour by DCE-MRI.
Analysis of the post-treatment DCE-MRI data acquired
3–14 days after MRIgFUS of small breast tumours
revealed that the ISI and PEI parameters correlate
strongly with the percentage of residual tumour deter-
mined by histopathology. The correlation coefficients for
the two parameters are very similar (0.749 for ISI and
0.778 for PEI) based on data from 26 treated tumours
whose volume of residual tumour determined by
histopathology varied between 0 and 90%. When data
for six tumours analysed less than 7 days after treat-
ment were not considered, the correlation coefficients
The British Journal of Radiology, April 2006
Effect of post-MRIgFUS delay on DCE-MRI evaluation of residual breast cancer
significantly improved, reaching 0.962 for ISI and 0.934
for PEI. These results stress the importance of the time
interval between MRIgFUS therapy and post-treatment
evaluation. In previous reports, the importance of the
time interval between treatment and DCE-MRI was
discussed for invasive treatments of breast lesions [19,
20, 22, 23]. In a study of 68 patients whose breast lesions
were treated by lumpectomy, an interval of 28 days after
treatment was recommended before the DCE-MRI
examination [23]. When conventional surgery was
performed, time intervals of 9–12 months were necessary
to obtain reliable data [19, 20, 22]. As conventional
surgical treatments cause a much larger perturbation of
the tissue, the fact that the interval necessary is much
longer than for MRIgFUS is consistent with the differ-
ence in severity between the two procedures.
For some patients the correlation between DCE-MRI
parameters and percentage of residual tumour volume
was especially poor, such as for patient 21 whose time-
signal intensity curves displayed abnormally high
enhancement leading the ISI and PEI parameters to
show as outliers on the correlation curves with the
percentage of residual tumour. This unexpected
enhancement is attributable to the fact that residual
tissue is not residual tumour, but rather consists of
enhancing benign tissue such as inflammation, necrosis,
fibrosis or oedema at the site of MRIgFUS. It is known
that the post-treatment contrast enhancement resulting
from such phenomena limits the positive predictive
value and specificity for the diagnosis of residual disease
[23, 27]. In our case, two arguments support this
explanation. The first argument is that patient 21 was
evaluated only 3 days after the MRIgFUS. When the six
patients who had their evaluation only 3 days after
treatment were excluded from the analysis, the results
were more reliable as assessed by higher correlation
coefficients and larger areas under the ROC curves. The
second argument in favour of non-tumour tissue for
patient 21 pertains to the shape of the time-signal
intensity curve at the post-treatment examination. The
pattern of this curve does not correspond to the expected
curve profile of a malignant tissue [21]; it displays a slow
rate of enhancement in comparison with the time-signal
intensity enhancement curve at the pre-treatment exam
and it continues to rise over the course of the examina-
tion. Malignancies typically display an irregular border
at the histological analysis; they enhance very rapidly
The British Journal of Radiology, April 2006
Figure 4. Receiver operating character-
istic curves for the increase in signal
intensity (ISI) parameter calculated from
the dynamic contrast-enhanced mag-
netic resonance imaging (DCE-MRI) data
for (a) 26 tumours analysed 3–14 days
after the MRI-guided focused ultrasound
surgery (MRIgFUS) and (b) 20 tumours
analysed 7–14 days after the MRIgFUS.
The areas under the curves were (a) 0.853
and (b) 0.986 and the p-values for
asymptotic significance were (a) 0.005
and (b) 0.001. A significant improvement
is obtained when only data for patients
examined 7 days or more after MRIgFUS
are considered.
and display a distinct early washout phase in the DCE-
MRI curves. For breast cancers, Figure 2a,c represents a
typical pattern. Previous studies have concluded that the
presence of contrast enhancement alone is not specific for
residual cancer, and that the shape of the enhancement
curve is a very important factor to consider [21, 28, 29].
Additional consideration of the morphology and internal
architecture of the excised mass and a critical analysis of
the time-signal intensity curves can be used to help
distinguish residual tumours from benign processes. In a
previous publication [3], histological slides were pre-
sented for a patient analysed 3 days after MRIgFUS
which clearly show these phenomena. Obviously, these
benign processes depend on the sensitivity of each
patient, and are strongly influenced by the time interval
between MRIgFUS and post-treatment evaluation.
Consequently, the accuracy of DCE-MRI for predicting
residual disease is affected by these processes.
Comparison of the ROC curves for all patients and for
the patients examined 7 days and more after MRIgFUS
shows a significant improvement with an area under the
curve of the ISI parameter increasing from 0.853 to an
almost perfect 0.986 for the latter group. The same trend
was observed for the PEI parameters although lower
values were obtained. These results confirm the impor-
tance of the post-treatment time interval and suggest that
the ISI parameter is a better parameter to evaluate
treatment outcome. The cut-off value for the determina-
tion of residual tumour was around 7% for ISI and 4% for
PEI.
The main limitation of this study is the low number of
patients with a high percentage of residual tumour.
Although our group has reported the largest number of
patients treated by MRIgFUS, the accuracy of our
correlations could be improved if more patient data
were available. Most of our patients showed less than
10% residual tumour. At low percentages, the accuracy
of the technique is reduced. This situation is unlikely to
improve as the treatment technique is increasingly
successful. Nonetheless, our results demonstrate a high
degree of specificity and sensitivity for the DCE-MRI
parameters.
In summary, our results demonstrate that DCE-MRI
parameters such as ISI and PEI are both useful in
determining the effectiveness of MRIgFUS in breast
carcinomas, provided the measurements are performed
sufficiently long after treatment, i.e. approximately
313

7 days, based on our data. Benign phenomena caused by
the treatment are responsible for abnormal enhancement
in the first few days. The accuracy of our analysis could
be improved by increasing the sample size and by
evaluating patients longitudinally at different time
intervals after MRIgFUS.
Acknowledgments
The authors thank Mr Martin Ruel, BSc, Dr Assia
Belblidia, MD, and the radiology technologists at Hôpital
Saint-Luc du CHUM for data acquisition. They are
grateful to Dr Sharon Thomsen, MD, for assistance with
histopathological analyses. The financial support from
InSightec, Inc., Haifa, Israel and Dallas, TX and from the
Canadian Institutes of Health Research, Ottawa, Canada
is acknowledged.
References
1. Moonen CTW, Quesson B, Salomir R, Vimeux FC, de Zwart
JA, van Vaals JJ, et al. Thermal therapies in interventional
MR imaging - focused ultrasound. Neuroimaging Clin N
Am 2001;11:737–47.
2. Hynynen K, Pomeroy O, Smith DN, Huber PE, McDannold
NJ, Kettenbach J, et al. MR imaging-guided focused
ultrasound surgery of fibroadenomas in the breast: a
feasibility study. Radiology 2001;219:176–85.
3. Gianfelice D, Khiat A, Amara M, Belblidia A, Boulanger Y.
MR imaging-guided focused US ablation of breast cancer:
histopathologic assessment of effectiveness – initial experi-
ence. Radiology 2003;227:849–55.
4. Gianfelice D, Khiat A, Amara M, Belblidia A, Boulanger Y.
Feasibility of magnetic resonance imaging-guided focused
ultrasound surgery as an adjunct to tamoxifen therapy in
high-risk surgical patients with breast carcinoma. J Vasc
Interv Radiol 2003;14:1275–82.
5. Gianfelice D, Khiat A, Amara M, Belblidia A, Boulanger Y.
MR imaging-guided focused ultrasound surgery of breast
cancer: correlation of dynamic contrast-enhanced MRI with
histopathologic findings. Breast Cancer Res Treat
2003;82:93–101.
6. Huber PE, Jenne JW, Rastert R, Simiantonakis, I, Sinn H-P,
Strittmatter H-J, et al. A new noninvasive approach in
breast cancer therapy using magnetic resonance imaging-
guided focused ultrasound surgery. Cancer Res
2001;61:8441–7.
7. Stewart EA, Gedroyc WM, Tempany CM, Quade BJ, Inbar
Y, Ehrenstein T, et al. Focused ultrasound treatment of
uterine fibroid tumors: safety and feasibility of a noninva-
sive thermoablative technique. Am J Obstet Gynecol
2003;189:48–54.
8. Tempany CM, Stewart EA, McDannold N, Quade BJ, Jolesz
FA, Hynynen K. MR imaging-guided focused ultrasound
surgery of uterine leiomyomas: a feasibility study.
Radiology 2003;226:897–905.
9. Padhani AR, Husband JE. Dynamic contrast-enhanced MRI
studies in oncology with an emphasis on quantification,
validation and human studies. Clin Radiol 2001;56:607–20.
10. Hayes C, Padhani AR, Leach MO. Assessing changes in
tumour vascular function using dynamic contrast-enhanced
magnetic resonance imaging. NMR Biomed 2002;15:154–63.
11. Knopp MV, Weiss E, Sinn HP, Mattern J, Junkermann H,
Radeleff J, et al. Pathophysiologic basis of contrast
enhancement in breast tumors. J Magn Reson Imaging
1999;10:260–6.
12. Hulka CA, Edmister WB, Smith BL, Tan L, Sgroi DC,
Campbell T, et al. Dynamic echo-planar imaging of the
314
A Khiat, D Gianfelice, M Amara and Y Boulanger
breast: experience in diagnosing breast carcinoma and
correlation with tumor angiogenesis. Radiology
1997;205:837–42.
13. Gibbs P, Liney GP, Lowry M, Kneeshaw PJ, Turnbull LW.
Differentiation of benign and malignant sub-1 cm breast
lesions using dynamic contrast enhanced MRI. Breast
2004;13:115–21.
14. Daldrup H, Shames DM, Wendland M, Okuhata Y, Link
TM, Rosenau W, et al. Correlation of dynamic contrast-
enhanced MR imaging with histologic tumor grade:
comparison of macromolecular and small-molecular con-
trast media. AJR Am J Roentgenol 1998;171:941–9.
15. Rydland J, Bjønerud A, Haugen O, Torheim G, Torres C,
Kvistad KA, et al. New intravascular contrast agent applied
to dynamic contrast enhanced MR imaging of human breast
cancer. Acta Radiol 2003;44:275–83.
16. Tsuboi N, Ogawa Y, Inomata T, Yoshida D, Yoshida S,
Moriki T, et al. Changes in the findings of dynamic MRI by
preoperative CAF chemotherapy for patients with breast
cancer of stage II and III: pathologic correlation. Oncol Rep
1999;6:727–32.
17. Drew PJ, Kerin MJ, Mahapatra T, Malone C, Monson JR,
Turnbull LW, et al. Evaluation of response to neoadjuvant
chemoradiotherapy for locally advanced breast cancer with
dynamic contrast-enhanced MRI of the breast. Eur J Surg
Oncol 2001;27:617–20.
18. Martincich L, Montermurro F, De Rosa G, Marra V,
Ponzone R, Cirillo S, et al. Monitoring response to primary
chemotherapy in breast cancer using dynamic contrast-
enhanced magnetic resonance imaging. Breast Cancer Res
Treat 2004;83:67–76.
19. Soderstrom CE, Harms SE, Farrell RS Jr, Pruneda JM,
Flamig DP. Detection with MR imaging of residual tumor in
the breast soon after surgery. AJR Am J Roentgenol
1997;168:485–8.
20. Viehweg P, Heinig A, Lampe D, Buchmann J, Heywang-
Köbrunner SH. Retrospective analysis for evaluation of the
value of contrast-enhanced MRI in patients treated with
breast conservative therapy. MAGMA 1998;7:141–52.
21. Kuhl CK, Mielcareck P, Klaschik S, Leutner C, Wardelmann
E, Gieseke J, et al. Dynamic breast MR imaging: are signal
intensity time course data useful for differential diagnosis
of enhancing lesions? Radiology 1999;211:101–10.
22. Heywang-Köbrunner SH, Schlegel A, Beck R, Wendt T,
Kellner W, Lommatzsch B, et al. Contrast-enhanced MRI of
the breast after limited surgery and radiation therapy. J
Comput Assist Tomogr 1993;17:891–900.
23. Frei KA, Kinkel K, Bonel HM, Lu Y, Esserman LJ, Hylton
NM. MR imaging of the breast in patients with positive
margins after lumpectomy: influence of the time interval
between lumpectomy and MR imaging. AJR Am J
Roentgenol 2000;175:1577–84.
24. Thomsen S. Mapping of thermal injury in biologic tissues
using quantitative pathologic techniques. Proc SPIE
1999;3594:82–95.
25. Sachs L. Applied statistics: a handbook of techniques. 2nd
edn. New York, NY: Springer-Verlag, 1984.
26. Liney GP, Gibbs P, Hayes C, Leach MO, Turnbull LW.
Dynamic contrast-enhanced MRI in the differentiation of
breast tumors: user-defined versus semi-automated region-
of-interest analysis. J Magn Reson Imaging 1999;10:945–9.
27. Orel SG, Reynolds C, Schnall MD, Solin LJ, Fraker KL,
Sullivan DC. Breast carcinoma: MR imaging before re-
excisional biopsy. Radiology 1997;205:429–36.
28. Orel SG, Schnall MD, LiVolsi VA, Troupin RH. Suspicious
breast lesions: MR imaging with radiologic-pathologic
correlation. Radiology 1994;190:485–93.
29. Heywang-Köbrunner SH. Contrast-enhanced M.R.I. of the
breast. 2nd edn. Berlin, Heidelberg: Springer-Verlag, 1995.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 315–318

Introduction of grids to mobile ICU radiography in a teaching

hospital
D W ANDERSON, MBBS (hons), FRANZCR

Monash Medical Centre, 246 Clayton Road, Clayton, Melbourne, Victoria, Australia 3168

ABSTRACT. The purpose of this study was to review the change in image quality before
and after introducing grid use routinely to our mobile X-ray service. This was studied in
the intensive care unit (ICU) setting, comparing images obtained over a 2 week period
prior to and after the introduction of the change in technique. We introduced a 6:1
Received 18 March 2005
grid with appropriate changes in exposure factors. No other alterations were made. Revised 7 September 2005
There were 133 patients in the preliminary group and 196 patients in the post-grid Accepted 15 September
group. We found a reduction in the proportion of images that were of non-diagnostic 2005
or barely diagnostic quality. Non-diagnostic examinations were reduced from 18% to
DOI: 10.1259/bjr/66912494
1%. Introducing grids to our mobile service resulted in improvement in image
diagnostic quality, largely by reducing the proportion of poor and unacceptable quality ’ 2006 The British Institute of
images. This effect does not appear to have been documented in the literature. Radiology

Monash Medical Centre (MMC) is a large teaching
hospital in suburban Melbourne, Victoria. We have
recently introduced grids to our mobile ward service.
Data were collected before and after the introduction of
grids to assess their impact on the diagnostic quality of
images obtained on the daily morning intensive care unit
(ICU) film round.
N Inability to follow the course of central lines, catheters,
drain tubes and other medical devices.
N Inability to identify the tip location of the above lines,
catheters and devices.
N Respiratory or other motion artefact of severe nature.
N Grid cut-off or other grid artefact of severe nature.
N Non-inclusion of the complete extent of the lung
parenchyma.

Methods Also contributing to image grading was a subjective

All patients examined during the morning ICU X-ray
round were collected for a 2 week period prior to the
introduction of a grid technique. Images were obtained
using GE AMX4 mobile machine and Agfa CR system.
1 month following the changed technique, data were
again collected for a 2 week period. Images were
obtained using a 6:1 grid landscape or portrait grid with
GE AMX4 mobile machine and Agfa CR system. Both
sets of images were viewed on BARCO monitors. Data
for the 4 week period following the introduction of the
new technique were not collected to allow staff to
become familiar with changed techniques, particularly
exposure alterations. This was thought desirable for
repeatable and reliable patient examinations. The images
were graded for diagnostic quality using a five point
grading system as follows:
assessment of lung parenchymal detail, both the ability
to satisfactorily characterize the nature and extent of
infiltrates as well as the ability to resolve normal
parenchymal architecture in areas of non-pathological
lung. The components of our scoring system are similar
to those used by others in the assessment of portable
radiographs [1–3].
Images were graded independently by two experi-
enced radiologists, each with more than 2 years
experience with the display system used. Images were
modified for window settings by each viewer to their
satisfaction with active changes made during review of
each image to fully interrogate each image to the
satisfaction of the reviewers. Differences in grade were
resolved by consensus.

(1) not of diagnostic quality

(2) poor, barely adequate diagnostic quality
(3) fair, acceptable diagnostic quality
(4) good, above average diagnostic quality
(5) excellent diagnostic quality

Images were determined ‘‘not of diagnostic quality’’ if

any of the following were present:

N Inability to see the thoracic spine through the cardiac Figure 1. Comparison chest radiograph (CXR) pre- and
shadow. post-grid.

The British Journal of Radiology, April 2006 315

 D W Anderson

A blinded interpretation of pre- and post-grid images

was not attempted as the use of a grid results in
discernable grid lines, rendering overt identification of
technique used.

Results
The initial image set performed before grid technique
(total 133 patients) produced the following results:

Grade 1: 24/133 (18%)

Grade 2: 55/133 (41%)
Grade 3: 40/133 (30%)
Grade 4: 13/133 (10%)
Grade 5: 1/133 (1%)

The second image set, performed following the

introduction of grid technique (total 196 patients)
produced the following results: Figure 3. Same patient as Figure 2 with grid.

Grade 1: 2/196 (1%)

Grade 2: 30/196 (15%)
Grade 3: 104/196 (53%)
Grade 4: 48/196 (25%)
Grade 5: 12/196 (6%)

The difference in group size was the result of

differences in patient load in the ICU for the morning
radiology service.
Figures 2–5 are comparative chest X-rays performed
on the same day with and without grids on two separate
patients.

Discussion
The initial data set produced a somewhat surprising
result in that nearly one in five images (18%) was not of
diagnostic quality. This was largely the result of not
being able to identify the position or tip of various
devices. Swan-ganz catheters and fine bore central lines Figure 4. Chest radiograph no grid.

Figure 2. Chest radiograph no grid. Figure 5. Same patient as Figure 4 with grid.

316 The British Journal of Radiology, April 2006

Introduction of grids to mobile ICU radiography
were most often affected. There was also a contribution
in many images from difficulties in reliably characteriz-
ing infiltrates. This was most notable in larger patients.
In combination with barely acceptable images (category
2, 41%), our ICU mobile service was producing images of
poor to unacceptable quality in nearly six out of 10
patients. Portable images had been acknowledged as of
lesser quality than imaging performed in department.
This had been considered a largely inevitable outcome
given the compromises involved in portable imaging.
The desire to investigate the impact of grids was
prompted by their effect when used on a very large
patient in the ICU in attempts to produce a diagnostic
image. That grid use can result in a significant improve-
ment in signal to noise ratio (SNR) at the cost of
increased dose is well recognized [4–6].
Our service had not conducted a formal review of
image diagnostic quality in any systematic fashion for
some years, instead relying on an ad-hoc notification and
correction of specific problems as they arose. Review had
not been performed since the introduction of our film-
less screen based CR reporting system. As such, we do
not have a prior benchmark for comparison with these
results.
The data set following introduction of grid technique
shows a significant reduction in the proportion of images
deemed not of diagnostic quality (18%.1%). There was
also a significant reduction in category two images
(41%.15%). As a result of these improvements, 78% of
images were deemed of acceptable or above average
diagnostic quality (53% category 3 + 25% category 4).
Review of the literature shows little has been published
quantifying the change in diagnostic viewing quality of
mobile ICU images performed with and without grids.
Two authors have stated that portable chest radiography
is significantly improved with use of a 6:1 or 8:1 grid [7,
8], but without comparative evidence.
Introduction of grids has also introduced some new
problems. All images contain discernable grid lines.
Whilst visible, these were not thought to impede
satisfactory image interpretation. Grid ‘‘cut-off’’ produ-
cing asymmetrical density was seen on approximately
30% of films. This was of variable severity but again did
not appear to significantly affect image interpretability.
No image was deemed category 1 as a result of grid cut-
off. The finding that grid cut-off was not a significant
problem is in keeping with the work of Grunert et al and
Ciccotosto et al [2, 5, 9].
A grid technique results in an increased radiation dose
to the patient. Rill et al stated that use of a grid resulted
in a 400–600% increase in dose to their phantom for a 60–
300% improvement in SNR [4]. Increasing tube potential
from 80 kVp to 100 kVp resulted in further 10%
increases in SNR and phantom dose. Similar dose
increases and decreases in scatter radiation were found
by Floyd et al [6]. Balanced against this is the critical
status of the typical ICU patient with a requirement for
the most accurate diagnostic information achievable.
Most patients in our ICU are older adults with multiple
co-morbidities. As such, the increased dose is acceptable.
No attempt was made to assess the clinical relevance of
changes to subjective image quality in this study as this
lay outside our scope.
The British Journal of Radiology, April 2006
The intent of this article is to provide a small measure
of our experience with introducing grids to our ICU
mobile service and their impact on subjective radiologist
assessed image quality, and the following limitations are
acknowledged:
(1) The sample sizes are relatively small
(2) No attempt was made to correlate image quality
with patient size (body mass index, BMI)
(3) No attempt was made to determine the effect of
radiographer experience on image quality. MMC is
a teaching hospital with staff of significantly
varying experience
(4) No attempt has been made to determine if the
subjective improvement in image quality has had
any relevant effect on clinical decision making or
patient outcome
(5) The study is unable to be performed in a blinded
fashion due to image effects resulting from the
intervention (grid use)
(6) The differences in sample size (patient load), may
reflect further unknown confounding factors in the
patients with independent effect on image quality.
We made no attempt at randomization
(7) The introduction of a change in technique may
have resulted in a change in the diligence of
radiography staff regarding positioning, exposures
and other factors
Optimizing image diagnostic quality is relevant in the
ICU setting as other authors have found unexpected
findings in 37–65% of patients [10–13]. Ekenmeyer et al
stated changes were initiated in diagnostic approach or
therapy as a result of CXR findings alone in 27% of
patients [11]. Malpositioned ETT, central lines or
catheters have been seen in 9–20% of patients [11, 13].
An American study from 1994 showed 12% of hospitals
using grids for mobile chest radiography. It was not stated
if this was routine or on a patient by patient basis [14].
Review of technique used by the other large public
hospitals in our city found only one of four institutions
were using grids for mobile ICU radiography and this was
reserved for very large patients only. Reasons given
included inconvenience and historical practice.
Conclusion
Presented is our experience with introducing grids to
our ICU mobile service. We found an improvement in
overall image quality, largely as a result of reduced
numbers of poor and unacceptable quality images. This
effect has not been documented in the literature to our
knowledge. From this preliminary work, further direc-
tions include a randomized study to confirm the positive
impact of grids on image quality and an assessment of
the impact of improved image quality with grid
technique on clinical decision making in the ICU setting.
References
1. Langevin PB, Hellein V, Harms SM, Tharp C, Cheung-
Seekit C, Lampotang S. Synchronization of radiograph film
317

exposure with the inspiratory pause. Am J Respir Crit Care
Med 1999;160:2067–71.
2. Ciccotosto C, Storto ML, Guidotti A, Ferrante R, Bonomo L.
Bedside thoracic radiography: a comparison between 3
different types of grid. Radiol Med (Torino) 1994;87:127–33.
3. Rottenberg GT, Chin RJ, Allen CM, Maggs P, Kessel D.
Portable chest radiology in intensive care: a comparison of a
new dual characteristic film-screen system (Insight) incor-
porating a flexible grid with a standard film-screen system.
Clin Radiol 1996;51:494–8.
4. Rill LN, Brateman L, Arreola M. Evaluating radiographic
parameters for mobile chest computed radiography: phan-
toms, image quality and effective dose. Med Phys
2003;30:2727–35.
5. Grunert JH, Boy B, Groenwold SK, Krahn-Peters V,
Hendrickx P. Use of multi-grid screens in intensive care
units. Aktuelle Radiol 1995;5:293–6.
6. Floyd CE Jr, Chotas HG, Ravin CE. Scatter-reduction
characteristics of an infinity-focussed gridded radiographic
cassette. Invest Radiol 1994;29:852–5.
7. BHR 1980; HHS Publ 80–8124: 202–6.
318
D W Anderson
8. Rossi RP, Harnisch BD, Hendee WR. Evaluation of an
automatic exposure control device for mobile radiography.
Radiology 1982;145:823–7.
9. Grunert JH, Boy B, Busche D, Groenewold SK, Herrmann
H, Krahn-Peters V, et al. Grids and high kilo-volt-peak-
setting in bedside chest radiographic examinations. JBR-
BTR 2000;83:296–9.
10. Janower ML, Jennas-Nocera Z, Mukai J. Utility and efficacy
of portable chest radiographs. AJR Am J Roentgenol
1984;142:265–7.
11. Ekenmeyer WB, Crapo RO, Calhoon S, Cannon CY, Clayton
PD. Efficacy of chest radiography in a respiratory intensive
care unit. A prospective study. Chest 1985;88:691–6.
12. Greenbaum DM, Marschall KE. The value of daily chest x-
rays in intubated patients in the medical intensive care unit.
Crit Care Med 1982;10:29–30.
13. Henschke CI, Pasternak GS, Schroder S, Hart KK, Herman
PG. Bedside chest radiography: diagnostic efficacy.
Radiology 1983;149:23–6.
14. Wandtke JC. Bedside chest radiography. Radiology
1994;190:1–10.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 319–325

Colour Doppler ultrasound patterns and clinical follow-up of

incidentally found hypoechoic, vascular tumours of the spleen:
evidence for a benign tumour
1 1 1
C GÖRG, MD, K GÖRG, MD, T BERT, MD and 2P BARTH, MD

1
Medizinische Universitätsklinik, Baldingerstraße, 35043 Marburg/Lahn, Germany and 2Institut für
Pathologie der Universitätsklinik, Baldingerstraße, 35043 Marburg/Lahn, Germany

ABSTRACT. Between January 1990 and January 2005, incidental hypoechoic, vascular
tumours of the spleen were identified in 13 patients using B-mode and colour Doppler
ultrasound (CDS). All lesions found were well demarcated, intrasplenically located, and
ranged in size between 1 cm and 4 cm. The increased vascular pattern on CDS was
confirmed in 9 of the 13 cases by contrast enhanced ultrasound (CES), while two
patients showed reduced vascularity on CES. In 10 patients, lesions were confirmed by
contrast enhanced CT. Histological examination was performed in three patients with
the diagnosis of capillary haemangioma (n52) and hamartoma (n51). In the remaining
cases, ultrasound follow-up was performed (range 4 months to 13 years) and
demonstrated no evidence of tumour growth in all but one patient. During a 4 year Received 1 February 2005
follow-up, one lesion increased in size from 1.0 cm to 1.5 cm and in the same patient an Revised 11 June 2005
additional 0.5 cm sized hypoechoic increased vascular lesion was also found. In the Accepted 17 August 2005
spleen a hypoechoic lesion with an increased vascular pattern incidentally found by
DOI: 10.1259/bjr/81529894
ultrasound most likely indicates a benign tumour with capillary haemangioma/
hamartoma as the most likely diagnosis. However, it should be emphasised that in all ’ 2006 The British Institute of
cases a careful ultrasound follow-up is warranted. Radiology

In recent years, B-mode ultrasound patterns of various
common splenic lesions including infarction [1, 2]
lymphoma [3], rupture [4, 5], metastases [6], cysts [7],
haemangioma [8] and other tumours [9] have been
described.
The poor correlation between ultrasound appearances
and specific splenic pathology causes difficulty in clinical
diagnosis. In liver lesions, colour Doppler ultrasound
(CDS) as well as contrast enhanced ultrasound (CES)
have improved diagnostic accuracy [10]. To date there
are few data regarding CDS and CES patterns of splenic
lesions [11–14].
This report reviews the spectrum of ultrasound
findings and the clinical follow-up we have observed
in 13 patients with incidentally found hypoechoic
vascular splenic lesions.
Patients and methods
Between January 1990 and January 2005, 13 patients
were found incidentally to have focal hypoechoic splenic
lesions with an increased vascular pattern on CDS at an
internal medicine centre. The patient ages ranged from
2 years to 76 years. The male–female ratio was 8:5.
Ultrasound and clinical data of these 13 study patients
were retrospectively evaluated.
The inclusion criteria for the study were:
N Incidentally found focal splenic lesion without clinical
symptoms or evidence of metastasised malignant
diseases.
N Hypoechoic splenic lesion using the surrounding
splenic tissue as an ‘‘in vivo’’ reference.
N Increased vascularity on colour-Doppler ultrasound
using the surrounding splenic vascularity as an ‘‘in
vivo’’ reference.
The following B-mode and CDS ultrasound para-
meters were retrospectively evaluated: splenic size,
sonomorphological pattern of the parenchymal lesions
(homogeneous vs inhomogeneous vs central echoic),
number (solitary vs multiple), configuration (round vs
oval), margin (smooth vs irregular), maximal size of
lesions and CDS patterns of vessels (diffuse vs radial).
Between 1990 and 1998, patients were examined by
Acuson 128 XP (Acuson, Mountain View, CA). From
1998, all new study patients and follow up examinations
were performed using an Acuson Sequoia (Acuson;
Siemens Medical Solutions, Mountain View, CA)
equipped with a 3.5 MHz curvilinear array and 8 MHz
linear array transducer. All ultrasound studies were
performed by a single observer CG with more than
20 years of experience. Colour Doppler settings were
optimized to achieve the greatest sensitivity for allowing
detection of low flow.
Since January 2004 all new study-patients and all
follow-up study-patients were (re)examined by CES with
a second generation contrast agent (SonoVueH; Bracco
SpA, Milan, Italy). Informed consent according to
legislation was obtained in each patient for CES
examination and statement of the local internal review
board was achieved.

The British Journal of Radiology, April 2006 319

 C Görg, K Görg, T Bert and P Barth

4 years, increased in size, one

Following baseline ultrasound, CES studies were

2 months, tumour resection

8 months, constant in size

4 months, constant in size

6 months, constant in size
6 months, constant in size
performed using an Acuson Sequoia equipped with

2 years, constant in size

4 years, constant in size

3 years, constant in size

3 years, constant in size
2 months, splenectomy

1 year, constant in size

contrast-specific, continuous-mode software operating at
a low acoustic pressure. Immediately after contrast

Follow-up/treatment
medium injection, the splenic lesions were observed for

13 years, biopsy
evidence of contrast uptake over a 5 min period. CES

more lesion
studies were analysed on the basis of review of
ultrasound unit-stored clips. CES parameters were
determined by only one observer (CG). For characteriza-
tion of potential interobserver variability, one other

US, ultrasound, CDS, colour Doppler ultrasound; CT, computed tomography; MRI, magnetic resonance imaging; CES, contrast enhanced ultrasound.
completely blinded observer (KG) with at least 20 years
of ultrasound experience retrospectively analysed CES

Additional
data on the basis of review of ultrasound unit-stored

imaging
images. A third reviewer (TB) decided in cases of

MRI
CT
CT

CT

CT
CT

CT

CT
CT
CT
CT
disagreement (2 of 18 CES patterns).

–
The following CES patterns of intrasplenic lesions

CES-parenchymal
using the normal splenic tissue as an ‘‘in vivo’’ reference
were considered: extent of enhancement during the

hypoechoic

hypoechoic

hypoechoic

hypoechoic
hypoechoic
hypoechoic
hypoechoic
arterial phase between 1–30 s and parenchymal phase

isoechoic

isoechoic
between 1 min and 5 min after injection (absent vs
hypoechoic vs isoechoic vs hyperechoic vs mixed

–
–

–
echogenicity). During the arterial phase, a distinction
between a central filling and a peripheral filling was

hyperechoic, peripheral filling

hyperechoic, peripheral filling

hyperechoic, peripheral filling

hyperechoic, peripheral filling
hyperechoic, peripheral filling
hyperechoic, peripheral filling
made.

hyperechoic, central filling

isoechoic, central filling

isoechoic, central filling

Results
The ultrasound data of the study patients are shown in
detail in Table 1. In all patients, splenic size was normal
(less than 11 cm 6 5 cm). Regarding echotexture,
CES-arterial

parenchymal lesions were homogeneous (n54), inhomo-

geneous (n52), and central echoic (n57) (star-like
pattern). Lesions were solitary (n511), multiple (n52),
–
–

–
round (n513) and with a smooth margin (n513). Two
patients had two splenic lesions on primary diagnosis.
Size of lesions ranged from 1 cm to 4 cm (Figures 1–3).
Table 1. Clinical and ultrasound findings in high vascular splenic tumours

diffuse

diffuse

diffuse

diffuse

CDS of increased vascular lesions showed a diffuse

radial
radial

radial

radial

radial

radial
radial
radial
radial
CDS

vessel pattern (Figures 1 and 2) or a radial (‘‘starfish-

like’’) vessel pattern in the remaining 9 patients
(Figure 3).
In two patients (Nos. 2 and 9) surgical diagnosis was
Size (cm)

1.5, 0.5

performed (one splenectomy and one tumour resection)

3, 2.5

and confirmed capillary haemangioma. In patient 1,

1.5

1.5

ultrasound-guided tumour biopsy was performed and

1
2
1

1
1
1
1
4

the diagnosis of hamartoma was made. All non-

surgically treated patients were subject to ultrasound
follow-up investigations which were performed at
inhomogeneous
inhomogeneous

multiple, central echoic

multiple, central echoic

3 month intervals during the first year and then once

homogeneous

solitary, homogeneous
solitary, homogeneous
solitary, homogeneous
central echoic

solitary, central echoic

solitary, central echoic

solitary, central echoic

solitary, central echoic

every year (patient 3 was lost to follow-up after 2 years,

patients 7 and 11 died).
B-mode ultrasound

In patient 9, the resection specimen consisted of

macroscopically normal splenic tissue surrounding a
well circumscribed but unencapsulated red-brown
solitary,
solitary,
solitary,
solitary,

tumour with a maximum diameter of about 3.5 cm.

Histologically the lesion showed evenly distributed
capillaries with inconspicuous round lumina surrounded
by a thick rim of connective tissue (Figure 1c).
Immunostaining (standard avidin-biotin-complex peroxi-
10 (Figure 3)
5 (Figure 2)

9 (Figure 1)

dase method) for CD 34 and CD 31 demonstrated the

lumina to be outlined by slightly hyperplastic endothe-
Patient

lia (Figure 1d). These findings prompted the diagnosis

11
12
13

of a sclerosing capillary haemangioma.

1
2
3
4

6
7
8

320 The British Journal of Radiology, April 2006

CDS of vascular tumours of the spleen

(a) (b)

(c) (d)

Figure 1. Patient No. 9. (a) B-mode ultrasound shows a well-demarcated solitary homogeneous hypoechoic splenic tumour. (b)
Colour-Doppler ultrasound shows a tumour with marked diffuse intratumoral flow signals. (c) The lesion is composed of small
inconspicuous round lumina surrounded by broad strands of collagen fibre-rich connective tissue (haematoxylin and eosin
6 200). (d) Immunohistochemistry (6 200) demonstrates a peculiar network of small capillaries lined by CD34 positive
endothelia (sclerosing capillary haemangioma).

The British Journal of Radiology, April 2006 321

 C Görg, K Görg, T Bert and P Barth

In patient 2 the spleen weight was 160 g. The cut
surface showed no abnormalities except for a soft spongy
dark-red nodule measuring 6 mm in diameter.
Histologically, the nodule consisted of irregularly thin
walled anastomosing capillaries lined by slightly hyper-
plastic endothelia. Immunohistochemistry (standard
avidin-biotin-complex peroxidase method) revealed the
endothelia to be CD 34 positive, whereas CD 8 was
negative. For these morphologic findings, diagnosis of a
splenic capillary haemangioma was made.
Contrast enhanced ultrasound was performed in nine
cases (Table 1). During arterial phase, hyperechoic
enhancement was seen in seven patients, while an
isoechoic pattern was seen in two patients. Three cases
showed marked central filling with contrast (Figure 2),
while the remaining patients demonstrated peripheral

(a) (b)

(c) (d)

Figure 2. Patient No. 5. (a) B-mode ultrasound shows a small well-demarcated hypoechoic central hyperechoic splenic tumour
(arrows). (b) Power-Doppler ultrasound shows a vascular tumour with marked diffuse intratumoral flow signals (arrows). (c) In
the early arterial phase (15 s), contrast enhanced ultrasound (SonoVueH) shows a central contrast filling indicating the increased
vascular lesion (arrow). (d) In the late parenchymal phase (4 min) contrast enhanced ultrasound (SonoVueH) shows a hypoechoic
contrast enhancement of the tumour in comparison with the surrounding normal splenic tissue (arrow). (Continued)

322 The British Journal of Radiology, April 2006

CDS of vascular tumours of the spleen

(‘‘ring sign’’) filling (Figure 3). All patients showed

complete filling in of the lesion by the end of the arterial
phase. In the parenchymal phase, all but two lesions were
hypoechoic in reference to the surrounding tissue (Figures
2 and 3). In patients 1 and 7 an isoechoic enhancement was
observed during the parenchymal phase (Table 1).
Additional imaging diagnostic procedures which
confirmed the intrasplenic tumour were MRI (n51) and
contrast enhanced CT (n510). CT revealed an increased
vascular lesion in 9 out of 10 patients. Patterns of arterial

Figure 3. Patient No.10. (a) B-mode ultrasound shows two

(e)
Figure 2. (Cont.) (e) Contrast CT shows a hyperintense
vascular splenic tumour (arrow).
well-demarcated hypoechoic central hyperechoic splenic
lesions. (b) Colour-Doppler ultrasound shows two vascular
tumours with marked radial intratumoral flow signals in
both lesions. (c) In the early arterial phase (11 s), contrast
enhanced ultrasound shows a peripheral contrast enhance-
ment of both tumours in comparison with the surrounding
normal splenic tissue (arrows). (d) In the parenchymal phase
(1 min), contrast enhanced ultrasound shows a complete
filling of both lesions with an isoechoic enhancement of
both tumours in comparison with the surrounding normal
splenic tissue. (e) In the late parenchymal phase (3 min),
contrast enhanced ultrasound shows a wash out phenom-
enon of both lesions with a hypoechoic enhancement of
both tumours in comparison to the surrounding normal
splenic tissue (arrows). (Continued)

(a) (b)

(b)

(c) (d) (e)

The British Journal of Radiology, April 2006 323


(f )
Figure 3. (Cont.) (f) Contrast CT shows a hyperintense peripheral ring enhancement of the splenic tumour (arrow). (g)
Contrast CT shows a complete filling of the splenic tumour (arrow).
phase enhancement on CT were concordant to those seen
with CES. No ‘‘parenchymal phase’’ was examined by
CT. In patient 1, no tumour was seen on CT examination.
In all 11 patients without surgical treatment, ultra-
sound follow-up was performed and revealed stable
lesions in 10 cases. In patient 4 the lesion increased in
size from 1.0 cm to 1.5 cm and a second increased
vascular lesion of 0.5 cm was seen. Splenectomy was
recommended, but the patient refused surgery; treat-
ment with ultrasound follow-up continued (Table 1).
Discussion
The spleen is regarded as ‘‘the forgotten organ’’ of the
abdomen because splenic pathology is relatively rare [15].
Secondary lesions to the spleen (i.e. lymphoma, infarction,
rupture and metastasis) are more common and often
demonstrate a hypoechoic pattern on B-mode ultrasound.
On CDS, splenic infarction and splenic rupture are
characterized by an absent vascularity. In patients with
splenic lymphoma and splenic metastases a reduced
vascularity is the preferred CDS pattern [2, 6, 9, 11]. On
CES splenic infarction as well as rupture is characterized
by an unenhanced pattern. Splenic lymphoma demon-
strates irregular peripheral enhancement, while metastasis
are as usually predominantly non-enhancing, typically
with a rim of peripheral enhancement [13]. In all other
lesions presented here CES showed a marked hyperechoic
arterial phase with variable patterns of contrast filling and
relatively hypoechoic enhancement during the parenchy-
mal phase. Besides characteristic imaging patterns, clinical
data and follow-up are important additives for definitive
diagnosis.
A clinical problem occurs with the incidentally found
focal hypoechoic intrasplenic tumour on ultrasound. B-
mode ultrasound as well as CT and MRI are often non-
specific and histological verification is often warranted.
The differential diagnosis of incidentally found
hypoechoic splenic tumours includes hamartoma and
324
C Görg, K Görg, T Bert and P Barth
(g)
haemangioma. Splenic hamartoma (also called splenoma
or nodular hyperplasia of the spleen) is a rare, benign
tumour which usually presents as a solitary lesion [16].
Splenic haemangioma presents histologically with two
different types: the more common cavernous type and
the rare capillary type.
For a better interpretation of increased vascular lesions
found by CDS, CES or CT, it is helpful to review
angiographic studies of splenic tumours. On angiogra-
phy, hamartomas have a characteristic arteriographic
pattern with irregular dilated tortuous vessels, with or
without aneurysmal dilatation, within a well demarcated
mass, with occasional vascular lakes or arteriovenous
shunting [17–21]. In contrast to the more common
cavernous haemangiomas, the capillary haemangiomas
appear as multiple hypervascular nodules fed by dilated
intrasplenic arteries without tumour vessels and vascu-
lar lakes [22]. Histologically, hamartomas resemble
normal red pulp with slit like vascular spaces lined with
endothelial cells [23]. In fact, some hamartomas with a
distinctive lobular pattern have been reclassified as red
pulp capillary haemangiomas [24].
On B-mode ultrasound, cavernous haemangiomas are
described with a predominantly hyperechoic pattern and
absent or reduced Doppler flow signals on CDS [9, 11, 13,
25]. Capillary haemangiomas as well as hamartomas
present with a hypoechoic pattern on B-mode ultrasound
[26, 27]. CES showed an increased vascular pattern
in both the arterial and parenchymal phase with
multiple radial blood-flow signals in the early arterial
phase in a patient with a hamartoma [25]. In another case
report, splenic hamartoma was markedly enhanced on
CES [28].
In our series of hypoechoic increased vascular splenic
lesions with clinical follow-up, the diagnosis of a benign
splenic tumour was most likely. It should be noted that
in one case, ultrasound follow-up examination revealed
a mild increase in size of the vascular lesion was seen
and an additional lesion found. In nine patients the
increased vascular pattern was confirmed by contrast CT
The British Journal of Radiology, April 2006
CDS of vascular tumours of the spleen
[29]. One lesion was not seen by CT (patient 1). In the
same patient, CES also showed an isoechoic enhance-
ment during the arterial and parenchymal phase. Catalano
et al [30] reported about two patients with small capillary
angiomas clearly visible in the conventional ultrasound
and isoechoic in comparison with splenic parenchyma in
all contrast phases of the CT examination (and therefore
undetectable). In all other lesions presented here, CES
showed a marked hyperechoic arterial phase with variable
patterns of contrast filling and a hypoechoic enhancement
during the parenchymal phase.
We suggest that with an incidentally found splenic
tumour the hypoechoic increased vascular pattern on
ultrasound indicates a benign vascular tumour. Capillary
haemangioma/hamartoma seems to be the most likely
diagnosis, but in all cases a careful ultrasound follow-up
is warranted.
References
1. Görg C, Schwerk WB. Splenic infarction: sonographic
patterns, diagnosis, follow-up, and complications.
Radiology 1990;194:803–7.
2. Görg C, Zugmaier G. Chronic recurring infarction of the
spleen: sonographic patterns and complications. Ultraschall
in Med 2003;24:245–9.
3. Görg C, Weide R, Schwerk WB. Malignant splenic
lymphoma: sonographic patterns, diagnosis, and follow-
up. Clin Radiol 1997;52:535–40.
4. McKenney KL, Nunez DB, McKenney MG, Asher J, Zelnick
K, Shipshak D. Sonography as the primary screening
technique for blunt abdominal trauma: experience with
899 patients. AJR Am J Roentgenol 1998;170:979–84.
5. Görg C, Cölle J, Görg K, Prinz H, Zugmaier G. Spontaneous
rupture of the spleen: ultrasound patterns, diagnosis, and
follow-up. Br J Radiol 2003;76:704–11.
6. Görg C, Schwerk WB. Sonographic findings of splenic
metastases. Imaging/Bildgebung 1991;58:26–8.
7. Dachman AH, Ros PR, Murari PJ, Olmstedt WW,
Lichtenstein JE. Non parasitic splenic cysts: a report of 52
cases with radiologic-pathologic correlation. AJR Am J
Roentgenol 1986;147:537–42.
8. Ros PR, Moser RP, Dachman AH, Murari PJ, Olmstedt WW.
Hemangioma of the spleen: radiologic-pathologic correla-
tion in 10 cases. Radiology 1987;162:73–8.
9. Görg C. The spleen. In: Meire M, Cosgrove D, Dewbury K,
Farrant P, editors. Abdominal and general ultrasound,
second edition. London, Edinburgh, New York: Churchill
Livingstone 2001;17:379–445.
10. Wermke W, Gassmann B. Tumor diagnostics of the liver
with echo enhancers: colour atlas. Berlin, Heidelberg, New
York: Springer Verlag, 1998.
11. Görg C, Schwerk WB. Color Doppler imaging of focal
splenic masses. Eur J Radiol 1994;18:214–9.
12. Catalano O, Lobianco R, Sandomenico F. Splenic trauma:
evaluation with contrast-specific sonography and a
The British Journal of Radiology, April 2006
second-generation contrast medium. J Ultrasound Med
2003;22:467–70.
13. Peddu P, Shah M, Sidhu PS. Splenic abnormalities: a
comparative review of ultrasound, microbubble-enhanced
ultrasound and computed tomography. Clin Radiol
2004;59:777–92.
14. Catalano O, Sandomenico F, Matarazzo I, Siani A. Pictoral
essay: sonography of the spleen. AJR Am J Roentgenol
2005;184:1150–6.
15. Mortelé KJ, Merjo PJ, Kunnen M. Tumoral pathology of the
spleen. In: DeSchepper AM, Vanhoenacker F, editors.
Medical imaging of the spleen. Berlin, Heidelberg, New
York: Springer Verlag, 2000:101–23..
16. Steinberg JJ, Suhrland MJ, Valensi OJ. The association of
splenoma with disease. Lab Invest 1985;52:65A.
17. Kishikawa T, Numaguchi Y, Watanabe K, Matsuura K.
Angiographic diagnosis of benign and malignant splenic
tumors. AJR Am J Roentgenol 1978;130:339–44.
18. Wexter L, Abrams ML. Hamartoma of the spleen.
Angiographic observation. AJR Am J Roentgenol
1964;92:1150–5.
19. Komaki G, Gombas OF. Angiographic demonstration of a
calcified splenic hamartoma. Radiology 1976;121:77–8.
20. Rösch J. Tumours of the spleen: the value of selective
arterography. Clin Radiol 1966;17:183–90.
21. Teates CD, Seale DC, Allen MS. Hamartoma of the spleen.
AJR Am J Roentgenol 1972;116:419–22.
22. Tada S, Shin M, Takashima T, Noguchi M, Nishro I. Diffuse
capillary hemangiomatosis of the spleen as a course of
portal hypertension. Radiology 1972;104:63–4.
23. Silverman ML, LiVolsi VA. Splenic hamartoma. Am J Clin
Pathol 1978;70:224.
24. Krishnan J, Danon AD, Frizzera G. Use of anti-factor VIII-
related antigen (F8) and Q Bend 10 (CD34) antibodies helps
classify the benign vascular lesions of the spleen. Mod
Pathol 1993;6:94A.
25. Niizawa M, Ishida H, Morikawa P, Naganuma H,
Masamune O. Color Doppler sonography in a case of
splenic hemangioma: value of compressing the tumor. AJR
Am J Roentgenol 1991;157:965–6.
26. Tang S, Shimizu T, Kikuchi Y, Shinya S, Kishimoto R,
Fujioka Y, et al. Color Doppler sonographic findings in
splenic hamartoma. J Clin Ultrasound 2000;28:249–53.
27. Fujii T, Obara T, Shudo R, Maguchi H, Saitoh Y, Ura H,
et al. Splenic hamartoma associated with thrombocytope-
nia. J Gastroenterol 1997;32:114.
28. Chou Y-H, Chiou H-J, Tui C-M, Chiou S-Y, Hsia C-Y, Tsay
S-H. Splenic hamartoma: presentation on contrast-
enhanced sonography. J Clin Ultrasound 2004;32:425–8.
29. Abbott RM, Levy AD, Aguilera NS, Gorospe I, Thompson
WM. From the archives of the AFIP: primary vascular
neoplasms of the spleen: radiologic-pathologic correlation.
Radiographics 2004;24:1137–63.
30. Catalano O, Lobianco R, Sandomenico F, D’Èlia G, Siani A.
Real-time contrast-enhancement ultrasound of the spleen:
examination technique and preliminary clinical experience.
Radio Med (Torino) 2003;106:338–56.
325
 The British Journal of Radiology, 79 (2006), 326–330

A review of current local dose–area product levels for paediatric

fluoroscopy in a tertiary referral centre compared with national
standards. Why are they so different?
M P HIORNS, A SAINI and P J MARSDEN

Radiology, Great Ormond Street Hospital, Great Ormond Street, London WC1N 3JH, UK

ABSTRACT. A prospective single centre study has been performed to assess dose–area
product (DAP) values in children having fluoroscopic examinations and to revise local
diagnostic reference levels (DRLs). DAP measurements for 2658 examinations
performed in a dedicated fluoroscopy room over a period of 21 months were analysed.
Data for the eight most commonly performed examinations (2215 cases) are presented.
DAPs (75th centile) for upper gastrointestinal studies and micturating cystograms are
substantially lower (by a factor of between 6 5 and 6 25) than the current national
reference doses (NRDs), with some of the median values being 50 times lower. The Received 5 October 2004
small DAP values in all examinations demonstrate the substantial reduction in dose and Revised 8 August 2005
consequent risk that can be achieved when both equipment performance and operator Accepted 15 August 2005
technique are optimized. Whilst we recognize that different institutions will have
DOI: 10.1259/bjr/36530782
differing practices, it is important that practitioners are aware of the range of DAPs
achievable and that NRDs do not necessarily represent best practice, and may falsely ’ 2006 The British Institute of
reassure. Radiology

The risks of ionizing radiation are higher in the
paediatric population than in adults [1, 2]. It is therefore
particularly important that the dose–area product (DAP)
used in imaging children should be as low as practicable
whilst providing the clinician with diagnostic informa-
tion [3]. The actual doses achieved will vary greatly
depending on the age, sex, body mass, body thickness
and cooperation of the child. The type of equipment and
its use by the operator will also affect dose levels. The
most recent national guidelines published by the
National Radiological Protection Board (NRPB) in 2002
which include paediatric procedures for the first time are
based on a review of dose measurements collected over a
5 year period (1996–2000). DAPs were measured from
3671 paediatric fluoroscopic examinations in up to 29
different rooms [4]. From the results, recommended
NRDs could only be defined for three paediatric
fluoroscopic examinations; micturating cystograms, bar-
ium swallows and barium meals. To our knowledge,
very little published data exists on the current range of
doses being delivered throughout the country.
The objectives of this study were to revise locally
established DRLs and to compare performance with the
current national guidelines.
Methods
Data were collected prospectively on consecutive
patients from September 2001 to May 2003. For each
examination performed the patients name, hospital
number, date, kV, mAs, screening time, DAP, examina-
tion type, and radiologist’s name were recorded. All
studies were performed on a Siemens Polystar digital
unit (installed 1999; Siemens, Erlangen, Germany). The
DAPs were recorded in cGycm2 using a PTW Diamentor
M4 DAP meter, which is specifically sensitive enough for
paediatric work (PTW, Freiberg, Germany) with a high
resolution of 0.01 cGycm2. The DAP meter is calibrated
by our radiation physicists on a yearly basis and a
tolerance of ¡3% is considered acceptable. Overall
uncertainty is within ¡25% as recommended in the
National Protocol [5]. Other routine quality assurance is
completed every 3 months.
The exact technique was at the discretion of the
radiologist. For the period surveyed, a total of nine
consultant radiologists and several specialist registrars,
of varying experience, performed the examinations.
Views obtained were tailored to the individual clinical
question but basic views were obtained in all patients,
these would include a mixture of grabbed images and
spot exposures. Typically, both an upper gastrointestinal
(GI) series and a micturating cystogram would comprise
between six and 10 images.
Results
Data were collected in 2658 cases on an individual
basis, but were then grouped by age for specific
examinations. Only one study were abandoned, during
the period and all other studies produced images of
diagnostic quality. All the data were analysed, but the
data for the eight most common examinations are
presented (2215 cases). As patient dose has been shown
to increase with body size [6], patients were assigned
into one of three groups; 0–12 months, 1–7 years
inclusive and 8 years plus. The case distribution for

326 The British Journal of Radiology, April 2006

DAP levels for paediatric fluoroscopy
Figure 1. Case distribution of examinations in our institu-
tion. GI, gastrointestinal; MCU, micturating cystourethrogra-
phy; FT, follow through; IVU, intravenous urogram; NJT,
nasojejunal tube.
our institution is shown (Figure 1) and data for the most
common examinations are listed (Table 1). The DAP
range, mean DAP, standard deviation, median DAP,
75th percentile and 90th percentile were all calculated.
These values were used to establish revised DRLs for our
institution (Table 2). The 75th percentile was used for the
DRL (bold type), but we have also added median values
in our local reference chart as this gives a good indication
of overall practice and is less skewed by outliers.
Radiologists are quickly able to see if their DAPs are
similar to those of their colleagues. Some of our
examination groups contained only small numbers, but
as a large referral centre this is a reflection that very few
of these examinations are done in any centre. In these
cases we include the figures as a guide only.
Discussion
In the UK, the adoption of the European Council
Directive 97/43 Euratom [7] through the Ionising
Radiation (Medical Exposure) Regulations [3] has made
the establishment of DRLs and their update through
clinical audit a mandatory requirement. National diag-
nostic reference levels have been established for a variety
of mainly adult examinations as a guide to what is
thought to represent ‘‘good and normal practice’’.
Individual hospitals can then compare their DAP
performance and modify their practice accordingly. The
low numbers of patients and their distribution over the
range of different ages and sizes limits the number of
paediatric DRLs currently available. Nevertheless, it is
important that hospitals in which a relatively low
number of studies are performed continue to assess
local DAP levels and have some guide to those used
elsewhere.
The NRDs as recommended by the NRPB are based on
the 75th percentile of their survey data and are shown in
Table 3 [4]. The NRPB stratified its data into five
standard sizes of children corresponding to newborn
The British Journal of Radiology, April 2006
babies, 1 year, 5 years, 10 years and 15 years old and
children were allocated according to the nearest mile-
stone. Dose levels, either in the form of entrance surface
doses (ESD) per radiograph or DAP per examination
were adjusted for body thickness or both the height and
weight to give a value (in cGycm2) corresponding to the
nearest standard size child. It can be seen that for all
types of examination there are only small differences
between the DAPs for the 1 year and 5 year old groups.
The NRPB has therefore recommended that the same
NRD be applied for both age groups. The establishment
of reference doses in paediatric radiology as a function of
patient size has been investigated for patients of the
same age groups as used by the NRPB and it has been
shown for plain film radiography in the trunk that,
although there may be a significant difference in size
between adjacent reference ages, normalization factors
for any intermediate size were unlikely to exceed a factor
of two [8, 9].
The practical disadvantage of applying a wide variety
of standard sizes for dose assessment is more evident in
non-specialist hospitals where patient numbers are small
and long periods of data collection would be necessary.
Since body size affects the dose received in any
radiological procedure it would seem ideal to assign
patients into narrow age groups that will not have an
unduly large range of body weight and size [10]. Whilst
we realise that the use of wider age groups may result in
a larger range of doses, and obscure the effects of
examination technique and equipment performance, the
age groups used in this study have been designed for
simplicity and for consistency with previous publica-
tions. Despite the use of a single dedicated fluoroscopy
suite, there is a wide variation in measured DAPs within
the individual age groups of all the examinations in our
study. Such variation is similar to that found by both the
NRPB and the Council of the European Communities
(CEC), but in these studies the differences were largely
attributed to variations in equipment and film–screen
combinations [4, 11]. In our study, the range reflects the
wider distribution of patient size, the technical difficul-
ties of imaging uncooperative children and the variation
of operator technique and experience (including training
specialist registrars). We have found that some operators
are consistently producing diagnostic quality studies at
DAP levels significantly lower than others.
In our institution barium meals and barium swallows
are combined as one procedure, the upper GI series. We
have compared our values with those given for both
barium meals and barium swallows (Table 3) and have
found that our closest equivalent DAPs for 75th
percentile are between 9 and 25 times lower (UGI in
patients 15 years, and ,12 months, respectively) and for
a micturating cystogram are 5 to 18 times lower (MCU in
patients of 8 years and older, and ,12 months, respec-
tively). Reviewing our range of DAPs shows that, for
example, in UGIs in children up to 1 year, even our 90th
centile value (14.8 cGycm2) is 13 times less that the NRPB
75th centile (with the total range being 0.06–
41.9 cGycm2).
The mean DAP results from the present study are also
substantially less than similar data from previous
studies, including another specialist centre [12, 13]
(Table 4). A further study published in 2000 [8],
327
 M P Hiorns, A Saini and P J Marsden

Table 1. Paediatric fluoroscopic dose–area products (DAPs) recorded over a period of 21 months
Examination Number of DAP range Mean DAP Standard Median DAP 75th Percentile 90th percentile
studies (cGycm2) (cGycm2) deviation (cGycm2) (cGycm2) (cGycm2)

Upper GI series
0–1 235 0.06–41.9 6.4 8.6 3.6 7.6 14.8
1–7 376 0.09–117.6 9.5 11.4 6.3 11.5 21
8+ 197 0.18–203.5 24.7 26 18.7 31.6 54.6
Dysphagia
swallow
0–1 116 0.3–39.2 12.4 8.6 10 15.7 23.9
1–7 246 0.1–80.6 13 12.7 9.3 16.6 27.5
8+ 84 1.3–76.8 18.9 16.2 13.5 22.8 40.5
Micturating
cystourethrography
0–1 165 0.02–41.9 3.8 5.3 1.8 4.9 9.7
1–7 94 0.7–48.8 8.9 15.8 4.3 10 17.3
8+ 36 4–645.9 44.2 105.4 21.8 41.6 67.1
Patalal screening
0–1 5 1.9–5.3 4.4 4 2 5.3 8.7
1–7 118 0.5–31.5 5.5 5.3 4 6.3 10.6
8+ 146 1.3–27.2 6.9 17.3 4.1 6.7 11
Barium follow
through
0–1 18 0.2–73.9 14.7 18.6 8.1 17.9 33.4
1–7 85 0.1–42.9 9.9 9.5 6.7 13.6 25
8+ 101 0.2–241.2 31.7 42.4 18.1 38.7 63
Contrast enema
0–1 59 0.1–42.9 5 7.5 2.2 5.8 12.9
1–7 25 0.7–50.5 10 10.1 7.4 12.3 15.9
8+ 2 6.3–178.6 92.5 121.8 92.5 135.3 161.4
Barium enema
0–1 21 0.7–32.8 11.2 9.1 9.5 16.4 24.1
1–7 11 2.8–76.3 15 23.1 5.1 8.9 40.9
8+ 3 20.1–49 30.5 16 22.5 35.8 43.7
Intravenous
urogrphy
0–1 3 3–7.7 5.9 2.5 6.9 7.3 7.54
1–7 47 0.9–39.2 10.2 10.4 6.4 13.8 26.9
8+ 22 7–135.5 38.5 30.4 35.3 47.3 64.8

collecting DAP data from 12 European hospitals,
investigated doses for micturating cystograms in patients
between neonate and 15 years and obtained third
quartile doses approximately 12 times higher than the
figures we have obtained for all age groups. A study in
Finland [14] considered a total of 217 patients having a
wide range (12 types) of fluoroscopic examinations and
derived third quartile DAP values for three of these
examination types, which again are approximately 12
times higher than our figures for comparable examina-
tions and age groups. However, it is important to note
that these studies were conducted in the 1990s using
conventional fluoroscopic equipment. The large differ-
ences in doses demonstrated are almost certainly due to
differences in equipment and operator technique.
The use of automatic post processing of data in digital
systems prior to image display does not provide the
operator with feedback on exposure levels and as such
has the potential to increase dose. In our experience this
is generally not the case and we have found that the use
of low frame rates and last image hold ability is
advantageous. A similar trend was observed by the
NRPB who found that in 8 of 16 types of examination in
adults, the use of digital spot imaging reduced mean
DAP values by between 20% and 50% compared with
conventional techniques but was not significant in the
remaining 8 types of examination [4].
We recommend that non-specialist X-ray departments
should assign responsibility for paediatric imaging to a
select group of interested radiologists and radiographers,
thus reducing any ‘‘learning curve’’ effects. Optimizing
examination technique is of utmost importance. Pulsed
fluoroscopy is an effective method of reducing dose and
this feature is now standard on modern machines. Our
fluoroscopy unit defaults to 15 pulses s21 but by select-
ing 3 pulses s21 the dose is immediately reduced by a
factor of five assuming the screening time is unchanged,
although the image will be more ‘‘steppy’’ due to
misregistration. (On our unit the decrease in dose is
directly proportional to the decrease in pulse rate, but on
some units the pulse length increases slightly as the
pulse rate decreases, which would therefore result in a
less marked overall fall in DAP). Our current practice is
to perform most screening at 3 pulses s21, increasing the
rate if the child is very mobile or uncooperative. Images
tend to be grabbed from the digital system, reserving
occasional full exposures for delineation of fine detail, in
difficult cases, or if there are unexpected findings. We

328 The British Journal of Radiology, April 2006

DAP levels for paediatric fluoroscopy

Table 2. Revised DRLs used in our institution. The 75th unless a younger child was particularly large for their
percentile values are demonstrated in bold. The values age. Younger patients do not require a grid and this
below represent the median dose–area product (DAP) for should not be used. We would not advocate the use of a
each examination. Children over the age of 8 years having unit with a fixed grid for use on children. We use 0.3 mm
contrast or barium enemas have been included together due of added copper filtration and this is left permanently in
to the low numbers in each group place. Operators are encouraged to use the median
DAP (cGycm2) DAP (cGycm2) DAP (cGycm2) values as appropriate ‘‘targets’’ during an examination.
DAP meters are particularly useful for assessing and
Age 0–1 Age 1–7 Age 8+
comparing the dose from screening procedures as ‘‘dose
Upper GI series 8 12 32 6 area’’ provides a more useful indication of overall
3.6 6.3 18.7 patient exposure than measurements of entrance surface
Dysphagia 16 17 23 dose at different locations [13]. Fluoroscopic screening
swallow 10 9.3 13.5 time is of limited use as a measurement of dose as it
Micturating 5 10 42 makes no allowance for the influence of dose rate or field
cystourethrogram 1.8 4.3 21.8
size and to the contribution of any spot/grabbed images.
Patalal screening 5 6 7
2 4 4.1 NRDs are limited in having to take into account the
Barium follow 18 14 39 variation in equipment and film–screen combinations
through 8.1 6.7 18.1 and as such give an idea of national practice, which is not
Contrast enema 6 13 necessarily the best practice. The use of local audit and
2.2 7.4 49 critical review to ensure best practice should not be
Barium enema 16 9 22.5 underestimated. Cook and co-workers [15] described
9.5 5.1 reductions in ESD of greater than 50% for routine
Intravenous 7 14 47 paediatric examinations.
urography 6.9 6.4 35.3
Whilst third quartile data is useful in setting National
Reference Doses, and hence National DRLs, their
application to establishing local DRLs is of limited value.
choose to use an overcouch tube whilst maximizing the
The Joint Working Party of the IPEM/BIR/RCR/NRPB/
distance between the tube and the image intensifier, with
CoR [16] promotes the use of the mean of the distribution
the table as low and close as possible to the image of room mean doses within an organization as the local
intensifier. Whilst this arrangement may slightly increase DRL. However, an alternative method of setting a local
the dose we have found that paediatric patients find this DRL in situations were the sample size is small may be to
arrangement much less frightening and this, combined display data as a histogram of DAPs for set age ranges
with easier access to the child for the operator and the and to exclude outliers beyond the limit of the curve
holders, reduces the length of the examination thereby using a mathematical ‘‘best fit’’ or agreed mathematical
reducing dose overall. Coning to a small field of view is model. The DRL may then be a better reflection of ‘‘good
achieved by the operator using a light beam diaphragm and normal practice’’. It would be practical to choose a
for guidance and we would consider this essential for a DAP based on the shape of the curve that would reliably
paediatric population. We use a low attenuation carbon include ‘‘normal practice’’ but would flag up the high
fibre table. A removable grid is available, and is dose outliers. This method would depend on having
generally only used on patients over the age of 8 years sufficient values to plot a meaningful histogram and the
development of a reliable mathematical model.
For everyday working practice, established median
Table 3. Recommended national reference doses (NRDs) for
values may give a better impression of typical local DAP
complete examinations on paediatric patients based on the
NRPB 2000 review. Recommended values for 1 year and values and are a useful way for an operator to compare
5 year old sizes are given in brackets. Comparisons are given their cases against those of their colleagues. We display
with the nearest local equivalent dose reference levels (DRLs) median and third quartile values for all tests in our
fluoroscopy room as a guide.
Examination Standard age Current NRD Local DRL
(years) (cGycm2) equivalent
(cGycm2)

Barium meal 0 70 8
Conclusions
1 200 (200) 8 We have collected data for 2658 paediatric fluoro-
5 200 (200) 12 scopic examinations. Analysis of the data, and the
10 450 32
derivation of third quartile DRLs, shows that our
15 720 32
Barium swallow 0 80 8 practice uses very much lower DAPs (and therefore
1 160 (150) 8 DRLs) than the NRDs published by the NRPB and those
5 130 (150) 12 published elsewhere in the literature. The variation
10 270 32 between data in our study and the NRDs suggests that
15 460 32 if the NRDs are used as a sole guide, many institutions
Micturating 0 40 5 will be falsely reassured and may be using greater doses
cystourethrogram 1 90 (100) 5 than necessary. Only strict attention to technique and
5 110 (100) 10 critical review of local reference levels will ensure best
10 210 42
practice. NRDs or national DRLs, however they are
15 470 42
determined, can only reflect ‘‘good and normal practice’’

The British Journal of Radiology, April 2006 329

 M P Hiorns, A Saini and P J Marsden

Table 4. Mean DAP values are compared with those from previous studies
Present study Aberdeen data [10] Newcastle data [12, 13]

Age range Examination Mean DAP Age range Examination Mean DAP Age range Examination Mean DAP
(years) (cGycm2) (years) (cGycm2) (years) (cGycm2)

0–1 UGI 6.4 0–1 BM 29 0–1 BM 130

MCU 3.8 MCU 17 MCU 168
1–7 UGI 9.5 1–5 BM 38 1–5 BM 223
MCU 8.9 MCU 15 MCU 262
8+ UGI 24.7 6–15 BM 73 6–15 BM 546
MCU 44.2 MCU 55 MCU 433
DAP, dose–area product; UGI, upper gastrointestinal series; BM, barium meal; MCU, micturating cystourethrography.

if current DAP data is submitted regularly to the NRPB
or other national/international bodies.
References
1. National Radiation Protection Board. Occupational, public
and medical exposure, Documents of the NRPB, Vol 4,
No. 2. Chilton, UK: National Radiation Protection Board,
1993.
2. United Nations Scientific Committee on the Effects of
Atomic Radiation. Sources, effects and risks of ionising
radiation, UNSCEAR 2000 Report, Vol. II: effects. New
York, NY: United Nations, 2000.
3. Department of Health. The Ionising Radiation (Medical
Exposure) Regulations 2000. London: Department of
Health, 2000.
4. National Radiation Protection Board. Doses to patients from
medical X-ray Examinations in the UK – 2000 Review,
NRPB-W14. Chilton, UK: National Radiation Protection
Board, 2002.
5. National Protocol for Patient Dose Measurement in
Diagnostic Radiology. IPSM/NRPB/CoR. Chilton, UK:
National Radiological Protection Board, 1992.
6. Montgomery A, Martin CJ. A study of the application of
paediatric reference levels. Br J Radiol 2000;73:1083–90.
7. Council Directive 97/43 Euratom of 30 June 1997 on health
promotion of individuals against the dangers of ionising
radiation in relation to medical exposure. Official Journal of
the European Communities, 1997.
8. Hart D, Wall BF, Shrimpton PC, Dance DR. The establish-
ment of reference doses in paediatric radiology as a
function of patient size. Radiat Prot Dosim 2000;90:235–8.
9. Hart D, Wall BF, Shrimpton PC, Bungay DR, Dance DR.
Reference doses and patient size in paediatric radiology.
NRPB – R318 November 2000.
10. Martin CJ, Farquhar B, Stockdale E, Macdonald S. A study
of the relationship between patient dose and size in
paediatric radiology. Br J Radiol 1994;67:864–71.
11. Council of the European Communities. European guide-
lines on quality criteria for diagnostic radiographic images
in paediatrics, EUR 16261. Luxembourg: Office for Official
Publications of the European Communities, 1996.
12. Chapple CL, Faulkner K, Lee REJ, Hunter EW. Results of a
survey of doses to paediatric patients undergoing radio-
logical examinations. Br J Radiol 1992;65:225–31.
13. Chapple CL, Faulkner K, Lee REJ, Hunter EW. Radiation
doses to paediatric patients undergoing less common
radiological procedures involving fluoroscopy. Br J Radiol
1993;66:823–7.
14. Servomaa A, Komppa T, Heikkila M, Parviainen T. Patient
doses in paediatric fluoroscopic examinations in Finland.
Radiat Prot Dosim 2000;90:239–43.
15. Cook JV, Kryriou JC, Pettet A, Fitzgerald MC, Shah K,
Pablot SM. Key factors in the optimisation of paediatric x-
ray practice. Br J Radiol 2001;74:1032–40.
16. Guidance on the Establishment and Use of Diagnostic
Reference Levels for Medical X-Ray Examinations. IPEM
Report 88. York, UK: Institute of Physics and Engineering in
Medicine, 2004.

330 The British Journal of Radiology, April 2006

The British Journal of Radiology, 79 (2006), 331–335

Peripheral dose from uniform dynamic multileaf collimation

fields: implications for sliding window intensity-modulated
radiotherapy
1
D S SHARMA, MSc, DipRP, 2ANIMESH, MSc, DipRP, 1S S DESHPANDE, MSc, DipRP, 3R D PHURAILATPAM,
1 4 4
MSc, DipRP, D D DESHPANDE, PhD, S K SHRIVASTAVA, MD and K A DINSHAW, FRCR

Departments of 1Medical Physics and 4Radiation Oncology, Tata Memorial Hospital, Dr. Ernest
Borges Marg, Parel, Mumbai, 400 012, India, 2Kirloskar Theratronics, Mumbai and 3Clinical Research
Centre, ACTREC, New Mumbai, India

ABSTRACT. The increase in the number of monitor units in sliding window intensity-
modulated radiotherapy, compared with conventional techniques for the same target
dose, may lead to an increase in peripheral dose (PD). PD from a linear accelerator was
measured for 6 MV X-ray using 0.6 cm3 ionization chamber inserted at 5 cm depth into
a 35 cm 6 35 cm 6 105 cm plastic water phantom. Measurements were made for field
sizes of 6 cm 6 6 cm, 10 cm 6 10 cm and 14 cm 6 14 cm, shaped in both static and
dynamic multileaf collimation (DMLC) mode, employing strip fields of fixed width
0.5 cm, 1.0 cm, 1.5 cm, and 2.0 cm, respectively. The effect of collimator rotation and
depth of measurement on peripheral dose was investigated for 10 cm 6 10 cm field.
Dynamic fields require 2 to 14 times the number of monitor units than does a static
open field for the same dose at the isocentre, depending on strip field width and field
size. Peripheral dose resulting from dynamic fields manifests two distinct regions
showing a crest and trough within 30 cm from the field edge and a steady exponential
fall beyond 30 cm. All dynamic fields were found to deliver a higher PD compared with
the corresponding static open fields, being highest for smallest strip field width and
largest field size; also, the percentage increase observed was highest at the largest out-
of-field distance. For 6 cm 6 6 cm field, dynamic fields with 0.5 cm and 2 cm strip field
width deliver PDs 8 and 2 times higher than that of the static open field. The
corresponding factors for 14 cm 6 14 cm field were 15 and 6, respectively. The factors
by which PD for DMLC fields increase, relative to jaws-shaped static fields for out-of-
field distance beyond 30 cm, are almost the same as the corresponding increases in the Received 16 June 2004
Revised 31 May 2005
number of monitor units. Reductions of 20% and 40% in PD were observed when the Accepted 19 July 2005
measurements were done at a depth of 10 cm and 15 cm, respectively. When the
multileaf collimator executes in-plane (collimator 90 ˚) motion, peripheral dose DOI: 10.1259/bjr/16208090
decreases by as much as a factor of 3 compared with cross-plane data. The knowledge
’ 2006 The British Institute of
of PD from DMLC field is necessary to estimate the increase in whole-body dose and the
Radiology
likelihood of radiation induced secondary malignancy.

Absorbed dose outside the primary radiation field
(peripheral dose; PD) is of clinical interest in estimating
out-of-field organ dose and subsequent long-term radia-
tion sequelae. Potential effects of PD such as cataract
formation, gonadal dysfunction and infertility, and
damage to fetus, and their threshold doses, have been
summarized by Fraass and van de Geijn [1]. The
potential damage to the fetus from PD sometimes creates
clinical dilemmas during decision making in radiation
therapy of pregnant patients [2–4].
Prior investigators have confirmed that increase in the
number of monitor units (MU) in tomotherapy, com-
pared with a conventional technique for the same
tumour dose, results in a higher whole-body dose [5–
7]. For a dose of 70 Gy to a head and neck tumour,
whole-body dose has been shown to increase from
between 0.2 Sv and 0.25 Sv for conventional therapy to
nearly 2 Sv for tomotherapy; this in turn may increase
the risk of radiation-induced secondary malignancies by
a factor of eight [8]. The increase in MU is a function of
complexity of intensity-modulation, its delivery techni-
que and collimator design of the treatment machine.
Therefore, PD needs to be measured separately for the
technique employed on the treatment machine. This
study was designed to measure PD from 6 MV X-rays
employing static and dynamic multileaf collimation
(DMLC), which is the basis for dynamic intensity-
modulated radiotherapy (IMRT). Dependence of PD on
out-of-field distance, strip field width, field size, direc-
tion of multileaf collimator (MLC) motion and depth of
measurement were investigated.
Methods and materials
Peripheral dose was measured for 6 MV X-ray using a
linear accelerator (LA) (Clinac 2100 C/D; Varian

The British Journal of Radiology, April 2006 331


Associates, Palo Alto, CA) equipped with an MLC
consisting of 26 leaf pairs, each leaf projecting to a
10 mm leaf width at the isocentre distance. All measure-
ments were carried out using a 0.6 cm3 Farmer type
ionization chamber (PTW, Friedberg, Germany) inserted
at 5 cm depth into a 35 cm 6 35 cm 6 105 cm plastic
water phantom (Nuclear Associates, USA) under iso-
centric conditions. Field sizes of 6 cm 6 6 cm, 10 cm 6
10 cm and 14 cm 6 14 cm defined by conventional jaws
were simulated in DMLC field mode, wherein strip
fields of constant width 0.5 cm, 1 cm, 1.5 cm and 2 cm
created by the opposing MLCs were moved with
constant speed from one bank to the other in a rectilinear
fashion perpendicular to the radial axis of the LA. Thus
for every open field, six sets of PDs were measured – two
from static open fields shaped by jaws and MLCs and
four from DMLC. When the collimator is set at 0 ˚ (IEC
61217) [9], and MLCs are used either in static or dynamic
mode, the upper (Y) jaws define the superior and inferior
borders and the lower (X) jaws are positioned 0.8 cm
distal to the most retracted leaf position on each side of
the field. The detail performance characteristic of this
MLC in dynamic dose delivery has been described
elsewhere [10–12].
The number of MU required to deliver 1 Gy at 5 cm
depth on the central axis was found by matching meter
readings from the different DMLC fields to that from the
corresponding, static jaw-shaped open field. These MUs
were used for the subsequent PD measurement along a
longitudinal axis to an out-of-field distance up to 60 cm.
Table 1. Monitor unit (MU) multiplication factor (ratio of
MU from dynamic multileaf collimation (DMLC) fields to that
from correspondingopen field) for 6 MV X-ray for different
DMLC fields having effective field sizes of 6 cm66 cm,
10 cm610 cm and 14 cm614 cm
Strip field MU multiplication factor for different DMLC fields
width (cm)
Field size Field size Field size
6 cm66 cm 10 cm610 cm 14 cm614 cm
0.5 6.96 10.93 13.75
1.0 3.98 6.73 8.94
1.5 2.59 4.70 6.49
2.0 1.89 3.53 4.96
Figure 1. Peripheral dose in phantom from 6 MV X-ray for 0 ˚ collimator and field sizes (a) 6 cm 6 6 cm; (b) 14 cm 6 14 cm at
5 cm depth, normalized to 100% on the central axis at depth of maximum dose of static open field. MLC, multilead collimator.
332
D S Sharma, Animesh, S S Deshpande et al
The factor by which planned MU increases for different
DMLC fields over that of the corresponding static fields
for the same dose is defined to be MU multiplication
factor.
The same measurements were repeated for 10 cm 6
10 cm field size with collimator at 90 ˚, wherein the MLC
executes in-plane motion. Variation in PD with depth
was investigated for 10 cm and 15 cm depth under
isocentric set up using 10 cm 6 10 cm field size having
1 cm strip field width.
Results
The MU multiplication factor for different DMLC
fields was found to increase with decreasing strip field
width and increasing field size (Table 1).
The measured PD data for 6 MV X-ray at 5 cm depth
and for 0 ˚ collimator angle are shown in Figure 1 as a
function of out-of-field distance. Each family of six
curves represents the PD data for jaws and MLC shaped
static field and four DMLC fields having different strip
field widths. The dose at the depth of maximum dose
(Dmax) on the central axis of every jaws-shaped static
field was normalized to 100%. All the PD data were
expressed as a percentage of the respective Dmax. PD
data of all DMLC fields were found to manifest two
distinct regions.
PD , 30 cm
The region extending up to 30 cm out-of-field distance,
shows crests and troughs of varying amplitude; this is
prominent for 6 cm 6 6 cm (Figure 1a) and is found to
grow progressively smaller for larger fields (Figure 1b).
PD . 30 cm
In this region PD curves from DMLC fields of different
strip field width are seen running parallel and main-
taining constant slope. The factor by which PD for
different DMLC fields increase over jaws-shaped static
field is observed to be almost the same as that of the
The British Journal of Radiology, April 2006
Peripheral dose from DMLC fields

corresponding increase in MU. Although the absolute

value of PD is comparatively small, the percentage
increase in PD is higher for this region, showing
maximum increase at 60 cm out-of-field distance.
As expected, PD of all DMLC fields was found to be
higher than the corresponding static open fields. For the
same field size the smaller strip field width is seen
delivering higher PD compared with larger strip field
width. Also for the same strip field width, PD consis-
tently increases with increasing field sizes. For 6 cm 6
6 cm field, DMLC with 0.5 cm and 2 cm strip field width
deliver a maximum of 8 and 2 times higher PD than that
of static open fields (Figure 1a). The corresponding factor
for 14 cm 6 14 cm field is 15 and 6, respectively
(Figure 1b). DMLC field with 0.5 cm strip field width Figure 2. Composite peripheral dose (PD) distribution of
delivers 1.5 and 3 times higher PD than that of 2 cm strip 6 MV X-ray measured at 5 cm depth with 0 ˚ and 90 ˚
field width at 1 cm and 60 cm out-of-field distance, collimator angle from 10 cm 6 10 cm static field shaped by
respectively. jaws and multileaf collimator (MLC), normalized to 100% on
As an exception, jaws-shaped static fields may deliver the central axis at depth of maximum dose of static open
more PD than certain DMLC fields with larger strip field field.
width. This exceptional phenomenon is observed for out-
of-field distances extending up to 15 cm for 6 cm 6 6 cm
and up to 7 cm for 14 cm 6 14 cm. Also, PD data of
MLC shaped static field is found to be less than that of
jaw-shaped static fields for out-of-field distances up to
30 cm, beyond which MLC shaped static fields show a
slight increase in PD over jaws-shaped fields.
PD data for static and DMLC fields measured with 90 ˚
collimation shows a reduction in value compared with
data measured with 0 ˚ collimation. The magnitude of
reduction is seen to depend on out-of-field distance and
strip field width. For 10 cm 6 10 cm field, maximum
reduction in PD data for jaws and MLC shaped static
field are found to be by a factor of 2 and 2.5, respectively
(Figure 2). For DMLC fields, a maximum reduction
factor of 3 is observed at 30 cm out-of-field distance for
the smallest strip field width of 0.5 cm (Figure 3). Similar Figure 3. Composite peripheral dose (PD) distribution of
6 MV X-ray measured at 5 cm depth with 0 ˚ and 90 ˚
characteristics were observed for larger strip field
collimator angles from 10 cm 6 10 cm field simulated in
widths, albeit with progressively lower PD. MLC shaped dynamic multileaf collimation (DMLC) mode using 0.5 cm
static and DMLC fields show no change in PD at 20 cm sweeping gap width, normalized to 100% on the central axis
from the field edge for these two collimator orientations. at depth of maximum dose of static open field.
PD measured under isocentric condition at 5 cm,
10 cm and 15 cm depth using 10 cm 6 10 cm fields
and 1 cm strip field width are shown in Figure 4. A
maximum reduction of 20% and 40% at 60 cm from the
field edge is observed for 10 cm and 15 cm depth,
respectively, as compared with the data measured at
5 cm depth.

Discussion
Pre-measured PD data from static fields [1, 2, 13–17],
and an empirical relationship [18], are available as a
guideline to evaluate approximate out-of-field organ
dose from conventional treatment. The qualitative and
quantitative behaviour of our jaw-shaped open field
(10 cm 6 10 cm, 0 ˚ collimation and 5 cm depth) PD data
is in agreement with other findings [2, 17]. But beyond
Figure 4. Peripheral dose (PD) in phantom from 6 MV X-rays
30 cm out-of field distance, our PD data for MLC shaped measured at 5 cm, 10 cm and 15 cm depth with target to
static field disagrees with that of Mutic et al [17] who chamber distance of 100 cm. Effective field size of 10 cm 6
have reported reduction in PD at all out-of-field 10 cm having 1 cm sweeping gap width and 0 ˚ collimator is
distances. However, when the collimator is rotated to used. PD is normalized to 100% on the central axis at depth
90 ˚ our data agrees with the findings of Mutic et al [17]. of maximum dose of static open field.

The British Journal of Radiology, April 2006 333


Although IMRT offers significant advantages for dose
conformality to irregular target volumes with sharp dose
gradient beyond the target, it also increases MU/Gy and
hence whole body dose. This may be a concern for long-
term radiation sequelae in patients treated with this
technique [8, 19]. Whole-body dose data are available
from tomotherapy [5–8]. However, the same data cannot
be applied to sliding window IMRT technique as the
collimator scatter and transmission, head leakage and
internal scatter, which are the main components of PD,
depend on the collimator design of treatment machine
and beam delivery technique. Our study was carried out
to find PD from different DMLC fields which closely
simulated sliding window IMRT. When static fields are
simulated in DMLC mode, PD was found to increase by
a factor of 2 to 15, similar to the increase in MU for
DMLC fields over corresponding static open fields.
Increase in MU due to the use of a universal wedge
have also shown to increase PD by a factor of 2–4 com-
pared with PD resulting from unwedged beam [14–16].
Even though the intensity of radiation is not modu-
lated in our dynamic fields, different strip field width
can be thought to represent the different complexity of
intensity profile required in IMRT. As the degree of
modulation increases, i.e. peaks and valleys in the
intensity profile are sought to be closely spaced, these
highly modulated beams require narrow MLC openings
leading to sharp increase in MU. With sliding window
IMRT, higher MU becomes an absolute corollary and the
planned MU tends to be larger in cases where the fluence
distribution is more modulated.
Even though PD from DMLC field is generally
observed to be more than that of jaw-shaped static field,
a slight reduction in PD is observed near the field edge
for DMLC field with larger sweeping gap width. This
may be attributed to reduction in collimator scatter and
transmission of DMLC fields offered by the stationary
portion of the MLC. The increase in MU for smaller
sweeping gap widths for the same field and dose,
increases both scatter and leakage dose leading to the
increase in PD.
Besides internal scatter and leakage dose, PD in the
proximal region (5–30 cm) seems to be a complex
interplay of leaf scatter, leaf end design, MLC placement
in treatment head, MLC motion and beam-on time.
Similar to our crest and trough effect appearing within
30 cm of the field edge, Greene et al [20] have reported a
sharp peak at 20 cm from the edge of 10 cm 6 10 cm
field due to maximum scatter coming from the face of the
beam-defining jaws, which disappears for large field
sizes as the point that sees maximum scatter from the
face of the jaws shifts into the primary beam itself. In our
case, this situation never realises completely, even for
large DMLC fields, probably because scatter radiation
arises both from the face of jaws and from the rounded
end face of the sweeping MLCs.
The crest and trough effect appearing prominently for
smaller dynamic fields may be clinically significant
especially when a critical organ is lying close to the
primary field. If an organ at risk is lying at the distance of
15 cm from the field edge (field size 6 cm 6 6 cm) it
would received 4.5 times more dose than the case where
this effect is not taken into account. Moreover, it is not
known if commercially available treatment planning
334
D S Sharma, Animesh, S S Deshpande et al
systems take into account this pattern of PD during
optimization and dose computation. Further study is
required to segregate the different components of scatter
and leakage. Inclusion of PD characteristics in dose
calculation algorithms is suggested.
Most published data for static field shows small
variation in PD with depth of measurement [1, 2, 15].
However, similar to the finding of Tatcher et al, our
measurement for DMLC field shows decrease in PD with
depth of measurement.
The largest field studied is 14 cm 6 14 cm as limited
by the maximum range of travel of individual leaves
relative to the carriage. While normalizing the dose at
depth of dose maximum to 100%, central axis depth dose
characteristics of the DMLC field of a particular size is
assumed to remain same as that for a static field. The
uncertainty associated with low dose measurement
especially at 60 cm with lower MU generally for the
static fields are not considered in the presented data. For
open fields where the MU is small, MU was scaled by a
factor of 2–3 to get good signal at largest out-of-field
distance. The reproducibility of meter reading for DMLC
fields was observed to be less than 2% during the
measurement.
As the IMRT field employed in the clinical situations
consists of a mixture of multiple strip fields having
different gap width (beamlet size), validation of the
phantom measured PD from uniform DMLC field
needs to be carried out using the clinically employed
fluence.
References
1. Fraass BA, van de Geijn J. Peripheral dose from mega-
voltage beams. Med Phys 1983;10:809–18.
2. Stovall M, Blackwell CR, Cundiff J, Novack DH, Palta JR,
Wagner LK, et al. Fetal dose from radiotherapy with photon
beams: Report of AAPM Radiation Therapy Task Group
No. 36. Med Phys 1995;22:63–82.
3. Sneed PK, Albright NW, Wara WM, Prados MD, Wilson
CB. Fetal dose estimates for radiotherapy of brain tumors
during pregnancy. Int J Radiat Oncol Biol Phys
1995;32:823–30.
4. Sharma DS, Rakesh J, Tambe CM, Animesh, Deshpande
DD. Effect of tertiary multileaf collimator (MLC) on foetal
dose during three-dimensional conformal radiation therapy
(3DCRT) of a brain tumour during pregnancy. Radiother
Oncol 2004;70:49–54.
5. Followill D, Geis P, Boyer A. Estimates of whole-body
equivalent dose produced by beam intensity modulated
conformal therapy. Int J Radiat Oncol Biol Phys
1997;38:667–72.
6. Mutic S, Low DA. Whole-body dose from tomotherapy
delivery. Int J Radiat Oncol Biol Phys 1998;42:229–32.
7. Meeks SL, Paulino AC, Pennington EC, Simon JH,
Skwarchuk MW, Buatti JM. In vivo determination of
extra-target doses received from serial tomotherapy.
Radiother Oncol 2002;63:217–22.
8. Verellen D, Vanhavere F. Risk assessment of radiation-
induced malignancies based on whole-body equivalent
dose estimates for IMRT treatment in the head and neck
region. Radiother Oncol 1999;53:199–203.
9. International Electrotechnical Commission, IEC Report
61217: Radiotherapy Equipment – Co-ordinates,
Movements and Scales. IEC, Geneva, Switzerland, 1997.
10. Chui CS, Spirou S, LoSasso T. Testing of dynamic multileaf
collimation. Med Phys 1996;23:635–41.
The British Journal of Radiology, April 2006
Peripheral dose from DMLC fields
11. LoSasso T, Chui CS, Ling C. Physics and dosimetric aspects
of a multileaf collimation system used in the dynamic mode
for implementing intensity modulated radiotherapy. Med
Phys 1998;25:1919–27.
12. Low DA, Sohn JW, Klein EE, Markman J, Mutic S, Dempsey
J. Characterization of a commercial multileaf collimator
used for intensity modulated radiation therapy. Med Phys
2001;28:752–6.
13. Greene D, Chu GL, Thomas DW, Taylor RJ. Dose
levels outside radiotherapy beams. Br J Radiol
1983;56:543–50.
14. Sherazi S, Kase KR. Measurements of dose from secondary
radiation outside a treatment field: effects of wedges and
blocks. Int J Radiat Oncol Biol Phys 1985;11:2171–6.
15. McParland BJ. Peripheral dose of two linear accelerators
employing universal wedges. Br J Radiol 1990;63:295–8.
The British Journal of Radiology, April 2006
16. Mutic S, Esthappan J, Klein EE. Peripheral dose distribu-
tions for a linear accelerator equipped with a secondary
multileaf collimator and universal wedge. J Appl Clin Med
Phys 2002;3:302–9.
17. Mutic S, Klein EE. A reduction in the AAPM TG-36 reported
peripheral dose distributions with tertiary multileaf collima-
tion. Int J Radiat Oncol Biol Phys 1999;44:947–53.
18. Mazonakis M, Damilakis J, Varveris H, Theoharopoulos N,
Gourtsoyiannis N. A method of estimating fetal dose
during brain radiation therapy. Int J Radiat Oncol Biol
Phys 1999;44:455–9.
19. Hall EJ, Wuu CS. Radiation-induced second cancers: The
impact of 3DCRT and IMRT. Int J Radiat Oncol Biol Phys
2003;56:83–8.
20. Greene D, Karup PGG, Sims C, Taylor RJ. Dose levels
outside radiotherapy beams. Br J Radiol 1985;58:453–6.
335
 The British Journal of Radiology, 79 (2006), 336–341

The DXL Calscan heel densitometer: evaluation and diagnostic

thresholds
J A THORPE, MSc, BSc and S A STEEL, MSc, BSc

Centre for Metabolic Bone Disease, Royal Hull Hospitals NHS Trust, Hull Royal Infirmary, Anlaby
Road, Hull HU3 2RW, UK

ABSTRACT. The DXL Calscan (Demetech AB) is a new dual energy X-ray absorptiometry
device for determining heel bone mineral density (BMD). The system is based on the
standard technique of dual energy X-ray absorptiometry (DXA), using a fan beam
configuration, but introduces an additional laser measurement of heel thickness
intended to improve accuracy. We have examined the utility, in vitro and in vivo
performance of the DXL Calscan and established triage thresholds based on the UK’s
National Osteoporosis Society guidelines on peripheral densitometry. The Calscan
proved convenient, easy to use and was stable over time and within a range of
operating temperatures. Short-term in vitro precision as %CV, with phantom
repositioning, was 0.75% and long term precision 0.73%. Precision in vivo, determined
from duplicate right heel scans of 67 subjects, was 1.19%. Effective radiation dose to
the patient was ,0.1 mSv per scan. 140 white females (70 osteoporotic and 70 non-
osteoporotic), aged 55–70 years underwent scans of both heels. Subjects were defined
as osteoporotic or non-osteoporotic on the basis of axial DXA (spine L2–L4 and total
hip). Triage thresholds for reassurance-referral or referral-treatment were 0.391 g cm22 Received 16 March 2005
and 0.306 g cm22 for non-dominant and 0.395 g cm22, 0.294 g cm22 for dominant Revised 8 August 2005
heel, respectively. The non-dominant heel proved slightly superior to the dominant for Accepted 15 August 2005
triage purposes. Of the seven non-osteoporotic subjects misclassified as osteoporotic by
DOI: 10.1259/bjr/22191429
Calscan of either heel, six had severe axial osteopenia. If operated by trained personnel
and used in appropriate populations exhibiting risk factors, the Calscan is well suited ’ 2006 The British Institute of
for use in the management of post-menopausal osteoporosis. Radiology

Measurement of bone mineral density (BMD) by dual
energy X-ray absorptiometry (DXA) is now well estab-
lished as the method of choice for osteoporosis assess-
ment [1–3]. BMD assessment of the lumbar spine and hip
by DXA represents the current gold standard due to the
greater associated morbidity and mortality of fractures at
these two sites, superior fracture prediction [4, 5] and
response to treatment [6]. In addition to axial assessment,
there are a variety of DXA devices available for
measuring BMD in the forearm, heel and hand.
The DXL Calscan (Demetech AB, Solna, Sweden) is a
new peripheral device for calcaneal BMD assessment,
based on fan beam DXA. The Calscan also incorporates a
laser measurement of heel thickness to improve the
accuracy of calcaneal BMD. Standard DXA assumes a
two compartment model of tissue masses, the first bone
and the second a composite of lean and adipose tissue at
an assumed constant ratio. This assumed ratio does not
allow for fluctuations in lean and adipose tissue propor-
tions that have been demonstrated to occur at the spine [7]
and are likely to occur at the calcaneus or elsewhere [8, 9].
This leads to calcaneal DXA providing a precise but
potentially inaccurate estimate of BMD, with the degree of
inaccuracy dependent on body mass index [8].
This study has been funded in part by Demetech AB, Solna,
Sweden. Additional funding for J Thorpe provided by the local
osteoporosis charity, OSPREY.
These inhomogeneities can be corrected by solving the
BMD equation as a three component model of bone, lean
and adipose tissue. Swanpalmer [10, 11] described how a
third X-ray energy could achieve this, but concluded that
a significantly higher photon count (and hence scan
times) would be required to maintain an acceptable
degree of precision.
Jonson [12] deduced that if the combined width of all
three components were known, e.g. the width of the heel,
the ratio of soft to lean tissue could then be derived and
corrected for. The laser heel width measurement on the
Calscan provides this additional dimension allowing the
derivation of BMD from a three component model,
whilst theoretically maintaining DXA precision [13].
As with other peripheral DXA (pDXA) devices [14],
the Calscan is smaller, portable, cheaper and has a lower
radiation dose than axial densitometers. However,
pDXA results at the calcaneus cannot be interpreted
using the WHO definition [15] and do not correlate
perfectly with bone density at either spine or hip. The
imperfect correlation can lead to pDXA misclassifying
subjects to the opposite diagnostic group to which they
would have been classified by axial DXA [16–18],
particularly for subjects with BMD scores close to
diagnostic thresholds. Hence considerable debate
remains over how such peripheral devices might best
be employed in the clinical setting. The UK-based
National Osteoporosis Society (NOS) has recently stated

336 The British Journal of Radiology, April 2006

Evaluation and diagnostic thresholds of the DXL Calscan
that current evidence supports the use of peripheral
devices in a triage rather than diagnostic role, and has
established a method for determining the required triage
thresholds [19, 20]. The aim of this study was to
determine the in vitro and in vivo operating conditions
of the Calscan and to establish triage referral thresholds
based on the NOS guidelines.
Materials and methods
The DXL Calscan bone densitometer (Figure 1) utilizes
a fan beam, dual energy X-ray source and a solid state
detector to perform a scan of the heel. A region of interest
is positioned automatically by the software to derive
BMD. A concomitant measure of heel thickness is
obtained using the laser reflection to correct for varia-
tions in the soft to lean tissue ratio. The Calscan, at 25 kg
in weight and 80 cm long by 43 cm wide and 33 cm tall,
is relatively compact and easy to transport and includes
wheels at one end and a carry handle. As with all X-ray
equipment, the Calscan is potentially subject to changes
in tube temperature after performing an acquisition, and
as a portable device it may also be subject to fluctuations
in performance due to environmental changes. To
counter this, the software (version 1.3.1) requires a
warm-up acquisition when the device is switched on
and a 4 min cooling down period after each acquisition.
In vitro methodology
Phantom based studies were conducted to test the
effect of temperature, stability following relocation and
to determine in vitro precision. The DXL-Calscan comes
with a manufacturer-supplied phantom, made from
shaped pieces of aluminium depicting the calcaneus,
embedded in acrylic. Short term precision was deter-
mined as percent coefficient of variation (%CV) from 30
scans with phantom repositioning and 30 without. The
device was operated according to the manufacturer’s
instructions. All measurements were taken on the same
day. Accuracy was calculated from comparison with the
phantom’s stated BMD of 0.347 g cm22, established from
a central reference machine, corresponding to a T-score
Figure 1. The Demetech DXL Calscan and phantom.
The British Journal of Radiology, April 2006
of 22.1. Long term in vitro precision was determined
using daily single phantom scans acquired over a period
of 6 months as part of routine quality assurance.
Sixteen tests were conducted to assess the effects of
ambient temperature, device movement or tube heating
on accuracy or precision. For each test the device was
disconnected, moved to a different room, warmed up
and a phantom scan acquired as soon as the warm-up
was complete. The device was given a further 30 min to
stabilize, then a second phantom scan was acquired and
the device powered down and allowed to cool for 30 min
before beginning the next test. Temperature was mea-
sured on an alcohol room thermometer throughout.
Electrical, laser safety and radiological protection sur-
veys were also carried out.
In vivo methodology
Subjects
Females attending for routine BMD of spine and hip
were approached for participation in this study. These
were referred on the basis of agreed local risk criteria
which are broadly in agreement with those of the Royal
College of Physicians [1]. The study was approved by the
local research Ethics Committee and informed consent
was obtained. All subjects were white and between the
ages of 55 years and 70 years (Table 1). A total of 140
women were recruited.
Axial DXA assessment
Subjects underwent BMD of lumbar spine and hip
using a GE-Lunar Prodigy (GE-Lunar, Madison, WI) as
part of their routine examination and clinical manage-
ment was determined on the basis of the results. In our
centre, DXA of the right hip is performed unless
contraindicated. For the purposes of this study, if the
lower of either L2–L4 spine or total hip BMD T-score
values was below 22.5, the subject was classified as
osteoporotic. Otherwise they were classified as non-
osteoporotic. When lumbar vertebrae showed clear signs
of degenerative changes, the individual vertebrae
affected were excluded from the lumbar spine results.
Four subjects had individual vertebrae excluded – one
osteoporotic and two non-osteoporotic subjects with
degenerative changes of L4 and one non-osteoporotic
subject with changes at L3. For eight other subjects (four
osteoporotic, four non-osteoporotic by final diagnosis),
two or more vertebrae on the same subject showed
Table 1. ‘‘Evaluation and Diagnostic Thresholds of the DXL
Calscan’’. Mean (standard deviation) subject demographic
variablesfor whole group, osteoporotic and non-osteoporo-
tic subjects, respectively
Whole group Osteoporotic Non-osteoporotic
n 140 70 70
Age (years) 62.7 (4.5) 63.2 (4.3) 62.2 (4.7)
Height (cm) 159.4 (6.5) 159.2 (7.0) 159.5 (6.0)
Weight (kg) 64.3 (10.7) 62.6 (11.1) 66.0 (10.1)
BMI (kg m22) 25.3 (3.9) 24.7 (3.9) 26.0 (3.9)
BMI, body mass index.
337

degenerative changes. For these eight subjects the spine
results were disregarded and diagnosis was made by
total hip DXA alone. In one case this caused the subject to
move from the osteoporotic to non-osteoporotic group.
Recruitment continued until 70 osteoporotic and 70 non-
osteoporotic subjects were enrolled.
Calcaneal DXA assessment
BMD of both heels was obtained using the DXL
Calscan. To determine in vivo precision, 67 of the 140
subjects underwent a repeat acquisition of the right heel,
with repositioning between each. The calcaneal regions
of interest (ROI) were manually checked and, if deemed
necessary, ROI position was corrected as per the user
manual instructions. For the repeat Calscan acquisition,
the second scan for each subject was analysed on a
separate day to the first to reduce the possibility of
operator bias during any ROI repositioning.
Results
Operational utility
The time from scan initiation to appearance of the
results is 94 s, with an additional 4 min required to allow
the X-ray tube to cool before another acquisition can be
taken. The Calscan is able to image either heel from the
same side of the device making it easier for the patient
and minimizing floor space required where both heels
are to be scanned. As for all equipment using ionizing
radiation, the Calscan requires a standard radiation
safety assessment but also an additional laser safety
assessment. The footwell of the Calscan was of an open
design, and had the advantage of allowing the operator
to manually assist the positioning of the heel. The open
design allowed easy access for the operator and was
comfortable for the patient, but did require some
attention to achieve the ideal positioning.
In vitro results and environmental effects
Short term in vitro precision (coefficient of variation)
was 0.76% CV (mean BMD 0.347¡0.0026 g cm22) with
phantom repositioning, 0.75% (0.347¡0.0025 g cm22)
without. Long term precision was 0.73%. The device
was accurate, with no measurable difference between
mean phantom BMD as measured on our machine
compared with that of the central reference machine.
Average phantom BMD and precision for the 32
environmental scans was 0.347 g cm22 and 0.62% CV.
Average phantom BMD and precision for the 16 scans
taken as soon as possible after a warm-up scan, i.e. after
the enforced 4 min cooling down period between scans,
was 0.347 g cm22 and 0.65% CV. For the 16 scans
acquired after the tube had been allowed to cool for half
an hour, average BMD was 0.348 g cm22 and 0.61% CV.
During the 32 scans of the 16 environmental tests, the
room temperature varied from 21.3 ˚C to 26.9 ˚C, the
upper value being slightly outside the manufacturer’s
recommended operating range of 15 ˚C to 25 ˚C.
338
J A Thorpe and S A Steel
Comparison of the results recorded at the 16 highest
temperatures (range 23.8 ˚C to 26.9 ˚C) with the results at
the 16 lowest (21.3 ˚C to 23.8 ˚C) did not change phantom
BMD or precision significantly. For the 16 higher
temperature scans, phantom BMD and precision was
0.347 g cm22 and 0.64% CV, respectively. For the 16
lower temperature scans, phantom BMD and precision
was 0.348 g cm22 and 0.58% CV, respectively.
Radiation and laser safety
The effective radiation dose to the patient was
,0.1 mSv per scan and a controlled area of 0.5 m was
defined around the device in order to comply with IRR
1999 [21]. At this distance, scatter dose to the operator
would not exceed annual dose limits, even at maximum
scan throughput. The laser assessment found the Calscan
laser itself to be class 2 by UK/European/US standards
and thus capable of causing eye damage, but the location
of the laser within the footwell removed the possibility of
accidental exposure and so the laser was deemed to be
safe (class 1), provided the Calscan outer casing was in
place. The permanent filtration and laser class were not
marked on the casing as is required to comply with UK/
EU standards [22, 23]. A laser warning label was added
to the Calscan and local rules were established that
reflective objects should be kept clear of the footwell, as
stated in the user manual. No other laser precautions
were deemed necessary. No safety problems with the
laser occurred during the project, but it was noticed that
opaque black hosiery could produce spurious BMD
results, although other hosiery did not.
In vivo results
Subject demographics and bone density results are
summarized in Tables 1 and 2, whilst coefficients of
determination (adjusted R2) between key variables are
shown in Table 3. Mean in vivo BMD of the right heel for
all 67 subjects (19 osteoporotic, 48 non-osteoporotic)
given repeat measurements was 0.357 g cm22 (range
0.186–0.518 g cm22, standard deviation 0.074 g cm22).
Mean absolute difference between paired results was
0.0046 g cm22 (range 0–0.018 g cm22). Calscan precision
for the 67 subjects as %CV (derived from root mean
square) was 1.19%.
Taking the osteoporotic and non-osteoporotic preci-
sion groups separately, mean BMD for the 19 osteopo-
rotic subjects was 0.299 g cm22 (0.186–0.437 g cm22, SD
0.065). Mean absolute difference was 0.0042 g cm22 (0–
0.012 g cm22). Precision was 1.30%CV. For the 48 non-
osteoporotic precision subjects mean BMD was
0.392 g cm22 (range 0.281–0.518 g cm22, SD 0.061).
Mean absolute difference was 0.0048 g cm22 (0–
0.0018 g cm22). Precision was 1.09%CV.
Establishing triage thresholds
In the revised NOS guidelines on peripheral DXA,
Blake et al [20] recommend the use of peripheral devices
in a triage role as an adjunct to axial DXA and suggest a
The British Journal of Radiology, April 2006
Evaluation and diagnostic thresholds of the DXL Calscan

Table 2. ‘‘Evaluation and Diagnostic Thresholds of the DXL Calscan’’. Mean (standard deviation) DXA bone density and T-score
results for whole group, osteoporotic and non-osteoporotic subjects, respectively
Whole group Osteoporotic Non-osteoporotic

Spine L2–L4 BMD 0.948 (0.173) 0.822 (0.070) 1.076 (0.152)

T-score 22.10 (1.45) 23.15 (0.59) 21.03 (1.26)
Total hip BMD 0.832 (0.122) 0.774 (0.092) 0.892 (0.121)
T-score 21.39 (1.02) 21.88 (0.77) 20.89 (1.01)
Non-dominant heel BMD 0.356 (0.064) 0.328 (0.054) 0.383 (0.062)
T-score 21.96 (0.97) 22.37 (0.82) 21.53 (0.93)
Dominant heel BMD 0.356 (0.064) 0.328 (0.050) 0.383 (0.065)
T-score 21.96 (0.97) 22.38 (0.75) 21.54 (0.99)
Units: BMD (g cm22); T-score (St Dev).
BMD, bone mineral density.

Table 3. ‘‘Evaluation and Diagnostic Thresholds of the DXL the dominant heel, the referral rates for the Calscan in
Calscan’’. DXA BMD Correlation (adjusted R2 value) this group were 52.9% (non-dominant) and 58.6%
Spine Total femur Dominant Non- (dominant), but with an error margin of ¡9% due to
L2–L4 heel dominant the small sample size. Of the seven non-osteoporotic
heel subjects misclassified as osteoporotic by Calscan of either
Spine L2–L4 1 0.379 0.285 0.276
non-dominant or dominant heels, six had severe osteo-
Total femur 1 0.331 0.350 penia (axial T-score ,22).
Dominant heel 1 0.905 A better estimate of the expected referral rate can be
Non-dominant 1 drawn from comparing our derived thresholds (and
heel confidence intervals) to the mean and standard deviation
of the Calscan reference data. The reference data are
drawn from a population of 993 Swedish women
method for defining the two triage thresholds required. between 15 years and 85 years of age [24] (381 between
The upper of the two thresholds is set at a point above 50 years and 69 years), albeit a population without
which only 10% of osteoporotic subjects would fall, known risk factors. If we assume a hypothetical referral
whilst the lower threshold is a point below which only group with an even distribution of subject ages from
10% of non-osteoporotic subjects would fall. Subjects 55 years to 70 years and the same spread and trend in
who fall above the upper threshold would be assumed BMD results as the Swedish reference group, the
non-osteoporotic, whilst subjects who fall below the expected mean referral rate at the non-dominant heel
lower threshold assumed osteoporotic. Subjects falling would be 36.7%. Adjusting for the distribution of ages
between the two would be recommended for referral for seen in our 140 subjects, the figure would be 36.8%.
axial DXA.
Using the 140 subjects in this study, the upper and
lower thresholds for Calscan for the non-dominant and
Discussion
dominant heels are shown in Figure 2. The proportion of
subjects in the equivocal group, and therefore requir- The Calscan proved reliable, precise, accurate and easy
ing axial DXA, is shown in Table 4. Based on these to use. Calscan performance was stable within a normal
thresholds of 0.391 g cm22 and 0.306 g cm22 for the range of room temperatures, and was not affected by
non-dominant and 0.395 g cm22 and 0.294 g cm22 for recent movement of the device. There was no difference

Figure 2. (a) DXL Calscan upper and lower triage thresholds for the non-dominant heel. (b) DXL Calscan upper and lower triage
thresholds for the dominant heel.

The British Journal of Radiology, April 2006 339


Table 4. ‘‘Evaluation and Diagnostic Thresholds of the DXL Calscan’’. Referral by number of subjects
Thresholds: Upper
Lower
Units: BMD (T-score). BMD in g cm22. T-score in standard deviations
Above upper threshold: reassure Non-osteoporotic
Osteoporotic
Between thresholds: refer Non-osteoporotic
Osteoporotic
Below lower threshold: treat Non-osteoporotic
Osteoporotic
Units: number of subjects
Referral rate
in performance when acquiring scans in quick succes-
sion, or with 30 min breaks between them, with the
enforced 4 min period between scans appearing suffi-
cient time for the X-ray tube to cool. Radiation dose to
patient and scatter dose to operator were low and the
device requires only a small controlled area. There is no
lower limit of applicability and the Ionising Radiation
Regulations still apply, requiring therefore the advice of
a radiation protection advisor, device risk assessment,
production of local rules, written procedures and
appropriate training of staff.
At 1.19% CV, precision in vivo for our 67 precision
subjects as a whole was slightly superior to the 1.24% for
the Calscan and 1.28% for the GE-Lunar Pixi reported by
Hakulinen et al [8], who performed repeat scans on 38
(18 male, 20 female) subjects with a mean (SD) age of
59.7 years (¡9.4 years). Although at 1.30% CV the
precision for our 19 osteoporotic precision subjects is
poorer than the 1.09% for the 48 non-osteoporotic
subjects, at 0.0042 g cm22 versus 0.0048 g cm22, the
mean absolute error per repeat measurement was
actually lower for the osteoporotic than for the non-
osteoporotic group, and so the difference in precision can
be explained by the difference in the mean BMD scores
for the two groups. As with all DXA systems, attainment
of good precision requires technical staff to be trained,
experienced and to practice good technique.
We found the coefficient of determination (R2) with
spine and total hip DXA to be 0.28 and 0.35 at the non-
dominant heel. Correlation at spine was lower than the
0.59 reported by Martini et al [25], or the 0.61 reported by
Hakulinen [8]. It is not clear if this is due to differences in
the sample groups and the small size of the Martini and
Hakulinen samples.
Using T-scores, the upper and lower triage thresholds
as defined by the NOS method for the Calscan were at
21.4 and 22.7, respectively, for the non-dominant heel.
These T-scores are only applicable to post-menopausal
white women aged 55–70 years who meet the normal
criteria for axial bone densitometry examination. As with
all T-scores, the exact threshold values depend on the
reference range. Were this to be changed, then the T-
scores would need to be recalculated from the under-
lying BMD scores of 0.391 g cm22 and 0.306 g cm22. In
addition, the T-score thresholds of any peripheral
devices employed in a triage role are likely to become
more negative with advancing subject age [20], but the
unreliability of spine DXA in subjects over 70 years of
age makes the calculation of peripheral threshold values
340
J A Thorpe and S A Steel
Non-dominant heel Dominant heel
0.391 (21.4) 0.395 (21.4)
0.306 (22.7) 0.294 (22.9)
30 31
8 7
33 32
41 50
7 7
21 13
52.9% 58.6%
problematic for such a group, without resorting to total
hip BMD alone.
There continues to be a growth in demand for bone
densitometry services through increased awareness of
health professionals and the public, rising healthcare
costs of fragility fractures and the development and
introduction of new bone protective treatments. The
provision and availability of such services, however,
remains patchy and inconsistent. In an area where
demand on axial DXA is exceeding capacity, peripheral
DXA could prove useful in a triage role to ensure best
and most cost effective use of this resource. However, a
comprehensive analysis of the resource implications of
such an approach is required. Applying the triage
thresholds to the population used for this evaluation
would suggest that over 50% would require referral for
axial DXA. As indicated, the study was not powered to
provide an accurate assessment of referral proportion
and the true figure is probably below 40%. Provided the
cost per case for the heel DXA measurement is less than
60% that of a spine and hip measurement, there should
be a net saving. Where the peripheral device is
community or primary care based, there may be an
increase overall in patients identified due to the more
accessible nature of the service which would reduce the
potential cost savings and also increase the burden on
the prescribing budget.
Where there is no access to axial DXA locally,
peripheral DXA may play a role in identifying those at
risk of fragility fracture provided it is used with care and
in appropriate populations with clearly identified clinical
risk factors. The Calscan device appears suitable for
either role using the thresholds derived in this study.
There is a high proportion (95%) of the more metaboli-
cally active trabecular bone in the calcaneus [15] which
would suggest that this site is sensitive to mechanisms
affecting bone metabolism. This, together with the
advantage of being a weight bearing bone, should better
reflect the changes occurring at the spine and hip than at
other peripheral sites. The moderate correlation between
the heel and axial sites observed in this study may be due
to sample bias as the subjects were drawn from those
attending for bone densitometry. The lack of agreement
observed generally between sites is also partly due to the
varying trabecular to cortical ratios with the spine being
50% trabecular and hip 40%.
There is no published evidence to date that patients
commenced on treatment on the basis of falling below
the lower triage threshold by pDXA could be monitored
The British Journal of Radiology, April 2006
Evaluation and diagnostic thresholds of the DXL Calscan
by pDXA. Ringe et al demonstrated promising results in
heel BMD with ibandronate [26], but they employed a
non-standard technique and do not compare the
observed 15% increase at 2 years with the least sig-
nificant change. It is also known that some bone
protective treatments are only effective in reducing
fractures in those defined osteoporotic by hip BMD
[27]. There are no data yet on effectiveness of treatments
in those targeted by the pDXA triage technique although
use of the derived lower pDXA threshold provides 90%
confidence that the patient would be found osteoporotic
by spine and hip, with almost all the remainder severely
osteopenic by spine or hip.
Use of peripheral devices in a triage role as an adjunct
to an established axial DXA service could bring
substantial benefits to both patient and healthcare
providers, and the Calscan is well suited for this
purpose. However, it should be operated only by
qualified personnel, used in selected populations and
results interpreted in conjunction with clinical risk
factors for fragility fracture.
Acknowledgments
The authors wish to acknowledge Vincent Mann and
Melanie Auty for their technical assistance in the
environmental measurements for this paper.
References
1. Royal College of Physicians. Osteoporosis: Clinical guide-
lines for prevention and treatment. London, UK: RCP, 1999.
2. U.S. Department of Health and Human Services. Bone
Health and Osteoporosis: A Report of the Surgeon General -
Chapter 8: Assessing the Risk of Bone Disease and Fracture.
U.S. Department of Health and Human Services, Office of
the Surgeon General, 2004.
3. National Osteoporosis Society. Position statement on the
reporting of dual energy x-ray absorptiometry (DXA) bone
mineral density scans. Bath, UK. NOS, August 2002.
4. Marshall D, Johnell O, Wedel H. Meta-analysis of how well
measures of bone mineral density predict occurrence of
osteoporotic fractures. Br Med J 1996;312:1254–9.
5. Stone KL, Seeley DG, Lui L-Y, et al. BMD at multiple sites
and risk of fracture of multiple types: long-term results
from the Study of Osteoporotic Fractures. J Bone Miner Res
2003;18:1947–54.
6. Eastell R. Treatment of postmenopausal osteoporosis. N
Engl J Med 1998;338:736–46.
7. Formica C, Loro ML, Gilsanz V, Seeman E. Inhomogeneity
in body fat distribution may result in inaccuracy in the
measurement of vertebral bone mass. J Bone Miner Res
1995;10:1504–11.
8. Hakulinen MA, Saarakkala S, Toyras J, Kroger H, Jurvelin
JS. Dual energy x-ray laser measurement of calcaneal bone
mineral density. Phys Med Biol 2003;48:1741–52.
9. Hausler KD, Rich PA, Barry EB. Water bath and contact
methods in ultrasonic evaluation of bone. Calcif Tissue Int
1997;61:26–9.
10. Swanpalmer J, Kullenberg R, Hansson T. The feasibility of
triple-energy absorptiometry for the determination of bone
mineral, Ca and P in vivo. Physiol Meas 1998;19:1–15.
The British Journal of Radiology, April 2006
11. Swanpalmer J, Kullenberg R, Hansson T. Measurement of
bone mineral using multiple-energy x-ray absorptiometry.
Phys Med Biol 1998;43:379–87.
12. Jonson R, Mansson LG, Rundgren A, Szucs J. Dual-photon
absorptiometry for determination of bone mineral content
in the calcaneus with correction for fat. Phys Med Biol
1990;35:961–9.
13. Swanpalmer J, Kullenberg R. A new measuring device for
quantifying the amount of mineral in the heel bone. Ann N
Y Acad Sci 2000;904:115–7.
14. Patel R, Blake GM, Fogelman I. Radiation dose to the
patient and operator from a peripheral dual x-ray absorp-
tiometry system. J Clin Densitom 1999;2:397–401.
15. WHO Technical Report Series 843. Assessment of fracture
risk and its application to screening for postmenopausal
osteoporosis. Geneva, Switzerland: World Health
Organization, 1994..
16. Grampp S, Genant HK, Mathur A, et al. Comparisons of
non-invasive bone mineral measurements in assessing age-
related bone loss, fracture discrimination and diagnostic
classification. J Bone Miner Res 1997;12:697–711.
17. Lu Y, Genant HK, Shepherd J, et al. Classification of
osteoporosis based on bone mineral densities. J Bone Miner
Res 2001;16:901–10.
18. Faulkner KG, Von Stetton E, Miller P. Discordance in
patient classification using T-scores. J Clin Densitom
1999;2:343–50.
19. National Osteoporosis Society. Position statement on the
use of peripheral x-ray absorptiometry in the management
of osteoporosis. Bath, UK: NOS, November 2004.
20. Blake GM, Chinn DJ, Steel SA, Patel R, Panayiotou E,
Thorpe J, et al. The Revised National Osteoporosis Society
Position Statement on Peripheral x-ray Absorptiometry: a
list of device specific thresholds for the clinical interpreta-
tion of pDXA examinations. Osteoporosis International; In
press.
21. The Ionising Radiations Regulations 1999. London, UK, The
Stationery Office Limited, ISBN 0 11 085614 7; 1999.
22. British Standards Institute (BSI). Medical electrical equip-
ment. Particular requirements for safety. Specification for x-
ray source assemblies and x-ray tube assemblies for medical
diagnosis. British Standards Publishing BS EN 60601-2-
28;1993.
23. European Committee for Electrotechnical Standardization
(CENELEC). Safety of laser products – Part 1: Equipment
classification, requirements and user’s guide. EN 60825-1;
1994.
24. Kullenberg R. Reference database for dual x-ray and laser
Calscan bone densitometer. J Clin Densitom 2003;6:367–72.
25. Martini G, Valenti R, Gennari L, Salvadori S, Galli B, Nuti R.
Dual x-ray and laser absorptiometry of the calcaneus:
comparison with quantitative ultrasound and dual-
energy x-ray absorptiometry. J Clin Densitom 2004;7:
349–54.
26. Ringe JD, Dorst A, Faber H, Ibach K, Preuss J. Three-
monthly ibandronate bolus injection offers favourable
tolerability and sustained efficacy advantage over two
years in established corticosteroid-induced osteoporosis.
Rheumatology 2003;42:743–9.
27. Cummings SR, Black DM, Thompson DE, Applegate WB,
Barrett-Connor E, Musliner TA, et al. Effect of alendronate
on risk of fracture in women with low bone density but
without vertebral fractures: results from the Fracture
Intervention Trial. JAMA 1998;280:2077–82.
341
 The British Journal of Radiology, 79 (2006), 342–346

SHORT COMMUNICATION

Time-dependent observer errors in pulmonary nodule detection

1 1 1
D MANNING, PhD, FInstP, S C BARKER-MILL, PhD, T DONOVAN, MSc and 2T CRAWFORD, PhD

1
School of Medical Imaging Sciences, St Martin’s College, Lancaster LA1 3JD and 2Department of
Psychology, Lancaster University, Lancaster, UK

ABSTRACT. The work was carried out to investigate differences in visual search
characteristics between groups of observers with different levels of experience in the
task of pulmonary nodule detection in chest radiology and we report here on these
differences in respect of time related decisions. Volunteer observers were divided into
three groups depending on their level of expertise. There were eight radiologists, eight
radiographers and eight novices. Their task was to detect pulmonary nodules in a test
bank of 120 digitized posteroanterior (PA) chest radiographs. Five of the eight
Received 2 March 2005
radiographers were tested twice: once before and once after a 6-month training Revised 24 August 2005
programme in interpretation of the adult chest radiograph. During each test session Accepted 30 September
the observers’ eye movements were tracked. Data on the observers’ decisions through 2005
Alternate Free Response Operating Characteristic (AFROC) methodology were
DOI: 10.1259/bjr/13453920
correlated to their eye-movement and fixation patterns. True negative decisions from
all observers were associated with shorter fixation times than false negative decisions. ’ 2006 The British Institute of
No correct negative decisions were made after fixations exceeding 3 s. Radiology

In previously reported studies we have investigated
observer experience [1, 2] and the effects of lesion
conspicuity [3] on performance in nodule detection from
plain chest radiology. Comparisons of diagnostic perfor-
mance through alternate free response receiver operating
characteristic (AFROC) [4] showed that after training and
extensive caseload experience radiographers’ detection
rate improved to approach that of the experts. AFROC is
a variation on receiver operating characteristic methodo-
logy that takes into account the location in the image of
the observer decision. Through this technique the
observer must indicate not only his decision on the
disease status of the image (positive or negative for
the presence of a pulmonary nodule) but also the
location of the lesion.
The observer groups were as follows:
N eight first-year student radiographers considered as
novices to chest interpretation tasks;
N eight experienced clinical radiographers before train-
ing in chest interpretation;
N five of the eight radiographers after their training in
chest interpretation; and
N eight radiologists.
Eye-tracking these observers gave insight into diffe-
rences between the groups in terms of their visual search
strategies and we concluded that, amongst other things,
the experts were more economical in their patterns of
search, carried out fewer fixations and spent less time on
the task. After a training period that included a
minimum of 500 cases but no specific instruction in
search patterns, the radiographers developed
spontaneously similar search strategies to those of the
radiologists. But both radiologists and trained radio-
graphers had a false negative error-rate in excess of 40%
[2]. We accept that the requirements of AFROC are
stringent (a false negative is defined by a missed lesion
rather than an incorrect case decision); but this is still a
significant miss-rate and in this paper we report our
observations on the relationship between the experience
of the observer, the duration of visual attention and the
probability of a false negative error in the task.
Eye movement, visual attention and visual scan
paths
The field of view in humans extends over 180 ˚ but only
the centre of the field provides sharp details.
Consequently, we move our eyes (and heads) to bring
interesting features into the centre. The pause over the
point of interest is known as a foveal fixation and has
duration of about 100–300 ms before we move our
interest to a new location in a fast jump called a saccade.
Fixations are characterized by multiple (clustering)
nature when we dwell extensively on a location because
the eyes do not remain stationary for long before losing
sensitivity. Saccades are too fast for information to be
gathered during their operation and so eye-tracking
experiments assume that fixations represent the location
of conscious attention. Viewers are not conscious of the
pattern of their saccades during an observation of a
scene, however, and it has been shown that scanpaths for
an individual are modified by the task presented [5]. The
path is the sequence in which image details are
scrutinized but fixation point clusters illustrate the parts
of the scene the observer find most interesting as

342 The British Journal of Radiology, April 2006

Short communication: Time-dependent observer errors
opposed to the parts that receive no attention at all. It has
been noted that the same details in an image can be
accessed using quite different scanpaths but similar
fixation clusters [6]. As a result, analysis of fixation
cluster location and dwell time gives a better opportunity
for performance comparisons than the scanpath
sequence if the research question is focused on whether
image features are being hit. However, scanpath data can
give fascinating insights into how individuals prioritise
locations of semantic interest although the analysis of
these paths is difficult and sometimes inconclusive.
Errors in radiology are known to be multifactorial but
a sizable proportion are known to be perceptual in origin
[7, 8]. This has led to interest in the possible sources of
perceptual error and the visual strategies of radiologists
in their task [9]. Early work on the visual dwell
behaviour of radiologists during film reading was
carried out by Kundel et al [10–12] who went on to later
describe survival curve methods for analysing time-
related decision-making in those observers [13, 14]. Our
aim here is to report on the timing of decisions, their
accuracy and their relationship to experience by the
survival analysis of decisions from eye-tracking data.
Aim
The aim of this study was to report an observed
relationship between the experience of the observer, the
duration of visual attention and the probability of
incorrect decisions in pulmonary nodule detection.
Materials and methods
Observers
Eight volunteers were recruited from a cohort of first
year radiography students who acted as the novice
group of film readers. Eight experienced radiographers
volunteered from a post-graduate course on chest image
interpretation at the commencement of their course. Five
of these radiographers then offered themselves for repeat
testing at the end of their 6 months training and 500 cases
of experience. Eight radiologists with extensive and
current experience in chest radiology volunteered to act
as the most experienced group of observers. All
observers gave their consent and all performance data
were made anonymous. Recruitment of all the observers
and the conduct of the experiments followed the ethical
guidelines for experiments involving volunteer human
subjects at St Martin’s College and Lancaster University.
At the time when the data for these experiments were
collected (in 2001) there was no requirement of COREC
approval for NHS employees who were acting as
volunteer participants in this type of research.
Detectable nodules
Each observer viewed a bank of 120 digitized chest
images of adults. The images contained 81 pulmonary
nodules agreed as significant in pathological appearance
from confirmed radiological reports. Nodules were
The British Journal of Radiology, April 2006
roughly circular and ranged in diameter size from
5 mm to 20 mm with varying degrees of measured
conspicuity [3]. Nine films contained more than one
nodule and 30 nodules were located in these multi-
nodule films. Normal films were included in the
observer task and the complete test bank was divided
into three sets of 40 images with prevalence-rates of 12%,
50% and 82%.
Observer performance measurement
Alternate free response operating characteristic meth-
odology (AFROC) was used [4]. This required observers
to indicate a location to a decision for a lesion and to
assign a score between 1 and 4 on their level of
confidence in that decision. A zero score was allocated
to all decisions of ‘‘no nodule present’’. In AFROC
methodology false positive decisions are treated in the
following way: the highest scoring false positive decision
is the only one recorded per image which avoids the
possibility of infinite values in summing false positive
responses.
Observer test sessions were never longer than 1 h to
avoid the effects of fatigue on performance. There was a
minimum 6-month interval between the before- and
after-training observer tests on the radiographers to give
an effective case-memory washout period.
Parameters
All eye-tracking data comparing the performance of
the observer groups were processed through the ASL
(Applied Science Laboratories, Bedford, MA), software
EYENALH. The measured parameter from the eye
tracking for use in the present study was the accumu-
lated dwell time at each decision point.
A fixation
We defined a fixation (visual dwell) as a point of gaze
remaining continuously within a 1 ˚ area for at least
100 ms. A 1 ˚ angle subtended from the eye to the image
at a viewing distance of 40 cm approximates to a circle
the size of a 2.5 cm diameter disc at arms length. Re-
fixations were summed to give cumulative fixation times
to clusters when there was overlap of the 1 ˚ areas up to a
5 ˚ area. These definitions are ones that are commonly
used in work of this kind [9, 14]. The observers were
allowed to search freely and no limit was imposed on the
duration of inspection for any given image.
A true negative decision was defined as a timed
fixation of a lesion-free zone of the chest image that
elicited a zero response from the observers on the
AFROC scale.
The dwell-time data for all fixations related to positive
and negative decisions were analysed through the
statistical package SPSSH to provide information on the
percentage survival of decisions over time.
These data allowed us to characterize the observers’
decisions in greater detail than true and false negative and
true and false positive outcome, giving an opportunity to
343
 D Manning, S C Barker-Mill, T Donovan and T Crawford

identify time-related features of decision outcomes. The their true positive decisions took more than 5000 ms of
time related information on decisions gave an opportunity visual scrutiny.
to analyse time differences between correct and incorrect In the tables we have highlighted the time values for
decisions whether positive or negative. the completion of all the observer decisions in their
outcome categories to draw attention to the differences in
duration between the true and false decisions in each
Results case. For all observers the false negative decisions took
them significantly longer than true negatives. For
positive decisions, correct decisions were made faster
Time related decisions: survival analysis
than incorrect ones except in the case of the novices who
We investigated how the four possible decision dwelt longer on genuine lesions before deciding they
outcomes of true and false positive (TP FP) and true were pathologies.
and false negative (TN FN) related to the duration of The survival curves show that for all observers 50% of
gaze through a survival analysis of the fixation data. The all their true negative decisions were made within
results are presented in Tables 1–4 and Figures 1–4. 1000 ms of gaze duration and the false negative decisions
Survival analysis is used in this context as a means of have dwell-times somewhere between the positives (true
showing the proportions of decisions that are completed and false) and the true negatives. This is consistent with
for each category (TP FP TN FN) at increasing the findings of others who have carried out this type of
accumulated time intervals of visual attention. It is analysis [14] but our results indicate extension along the
analogous to the cell survival curves used in radio- time axis with decreasing levels of experience.
biology to indicate the proportion of cells surviving
increasing radiation doses [15]. So in Table 1 and
Figure 1 for example, 50% of all true positive decisions Discussion
made by the radiologists survived 2200 ms of accumu-
lated visual attention on the tumour location but none of The aim of this work was to analyse eye-tracking data
to classify the false negative errors made when observer
groups with different levels of experience are asked to
Table 1. Radiologists detect pulmonary nodules. The different levels of
experience in the groups give some insight into how, if
Survival duration (ms) for each decision group
at all, errors vary with expertise. The time related data
TN FN TP FP % surviving provided measurable differences in the fixation times
0 0 0 0 100
associated with observers making correct or incorrect
200 220 400 220 95 decisions.
220 240 455 250 90
450 500 1000 750 80
680 700 1680 1700 70 Survival analysis
750 800 1800 1800 60
790 1000 2200 2200 50 Observers’ decisions are made over time periods that
820 1250 2500 2700 40 can be related to the duration of visual fixations from
900 1600 3000 3500 30 eye-tracking data. Mean or median dwell times over
1010 2000 3500 4000 20 trials and between readers can be calculated but dwell
1800 3000 4500 4700 10
times are not normally distributed, making statistical
2250 4750 5000 5500 0
comparisons difficult to interpret. The technique of
TN, true negative; TP, true positive; FN, false negative; FP, survival analysis has been found useful in these
false positive. circumstances [13, 14] and this operation on the data
demonstrated some consistent findings for categories of
decisions.
Table 2. Radiographers post training in chest interpretation
Figures 1 and 2 show the family of decisions survival
Survival duration (ms) for each decision group curves for radiologists and radiographers with chest
TN FN TP FP % surviving interpretation training derived from the data in Tables 1
and 2. Virtually all the true negative decisions made by
0 0 0 0 100 these, the most experienced observers, were made within
200 220 300 500 95 2 s of cumulative fixation time on an image feature. False
300 330 800 600 90
negative decisions were characterized by longer dwell
450 500 900 800 80
500 500 1200 1000 70 times but no negative decisions (either true or false) were
500 550 1850 2000 60 made by radiologists after 4.75 s fixation. Figures 3 and 4
550 700 2500 4000 50 and Tables 3 and 4 show similar findings for the low
600 1100 3000 5000 40 experience groups but there is a tendency towards longer
800 2000 4400 5500 30 fixation times for incorrect negative decisions with
1000 2500 5000 6000 20 decreasing levels of experience. In the case of the novices
1500 4000 6000 7000 10 (Table 4) lesions were missed after a cumulative gaze
3000 5000 6600 8000 0 duration of up to 8 s. The time difference in the
TN, true negative; TP, true positive; FN, false negative; FP, proportions of true and false negative decisions made
false positive. by these observers can be summarized by saying that in

344 The British Journal of Radiology, April 2006

Short communication: Time-dependent observer errors

Table 3. Radiographers pre training in chest interpretation

Survival duration (ms) for each decision group

TN FN TP FP % surviving

0 0 0 0 100
300 300 350 350 95
330 400 800 820 90
330 500 1100 1200 80
400 600 2000 2000 70
560 1200 3000 3100 60
700 1600 3300 3800 50
850 2000 3800 4800 40
1000 2400 4200 5500 30
1200 3000 5100 6500 20
1600 4500 6500 8000 10
2400 6500 8200 9000 0
TN, true negative; TP, true positive; FN, false negative; FP,
false positive.

Table 4. Novices
Survival duration (ms) for each decision group
Figure 1. Time-related decisions for radiologists. Compared
TN FN TP FP % surviving with less experienced observers the radiologists made their
0 0 0 0 100 decisions faster. The separation between the positive and
300 300 500 400 95 negative decision pair curves is closer than for all other
400 400 1000 500 90 groups.
500 500 1700 800 80
550 600 2000 1200 70
600 1000 2600 1500 60
680 1500 3800 2000 50
800 1800 4500 3000 40
900 2200 5500 4500 30
1500 3500 6500 5600 20
2200 5000 9000 8000 10
3000 8000 12000 10000 0
TN, true negative; TP, true positive; FN, false negative; FP,
false positive.

this experiment, all negative decisions made after gaze

duration of 3 s were incorrect.
It seems that correct negative decisions (correct deci-
sions that normal features are not nodules) tend to be
made rapidly after fixation occurs and that this is a feature
of expert performance. Conversely, incorrect negative
decisions are characterized by extended gaze duration
especially in novice readers. When areas of the images
hold their attention for several seconds of accumulated
fixation time, observers show a semantic interest in the
appearances. This suggests that they are suspicious of the
feature and they are operating on the information at a
perceptual/cognitive level. These errors are not failures of
detection but of recognition and decision and can be Figure 2. Time-related decisions for radiographers post-
explained partly by the visual ambiguities of the image training. After experience and training over 6 months and
and partly by the level of experience of the reader. The 500 cases the radiographers have speeded their decision
finding may be important because of: (a) its potential for making and markedly reduced the separation of their true
reducing false negative error if visual dwell information is and false negative curves. Compare with Figure 3.
fed back to the reader; and (b) its potential use in training
schemes for radiologists and monitoring the effects of that their negative decisions are more likely to be incorrect
caseload experience. when they are made after a period of indecision over a
If these results are reproducible in other settings and our particular image feature. In short, we suggest informing
interpretation of their meaning is correct there are several observers that if a feature looks suspicious enough to
ways that they may help in the improvement of observer warrant more than 2 s of their attention it is probably not
performance. A simple expedient is to inform film readers innocent. More sophisticated aids linked to these findings

The British Journal of Radiology, April 2006 345


Figure 3. Time-related decisions for radiographers
pre-training.
Figure 4. Time–related decisions for novices. Novices show a
similar trend to the experienced observers although their
positive decision curves are reversed and more widely
separated. All their decisions extend further along the time
axis.
might involve computer aided feedback to observers in
real-time to give indication of the gaze duration for
individual locations in the image.
Conclusion
The data and their analysis from an eye-tracking
experiment have given insights into the errors made by
observers with different levels of experience. The most
notable outcomes of this are that the duration of fixation
on a feature in the chest image may be an effective
346
D Manning, S C Barker-Mill, T Donovan and T Crawford
discriminator in predicting whether a negative decision
will be correct or incorrect.
We consider this observation to be potentially impor-
tant for feedback strategies for education purposes and
to aid performance.
Acknowledgment
We wish to thank the radiologists, radiographers and
students, all of whom are associated with the School of
Medical Imaging Sciences at St Martin’s College,
Lancaster, UK, who kindly agreed to take part in this
study. Eye-tracking equipment for this research was
supplied through a grant from the Peter Barker-Mill
Memorial Trust.
References
1. Manning DJ, Ethell S, Crawford T. An eye-tracking AFROC
study of the influence of experience and training on chest x-
ray interpretation. Medical imaging 2003; Image Perception
Observer Performance and Technology Assessment. Editors
Dev Chakraborty and Elizabeth Krupinski, Proc SPIE
2003;5034:257–266.
2. Manning DJ, Ethell SC, Crawford T, Donovan T. How do
radiologists do it? The influence of experience and training
on searching for chest nodules. Radiography Available
online 23 March 2005. Item 49, www.sciencedirect.com/
science/journal/10788174.
3. Manning DJ, Ethell SC, Donovan T. Detection or decision
errors? Missed lung cancer from posteroanterior chest
radiograph. Br J Radiol 2004;77:231–5.
4. Chakraborty DP. Statistical power in observer performance
studies: comparison of the receiver operating characteristic
and free response method in tasks involving localization.
Acad Radiol 2002;9:147–56.
5. Yarbus AL. Eye movements and vision. New York, NY:
Plenum Press, 1967.
6. Kundel HL. Reader error, object recognition and visual
search. In: Chakraborty D, Eckstein M, editors. Medical
imaging 2004: image perception, observer performance
and technology assessment. Proceedings of SPIE 2004;5372:
1–9.
7. Fitzgerald R. Error in radiology. Clin Radiol 2001;56:938–46.
8. Chief Medical Officer. Learning from failure: evidence and
experience. An Organisation with a Memory. London:
Stationery Office, 2000:1–7.
9. Gale AG. Human response to visual stimuli. In: Hendee
WR, Wells PNT, editors. The perception of visual informa-
tion, 2nd edition. New York, NY: Springer, 2000:134.
10. Kundel H, Nodine C, Carmody D. Visual scanning, pattern
recognition and decision making in pulmonary nodule
detection. Invest Radiol 1978;13:175–81.
11. Kundel H, Nodine C, Krupinski E. Searching for lung
nodules: visual dwell indicates locations of false positi-
veand false negative decisions. Invest Radiol 1989;24:472–8.
12. Kundel H, Nodine C, Toto L. Searching for lung nodules:
the guidance of visual scanning. Invest Radiol 1991;26:
777–81.
13. Anderson S, Auquier A, Hauk WW, Oakes D, Vandaele W,
Weisberg HI. Statistical methods for comparative studies.
New York, NY: John Wiley & Sons, 1980.
14. Kundel H. Visual search in medical images. In: Beutel J,
Kundel H, Van Metter R, editors. Handbook of medical
imaging Vol 1 Physics and psychophysics. Bellingham WA:
SPIE, 2000:837–58.
15. Motulsky H. Intuitive biostatistics. Oxford: Oxford
University Press Inc., 1995:Chapter 6.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 347–352

SHORT COMMUNICATION

Study of scattered radiation for in-air calibration by a multiple-

distance method using ionization chambers and an HDR 192Ir
brachytherapy source
1 2
N P PATEL, MSc, DipRP, B MAJUMDAR, MSc, PhD, PDC and 3V VIJAYAN, MSc, DipRP, PhD

1
Department of Medical Physics, Acharya Harihar Regional Cancer Centre, Cuttack-753007, India,
2
Department of Physics, Government College of Science, Raipur-492010, India and 3Health Physics
Unit, Institute of Physics, Bhubaneswar-751005, India

ABSTRACT. The aim of this study is to estimate the room-scatter correction when
measuring air kerma rate of an HDR 192Ir brachytherapy source by in-air calibration. The
variation in scattered radiation due to the specially designed jig and from the room
walls was also studied. Two therapy ion chambers of volume 0.1 cm3 and 0.6 cm3 were
used in the present study. Air kerma was measured by placing the source at several
distances between 10 cm and 20 cm from the chamber. The scatter radiation was
determined by superimposing the theoretically derived model curve of known scatter
(based on the inverse square law) over the plot of measured air kerma strength values.
The scatter radiation was estimated in terms of percentage of the primary radiation at
10 cm measurement distance. The scatter estimated by the 0.6 cm3 chamber at two
Received 7 December 2004
positions was 0.33% and 0.59%, respectively. Similarly the scatter estimated at two Revised 28 September
other positions by the 0.1 cm3 chamber was 0.58% and 1.11%. This variation in scatter 2005
with position as well as with the chamber was due to the varying scatter contribution Accepted 3 October 2005
from components of the measurement set-up. The scatter radiation becomes constant at a
DOI: 10.1259/bjr/54848398
distance greater than 100 cm from the walls of the room. We conclude that a fixed
chamber with changing source positions should be used in multiple-distance ’ 2006 The British Institute of
measurement of air kerma rate when using a measurement jig. Radiology

Calibration of brachytherapy sources is an end-user
requirement prior to clinical use. The recommended
quantity for specifying a brachytherapy source is the air
kerma strength (AKS) [1]. Air kerma strength is defined
as the product of air kerma rate at a calibration distance,
d, in free space, K(d), measured along the transverse
bisector of the source, and the square of the distance, d. A
source calibration accuracy of ¡3% relative to existing
AKS standards seems reasonable [2]. At present, high
dose rate 192Ir sources are widely used for brachytherapy
treatment due to their high specific activity and low
effective photon energy.
Calibration of an HDR 192Ir source is performed by in-air
measurement or by a solid phantom technique using a
thimble chamber. An alternative method is to use a well-
type chamber. An IAEA technical document has discussed
several factors including measurement distance, chamber
size, positioning uncertainty, charge leakage and room
scatter to be considered for in-air measurement [3]. The
measurement distance should be chosen such that the
combined uncertainty in source calibration due to other
factors can be minimized. The optimum measurement
distance estimated for a Farmer-type chamber with an
HDR 192Ir source is 16 cm [4]. Experimental studies have
estimated the room scatter using a large volume spherical
chamber (3.6 cm3) by multiple-distance measurement from
10 cm to 40 cm in a specially designed jig [5, 6]. The ESTRO
guidelines recommend minimizing the positioning uncer-
tainty and scatter contribution using a calibration jig with
source positions at a distance of 10 cm from both sides of a
centrally placed 0.6 cm3 Farmer-type chamber [7]. The
Monte Carlo study found that the assumption of constant
room scatter over the measurement distance is not valid
and that in-air measurement needs correction for accurate
estimation of AKS [8].
The calibration jig is designed to hold the ionization
chamber and source applicator for in-air measurement.
Although several factors that might affect the source
calibration have been taken care of in designing the jig,
the contribution of scattered radiation seems significant.
The aim of the present study is to estimate the scattered
radiation by a multiple-distance method using therapy
level ionization chambers. The variation in scatter
radiation inside the jig and from the walls of the
measurement room is discussed.
Methods and materials
We have used a slightly different method in the
present study to estimate the scattered radiation from

The British Journal of Radiology, April 2006 347


that used by earlier experimental studies [5, 6]. In air
measurement, the measured air kerma (Md) at distance d
is the sum of the primary (Mp) and scatter (Ms) radiation,
i.e. Md5Mp+Ms. The primary radiation (Mp) follows the
inverse square law and the scatter radiation (Ms) is a
constant. The product of Md and the square of the
distance d is a variable quantity which is denoted by f(d)
and expressed by Equation (1).


f ðd Þ~Md ðd Þ2 ~ Mp zMs d 2
The f(d) values for different source to chamber distances
were normalized with respect to the f(d) value at
minimum measurement distance and plotted against d.
The scatter (Ms) shown in Equation (1) is unknown but
can be determined by generating an equivalent curve
from theoretical calculation.
In the theoretical calculation, the air kerma from
primary radiation Ep at minimum measurement distance
was taken to be one and the inverse square law was used
to calculate the air kerma at all other true measurement
distances d95d+D, where D is the offset error in
measuring the initial nominal distance d. The offset error
D is due to the offset positioning of the chamber (č) and
source (š), i.e. D5č+š. The variable quantities F(d) were
determined using Equation (2):


F ðd Þ~ Ep zEs d 2
The scatter value Es was assumed constant over the
range of measurement distance. The curve of normalized
F(d) values vs distance d was generated by choosing the
value of Es such that:
h i

 
Md ðd Þ2 & Ep zEs d 2 plot vs: d
plot vs: d
then Ms5Es
The condition as described in Equation (3) was
accomplished when the theoretical curve F(d) was fitted
to the measured normalized f(d) curve. It is evident from
Equation (2) that for a constant value of Es, different
curves of F(d) vs distance can be generated for primary
radiations Ep with different offsets D. In the present
study, nominal distances d for measurement of air kerma
(Md) were chosen between 10 cm and 20 cm. Calculation
was performed using Excel Workbook (Microsoft
Corporation).
A simple mathematical formula was derived to show
the relationship between the scatter levels and ratio of air
kerma for constant distance at two different positions
inside the room. If r is the ratio of measured air kerma
from unknown to known scatter position, air kerma from
primary radiation (Mp) and the scatter at two positions
are (Ms) and (Ms9), respectively, we can express the
unknown scatter by


Ms0 ~r Mp zMs {Mp
The GammaMed Plus brachytherapy unit with an
HDR 192Ir source supplied by MDS-Nordion Haan
348
N P Patel, B Majumdar and V Vijayan
ð1Þ
Figure 1. Measurement set-up shows the jig with a 0.1 cm3
chamber and brachytherapy unit.
GmbH, Germany (at present Varian Medical Systems)
was used in this experiment. Two different therapy ion
chambers (volume 0.6 cm3 and 0.1 cm3) with UNIDOS
dosimeter from PTW – Freiburg, Germany were used.
The wall thickness and build up cap (60Co beam) of the
0.6 cm3 chamber were 0.054 g cm22 and 0.547 g cm22,
respectively. The wall thickness of the 0.1 cm3 chamber
without any build up cap was 0.12 g cm22.
Figure 1 shows the specially designed measurement
jig. The dimensions of the jig were about 30 cm 6 20 cm
6 27 cm and it was made up of non-scattering low Z
ð2Þ
materials of acrylic plates and wooden frames. Figure 2a
shows the lateral view of the measurement set-up. The
chamber holder was mechanically fixed at positions A
and B in the jig whereas the applicator holder was
shifted linearly along the track. A fine laser beam was
projected over the jig to verify the sagittal, transverse and
coronal cross-sections planes. Two parallel scale systems
aligned with the help of laser beams were used to
ð3Þ
determine the measurement distance and the vertical
position of the source applicator. A laser beam and
magnifying glass were used to determine the offset
position (č) of the central axis of the chamber with
the scale system. Figure 2b shows the positioning of the
centres of the chamber and source with respect to the
measuring scale. The source applicator (inner and outer
diameter of 1.35 mm and 1.65 mm, respectively) can be
placed at any scale pointer with an accuracy of about
¡0.01 cm.
A plan view and the dimensions of the brachytherapy
treatment room used in the present study are shown in
Figure 3. The wall of the treatment room is made of
concrete (density 2.35 g cm22). The figure indicates the
location of measurement at the centre of the room and
other locations used to study single and double wall
scattering effects. The contribution to scattering from the
walls at the centre of the room was assumed negligible.
The locations of measurement for the wall effect study
were at various distances from 12.5 cm to 125 cm from
the single wall (square marks) and as well as from two
walls (circle marks) as shown in the Figure 3.
First, the jig was placed at a height of 120 cm from the
ð4Þ
floor at the centre of the room. The 0.6 cm3 chamber was
fixed at position A in the jig. The source applicator was
placed 1 cm from the chamber. Successive measure-
ments were taken by vertical scanning of the source to
The British Journal of Radiology, April 2006
Short communication: Scattered radiation for in-air calibration
Figure 2. (a) Lateral view of measurement set-up. The chambers 0.6 cm3 and 0.1 cm3 are placed at positions A and B,
respectively. The source position c is for the 0.6 cm3 and c’ for the 0.1 cm3 chamber that means the heights of the effective points
of chambers differed by 1.4 cm. (b) The offsets in positioning of the centre of chamber (č) and source (š). The total offset in the
setup distance is given by D52č+š.
determine the reference dwell position where the
chamber shows maximum response. In the vertical
scanning, the source was moved to dwell positions
(stepping size of 0.2 cm) along the applicator. The
reference dwell position where the chamber shows
maximum response lies on the transverse axis that
passes through the centre of the source as well as
chamber. For measurement of air kerma, the source
applicator was placed at measuring distance and the
source was moved to the reference dwell position. Air
kerma was measured for 100 s. A similar procedure was
followed for measurements at other distances. The
nominal distances (d) were 10 cm, 12 cm, 14 cm, 16 cm,
18 cm, and 20 cm. At each nominal distance, five source
offset positions (š5–0.4, –0.2, 0.0, +0.2, and +0.4 cm) were
taken. Air kerma measurements at offsets of 10 cm
distance were repeated as their readings were used for
normalization. Hence 30 values of air kerma at different
source to chamber distance were measured. Measured
values were corrected for air attenuation and non-
uniformity due to chamber size [9]. Transit time
correction was not required as the dosemeter timer was
used to collect the charge during an interval after the
source stopped moving. The leakage current was
negligible. A change in temperature and pressure was
monitored and successive corrections were made. The
offset of the chamber position (č) was determined to
calculate the total offset (D) as shown in Figure 2b. The
scattered radiation Es was estimated as per the metho-
dology described in the text using a computer. A similar
procedure was followed to estimate the scatter for the
0.6 cm3 chamber at position B. In the case of the 0.1 cm3
chamber, a similar procedure (except the stepping
distance was 0.1 cm and air kerma measurement period
200 s) was used to estimate the scatter at positions A and B.
The jig was placed at different locations (circle and
square marks) successively as shown in Figure 3 to study
the scatter effects from the walls of the room. A similar
procedure was followed to estimate the scattered
radiation at every location by taking 18 air kerma
measurements (only for three offset positions of the
source) with the 0.6 cm3 chamber placed at position A in
the jig. The scatter radiation vs distance from the single
and double wall was determined.
The British Journal of Radiology, April 2006
The relationship given in Equation (4) was verified
experimentally. The estimated scattered radiation Ms at
the centre of the room for the 0.6 cm3 chamber at position
A in the jig was taken as the reference value. The
estimated scattered radiation at other locations (taken as
Ms9) was used to calculate the ratios of air kerma r
(relative ionization) for the locations to the centre of the
room. To verify the accuracy of calculated relative
ionization (r), the air kerma for constant source to
chamber distance was measured at the centre of the
room and the other locations.
Results
The maximum response of the 0.6 cm3 and 0.1 cm3
chambers was found at different source dwell positions
6.8 cm and 5.4 cm from the end of the source applicator,
respectively. This means that the heights of the effective
centres of the chambers from the base of the jig differed
by 1.4 cm. Reproducibility (n55) of our measurement by
repositioning of the source applicator was within 3%,
1%, and 0.1% at measurement distances of 1 cm, 5 cm,
and 10 cm, respectively. The linearity of the dosemeter
within the range of 10–200 s was 0.01% for the 0.6 cm3
chamber and 0.02% for the 0.1 cm3 chamber.
Figure 4 shows the plot of normalized f(d) values from
measured air kerma (dots) and theoretically derived F(d)
(lines) vs nominal distance d for different offsets (D) for
the 0.6 cm3 chamber at position B. It shows that the plot
of measured f(d) values are best fitted by the theoretical
curves of corresponding offsets (D) with varying scatter
components Es of the primary at a measurement distance
of 10 cm. The mean of these scattered components Es was
0.59¡0.06%. This means that the measured air kerma has
a contribution from scatter (Es) of about 0.59% of the
primary radiation at 10 cm distance and the scatter
quantity is assumed to be constant over the range of
measurement distance.
The estimated scattered radiation in terms of percen-
tage of primary radiation at 10 cm measurement distance
for both the chambers at positions A and B is shown in
Table 1. These results were obtained with the jig placed
at the centre of the room. The uncertainty in the
349

positioning of chamber was about ¡0.02 cm, which
resulted in uncertainty of ¡0.1% in the estimation of
scattered radiation. The scatter measured by the 0.6 cm3
chamber at position A was 0.33% and when the chamber
was moved inside the jig by 2 cm at position B, the
measured scatter level increased to 0.59%. Similarly the
scatter radiation measured by the 0.1 cm3 chamber at
positions A and B was 0.58% and 1.11%, respectively.
Even if the uncertainty in chamber position is taken into
account, our results show a variation in the scattered
radiation over the different measurement positions
inside the jig. There is also a significant difference
Figure 4. The plot of normalized f(d) values from measured
air kerma (dots) and theoretically derived F(d) (lines) vs
nominal distance d for the 0.6 cm3 chamber at position B.
The measured f(d) curves were fitted by theoretical curves
F(d) of corresponding offsets (D) with varying scattered
components (Es). The offsets (D) in nominal distance were
+0.3 cm, +0.1 cm, 20.1 cm, 20.3 cm, and 20.5 cm, respec-
tively. The percentage value in brackets is the scatter
component. The mean of scattered components was
0.59¡0.06% at the centre of the room for the 0.6 cm3
chamber at position B in the jig.
350
N P Patel, B Majumdar and V Vijayan
Figure 3. Plan view of the brachytherapy
treatment room. Not to scale.
between the results for the 0.1 cm3 and 0.6 cm3 chamber.
It should, however, be noted that the centres of the
chamber volume were located at different heights. An
additional chamber holder, which was very close to the
chamber volume in measurement with 0.1 cm3, could
have increased the scattered contribution.
Figure 5a shows the scattered radiation as a function of
distance from a wall in single and double wall scattering,
estimated by multiple-distance measurement using the
0.6 cm3 chamber at position A. It is observed that the
scattered radiation becomes constant at about 0.33% at
distances greater than about 100 cm from the walls for
both single and double wall scattering. A steady increase
in scattered radiation was observed below 50 cm from a
single wall and 75 cm from the double wall. At 12.5 cm
from the wall, scattered radiation was estimated at 2.0%
and 2.7% of the primary for single and double wall,
respectively.
The deviation in measured relative ionization from
calculated relative ionization as a function of the distance
from a single and double wall for the 0.1 cm3 and 0.6 cm3
chamber is shown in Figure 5b. The calculated relative
ionizations were determined from the estimated scat-
tered radiation at different locations using the 0.6 cm3
chamber as shown in Figure 5a. Results show that
calculated relative ionization is in good agreement with
measured relative ionization with most of the deviations
within ¡0.5% for both the chambers.
Table 1. Estimation of scattered radiation for the chambers
Type of the chamber Estimated value of scattered radiation
for n55 (in % of primary radiation at
d 510 cm)
Position A Position B
3
0.6 cm chamber 0.33 (¡0.04) 0.59 (¡0.06)
0.1 cm3 chamber 0.58 (¡0.05) 1.11 (¡0.04)
The British Journal of Radiology, April 2006
Short communication: Scattered radiation for in-air calibration
Figure 5. (a) Estimated scattered radiation as a function of distance from the wall in double and single wall scattering by
multiple-distance measurement using the 0.6 cm3 chamber at position A inside the jig. (b) Deviation in measured relative
ionization from the calculated value as a function of distance from wall for both the chambers (DW, double wall; SW, single
wall). The measurement distances were 12 cm, 16 cm, and 18 cm. The calculated relative ionizations were determined from
estimated scattered radiations using the 0.6 cm3 chamber.
Discussion
The chamber used in the present study was fixed and
the source applicator was moved to different positions.
The accuracy of positioning of the source applicator was
about ¡0.01 cm. Displacement of the source inside the
applicator was crucial in the measurement as it has a
smaller diameter (0.9 mm) than the inner diameter of the
applicator. Displacement of the source was not observed
during reproducibility measurements at any reference
position when performed by repositioning the applica-
tor. However, when the applicator was moved to another
position, the source was displaced laterally due to
movement of the source guide tube. The maximum error
in setup distance may increase by up to ¡0.032 cm as a
result of the combined effects of source displacement and
applicator positioning. In order to minimize the uncer-
tainty in the estimation of the scattered radiation due to
positioning error, we have considered five offset posi-
tions of the source. The mean of scattered values from
five offsets was calculated. The standard deviations in
the measured f values (after correction of scatter) were
found to be well within 0.2% and 0.3% for the 0.6 cm3
and 0.1 cm3 chamber, respectively.
Measurements were also compared with the three-
equation solution technique described earlier by other
authors [5, 6]. It was observed that the shape of the curve
derived from Equation (3) contains a unique pair of scatter
(Es) and offset (D) values. The solution technique was
found to be very sensitive to measurement error. For
example, a deviation of ¡0.2% in air kerma values from
primary radiation (without any scatter contribution) at five
distances gives the scatter value from –0.8% to 0.65% using
the solution technique. However, in the present analysis,
the estimated scattered radiation was found to be well
within ¡0.1%. IAEA guidelines recommend that offset
values determined from the solution of the three equations
should not vary by more than ¡0.1 cm [3]. This limit can
be retained only if the deviation in measured f values (after
correction of scatter) is about ¡0.03%. Therefore, in the
solution technique, separations between measurement
distances must be very precise. In the present method,
The British Journal of Radiology, April 2006
variation in separation is managed by curve fitting but it is
essential to know the measurement distance for accurate
estimation of scatter radiation.
The air kerma at any point is the sum of primary and
scatter radiation. If the multiple-distance method is used
then constancy of scatter over the measurement distance is
essential for its precise estimation. Our results show that in
addition to the constant scatter contribution from the
room, the components of the jig also contribute signifi-
cantly to scatter, which varies from position to position
inside the jig. A large difference in scatter values between
positions A and B shows a high gradient scatter field over
the inner to outer area of the jig. The curves in Figure 4
show that the measured f(d) values of each offset were in
‘‘best fit’’ with the theoretical curve F(d), if the scatter
gradient over the central area of the jig is assumed to be
low. This means the variation in scatter contribution to the
chamber from multiple source positions is negligible.
Therefore, a measurement set-up with a fixed chamber
position should be chosen to minimize the variation in
scatter inside the jig for the multiple-distance method.
The optimum distance for a Farmer-type chamber to
give minimum combined uncertainty effect from several
factors was found at 16 cm [4]. IAEA recommends
measurement at multiple distances between 10 cm and
40 cm [3]. In general, the calibration jig is required for the
multiple-distance measurement, thus requiring estima-
tion of scattered radiation for the local conditions. In case
of the Farmer-type chamber, the measurement distances
should be selected around the optimum distance of
16 cm and a separation of 2 cm between distances is
sufficient to give satisfactory results for the multiple-
distance method. If a larger range of measurement
distances is used, the variation in scatter from the jig
could enhance the uncertainty in estimation of scatter
radiation as well as air kerma rate.
Conclusion
This study suggests that a fixed chamber with changing
source positions should be used in multiple-distance
351

measurement of air kerma rate when using a measure-
ment jig. The source should be placed at different
positions over a short-range around the optimum
measurement distance of 16 cm for the Farmer-type
ionization chamber. Scattered radiation estimated from
above method remains unchanged for the reproducible
measurement set-up, which makes the procedure of
frequent source calibration very simple.
References
1. American Association of Physicists in Medicine.
Specification of brachytherapy source strength. AAPM
Report 21. New York, NY: AAPM, 1987.
2. Nath R, Anderson LL, Meli JA, Olch AJ, Stitt JA, Williamson
JF. Code of practice for brachytherapy physics: Report of the
AAPM Radiation Therapy Committee Task Group No. 56.
Med Phys 1997;24:1557–98.
3. IAEA-TECDOC-1274. Calibration of photon and beta ray
sources used in brachytherapy - Guidelines on standardized
352
N P Patel, B Majumdar and V Vijayan
procedures at Secondary Standards Dosimetry Laboratories
(SSDLs) and hospitals. Vienna: IAEA, 2002.
4. DeWerd LA, Ezzell GA, Williamson JF. Calibration principles
and techniques. In: Nag S, editor. High dose rate brachyther-
apy. New York: Futura Publishing Company Inc., 1994.
5. Goetsch SJ, Attix FH, Pearson DW, Thomadsen BR.
Calibration of 192Ir high-dose-rate afterloading systems.
Med Phys 1991;18:462–7.
6. Stump KE, DeWerd LA, Micka JA, Anderson DR. Calibration
of new high dose rate 192Ir sources. Med Phys 2002;29:1483–8.
7. ESTRO Booklet No. 8: A Practical guide to quality control of
brachytherapy equipment. Mounierlaan 83/12 – 1200
Brussels (Belgium): ESTRO 2004.
8. Palani Selvam T, Govindarajan KN, Nagarajan PS,
Sethulakshmi P, Bhatt BC. Monte Carlo aided room scatter
studies in the primary air kerma strength standardization of
a remote afterloading 192Ir HDR source. Phys Med Biol
2001;46:2299–315.
9. Kondo S, Randolph ML. Effect of finite size of ionization
chambers on measurements of small photon sources. Radiat
Res 1960;13:37–60.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 353–355

SHORT COMMUNICATION

Implications of quality adjusted survival for clinical trials in

radiation oncology
B JONES, MD, FRCR, MedFIPEM

Birmingham Cancer Centre, University Hospital Birmingham, Birmingham B15 2TH, UK

ABSTRACT. Clinical trials in radiotherapy sometimes compare changes in radiation dose

distribution using different radiation techniques. The use of quality adjusted survival can, Received 23 May 2005
in special circumstances, reduce the requirement of large patient numbers in order to show Revised 27 July 2005
a significant difference in overall outcome. The provisos are that marginal improvements Accepted 7 October 2005
in survival or tumour control endpoints and a reduction in toxicity scores are present. The
DOI: 10.1259/bjr/15389891
converse findings would also be amenable to this approach. Random sampling methods
are used to construct a patient population where the first set of conditions is met. Further ’ 2006 The British Institute of
work is necessary to refine the absolute indications for this technique. Radiology

The results of modern clinical trials in radiotherapy
frequently show marginal changes in survival, which
necessitate very large numbers of patients to demon-
strate even a trend of improvement in outcomes. For
example, the four-armed trial of accelerated and hyper-
fractionated radiotherapy conducted in the USA by Fu
and colleagues [1] involved 1113 entered patients, with
statistical significance probability values ranging from
0.045 to 0.067 for local tumour control and disease free
survival. Separate analysis of survival and side effect
statistics is the norm, often with much less detail given in
terms of reporting the latter. The costs of clinical trials –
in terms of human and financial resources provided by
the government and charitable organizations, as well as
the restrictions imposed upon patients – are considerable
in advanced countries such as the UK. Better ‘‘value for
money’’ would be achieved if trial costs could be
curtailed – this would at least allow more trials to be
conducted. The concept of quality adjusted survival is
not new [2] and could be more frequently applied in
clinical trial analysis, although it is most frequently used
in cost benefit studies by health economists.
Clinical trials in oncology are usually concerned with
the duration of local tumour control, survival times (both
continuous variables) and the severity of side effects
(normally graded as discrete variables). In the testing of
new radiation therapy techniques, for example if dose
distribution is changed while delivering the same total
tumour dose, it might be that the same or similar local
control and survival can be expected but that the quality
of life may be the intended benefit [3]. Another example
would be the delivery of a slightly higher radiation dose
to the tumour while at the same time reducing the dose
to normal tissues by means of charged particle beams [4],
with a modest increase in tumour control and a
reduction in side effects.
There are potential ethical difficulties about the
conduct of such trials where the dose distributions
(obtained using the predictions of the laws of physics)
are judged to be so much better in the case of the new
treatment [5]. However, past experience with ‘‘new’’
forms of radiotherapy required trials to demonstrate that
theoretically predicted gains were not realised in
practical situations [6, 7], although in these studies there
was no change in dose distribution. There is also a
dilemma when new treatments are phased in slowly
owing to reduced treatment capacity: in such cases there
might be opportunities to do randomized studies.
It would be advantageous if subtle changes in tumour
control, survival and simultaneous improvements in
quality of life could be found from trials that incorporate
smaller numbers of patients; results should be available
sooner and the financial costs of trials would be
considerably reduced.
The present paper aims to demonstrate the potential
advantage of quality-adjusted survival in a simulated
clinical trial where the necessary conditions of marginal
improvements of disease free survival and reduced
complications occur.
Methods
The disease free survival times of a cancer patient
population are simulated by random sampling techni-
ques for the two arms of a trial that compares
conventional X-ray therapy (XRT) with charged particle
beam (CPB) therapy given to slightly higher dose. The
assumptions made are given in Table 1. This example is
not meant to advance the cause of any particular form of
therapy, but merely used to demonstrate the statistical
principles. Mathematica and GraphPad Prism software
are used to obtain the results.
The quality-adjusted survival is calculated by multi-
plying the actual survival by a factor F defined as:
5{grade
F~
5

The British Journal of Radiology, April 2006 353


Table 1. Assumptions made in modelled population
XRT CPB
Mean survival in years with standard 4 (2.7) 4.5 (3)
deviation in parentheses for a
log-normal distribution
Number of ‘‘cured’’ 10 year survivors 23 30
not included in above distribution
Grade 0 toxicity % 40 60
Grade 1 toxicity % 40 30
Grade 2 toxicity % 15 7 The relatively simple exercise presented in this paper
Grade 3 toxicity % 4 2
Grade 4 toxicity % 1 1
XRT, conventional X-ray therapy; CPB, charged particle beam
therapy.
In this way, toxicity scores of 0, 1, 2, 3 and 4,
respectively, have F51, 0.8, 0.6, 0.4, 0.2.
The assumptions made are given in Table 1, including
the cumulative toxicity scores.
For the results shown in Figure 1, the survival log rank
test provides p50.062 for a one tailed study. The side
effect profile is separately analysed to give x2 (with four
degrees of freedom)59.04, p50.061. When the quality-
adjusted survival is assessed, then p,0.001, which is a
substantial reduction. Essentially this is the same
approximate result as would be obtained by multiplying
the two probabilities, i.e. the probability of the null
hypothesis being accepted for the uncorrected survival
and for the side effect profile being the same in both arms
of the trial (p50.061 6 0.001¡le;0.001).
Rather than use the classical x2 test with n–1 degrees of
freedom, where n is the number of categorical variables,
the x2 test for trend can be used with only one degree of
freedom, which inevitably reduces the p-value. By use of
this test, the results given in Table 1 yield a p50.003,
which is significant without combining with the disease
free survival results. However, the necessary condition
for the use of this test is that the outcome variable must
be well ordered; this condition will not be satisfied in
trials where there might be less grade 0 than grade 1
toxicity.
Figure 1. Kaplan-Meier plot of a simulated clinical trial
comparing conventional X-ray therapy (XRT) with charged
particle beam (CPB) therapy. There are 100 patients in each
treatment arm and the assumptions made are given in
Table 1.
354
B Jones
Discussion
At present there is some dissatisfaction with the
constraints imposed by standard clinical trials in radio-
therapy [8]. A wider spectrum of research methods, such
as observational studies and mathematical modelling, is
required to solve at least some of the problems
confronted in a complicated discipline such as radiation
oncology, where trials may not be appropriate, or
considered unethical [9].
demonstrates proof of principle that fusion of survival
with quality of life end points may very substantially
reduce the numbers of patients required to show that a
new treatment confers significant overall benefits. The
number of patients required to confirm marginal benefits
in radiation oncology is discussed further by Bentzen
[10]. To use the quality adjusted approach, certain
safeguards would be required: e.g. that the ‘‘survival’’
trend is superior, with a suggested p-value of ,0.1; also
that the side effect profile is directionally correct, i.e. that
there would be reduced numbers of toxicity scores in
each toxicity category .0 for the test treatment with
statistical confirmation of a trend with p,0.1 Further
sensitivity analysis along these lines should be per-
formed by expert statisticians, with more complex
calculations of the expected numbers required to achieve
the required level of statistical confidence using the
adjusted survival [10, 11]; recommendations of accepta-
ble p-value ranges from the primary end point analyses
might also be helpful. Quality adjusted survival must be
regarded as either a primary or a secondary endpoint, to
be used in special circumstances: this technique might
even have a place where survival is unchanged, since
further endpoints such as local tumour control or disease
free survival may show significant differences when
adjusted for quality of life.
The quality survival concept has been applied else-
where in oncology and some statisticians have already
advocated this approach [11]. For example, in leukaemia
[12], where four clinical states (viz. toxicity, treatment
free of toxicity, no treatment nor symptoms, relapse)
were defined. The average times spent in each state were
weighted by utility coefficients that reflect relative value
according to quality of life. This represents a much
more complex process, which also acknowledges the
reversible nature of some of the treatment-related side
effects.
However, in radiotherapy the serious side effects tend
to be chronic and cumulative with time; consequently,
such a sophisticated approach would seem to be
unnecessary. The use of extant grading systems in
oncology can be used in toxicity grade allocation [13]
and in crude quality adjusted survival. Actuarial assess-
ment of each side effect grade could be used, but again
the numbers of patients and events may be so small that
no significant differences would be found, as would be
the case for grades 3–4 in Table 1.
The present emphasis on obtaining improved survival
in clinical trials ignores the quest for testing improved
quality of life. The main funding agencies do not
normally support or encourage: primary quality of life
endpoint studies in which there is no expectation of
enhanced tumour control; observational studies of more
The British Journal of Radiology, April 2006
Short communication: Quality adjusted survival
complex or newer forms of therapy; large multicentre
non-randomized studies that allow convergence of
technique and good quality assurance as a prerequisite
for multicentre trials. Testing of new radiotherapy
techniques – such as charged particle beams – that are
specifically designed to reduce side effects, yet either
maintain or increase tumour control, might need to
utilize quality based end points in trial analysis given
that the conventional clinical trial end points may not be
sufficiently sensitive where subtle changes in multiple
outcomes occur.
It is hoped that this paper will stimulate discussion
and lead to better and more cost effective studies within
the UK National Cancer Research Institute as well as in
other countries.
References
1. Fu KK, Pajak TF, Trotti A, Jones CU, Spencer SA, Phillips
TL, et al. A Radiation Therapy Oncology Group (RTOG)
phase III randomized study to compare hyperfractionation
and two variants of accelerated fractionation to standard
fractionation radiotherapy for head and neck squamous cell
carcinomas: first report of RTOG 9003. Int J Radiat Oncol
Biol Phys 2000;48:7–16.
2. Nord E. Cost–value analysis in Health care: making sense
out of QALY’s. Cambridge University Press, 1999.
3. Dearnaley DP, Khoo VS, Norman AR, Meyer L, Nahum A,
Tait D, et al. Comparison of radiation side-effects of
conformal and conventional radiotherapy in prostate
cancer: a randomised trial. Lancet 1999;353:267–72.
The British Journal of Radiology, April 2006
4. Jones B, Rosenberg I. Particle therapy Cooperative
Oncology Group (PTCOG40), Institute Curie 2004. Br J
Radiol 2005;78:99–102.
5. Suit H, Goldberg S, Niemerko A, Trofimov A, Adams J,
et al. Proton beams to replace photon beams in radical dose
treatments. Acta Oncologica 2003;42:800–8.
6. Errington RD, Ashby D, Gore SM, et al. High energy
neutron treatment for pelvic cancers: study stopped
because of increased mortality. Br Med J 1991;302:1045–51.
7. Maor MH, Errington RD, Caplan RJ, et al. Fast neutron
therapy in advanced head & neck cancer: a collaborative
internal randomised trial. Int J Radiat Oncol Biol Phys
1995;32:599–604.
8. Munro AJ. The conventional wisdom and the activities of
the middle range. Br J Radiol 2005;78:381–3.
9. Jones B, Dale RG, Carabe A. Conventional wisdom and
activities of the middle range. Br J Radiol 2005;78:1119.
10. Bentzen SM. Towards evidence based radiation oncology:
improving the design, analysis, and reporting of clinical
outcome studies in radiotherapy. Radiother Oncol
1998;46:5–18.
11. Billingham LJ, Abrams KR. Simultaneous analysis of
quality of life and survival data. Stat Methods Med Res
2002;11:25–48.
12. Levy V, Porcher R, Delabarre F, Leporrier M, Cazin B,
Chevret S. French Cooperative CLL Group. Evaluating
treatment strategies in chronic lymphocytic leukemia: use
of quality-adjusted survival analysis. J Clin Epidemiol
2001;54:747–54.
13. Davidson SE, Burns MP, Routledge JA, Swindell R, Bentzen
SM, West CM. Assessment of morbidity in carcinoma of the
cervix: a comparison of the LENT SOMA scales and the
Franco-Italian glossary. Radiother Oncol 2003;69:195–200.
355

Correspondence
(The Editors do not hold themselves respondible for opinions expressed by correspondents)
Response to ‘‘Radiation dose measurement
and optimization’’
The Editor—Sir,
In his Letter to the Editor on ‘‘Radiation dose measure-
ment and optimization’’ in the September 2005 issue of
this journal [1], Dr Moores raised a number of rather
paradoxical objections to the use of dose–area product
(DAP) measurements for monitoring radiation doses to
patients in diagnostic radiology. His arguments that
DAP meters are unsuitable because:
(i) they do not measure absorbed dose to a specified
medium,
(ii) they combine two quantities (entrance surface dose
and entrance field area) into one, where the
contributions from each are indistinguishable,
(iii) male patients have much higher DAP values on
average than female patients,
are based on a fundamental misconception of the role of
diagnostic reference levels (DRLs) in the optimization of
medical exposures.
DRLs are essentially investigation levels that, if
exceeded, trigger a detailed investigation into the causes
of unusually high exposures. They are just the first step
in the optimization process – a means for identifying
those situations in most urgent need of investigation
because the doses appear to be towards the top end of
the distribution seen on a national scale. The choice of
DAP as one of the quantities in which DRLs are
expressed is based mostly on the practicability of making
DAP measurements during radiographic and fluoro-
scopic examinations in the clinical environment and the
widespread installation of DAP meters on X-ray imaging
equipment throughout the health service.
There are three essential requirements for the dose
quantities to be used in DRLs; they should be unam-
biguously defined so that everyone clearly understands
what is to be measured, they should be capable of
simple, direct measurement (or calculation from indirect
measurements) with readily available dosemeters of
sufficient precision and accuracy, and they should
provide an indication of the typical dose received by
patients examined in a particular facility from a
particular type of X-ray examination. Clear guidance
on how DAP meters can be used to meet all these
requirements is provided in the National Protocol for
Patient Dose Measurements in Diagnostic Radiology [2] and
in Guidance on the Establishment and Use of DRLs for
Medical X-ray Examinations [3] published by IPEM. Other
practical dose quantities will be needed for expressing
DRLs in different imaging modalities such as CT,
mammography and dental radiography. These have
been clearly defined in many publications, including
IPEM report 88 [3]. It is, therefore, rather puzzling to see
Dr Moores describing this wide variety of dose quantities
356
The British Journal of Radiology, 79 (2006), 356–358
as an intolerable situation, most likely due to ‘‘the lack of
a clear understanding of the meaning and relevance of
the term ‘‘optimization’’ in the field of diagnostic
radiology’’.
Dr Moores is correct in pointing out that the full
optimization process (i.e. maximizing the ratio of benefit
to harm) in diagnostic radiology, requires assessment of
both the diagnostic efficacy of the examination and the
resulting radiation risk. However, this is the purpose of
the detailed investigation that is triggered by exceeding
the DRL; it was never intended that the dose measure-
ments made to determine whether the DRL is being
exceeded should, by themselves, provide all the informa-
tion necessary for complete optimization.
One of the first investigations to carry out on
exceeding a DRL might well be to see if the sample of
patients on whom DAP measurements were made was
biased towards excessively large patients, since the DRLs
are set for average-sized patients. The fact that males
tend to be generally larger than females and conse-
quently will on average have higher DAP values, is no
reason, per se, for not using DAP measurements to trigger
such investigations. A further investigation might be
necessary to determine whether the excessive DAP
values were due to the use of large entrance surface
doses or large field areas (or both).
Optimization in diagnostic radiology does not mean
simply maximizing image quality and minimizing
patient dose, rather it requires radiologists to determine
the level of image quality that is necessary to make the
clinical diagnosis and then for the dose to be minimized
without compromising this image quality. Radiologists,
of course, subjectively assess the adequacy of their
images every time they report on them, and it can be
argued that no one else is in a better position to do so.
What radiologists cannot do intuitively is to assess the
patient dose, which is why practical patient dose
monitoring techniques and DRLs are needed to trigger
more detailed optimization investigations, if images
that are locally considered to be ‘‘diagnostically accep-
table’’ are being obtained with excessively high patient
doses.
Dr Moores’ plea for the development of ‘‘a scientific
framework for the quantification of diagnostic outcome
and therefore clinical benefit’’ will, no doubt, have an
important role in these detailed optimization studies. But
it is difficult to see how such a framework would
eliminate the ‘‘general ignorance within the medical
profession concerning the levels of radiation employed
in diagnostic radiology’’, as he claims. Rather, the
establishment and implementation of DRLs using prac-
tical dose quantities such as DAP, has probably done
more to raise awareness of patient doses in the radiology
professions than any other national initiative. The
halving of the DRLs that has been achieved for many
common types of X-ray examination in the UK over the
past 20 years [4] with no apparent loss of diagnostic
The British Journal of Radiology, April 2006
Correspondence
efficacy, provides ample evidence for the effectiveness of
this simple first step in the optimization process.
Yours etc.,
B F WALL
Radiation Protection Division
Health Protection Agency
Chilton
Didcot
Oxon OX11 0RQ
UK
(Received 18 October 2005 and accepted 25 October 2005)
DOI: 10.1259/bjr/34749787
References
1. Moores BM. Radiation dose measurement and optimization.
Br J Radiol 2005;78:866–8.
2. Dosimetry Working Party of IPSM. National Protocol for
Patient Dose Measurements in Diagnostic Radiology.
Chilton, NRPB, 1992.
3. IPEM Report 88. Guidance on the Establishment and Use of
DRLs for Medical X-ray Examinations, York, IPEM, 2004.
4. Hart D, Hillier MC, and Wall BF. Doses to Patients from
Medical X-ray Examinations in the UK – 2000 Review.
NRPB-W14, www.hpa.org.uk/radiation, 2002 [Accessed 25
November 2005].
Authors’ reply
The Editor—Sir,
I wish to thank B Wall for his comments on my letter and I
would like to respond, first of all, to paragraphs 4 and 5 of
his letter. Here, he appears to recommend an unproven
optimization strategy for female patients, which is based
upon the use of dose reference levels (DRLs) that are
dependent upon the exposure of male patients. I believe
that such an approach severely disadvantages female
patients. It either ignores or neglects the well-known and
significant differences in both entrance surface dose (ESD)
and dose–area product (DAP) values for female patients,
compared with those of males, as well as their own distinct
optimization requirements.
A comparison of male and female ESD and DAP
values measured during a national survey of doses to
patients in the UK [1] and the corresponding DRLs
presented in the National Protocol for Patient Dose
Measurements [2] indicates that National and European
adult patient DRLs are governed by the levels of
exposure to male patients.
The reason for this lies, to a large extent, in the
recommendation that the mean weight of an adult patient
sample, in any ESD/DAP audit, should lie within 5 kg of
70 kg [2]. In the national patient dose survey [1], the
average weight of female patients was roughly 60 kg
whilst that of the male patients was 70 kg. Similar weight
differences for female and male patients have also been
noted in studies involving fluoroscopy [3] and radio-
graphy [4]. Also, the weights of female and male adult
voxel phantoms, developed for computational radiation
protection studies, have been designed around similar
The British Journal of Radiology, April 2006
weight differences [5–7]. The decision to employ a single-
weight dependent DRL value has never been scientifically
justified as an effective basis for optimization of practices
in respect of female patients.
As I indicated in my letter, results of the 1985 national
survey [1] demonstrate that the mean ESD for female
patients undergoing a PA chest X-ray examination was
30% lower than that of the male patients. Such
differences in ESD were also noted, for example, for
the AP abdomen (27%), lat thoracic spine (33%),
lat lumbar spine (25%) examinations. Other studies have
also shown similar differences in ESD [4], for PA chest
(70%), knee (100%) and AP skull (70%), examinations.
The mean DAP value for female patients was
significantly lower than that of male patients [1] for the
PA chest (86%), thoracic spine (22%), barium meal (39%),
IVU (43%), cholangiography and cholecystography
(43%) examinations. Other studies have shown similar
differences in DAP values for female patients [3]; barium
meal (36%), barium swallow (45%) etc. In the case of the
DAP values, such differences can arise from unknown
variations in the ESD and field sizes employed.
Under present circumstances, local DRLs derived for
patients in a women’s hospital would not be relevant to
females in a mixed-sex patient group of 70 kg average
weight. Different audit standards, therefore, can exist for
different female populations. Similarly, female DAP
values for a PA X-ray examination of the chest may
exceed their own population norm by up to 86% [1], due to
unknown variations in either ESD or field size, before the
need for an investigation is indicated. Such an indication
may be dependent upon the patient mix, including
weights, in the audit sample, as well as specifically
radiation protection considerations. Why should easily
demonstrable, optimized practices for female patients be
hidden under patient sampling/weight considerations
involving male patients?
Besides exposure differences for female patients, other
factors merit consideration in respect of optimization
strategies for female patients:
N For many examinations, the female breast may, or
could potentially, be irradiated, depending upon field
size and projection, as well as technique.
N A large percentage of female patients have the
potential to be pregnant, which may then involve
the exposure of radiosensitive third parties.
N Computed effective doses per unit of entrance surface
air kerma (ESAK) are significantly different, based
upon voxel phantom studies [7].
N Optimization of tube voltage and dose for digital
chest radiography are different [8].
N The relative frequencies of different types of examina-
tions are different for female patients [9].
N The distribution of examinations contributing to the
genetically significant dose (GSD), are different [9].
N The risks of fatal cancer attributable to fluoroscopy are
different [10].
N Organ doses during CT examinations are different
[11].Parent not defined
Given such extensive differences, why are female
patients not considered to be a separate exposed group
357

with its own DRL values, so that their own unique safety
and imaging needs can demonstrably be given full and
direct consideration in respect of radiation safety?
The main reason for the widespread installation of
DAP meters on X-ray imaging equipment throughout the
NHS [paragraph 2 of B Wall’s comments] is due to the
fact that funds were made available from central
government (presumably on the advice of the NRPB),
in the early 1990s, for the purchase and distribution of
200 DAP meters. Prior to this, extensive medical physics
effort had been expended on developing the use of
calibrated X-ray source data for patient dose assess-
ments. This effort included the background science [12]
and calibration methods [13]. Such an approach harmo-
nises patient dose assessments in both radiotherapy and
diagnostic radiology and is based upon over 50 years of
scientific effort. Unfortunately this was ignored in the
National Protocol [2].
The reduction in patient doses in diagnostic radiology
over the past 20 years [paragraph 7 of B Wall’s comments]
is due to the introduction of rare earth intensifying screens
in radiography and sodium iodide phosphors in fluoro-
scopic units. These initiatives have been supported by the
introduction of low attenuation materials in table-tops and
Bucky stands. The UK national dose audits have faithfully
recorded their well-documented effects. Our manufacturer
colleagues are to be commended upon their contribution
to patient dose reduction.
Yours etc.,
B M MOORES
Co-ordinator
Radiological Unification Strategies [RADIUS] Group
EC Radiation Protection Research Programme
(Received 18 November 2005 and in final form 4 January 2006, accepted
11 January 2006)
DOI: 10.1259/bjr/71189157
References
1. Shrimpton PC, Wall BF, Jones DG, Fisher ES, Hillier MC,
Kendall GM. A national survey of doses to patients
358
Correspondence
undergoing a selection of routine x-ray examinations in
English hospitals. NRPB-R200. Chilton, UK: National
Radiological Protection Board, 1986.
2. National protocol for patient dose measurements in
diagnostic radiology. Chilton, Didcot: National
Radiological Protection Board, 1992.
3. Rowley KA, Hill SJ, Watkins R, Moores BM. An investiga-
tion into the levels of radiation exposure in diagnostic
examinations involving fluoroscopy. Br J Radiol 1987;60:
167–73.
4. Gallini RE, Belletti S, Berna V, Giugni U. Adult and child
doses in standardised x-ray examinations. Radiat Prot
Dosim 1992;43:41–7.
5. Zankl M, Fill U, Petoussi-Henss N, Regulla D. Organ dose
conversion coefficients for external photon irradiation of
male and female voxel models. Phys Med Biol
2002;47:2367–85.
6. Kramer R, Vieira JW, Khoury HJ, Lima FRA, Fuelle D. All
about MAX: a male adult voxel phantom for Monte Carlo
calculations in radiation protection dosimetry. Phys Med
Biol 2003;48:1239–62.
7. Kramer R, Khoury HJ, Vieira JW, Loureiro ECM, Lima VJM,
Lima FRA, et al. All about FAX: a female adult voxel
phantom for Monte Carlo calculation in radiation protec-
tion dosimetry. Phys Med Biol 2004;49:5203–16.
8. Pascoal A, Patel R, Lawinski CP, Tabakov S. Optimization
of tube voltage and dose for digital chest radiography – a
study addressing patient size. In: Proceedings of UK
Radiological Congress, 6–8 June, 2005. London, UK:
British Institute of Radiology, 2005:2.
9. Wall BF, Rae S, Darby SC, Kendall GM. The NRPB survey:
methods and results. Dosimetry in diagnostic radiology,
CRS 40. Chapter 6, 44–55. The Hospital Physicists
Association 1984.
10. Goodenough DJ. Lessons learned in radiology. Proceedings
of an international conference held in Malaga, Spain, 26–30
March 2001, 145–155. International Atomic Energy Agency,
2001.
11. Faulkner K, Moores BM. Radiation dose and somatic risk
from computed tomography. Acta Radiologica 1987;28:
483–8.
12. Harrison RM. Dose measurement in vivo and in vitro. CRS
40. Chapter 2, 8–17. The Hospital Physicist’s Association.
1984.
13. TGR 32 Measurement of the performance characteristics of
diagnostic X-ray systems used in medicine. Part 1: X-ray
tubes and generators. The Hospital Physicist’s Association.
1984.
The British Journal of Radiology, April 2006
The British Journal of Radiology, 79 (2006), 359–361

CASE OF THE MONTH

The angel in the marble

1 2
D J A BUTTERISS, BSc, MBBS (Hons), FRCR, M CLARKE, FRCS, FRCOphth and 1D BIRCHALL, MB, BChir, FRCR

1
Department of Neuroradiology, Regional Neurosciences Centre, Newcastle General Hospital,
Newcastle upon Tyne and 2Ophthalmology Department, Royal Victoria Infirmary, Newcastle upon
Tyne, UK

Received 16 May 2005

Accepted 27 June 2005

DOI: 10.1259/bjr/33391462

’ 2006 The British Institute of

Radiology

A 56-year-old woman presented to the ophthalmology
department complaining of subacute painless decrease in
visual acuity in her right eye. She had a previous history
of longstanding poor visual acuity in the left eye, which
Address correspondence to: Dr D Birchall, Department of
Neuroradiology, Regional Neurosciences Centre, Newcastle General
Hospital, Westgate Road, Newcastle upon Tyne NE4 6BE, UK.
had not been formally investigated. No other significant
history was elicited. Fundoscopy revealed atrophy of the
left optic disc and swelling of the right disc. No other
significant abnormality was found.
CT (Figure 1) and MRI (Figure 2) were performed on
this patient. What are the abnormalities? What is the
diagnosis in this case? What are the potential causes of
optic atrophy in this condition?

(a) (b) (c)

Figure 1. (a) CT shows buckling of the right optic nerve and diffuse calvarial hyperostosis. (b) There is generalized hyperostosis
of the skull base. (c) There is bilateral narrowing of the optic canals and superior orbital fissures.

The British Journal of Radiology, April 2006 359

 D J A Butteriss, M Clarke and D Birchal

(a) (b)

(c) (d)

Figure 2. (a) Axial T2 image shows dilatation of the optic nerve sheath on the right, and optic nerve buckling. Diffuse
hyperostosis is again evident. (b) Coronal short tau inversion recovery (STIR) imaging through the orbits confirms optic nerve
sheath dilatation on the right, with increased signal within the nerve itself. The left optic nerve is atrophic. (c) More posterior
the coronal STIR image shows compression of the optic nerves within the optic canals (arrows). (d) Sagittal T2 image
demonstrates skull base and upper cervical hyperostosis, with distortion of the cervico-medullary junction and of the pituitary
fossa.

Imaging findings stenosed bilaterally, particularly on the left. The right

optic nerve is buckled. The intracranial contents are
CT demonstrates marked thickening and sclerosis of otherwise normal. In the absence of other causes of
the calvarium and skull base. The optic canals are diffuse hyperostosis, these appearances are consistent

360 The British Journal of Radiology, April 2006

Case of the month: The angel in the marble
with osteopetrosis. MRI shows left sided optic atrophy.
On the right, there is dilatation of the optic nerve sheath
and increased T2 signal within the optic nerve, appear-
ances consistent with decompensating subacute optic
nerve compression at the orbital apex. Sagittal imaging
shows clival and skull base hyperostosis with distortion
of the medulla and proximal cervical spinal cord, and
compression of the pituitary gland.
Discussion
Osteopetrosis, also known as Albers-Schonberg dis-
ease or marble bone disease, is an uncommon genetic
disorder characterized by increased bone thickness and
density, and is caused by impaired osteoclast maturation
or function resulting in decreased bone resorbtion and
modelling. Several forms have been identified, the most
common being the infantile autosomal recessive type
and the delayed autosomal dominant type. The recessive
form is also known as malignant osteopetrosis and
presents in early childhood with fractures and the
sequelae of pancytopaenia due to marrow space oblit-
eration. Neurological complications are common, and
are secondary to raised intracranial pressure (ICP) due to
calvarial expansion, and narrowing of the cranial
foramina resulting in cranial nerve, spinal cord and
vascular compression. The delayed form usually pre-
sents in early adulthood, although it demonstrates
increasingly early onset in affected familial series. It
may be asymptomatic and diagnosed incidentally on
radiological investigations, but usually presents with
easy fracturing or with neurological complications such
as raised ICP or cranial nerve palsies. It is divided into
two forms based on radiological and clinical criteria:
type 1 demonstrates diffuse calvarial thickening and
sclerosis with diffuse sclerosis of the axial skeleton; type
2 has less marked calvarial thickening, but the skull base
is more severely affected than in type 1 and there is
vertebral endplate sclerosis and bone-within-a-bone
appearance [2, 3].
The British Journal of Radiology, April 2006
Overall, the most common neurological complication
of osteopetrosis is visual loss and optic atrophy. There
are three possible causes of visual loss [4]:
(1) Optic nerve compression secondary to stenosis of
the optic canals.
(2) Papilloedema due to raised ICP.
(3) Primary retinal degeneration.
In this report, the likely cause of visual loss and optic
atrophy is stenosis of the optic canals. The late presenta-
tion is an unusual feature of the case.
Secondary optic atrophy is treatable, management
dependent on the cause. Optic canal stenosis is treated
by orbital deroofing, and papilloedema treated by
reducing ICP, either by ventricular shunting, or by
craniectomy alone or in conjunction with cranial vault
expansion [5].
Acknowledgment
Michelangelo, ‘‘I saw the angel in the marble and
carved until I set it free.’’
References
1. Hall CM. International Nosology and Classification of
Constitutional Disorders of Bone (2001). Am J Med Genet
2002;113:65–77.
2. Anderson PE, Bollerslev J. Heterogeneity of autosomal
dominant osteopetrosis. Radiology 1987;164:233–5.
3. Cure JK, Key LL, Goltra DD, VanTassel P. Cranial MR
imaging of osteopetrosis. AJNR Am J Neuroradiol 2000;21:
1110–5.
4. Stewart CG. Neurological aspects of osteopetrosis.
Neuropathol Appl Neurobiol 2003;29:87–97.
5. Vanier V, Miller NR, Carson BS. Bilateral visual improve-
ment after unilateral optic canal decompression and cranial
vault expansion in a patient with osteopetrosis, narrowed
optic canals, and increased intracranial pressure. J Neurol
Neurosurg Psychiatry 2000;69:405–6.
361
BJR
The British Journal
of Radiology

May 2006
Volume 79
Issue 941
May 2006, Volume 79, Issue 941
● PET poised to alter the current paradigm for response assessment
of non-Hodgkin’s lymphoma

● Acute adverse reactions to magnetic resonance contrast

media – gadolinium chelates

● Observer variation in the evaluation of lumbar herniated discs

and root compression: spiral CT compared with MRI

● Initial experiences of using an automated volumetric measure of

breast density: the standard mammogram form

● Occupational radiation doses in interventional cardiology: a

15-year follow-up

● A four-dimensional computer simulation model of the in vivo

response to radiotherapy of glioblastoma multiforme: studies on
the effect of clonogenic cell density

● A quantitative study of IMRT delivery effects in commercial

planning systems for the case of oesophagus and prostate
tumours

● Effect of ticlopidine in the prevention of radiation enteropathy

● Radioprotective effects of hesperidin against gamma irradiation

in mouse bone marrow cells

● Characteristics and predictive factors of early-onset diarrhoea

during pelvic irradiation

● Optical density variations in CT films and their effect on image

quality

● Assessment of environmental disturbances to the static magnetic

field in magnetic resonance installations

● An investigation of search pattern extent in the threshold

contrast detection task

● Imaging pulmonary embolism in pregnancy: what is the most

appropriate imaging protocol?
 ● Primary subcutaneous sacrococcygeal ependymoma: a case
report and review of the literature

● Correspondence

● A catastrophic complication
The British Journal of Radiology, 79 (2006), 365–367
COMMENTARY
PET poised to alter the current paradigm for response assessment
of non-Hodgkin’s lymphoma
1
A GUERMAZI, MD and 2M E JUWEID, MD
1
Synarc Inc. and University of California, San Francisco, CA and 2University of Iowa Hospitals and
Clinics, Iowa City, IA, USA
Non-Hodgkin’s lymphoma (NHL) is a group of
heterogeneous tumours originating in lymphoid tissue.
The outcome of patients with lymphoma is, in general,
better than non-lymphoid malignancies. Indeed, aggres-
sive NHLs, which make about two-thirds of the NHL
subtypes, are potentially curable (about 60% of patients
are cured) with conventional and high-dose therapy.
More than 70% of patients with newly diagnosed
lymphoma respond well to chemotherapy alone or to
chemoimmunotherapy with or without radiotherapy. In
addition to the availability of new treatments, accurate
staging and risk stratification with appropriate selection
of treatments, as well as improved assessment of
response to therapy, have resulted in a higher success
rate in the treatment of lymphoma. The increasing
number of treatment options available for patients with
lymphoma currently being tested in clinical trials clearly
requires the availability of standardized guidelines for
assessment of response to therapy, which should ensure
comparability of success rates among these trials.
The International Workshop Criteria (IWC) for
response assessment of NHL proved to be quite useful
for standardized assessment of response to various
lymphoma treatments tested in a significant number of
clinical trials [1]. These 1999 criteria rely on clinical,
biochemical, histopathological and imaging studies,
especially CT. This first attempt at standardization of
response criteria in NHL proved to be a significant step
forward; yet with the advent of functional imaging with
positron emission tomography (PET), particularly using
fluorodeoxyglucose (FDG), the limitations of these
criteria have become increasingly apparent.
Six years after their publication in the Journal of Clinical
Oncology, it was clear that the IWC were in need of
Address correspondence to: Ali Guermazi. MD, Synarc Inc., 575
Market Street, 17th Floor, San Francisco, CA 94105, USA.
The British Journal of Radiology, May 2006
Received 14 September
2005
Revised 16 December 2005
Accepted 16 January 2006
DOI: 10.1259/bjr/31368041
’ 2006 The British Institute of
Radiology
revision. Dr Bruce Cheson from Georgetown University
in Washington DC, who organized the first International
Workshop to Standardize Response Criteria for NHL in
1998, organized a meeting together with Dr Volker Diehl
from the University of Cologne, Germany, called the
International Harmonization Project (IHP), at which an
invited international group of participants with clinical,
histopathological and imaging expertise in NHL delib-
erated on the need for harmonization of clinical trial
parameters and revision of the currently utilized
response criteria in lymphoma. The meeting took place
in June 2005 at the 9th International Conference on
Malignant Lymphoma in Lugano, Switzerland. Six
subcommittees were formed — response criteria, clinical
features, endpoints, pathology/immunohistology, pae-
diatrics and imaging. Each subcommittee presented the
results of their deliberations.
The imaging subcommittee, including the authors of this
commentary among others, stressed the importance of the
introduction of new technologies, such as PET and PET/
CT. Indeed, one of the most vexing clinical dilemmas in
response assessment and management of NHL is the
detection of active disease in a residual mass by conven-
tional imaging (usually CT), following treatment despite
resolution of clinical symptoms and normalization of
laboratory tests (i.e. in patients with an otherwise clinical
complete response). Such masses, usually seen in the
mediastinum and abdomen/pelvis, may be seen in up to
40% of patients with aggressive NHL and an even higher
percentage of patients with Hodgkin’s disease. Yet the vast
majority of these masses represent necrosis and/or fibrosis
and only a small percentage represents or contains residual
active lymphoma. It is also noteworthy that the frequency
of occurrence of these residual masses correlates with the
size of tumour mass(es) at diagnosis, with a much more
common occurrence of residual masses in patients with
initially bulky compared with non-bulky lymphoma.
365

Obviously, characterization of these residual masses is
important: residual lymphoma portends a poor prognosis
and requires further salvage therapy, while an inactive
residual mass usually confers an excellent prognosis
without further treatment.
Residual masses can be assessed with biopsy/fine
needle aspiration or even surgery. However, these
approaches, while generally accurate, are quite invasive
and associated with some morbidity. Non-invasive and
yet accurate approaches for assessment of these masses
are clearly warranted.
Conventional morphologic imaging (e.g. CT and con-
ventional MRI) alone do not allow discrimination between
active tumour and fibrosis and/or necrosis in residual
masses. Most CT studies report poor specificity in the
characterization of residual masses (40–50%). Conventional
MRI has been shown to be slightly more reliable in this
setting, but its performance is still unsatisfactory. Gallium
scintigraphy (GS) has been shown to be substantially more
accurate than CT and MRI for characterization of these
masses, particularly in patients with aggressive lymphoma,
with less reliable results in low-grade or follicular
lymphoma that are substantially less gallium-avid.
Importantly, GS is not well standardized and is also not a
convenient imaging modality for both the patients and
referring physicians because of the frequent requirement
for imaging at multiple time points over several days to
more than a week. The limited availability of the gallium-67
isotope, because of limited production and distribution,
poses another significant problem. All of these issues in
addition to the only moderate accuracy (about 60–70%)
resulted in a relatively limited utilization of GS in both the
USA and Europe.
The advent of PET promises to solve many of the
problems associated with the use of GS, being much more
convenient to use with less than 2 h required for a PET
study to be completed. More importantly, PET clearly has
higher accuracy than GS and morphologic imaging for
assessment of most subtypes of lymphoma, including
follicular histology. PET alone or coupled with CT (PET/
CT) has demonstrated a high accuracy (i.e. . 80%) for
assessment of response to therapy in patients with
aggressive NHL, particularly with respect to characteriza-
tion of residual masses, and it is now generally recognized
that PET is the most accurate non-invasive method for
differentiating necrosis and/or scar from viable tumour in
residual masses [2–6]. Largely because of the recognized
high accuracy of PET in characterizing the nature of
residual masses following treatment of lymphoma, the
IHP workshop has unequivocally agreed that PET or PET/
CT should be recommended in assessment of response or
restaging of aggressive NHL and Hodgkin’s lymphoma
(HL) at the completion of therapy.
This decision is supported by several persuasive studies
by various groups of investigators who evaluated the
predictive value of PET following first-line therapy of
patients with aggressive NHL and/or HL compared with
CT or conventional staging methods [3–5]. More recently,
Juweid et al [6] went one step further and compared a
proposed response classification based on integration of
FDG-PET into the IWC (IWC+PET) with the IWC alone in
patients with aggressive NHL. This retrospective study
included 54 patients with aggressive NHL (87% diffuse
large cell) who underwent PET and CT 1–16 weeks
366
A Guermazi and M E Juweid
(median 4 weeks; 83% within , 3–11 weeks) after 4–8
cycles of chemotherapy (84% received 6–8 cycles) mostly
with cyclophosphamide, doxorubicin, vincristine and
prednisone (CHOP) with or without rituximab. These
patients were assessed for complete response (CR),
unconfirmed complete response (CRu), partial response
(PR), stable disease (SD) and progressive disease (PD) by
the IWC and by specifically defined IWC+PET-based
response designations, derived by integrating the PET
findings into the previously defined IWC designations.
Progression-free survival (PFS) was used as the primary
outcome measure to compare the IWC-assigned and
IWC+PET-assigned response designations. The investiga-
tors found that there was a strikingly higher proportion of
patients with CR by IWC+PET compared with IWC alone
(35/54 versus 17/54, respectively), with no patients
designated as CRu by IWC+PET compared with 7 with
IWC alone, because such patients were either designated as
CR if PET was negative (fibrosis/necrosis) or PR if PET was
positive (residual tumour). This latter finding is significant
since, in fact, several studies have shown a substantial rate
of CRu designations by the predominantly CT-based IWC
classification due to the findings of residual masses by CT.
Thus, using the IWC+PET classification, the vast majority
of these CRus would be re-classified as CR or, less
frequently, as PR. The most critical finding of the study
by Juweid et al, however, was that based on PFS as an
outcome measure, the IWC+PET-based response classifica-
tion was more accurate than IWC-based classification in
predicting the ‘‘expected’’ outcome of patients in the
various response categories. For example, whereas two-
thirds of the patients designated as PR by IWC+PET
progressed within a median of only 8.5 months, almost 60%
of patients with a PR designation by IWC were progres-
sion-free at a median of nearly 3 years post-therapy.
Interestingly, in the 35 patients with CR by IWC+PET, PFS
was identical to that in the only 17 patients with CR by IWC
(31.5+ months), although slightly more than half of the
IWC+PET CRs were actually thought to be in PR or even
only having SD. This indicates that the ‘‘quality’’ of a CR by
IWC+PET, in terms of conferring a much more favourable
prognosis compared with PR or SD, was very similar if not
equal to the quality of a CR by IWC. The potential
implications for the management of patients with aggres-
sive NHL are apparent: whereas patients with PR or SD by
IWC+PET should be evaluated for persistent disease and, if
confirmed, considered for additional therapy, patients with
CR by IWC+PET may be safely observed. In contrast, at
least the PR designation by IWC, assigned to slightly more
than one-third of the patients in this retrospective study,
provides incorrect characterization of the actual response to
treatment in an unacceptably high fraction (,60%) of
patients with such designation, casting serious doubts
about its usefulness.
Obviously, the study by Juweid et al [6] had its
limitations: the retrospective nature of the evaluation
performed; the relatively small number of patients eval-
uated; and the somewhat variable timing of PET following
treatment and variable number and type of chemotherapy
cycles administered. Yet this ‘‘proof-of-principle’’ study,
along with similar evidence of superiority of PET
compared with CT, were convincing enough that the IHP
adopted IWC+PET-based definitions of CR, PR, SD and PD
similar to those presented in the study by Juweid et al [6],
The British Journal of Radiology, May 2006
Commentary: Response assessment of NHL by PET
with a proposal to completely eliminate the CRu designa-
tion. These new definitions were recently presented by Dr
Bruce Cheson at the 2005 Annual meeting of the American
Society of Hematology in Atlanta, Georgia [7]. In addition
to the proposed elimination of the CRu designation, a
noteworthy change compared with the criteria proposed
by Juweid et al [6] is that the criteria will now take into
consideration that some subtypes of NHL, such as
extranodal marginal zone lymphoma (MZL) of mucosa-
associated lymphoid tissue (MALT) are frequently not
FDG-avid, hence assessment will still be based on conven-
tional imaging. The criteria will now also be applicable to
patients with typically FDG-avid indolent NHLs, such as
follicular NHL, and also to patients with HL. The details of
these revised criteria will be reported in an upcoming
manuscript, to be submitted for publication shortly. A
more imaging-focused report of the imaging committee of
the IHP will follow subsequently.
Despite the clearly favourable contribution of PET or
PET/CT on response assessment of NHL, it should be
noted that, like any other modality, PET or PET/CT are
not perfect, as shown by several studies including the
study of Juweid et al [6]. In the latter study, one-third of
the patients (4/12) with a PR designation by IWC+PET
based on positive PET findings suggesting residual
disease remained progression-free for more than 3 years.
False-positive PET findings at the site of residual masses
may occur due to thymic rebound in the vicinity of a
residual mass in the superior mediastinum, sometimes
mistaken as uptake in the mass, or post-therapy
inflammatory changes, the latter substantially more
frequent following radiation than after chemotherapy
or chemoimmunotherapy [5, 8]. False-positive findings
outside the site of residual masses also may be caused by
rebound thymic hyperplasia, or infectious or inflamma-
tory processes including sarcoidosis.
With PET/CT there is a significant improvement in the
diagnostic accuracy of PET, principally because of the
more accurate anatomic localization of the PET findings
provided by the almost-simultaneously acquired CT. The
generally precise registration of the PET and CT images
largely eliminates the problem of false-positive physio-
logic FDG uptake, such as that in muscle or brown fat,
which is highly variable between patients [8]. In fact, a
recent study on the staging and re-staging of 73
lymphoma patients showed that PET/CT was superior
to PET alone with reported accuracies of 93% and 83%,
respectively [9]. To further improve the diagnostic
accuracy of PET, it will also be critical to address the
issue of standardization of PET technique (e.g. timing
after last treatment, timing of scanning after the injection)
in future studies.
Finally, it is important to note that PET and PET/CT are
likely to be useful not only for response assessment after
completion of therapy, but also for initial staging and early
response assessment after only a few cycles (e.g. 1–4) of a 6–
8-cycle chemotherapy or chemoimmunotherapy regimen.
In fact, several studies have shown that an interim
assessment with PET after 2–4 cycles of chemotherapy is
at least as accurate an assessment of disease response in
terms of prediction of progression free survival as an end of
treatment assessment [8, 10–12].
The British Journal of Radiology, May 2006
In summary, PET or PET/CT is poised to alter the
response assessment paradigm in patients with lym-
phoma which is currently largely based on conventional
imaging and has, in fact, already resulted in a significant
and important modification of the currently widely used
and accepted IWC for response assessment of lym-
phoma. The full report of this revision is eagerly awaited.
References
1. Cheson BD, Horning SJ, Coiffier B, Shipp MA, Fisher RI,
Connors JM, et al. Report of an international workshop to
standardize response criteria for non-Hodgkin’s lympho-
mas. NCI Sponsored International Working Group. Jô Clin
Oncol 1999;17:1244–53.
2. Reske SN. PET and restaging of malignant lymphoma
including residual masses and relapse. Eurô Jô Nucl Med
Mol Imaging 2003;Suppl. 1:S89–96.
3. Jerusalem G, Beguin Y, Fassotte MF, et al. Whole-body
positron emission tomography using 18F-fluorodeoxyglu-
cose for posttreatment evaluation in Hodgkin’s disease and
non-Hodgkin’s lymphoma has higher diagnostic and
prognostic value than classical computed tomography scan
imaging. Blood 1999;94:429–43.
4. Spaepen K, Stroobants S, Dupont P, et al. Prognostic value
of positron emission tomography (PET) with fluorine 18
Fluorodeoxyglucose ([18F]FDG) after first-line chemother-
apy in non-Hodgkin’s lymphoma: is [18F]FDG-PET a valid
alternative to conventional diagnostic methods? Jô Clin
Oncol 2001;19:414–9.
5. Naumann R, Vaic A, Beuthien-Baumann B, et al. Prognostic
value of positron emission tomography in the evaluation of
post-treatment residual mass in patients with Hodgkin’s
disease and non-Hodgkin’s lymphoma. Brô Jô Haematol
2001;115:793–800.
6. Juweid ME, Wiseman GA, Vose JM, Ritchie JM, Menda Y,
Wooldridge JE, et al. Response assessment of aggressive
non-Hodgkin’s lymphoma by integrated international
workshop criteria and fluorine-18-fluorodeoxyglucose posi-
tron emission tomography. Jô Clin Oncol 2005;23:4652–61.
7. Cheson BD, Pfistner B, Juweid ME, Specht L, Rosen ST,
Gascoyne R, et al. Revised response criteria for malignant
lymphomas from the members of the International
Harmonization Project (IHP) of the Competence Network
Malignant Lymphoma. Blood 2005;106:108, (Abstract).
8. Juweid ME, Cheson BD. Role of PET in lymphoma. J Clin
Oncol 2005;23:4577–80.
9. Allen-Auerbach M, Quon A, Weber WA, Obrzut S,
Crawford T, Silverman DH, et al. Comparison between
2-deoxy-2-[18F]fluoro-D-glucose positron emission tomo-
graphy and positron emission tomography/computed
tomography hardware fusion for staging of patients with
lymphoma. Mol Imaging Biol 2004;6:411–6.
10. Spaepen K, Stroobants S, Dupont P, et al. Early staging
positron emission tomography (PET) with fluorine 18
fluorodeoxyglucose ([18F]FDG) predicts outcome in patients
with aggressive non-Hodgkin’s lymphoma. Blood
2001;98:726a, (Abstract).
11. Jerusalem G, Beguin Y, Fassotte MF, et al. Persistent tumor
18F-FDG uptake after a few cycles of polychemotherapy is
predictive of treatment failure in non-Hodgkin’s lym-
phoma. Haematologica 2000;85:613–8.
12. Fields PA, Mikhaeel G, Hutchings M, van der Walt J,
Nunan T, Schely SA. The prognostic value of interim
positron emission tomography scans combined with
immunohistochemical data in diffuse large-B-cell lym-
phoma. Hematologica 2005;90:1711–3.
367
 The British Journal of Radiology, 79 (2006), 368–371

Acute adverse reactions to magnetic resonance contrast

media – gadolinium chelates
1
A LI, MBChB, FRCR, 2C S WONG, MBChB, 2M K WONG, MBBS, FRCR, FHKCR,
2
C M LEE, MBBS, FRCR, FHKCR and
2
M C AU YEUNG, MBBS, FRCR, FHKCR

1
Department of Radiology and Organ Imaging, United Christian Hospital, 130 Hip Wo Street, Kwun
Tong, Hong Kong, SAR and 2Department of Radiology, Princess Margaret Hospital, Hong Kong, SAR

ABSTRACT. The objective of this study was to evaluate the clinical safety of intravenous
gadolinium-based contrast media used in patients who underwent MRI at a single
institution. Acute adverse reactions to intravenous gadolinium-based contrast media
used for MRI at the Princess Margaret Hospital, Hong Kong, SAR, from January 1999 to
November 2004 were recorded in an incidence log book. The medical records of
patients’ demographics were retrospectively reviewed and the nature, frequency and
severity of the adverse reactions were investigated and documented. The incidence of
acute adverse reactions to intravenous gadolinium-based contrast media was 0.48% (45
patients with 46 adverse reactions). The severity of these adverse reactions were 96%
mild, 2% moderate (one patient developed shortness of breath that required oxygen
supplementation and intravenous steroidal management) and 2% severe (one patient
developed an anaphylactoid reaction, but successfully recovered through timely
resuscitation). No patients were recorded as having contrast extravasation and none
died as a result of any adverse reaction. Among the 45 patients who developed adverse
reactions, three patients (6.7%) had prior adverse reactions to iodinated contrast
media, three (6.7%) had prior reactions to a different gadolinium-based contrast
agent, one (2%) had asthma and nine (20%) had a history of drug/food allergy. Overall, Received 9 May 2005
41% of the adverse reactions were not documented in the final MRI report or the Revised 24 August 2005
clinical medical records. Gadolinium-based contrast media are safe and well tolerated Accepted 17 October 2005
by the vast majority of patients. In our study, the adverse reaction rate (0.48%) and the
DOI: 10.1259/bjr/88469693
incidence of severe anaphylactoid reaction (0.01%) concur with those reported in the
literature. Although most of the symptoms are mild and transient, these adverse ’ 2006 The British Institute of
reactions must be accurately documented and managed. Radiology

MRI is often considered a safer alternative to CT for
patients who cannot tolerate iodinated contrast media.
The intravenous gadolinium-based contrast media is
widely used in MRI for a number of reasons: it improves
sensitivity of lesion detection, and provides better
diagnostic specificity and more accurate depiction of
the extent of the disease involvement. Although the total
adverse reaction rates are in the range of 0.17% to 2.40%,
many of the problems that lead to concerns about the use of
iodinated contrast agents also exist in the use of gadoli-
nium-based contrast media [1–3]. The purpose of this
report is to quantify the frequency and manifestations of
adverse reactions associated with the use of gadolinium-
based contrast media in our regional hospital.
Materials and methods
From January 1999 to November 2004, 18 142 MRI
examinations were performed at the Princess Margaret
Hospital, Hong Kong, SAR (including those performed
on patients referred from Yan Chi Hospital and Caritas
Medical Center, Hong Kong, SAR). Gadolinium-based
contrast media was used in 9528 (52.8%) examinations.
All adverse reactions that occurred in patients within 1 h
of intravenous administration of the gadolinium-based
contrast media were documented in an incidence log
book by the nursing staff or the responsible radiologist.
This information was either provided voluntarily by the
patients or observed by the MRI radiographers, nurses or
radiologists. The following details were recorded in the
incidence log book: demographic data of the patients
(name, medical record number, sex and age); medical
history (e.g. asthma, or previous allergic reactions such as
allergies to iodinated or gadolinium-based contrast
media); severity and nature of adverse reactions; treat-
ments given and actions taken; the name and signature
of the reporter and responsible radiologists. This
information, the medical records and the MRI radiology
reports relevant to the patients were reviewed.
An adverse reaction was defined to be any unfavour-
able or unintended alteration in the clinical status that
was temporally associated with the use of contrast
media, even if it was not considered to be related to
administration of the drug. The sensations of hot or cold,
transient sensations of taste or mild pain at the injection
site were excluded as adverse reactions. The severity of
the reaction was subdivided into mild, moderate and
severe. Mild reactions were defined as symptoms that
were transient, or that required minimal or no therapy.

368 The British Journal of Radiology, May 2006

Adverse reactions to MR contrast media
Moderate reactions were more severe in nature or
occurred for a longer duration than mild reactions;
therapy was required but the patient was not considered
critical. Patients who had severe reactions required
resuscitation and their condition was considered to be
possibly life-threatening.
Results
A total of 45 (0.48%) of the 9528 patients were recorded
as having adverse reactions to the gadolinium-based
media. One patient developed repeated mild contrast
reactions in two separate MRI studies (counted as one
patient in our study). The mean age of the study
population was 42 years (median 42 years, range 12–79
years). Five patients (11%) were in the paediatric age
group (16 years or younger). 58% of the patients were
female (n526).
Table 1 displays the adverse reactions that occurred in
patients with a history of potentially predisposing
conditions. Among the 45 patients who developed
adverse reactions to gadolinium-based contrast media,
8 patients (18%) had previous exposure to an iodinated
contrast media (one patient had a prior adverse reaction
to the iodinated contrast media); 34 patients (75%) had
no previous exposure; status was unknown in 3 patients
(7%). Nine of the 45 patients (20%) who developed
adverse reactions had prior exposure to gadolinium-
based contrast media (three patients had prior adverse
reactions to the gadolinium-based contrast media); 33
patients (73%) had no prior exposure; the contrast
history was unknown in three patients (7%). Nine of
the 45 patients (20%) had a history of allergies (eczema,
to drugs, alcohol, etc.) and one patient (2%) had a history
of asthma.
Table 2 displays the type, number, and severity of
adverse reactions that occurred after the use of intrave-
nous gadolinium-based contrast media. Gadoterate
meglumine (Dotarem; Guerbet, France) was used in
three patients, Gadodiamide (Omniscan; Nycomed, New
York, NY) in 23 patients and Gadopentetate dimeglu-
mine (Magnevist; Berlex, Schering AG, Germany) in 19
patients. The dose was administered according to the
manufacturers’ instructions. Two of the patients who
developed adverse reactions underwent MR angiogra-
phy, which involved the use of power injectors to
administer the contrast agent at a rate of 2 ml per
second. In all other cases, the contrast media was
manually administered by injection.
Table 1. The medical history of patients who had adverse
reactions
Parameters Patient history
Yes No Unknown
Asthma 1 (2%) 37 (82%) 7 (16%)
History of allergy 9 (20%) 35 (78%) 1 (2%)
MRI contrast agent
Previous exposure 9 (20%) 33 (73%) 3 (7%)
Adverse reaction 3 has a total adverse reaction rate in the range of 0.17–
Iodinated contrast agent
Previous exposure 8 (18%) 34 (75%) 3 (7%)
Adverse reaction 1
The British Journal of Radiology, May 2006
Table 2. The type and severity of the adverse reactions
Adverse reactions Severity (number of patients)
Urticaria Mild (4)
Rash Mild (11)
Nausea/vomiting Mild (18)
Vasovagal attack Mild (2)
Dizziness Mild (3)
Confusion Mild (1)
Dyspnoea Mild (1), Moderate (1)
Chest discomfort Mild (1)
Palpitation Mild (2)
Anaphylactoid shock Severe (1)
Total 45
None of the patients in our study with a history of
adverse reactions to gadolinium-based (three patients) or
iodinated contrast media (one patient) had pre-medica-
tion to prevent recurrence of such adverse reactions.
Among these 45 patients, the majority of adverse
reactions to gadolinium-based contrast media were mild
and transient (96%) (Table 2). The most frequent mild
reaction was nausea and vomiting (40%) and 33% of
patients were affected by urticaria and a rash. All
patients with nausea, vomiting and urticaria responded
to conservative management, and the seven patients who
developed a rash were treated with a stat dose of oral
diphenhydramine. One patient developed dyspnoea that
was treated with oxygen supplementation, intravenous
steroids and fluids, and further inpatient management.
The patient was subsequently discharged uneventfully.
This patient did not have a history of respiratory
disorders such as asthma or chronic obstructive airway
disease—bronchospasm was the presumed cause of the
dyspnoea. Among the 9528 patients who underwent
contrast MRI, one patient (0.01%) developed an anaphy-
lactoid reaction. This 77-year-old patient was referred for
outpatient MRI for follow up of a known malignant
melanoma of the nostril that had intracranial extension.
The patient had undergone a contrast enhanced MRI
study using gadodiamide without complications, but
developed anaphylactoid shock shortly after injection of
gadopentate dimeglumine. The patient did not have a
history of asthma, allergies, or previous sensitivities to
drugs including iodinated contrast media. The patient
was successfully resuscitated (by securing the airways,
use of oxygen, and intravenous diphenhydramine,
steroids, adrenaline and fluids) and was hospitalized
for intensive care and later discharged uneventfully. No
contrast extravasation was recorded for any patient and
none died as a result of an adverse reaction.
Discussion
Gadolinium-based contrast agents constitute the lar-
gest group of MRI contrast media and are considered to
be very safe and well tolerated. Gadopentetate dimeglu-
mine, the first agent approved by the Food and Drug
Administration (FDA) for clinical use in the USA (1988),
2.40% [1–3]. Nelson et al [1] reported the adverse reaction
rate in a patient population of 15 496 given gadopente-
tate dimeglumine was 2.4%. The overall adverse reaction
369

rate was 1.2% if patients voluntarily reported their
symptoms instead of being directly questioned.
The safety of the gadolinium chelates is largely based
on their stability in vivo. The chelates are designed to
bind gadolinium ions tightly, preventing the possible
release of free gadolinium ions, which are extremely
toxic to the body. The differences between the gadoli-
nium chelates in terms of chemical stability and
transmetallation gives rise to some theoretical concerns.
Transmetallation refers to the substitution of the chelates
by copper and zinc ions in the body leading to the release
of free gadolinium ions. Both gadopentetate dimeglu-
mine and gadodiamide can inhibit the zinc-dependent
angiotensin-converting enzyme (ACE) in the body and
may also give rise to transmetallation [4–7]. To our
knowledge, no harmful effects in humans caused by free
gadolinium ion deposition resulting from the clinical use
of these agents have been reported. All commercially
available gadolinium chelates have comparable adverse
reaction profiles including mild reactions such as nausea
and severe anaphylactoid reactions [4]. For reference, the
incidence of urticaria in clinical trials for gadopentetate
dimeglumine is 0.3% in 1068 patient [8], for gadodiamide
it is 0.7% in 439 patients [9], and for gadoterate
meglumine it is 0.4% in 518 patients [10]. These figures
are not statistically differentiable.
Numerous cases of severe anaphylactoid reactions
associated with the use of gadolinium-based contrast
media have been reported [11–13]. For example, a fatal
reaction to gadopentetate dimeglumine was reported in
1995. This patient did not have a history of asthma or
previous allergic reactions to iodinated contrast media
[14]. The incidence rate of severe anaphylactoid reaction
in our series was 0.01%. This concurs with a large series
by Murphy et al [3], in which two severe reactions to
gadopentetate dimeglumine out of 21 000 patients were
reported. For reference, in another study by Caro et al,
the risk of a life-threatening event with iodinated
radiographic contrast media was 0.031% for low-osmo-
larity contrast media, and 0.157% for conventional ionic
contrast media [15]. Our results again confirm that
gadolinium-based contrast media is safer than iodinated
contrast media. In 1994, Witte et al reported a case of life-
threatening anaphylactoid reaction that occurred after
the administration of intravenous gadoteridol (Prohance;
Bracco Diagnostics, Princeton, NJ) in a patient who had
previously received gadopentetate dimeglumine [16]. To
our knowledge, our study is the first reported case in the
English language literature of a patient who developed
severe anaphylactoid reaction after intravenous admin-
istration of gadodiamide without having exhibited any
adverse reaction with the prior use of gadopentetate
dimeglumine. Although the incidence of life-threatening
reactions to gadolinium-based contrast media is low,
anaphylactoid reactions are among the worst adverse
reactions that might occur with the use of MRI contrast
agents. Therefore, it is prudent to provide sufficient
training for personnel and to have available adequate
resuscitation facilities in the MRI suite if these agents are
to be used.
Nelson et al reported that the likelihood of adverse
reactions to gadolinium-based contrast media is about
eight times higher in patients who have had previous
reactions to gadolinium-based contrast media. The
370
A Li, C S Wong, M K Wong et al
severity of a second adverse reaction tends to be more
pronounced than the first [1]. In our series, all three
patients who had prior adverse reactions to gadolinium-
based contrast media developed adverse reactions after
repeated use of such an agent. These three patients were
not pre-medicated to reduce risk of contrast reactions. In
an attempt to reduce the risk of a reaction to the contrast
media, a different gadolinium-based media (gadodia-
mide) was administered instead of the previously used
gadopentate dimeglumine. However, adverse reactions
still developed despite the use of a different gadolinium-
based contrast media, but these reactions were mild and
transient (rash and nausea). The frequency of adverse
reactions among patients who have a history of adverse
reactions to gadolinium-based contrast media cannot be
determined from our data. However, our observations
may support the findings of the study by Nelson et al.
The role of pre-medication with corticosteroids as a
preventive measure in patients with prior reactions to
gadolinium-based contrast media has not been well
established. In a study by Murphy et al [3], two patients
were pre-medicated before the gadolinium contrast
agent was administered. One patient had a single oral
dose of 50 mg diphenhydramine 1 h before the proce-
dure, and the second patient had an oral dose of 50 mg
diphenhydramine 1 h before the procedure and 20 mg
prednisolone every 6 h (starting 13 h before the exami-
nation). Both patients had the same adverse reaction as
with the previous gadolinium exposure. The second
patient returned for a further MRI study and was given
the same pre-medication, but a different contrast agent
from the previous one. He did not develop any
complications. Murphy et al [3] recommends the use of
12–24 h pre-medication with corticosteroids and subse-
quent use of a different or lower osmolar MRI contrast
agent as a preventative measure among this group of
patients. The evidence of benefits of such pre-medication
has not been well established. Until more data are
available, gadolinium-based MRI contrast media should
be used with caution in patients with a history of adverse
reactions, in particular severe reactions.
The frequency of adverse reactions to gadolinium-
based contrast agents is about 2.3–3.7 times higher in
patients with a history of reactions to iodinated contrast
material [1, 2]. In our study, three of the 45 patients who
experienced adverse reactions to gadolinium-based con-
trast media had also had previous reactions to iodinated
contrast media, but the rash and nausea symptoms were
mild and transient. Again, administration of gadolinium-
based contrast media among these types of patients should
be made with caution, because the role of steroid pre-
medication has not been well established.
Patients with asthma and various allergies are also at
greater risk of adverse reactions. Nelson et al [1] reported
that the frequency of adverse reactions is 1.5 and 1.9 times
higher, respectively, compared with patients who have
none of these allergies. Among the 45 patients in our study
who had adverse reactions (Table 1), one patient had
asthma and nine patients had a prior allergy (2 urticaria, 2
seafood, 1 eczema, 3 alcohol and 1 aspirin related).
In this study, 19 out of 46 cases of contrast adverse
reactions (41%) were not documented in the final MRI
report or the clinical medical records. The adverse
reactions in these cases were all mild and transient.
The British Journal of Radiology, May 2006
Adverse reactions to MR contrast media
Proper documentation of the adverse reaction is not only
essential for medico-legal purposes, but also more
importantly for the interest and well being of the patient.
If precautions could be taken before the repeated
administration of gadolinium-based contrast media, then
the awareness and management of adverse reactions
would be more efficient. Therefore, the accurate record-
ing of the reactions associated with gadolinium-based
contrast media should be as thorough as they are for
reactions associated with other drugs.
Several limitations were identified in the design of this
study. The design was a retrospective review of the
incidence log book, which was originally intended for
documenting all acute adverse reactions to the gadoli-
nium chelates for medical records and for departmental
audit. Although clear instructions were given to staff in
the department, all acute adverse reactions may not have
been documented, in particular those that were trivial in
nature. Voluntary reporting by patients was also likely to
lead to under-reporting. Additional minor reactions
might have been reported if patients were questioned
directly or by way of a questionnaire. Only acute adverse
reactions were recorded, occurring within 1 h between
the gadolinium-based contrast media injection and the
patient leaving the department after observation. In the
study by Nelson et al, 44.9% of adverse reactions
occurred more than 1 h after the gadolinium-based
contrast media injection. Those reactions were all mild
and transient, with no severe or life-threatening events
recorded [1]. In such circumstances, the incidence of
acute adverse reactions in our patient population still
concurs with those reported in the literature.
In conclusion, gadolinium-based contrast media is safe
and well tolerated by the vast majority of patients.
Although most adverse reactions are mild and transient,
life-threatening anaphylactoid reactions do occur.
Personnel should be well trained and resuscitation
facilities should be readily available in the MRI examina-
tion room. Accurate documentation and cautious
management of all adverse reactions related to gadoli-
nium-based contrast media should be made in the same
manner as for other drug reactions.
References
1. Nelson KL, Gifford LM, Lauber-Huber C, Gross CA, Lasser
TA. Clinical safety of gadopentetate dimeglumine.
Radiology 1995;2:439–43.
The British Journal of Radiology, May 2006
2. Niendorf H, Dinger J, Haustein J, Cornelius I, Alhassan A,
Clauss W. Tolerance data of Gd-DTPA: a review.
Eur J Radiol 1991;13:15–20.
3. Murphy KJ, Brunberg JA, Cohan RH. Adverse reactions to
gadolinium contrast media: a review of 36 cases. AJR
Am J Roentgenol 1996;167:847–9.
4. Runge VM. Safety of approved MR contrast media for
intravenous injection. J Magn Reson Imaging
2000;12:205–13.
5. Wedeking P, Kumar K, Tweedle MF. Dissociation of
gadolinium chelates in mice: relationship to chemical
characteristics. Magn Reson Imaging 1992;10:641–8.
6. Tweedle MF, Wedeking P, Kumar K. Biodistribution of
radiolabeled, formulated gadopentetate, gadoteriodol,
gadoterate, and gadodiamide in mice and rats. Invest
Radiol 1995;30:372–80.
7. Corot C, Idee JM, Hentsch AM, et al. Structure-activity
relationship of macrocyclic and linear gadolinium chelates;
investigation of transmetallation effect on the zinc-depen-
dent metallopeptidase angiotensin-converting enzyme.
J Magn Reson Imaging 1998;8:695–702.
8. Goldstein HA, Kashanian FK, Blumetti RF, et al. Safety
assessment of gadopentetate dimeglumine in U.S. clinical
trials. Radiology 1990;174:17–23.
9. Sze G, Brant-Zawadzki M, Haughton VM, Maravilla KR, et
al. Multicenter study of gadodiamide injection as a contrast
agent in MR imaging of the brain and spine. Radiology
1991;181:693–9.
10. Oudkerk M, Sijens PE, Van Beek EJ, Kuijpers TJ. Safety and
efficacy of dotarem (Gd-DOTA) versus Magnevist (Gd-
DTPA) in magnetic resonance imaging of the central
nervous system. Invest Radiol 1995;30:75–8.
11. Tishler S, Hoffman JC Jr. Anaphylactoid reactions to i.v.
gadopentetate dimeglumine. Am J Neuroradiol
1990;11:1167–9.
12. Tardy B, Guy C, Barral G, Page Y, Ollagnier M, Bertrand JC.
Anaphylactic shock induced by intravenous gadopentetate
dimeglumine. Lancet 1992;339:494.
13. Sahni H, Jayakumar PN, Lalla RK, Vasudev MK, Kovoor
JME. Anaphylactoid reaction following intravenous gado-
linium dimeglumine - a case report. Ind J Radiol Imag
2002;12:1:81–2.
14. Jordan RM, Mintz RD. Fatal reaction to gadopentetate
dimeglumine. AJR Am J Roentgenol 1995;164:743–4.
15. Caro JJ, Trindade E, McGregor M. The cost-effectiveness of
replacing high-osmolarity with low-osmolarity contrast
media. AJR Am J Roentgenol 1992;159:869–74.
16. Witte RJ, Anzai LL. Life-threatening anaphylactoid reaction
after intravenous gadoteridol administration in a patient
who had previously received gadopentetate dimeglumine.
Am J Neuroradiol 1994;15:523–4.
371
 The British Journal of Radiology, 79 (2006), 372–377

Observer variation in the evaluation of lumbar herniated discs

and root compression: spiral CT compared with MRI
1
J C VAN RIJN, MD, 2N KLEMETSO, MD, 1J B REITSMA, MD, PhD, 1P M BOSSUYT, PhD, 2F J HULSMANS, MD,
4 2 3 2
PhD, W C PEUL, MD, PhD, G J DEN HEETEN, MD, PhD, J STAM, MD, PhD and C B L M MAJOIE, MD, PhD

Departments of 1Clinical Epidemiology and Biostatistics, 2Radiology and 3Neurology, Academic

Medical Center, Meibergdreef 9, 1105 AZ, Amsterdam and the 4Department of Neurosurgery,
Leiden University Medical Center, Albinusdreef 2, 2333 ZA, Leiden, The Netherlands

ABSTRACT. Spiral CT is considered the best alternative for MRI in the evaluation of
herniated discs. The purpose of this study was to compare radiological evaluation of
spiral CT with MRI in patients suspected of herniated discs. 57 patients with lumbosacral
radicular syndrome underwent spiral CT and 1.5 T MRI. Two neuroradiologists
independently evaluated 171 intervertebral discs for herniation or ‘‘bulge’’ and 456
nerve roots for root compression, once after CT and once after MRI. We compared
interobserver agreement using the kappa statistic and we performed a paired
comparison between CT and MRI. For detection of herniated or bulging discs, we
observed no significant difference in interobserver agreement (CT kappa 0.66 vs MRI
kappa 0.71; p50.40). For root compression, we observed significantly better
interobserver agreement at MRI evaluation (CT kappa 0.59 vs MRI kappa 0.78; p50.01). Received 16 August 2005
In 30 of 171 lumbar discs (18%) and in 54 of 456 nerve roots (12%), the observers Revised 23 November 2005
disagreed on whether CT results were similar to MRI. In the cases without Accepted 9 December 2005
disagreement, CT differed from MRI in 6 discs (3.5%) and in 3 nerve roots (0.7%). For
DOI: 10.1259/bjr/26216335
radiological evaluation of lumbar herniated discs, we found no evidence that spiral CT
is inferior to MRI. For evaluating lumbar nerve root compression, spiral CT is less ’ 2006 The British Institute of
reliable than MRI. Radiology

In a series of comparative radiological studies on the
evaluation of lumbar herniated discs, MRI was found to
be ‘‘slightly better’’ than [1–4] or equal to [5–8] CT. We
found one study suggesting CT to be more accurate than
MRI [9], but overall, MRI is preferred and CT seems to
have comparable capacity to diagnose herniated discs.
Therefore, current guidelines designate MRI as the first-
choice investigation and suggest CT as the alternative in
the evaluation of the lumbar back if MRI is contra-
indicated or unavailable [11–13].
Besides herniated discs, the direct evaluation of nerve
roots by MRI has been considered an important asset to
facilitate decision making in patients with back pain [14–
17]. No evidence, however, is available on the capabilities
of spiral CT to evaluate lumbar nerve root compression.
The purpose of our study was to compare spiral CT
with MRI in a series of patients suspected of lumbar
herniated discs. For each technique we assessed inter-
observer agreement in detecting herniated discs, bulging
discs and nerve root compression. We also performed a
paired, direct comparison between spiral CT and MRI.
Methods
This study was conducted between June 1999 and June
2000, as part of a larger project on the diagnostic process
of patients with lumbosacral radicular pain at the
University of Amsterdam, The Netherlands.
The Institutional Review Board approved the study
protocol.
Patients were recruited from the neurology outpatient
department. Those eligible were patients referred by
their general practitioner with lumbosacral radicular
syndrome (LRS) with suspected disc herniation at the
levels L3–L4 to L5–S1 in whom conservative treatment
had been unsuccessful. LRS was defined according to the
national general practitioners’ guideline and the con-
sensus statement on diagnosis and treatment of LRS
defined by the Dutch Neurology Society [18]. The
hallmark of this definition is continuous monoradi-
cular or multiradicular pain below the knee with a
primary suspicion of disc herniation. Excluded were
patients younger than 18 years or older than 70 years,
pregnant women, patients with a previous history of
lumbosacral herniation or lumbosacral surgery, as well
as patients with contraindication for MRI. Eligible
patients received written and oral information about
the study.
After the neurologist had confirmed the diagnosis of
LRS, consenting patients were subjected to spiral CT and
MRI within 1 week. No specific treatment was given
within this period.

372 The British Journal of Radiology, May 2006

Spiral CT compared with MRI in the evaluation of suspected lumbar herniated discs
Imaging techniques
Spiral CT
Lumbar CT examinations were performed on a 2-slice
CT-Twin scanner (Philips Medical Systems, Best, The
Netherlands). Helical CT-scans were made with 120 kV,
265 mAs (effective dose: 6.4 mSv), table feed of
1.4 mm s-1, 2 mm61 mm collimation (effective slice
width, 1.1 mm) and 0.5 mm increment from the level
L3 to the bottom of S1. The gantry angle was aligned
through the disc space of L4–L5. In addition to the
original axial images with the bone window (level
400 HU/window 1600 HU), reformatted 4 mm axial
sections of all scanned interspaces were made parallel
to the interspaces using the soft tissue window (level
50 HU/window 180 HU). Also reformatted ‘‘0 mm’’
(i.e. slice width in the order of 1 mm) sagittal sections
were made using the bone window, and 4 mm sagittal
sections using the soft tissue window.
The quality of the CT-images used in this study,
obtained with a 2-slice CT-scanner, is comparable with
the quality of more modern, multislice CT-scanners. The
effective slice width in this study was 1.1 mm, which is
only slightly more than the effective slice width of
multislice CT-scanners. The effective slice width is
usually 25–30% higher than the collimation width and
thus in the order of 0.9–1.0 mm for, e.g. a collimation of
16 mm60.75 mm [19, 20].
MRI
Lumbar MR examinations were performed with a 1.5 T
Signa LX Scanner (General Electric Medical Systems,
Milwaukee, WI) using a dedicated lumbar spine surface
coil. The protocol included sagittal spin-echo T1
weighted (repetition time (TR) 500 ms, echo time (TE)
14 ms) and proton density/T2 weighted (TR 3500 ms, TE
20–120 ms) fast spin echo images with 4 mm slice
thickness, 0.5 mm intersection gap, 2006512 matrix
and 29 cm629 cm field of view. In addition, axial
spin-echo T1 weighted (TR 520 ms, TE 12 ms) and fast
spin-echo T2 weighted (TR 4500 ms, TE 120 ms) images
were obtained from the increment of L3 to the bottom
of S1 with 4 mm slice thickness, 0.5 mm inter-
section gap, 2006256 matrix and 15 cm615 cm field
of view. Axial images were obtained without
angulation.
Image evaluation
Two experienced neuroradiologists (CB and FJ) inde-
pendently evaluated all CT and MR images with
knowledge of the side and level of symptoms. Per
observer, the images were evaluated in two independent
sessions. There was at least 6 weeks between these
sessions. The images of either CT or MRI were presented
per patient. It was randomly allocated whether CT or
MRI was evaluated in the first session.
Three lumbar discs were examined per patient at
levels L3–L4 to L5–S1. Each disc was scored for the
The British Journal of Radiology, May 2006
presence of a herniation. No distinction between protru-
sion and extrusion was made; both were considered
‘‘herniated disc’’. If no herniation was detected, readers
evaluated the presence of a bulging disc. The definition
of a bulging disc was according to the description by
Jensen [21]: ‘‘circumferential symmetric extension of the
disc beyond the interspace’’.
Nerve roots L3 through S1 were evaluated per side
(eight nerve roots per patient) within the scanned area. A
nerve root was defined as the part of the nerve between
the central canal and the ganglion. As the scan protocol
for both techniques was from the increment of vertebra
L3 through the bottom of S1, the radiologists were not
able to evaluate a minimal proximal part of root L3 and a
minimal distal part of root S1.
A five-point scale was used per nerve root, anchored
as definitely no root compression, possibly no root
compression, indeterminate, possibly root compression,
and definitely compression. For the analyses, these
responses were dichotomized as ‘‘root compression’’
(possibly or definitely) or ‘‘no root compression’’ (all
other categories). MRI examples of the different cate-
gories have been presented in an earlier report on the
same series of patients [22].
Data analysis
Per technique, we calculated interobserver agreement
for the evaluation of herniated discs and bulging discs by
constructing 363 tables. Data were matched for lumbar
level (L3–L4 to L5–S1). Presence of nerve root compres-
sion was analysed using a 262 table. These data were
matched for each side (left and right) and for each nerve
root (L3 to S1).
To acknowledge possible correlation between inter-
vertebral discs and nerve roots within one patient,
bootstrap sampling was used to calculate the
unweighted interobserver kappa statistics and stan-
dard-errors for each contingency table [23]. A two-sided
z-test was then used to compare the kappa statistics of
CT with MRI.
To differentiate between CT versus MRI differences
due to disagreement and ‘‘true’’ differences (no dis-
agreement amongst observers), we performed a paired
comparison of the results of CT and MRI. We limited the
analysis to the detection of herniated discs. Therefore,
the readings for each technique were recoded to
‘‘herniated disc’’ or ‘‘no herniated disc’’. The latter
category comprised bulging discs and discs reported as
having no abnormality.
The results were coded per technique as follows:
(1) both observers agreed on the presence of a Hernia
Nuclei Pulposi (HNP), (2) observers disagreed or (3) both
observers agreed on the absence of a HNP. Subsequently,
a 363 table including the results of both observers was
constructed to compare both techni-
ques on a per patient basis. The McNemar test for
paired data was used to detect differences between
modalities.
All calculations were performed with SAS 9.0 and
SPSS 11.0. Values lower than p.0.05 were considered to
indicate statistical significance.
373
 J C van Rijn, N Klemetso, J B Reitsma et al

Results root, both readers reported root compression at MRI but

64 consecutive eligible patients with LRS were
identified. Three patients did not undergo MRI because
of claustrophobia. Four more patients had incomplete
data of spiral CT and MRI, leaving 57 patients with 171
lumbar discs and 456 lumbar roots for analysis.
Table 1 shows the interobserver agreement. There was
no significant difference between CT and MRI with
respect to interobserver agreement of herniated disc and
bulging disc evaluation, but for root compression the
kappa statistic was significantly higher for MRI. Figure 1
shows an example of a case in which CT and MRI agree
on a HNP with root compression without any inter-
observer disagreement.
Table 2 shows the results of the paired comparison of
CT versus MRI in detecting herniated discs and root
compression. Of 171 discs, observers disagreed on
herniation at CT evaluation in 20 discs (12%) compared
with 12 discs (7%) at MRI. The difference was not
significant (McNemar, p50.22).
In 141 (83%) cases there was no disagreement at CT
evaluation as well as no disagreement at MRI evaluation.
In three of these cases both readers reported a HNP at
CT, but not at MRI evaluation, and three times both
readers reported a HNP at MRI but not at CT evaluation.
Observers disagreed on the presence of root compres-
sion when evaluating CT in 40 of 456 nerve roots (8.8%)
compared with 22 roots (4.8%) when evaluating MRI.
The difference was significant (McNemar p50.03). An
example of a case in which observers disagreed on root
compression at CT but not on MRI is presented in
Figure 2.
In 402 roots (88%) there was no disagreement at CT
evaluation as well as no disagreement at MRI evaluation.
Of these cases, twice both readers reported root
compression at CT but not at MRI evaluation. In one
not at CT evaluation (case is presented in Figure 3).
Discussion
In this study we found no evidence that spiral CT is
inferior to MRI in the evaluation of herniated lumbar
discs and bulging discs, but we observed better inter-
observer agreement at MRI with respect to the evaluation
of nerve root compression.
The paired comparison revealed that in the vast
majority of the observed differences between spiral CT
and MRI there was interobserver disagreement involved
(either at CT or at MRI or at both evaluations). In these
cases, we cannot discriminate between differences due to
reader variability or due to ‘‘real’’ difference between
modalities.
The observed differences between CT and MRI with-
out any disagreement indicate possible ‘‘real’’ differ-
ences. However, after re-evaluation of these cases we can
conclude that all HNPs and all cases of root compression
were detectable on the CT images as well as on the MR
images.
It remains impossible to determine which one of these
imaging techniques corresponded best with ‘‘the truth’’
because there is no proper reference standard available
for this study. However, we observed no particular
pattern in the distribution of discordant readings: the
differences between spiral CT and MRI were divided
equally amongst both techniques. This supports the
assumption that spiral CT and MRI are comparable in
herniated disc evaluation.
Observer disagreement is a general obstacle in the
radiological evaluation of lumbar intervertebral discs,
both for spiral CT as well as for MRI. In an earlier
published study in the same series of patients, we

Table 1. Spiral CT and MRI interobserver kappa-statistics with 95% confidence intervals
Lumbar disc evaluation

Spiral CT MRI

Observer 2 Observer 2

Observer 1 HNP Bulge No HNP Total Observer 1 HNP Bulge No HNP Total

HNP 49 10 6 65 HNP 52 3 1 56
Bulge 3 24 9 36 Bulge 7 20 4 31
No HNP 1 9 60 70 No HNP 1 15 68 84
Total 53 43 75 171 Total 60 38 73 171
Kappa 0.66 (0.56–0.75) Kappa 0.71 (0.63–0.80)
Difference50.05; p50.40
Root compression

Spiral CT MRI

Observer 2 Observer 2

Observer 1 RC No RC Total Observer 1 RC No RC Total

RC 35 16 51 RC 45 10 55
No RC 24 381 405 No RC 12 389 401
Total 59 397 456 Total 57 399 456
Kappa 0.59 (0.47–0.70) Kappa 0.78 (0.69–0.87)
Difference50.19; p50.01
HNP, Hernia Nuclei Pulposi; RC, root compression.

374 The British Journal of Radiology, May 2006

Spiral CT compared with MRI in the evaluation of suspected lumbar herniated discs

Figure 1. 22-year-old male with left-sided lumbosacral radicular syndrome of L5 and paresis of the anterior tibial muscle and
with hypaesthesia of L5 and S1. Images show a left-sided herniated disc at level L4–L5 with root compression of L5 reported at CT
as well as at MRI by both observers (arrows). (a) Axial CT L4–L5. (b) Axial T2 weighted MRI L4–L5.

investigated the possible causes of interobserver dis-
agreement in MRI evaluation [22]. We identified a lack of
consensus on nomenclature of bulging discs as the main
cause of disagreement. Other causes were additional
pathology (e.g. spondylolisthesis or collapsed vertebral
bodies) and small sized lesions, which were believed to
be missed more easily.
The disagreement on nerve root evaluation can for a
large part be attributed to the lack of a proper
standardized and reproducible method to categorize
root compression. We chose to use a five point scale to
evaluate root compression. Compared with a dichoto-
mous test result, the advantage of this method was that it
resembled clinical practice because the observers were
given the opportunity to express their uncertainty. A
limitation of this method was that we were not able to
retrospectively retrieve the decisive items to get to a
radiological diagnosis.
Future studies should focus on reducing observer
disagreement through the development of standardized
radiological nomenclature for lumbar back abnormal-
ities. Development of universal selection criteria to
identify specific target groups of patients with back pain
and more specific imaging strategies are needed to
improve efficiency in decision making. This is also
supported by the recent MR-based study by

Table 2. Paired comparison of spiral CT versus MRI

Herniated disc detection

MRI

CT HNP Disagree No HNP Total

HNP 43 4 3 50
Disagree 6 2 12 20
No HNP 3 6 92 101
Total 52 12 107 171
Root compression

MRI

CT RC Disagree No RC Total

RC 32 1 2 35
Disagree 12 8 20 40*
No RC 1 13 367 381
Total 45 22* 389 456
HNP, both observers agreed on the presence of a HNP; No HNP, both observers agreed on the absence of a HNP.
*Significantly more disagreement at CT evaluation (McNemar p50.03).
HNP, Hernia Nuclei Pulposi; RC, root compression.

The British Journal of Radiology, May 2006 375

 J C van Rijn, N Klemetso, J B Reitsma et al

Figure 2. 43-year-old male with left-sided lumbosacral radicular syndrome of S1. At both CT and MRI there is disagreement on
root compression of L5 on the opposite side of signs and symptoms. (a) Axial CT L4–L5. Observers disagree on whether root L5 on
the right (asymptomatic side), is compressed (arrow). (b) Axial T2 weighted MRI L4–L5. Observers agree on compression of L5 on
the right (asymptomatic side) (arrow).

Cihangiroglu in 2004 [24]. To investigate the additional
value of myelography (either conventional or by CT or
MR imaging) might also be a valuable goal for future
studies. Especially in the evaluation of nerve root
impingement, it is suggested that myelography can play
an important role [25].
In our study, the observers were not blinded for side
and level of symptoms. This was done to mimic practice.
As part of a larger project, we also obtained blinded
evaluations of the same images. We observed no
significant differences between blinded and not blinded
evaluations [22]. After re-evaluation of the discordant

Figure 3. 58-year-old male with left-sided lumbosacral radicular syndrome of L3. (a) Axial CT L3–L4. Both observers reported
‘‘definitely no compression’’ of L3 on the left (arrow). (b) Axial T2 weighted MRI L3–L4. Both observers reported ‘‘definitely
compression’’ of L3 on the left (arrow).

376 The British Journal of Radiology, May 2006

Spiral CT compared with MRI in the evaluation of suspected lumbar herniated discs
results in the present study, we have no reason to believe
that the evaluations were influenced by knowledge of
suspected side and level of the LRS.
Our study was limited to the evaluation of herniated
discs, bulging discs and root compression. MRI has
better qualities to depict all surrounding soft tissue in the
lumbar area. This property of MRI is one of the major
reasons for neurosurgeons to use MRI to determine
whether a patient is a candidate for surgical intervention.
We did not investigate to what extent spiral CT might be
useful within surgical triage.
In summary, we recommend MRI as the investigation
of choice in patients suspected of herniated discs.
Nevertheless, a significant number of patients are not
able to undergo MRI due to claustrophobia or other
contraindications [26, 27]. In these patients, spiral CT
seems to be an excellent alternative. However, for the
evaluation of root compression, spiral CT appears to be
not as reliable as MRI.
References
1. Goscinski I, Ulatowski S, Urbanik A. Comparison of the
clinical usefulness of magnetic resonance (MR), computer
tomography (CT) and radiculography (R) in diagnostic
lumbar discopathy. Przegl Lek 2001;58:885–8.
2. Janssen ME, Bertrand SL, Joe C, et al. Lumbar herniated
disk disease: comparison of MRI, myelography, and post-
myelographic CT scan with surgical findings. Orthopedics
1994;17:121–7.
3. Forristall RM, Marsh HO, Pay NT. Magnetic resonance
imaging and contrast CT of the lumbar spine. Comparison
of diagnostic methods and correlation with surgical find-
ings. Spine 1988;13:1049–54.
4. Pevsner PH, Ondra S, Radcliff W, et al. Magnetic resonance
imaging of the lumbar spine. A comparison with computed
tomography and myelography. Acta Radiol Suppl
1986;369:706–7.
5. Albeck MJ, Hilden J, Kjear L, et al. A controlled comparison
of myelography, computed tomography and magnetic
resonance imaging in clinically suspected lumbar disc
herniation. Spine 1995;20:443–8.
6. Tullberg T, Grane P, Rydberg J, et al. Comparison of
contrast-enhanced computed tomography and gadolinium-
enhanced magnetic resonance imaging one year after
lumbar discectomy. Spine 1994;19:183–8.
7. Thornbury JR, Fryback DG, Turski, PA, et al. Disk-caused
nerve compression in patients with acute low-back pain:
diagnosis with MR, CT myelography and plain CT.
Radiology 1993;186:731–8.
8. Jackson RP, Cain JE, Jacobs RR, et al. The neuroradio-
graphic diagnosis of lumbar herniated nucleus pulposus: II.
A comparison of computed tomogrpahy (CT), myelogra-
phy, CT-myelography, and magnetic resonance imaging.
Spine 1989;14:1362–7.
The British Journal of Radiology, May 2006
9. Dellerud R, Johansen JG, Johnsen UL, et al. Differentiation
between contained and noncontained lumbar disk hernias
by CT and MR imaging. Acta Radiol 1995;36:491–6.
10. Eysel P, Rompe JD, Schaub T, et al. Value of imaging
techniques of lumbar intervertebral disk prolapse. Z Orthop
Ihre Grenzgeb 1994;132:371–6.
11. Patel N. Surgical disorders of the thoracic and lumbar
spine: a guide for neurologists. J Neurol Neurosurg
Psychiatry 2002;73(Suppl. I):i42–8.
12. Milette PC. Classification, diagnostic imaging, and imaging
characterization of a lumbar herniated disk. Radiol Clin N
Am 2000;38:1267–92.
13. Herzog RJ. The radiologic assessment for a lumbar disc
herniation. Spine 1996;21(24S):19S–38S.
14. Vroomen PC, De Krom MC, Wilmink JT, et al. Diagnostic
value of history and physical examination in patients
suspected of lumbosacral nerve root compression. J
Neurol Neurosurg Psychiatry 2002;72:630–4.
15. Pfirrmann CW, Dora C, Schmid MR, et al. MR Image-based
grading of lumbar nerve root compromise due to disk
herniation: reliability study with surgical correlation.
Radiology 2004;230:583–8.
16. Nygaard OP, Jacobsen EA, Solberg T, et al. Nerve root signs
on postoperative lumbar MR imaging. A prospective cohort
study with contract enhanced MRI in symptomatic and
asymptomatic patients one year after microdiscectomy.
Acta Neurochir (Wien) 1999;141:619–23.
17. Jarvik JJ, Hollingworth W, Haegerty P, et al. The long-
itudinal assessment of imaging and disability of the back
(LAIDBack) study. Spine 2001;26:1158–65.
18. Stam J. Consensus on diagnosis and treatment of lumbo-
sacral root entrapment syndromes. Ned Tijdschr Geneeskd
1996;140:52.
19. Prokop M. General principles of MDCT. Eur J Radiol
2003;45:S4–S10.
20. Prokop M. Multislice CT: technical principles and future
trends. Eur Radiol 2003;13 (Suppl. 5):M3–13.
21. Jensen MC, Brant-Zawadski MN, Obuchowski N, et al.
Magnetic resonance imaging of the lumbar spine in people
without back pain. N Engl J Med 1994;331:69–73.
22. Van Rijn JC, Klemetso N, Reitsma JB, et al. Observer
variation in MRI evaluation of patients suspected of lumbar
disk herniation. AJR Am J Roentgenol 2005;184:299–303.
23. Efron B, Tibshirani RJ. An introduction to the bootstrap.
Chapman & Hall, 1993.
24. Cihangiroglu M, Yildirim H, Bozgeyik Z, et al. Observer
variability based on the strength of MR scanners in the
assessment of lumbar degenerative disc disease. Eur J
Radiol 2004;51:202–8.
25. Bartynski WS, Lin L. Lumbar root compression in the
lateral recess: MR imaging, conventional myelography, and
CT myelography comparison with surgical confirmation.
AJNR Am J Neuroradiol 2003;24:348–60.
26. Melendez JC, McCrank E. Anxiety-related reactions asso-
ciated with magnetic resonance imaging examinations.
JAMA 1993;270:745–7.
27. Kanal E, Shellock FG, Talagala L. Safety considerations in
MR imaging. Radiology 1990;176:593–606.
377
 The British Journal of Radiology, 79 (2006), 378–382

Initial experiences of using an automated volumetric measure of

breast density: the standard mammogram form
1 2 3
M JEFFREYS, PhD, R WARREN, FRCR, R HIGHNAM, PhD and 4G DAVEY SMITH, DSc

1
Centre for Public Health Research, Massey University, Private Box 756, Wellington, New Zealand,
2
Addenbrooke’s Hospital, Cambridge, 3Mirada Solutions Ltd, Oxford and 4Department of Social
Medicine, University of Bristol, UK

ABSTRACT. Limitations of area based measures of breast density have led several
research groups to develop volumetric measures of breast density, for use in predicting
risk and in epidemiological research. In this paper, we describe our initial experiences
using an automated algorithm (standard mammogram form, SMF) to estimate the
volume of the breast that is dense from digitized film mammograms. We performed
analyses on 3816 mammograms of 626 women, who were part of the Glasgow Alumni
Cohort and had mammograms taken within the Scottish Breast Screening Programme
between 1989 and 2002. Absolute volume of dense breast tissue (SMF volume) and the
percentage of the volume of the breast that is dense (SMF%) were calculated. The
median (interquartile range) of SMF volume was 66 cm3 (48 to 98), and of SMF% was
23.4% (18.6 to 29.7). SMF%, but not SMF volume, was positively related to a six
Received 6 April 2005
category classification (SCC) of visually assigned area-based breast density (increase in Revised 16 July 2005
ln(SMF%) per category increase in SCC: 0.04% (95%CI: 0.03–0.05). The SMF algorithm Accepted 1 September
produced lower SMF volume for craniocaudal (CC) compared with mediolateral oblique 2005
(MLO) views, but CC/MLO differences for SMF% were small. The mean right/left
DOI: 10.1259/bjr/24769358
difference for ln(SMF volume) was 20.027 cm3 (95% confidence interval (CI) 20.044 to
20.009) and of ln(SMF%) was 0.005% (95% CI 20.008% to 0.019%). We present these ’ 2006 The British Institute of
initial data as a background for future analytical work using SMF. Radiology

The magnitude of the relationship between breast
density and breast cancer [1] has led to recognition that
breast density may have potential use as a biomarker for
breast cancer risk [2]. Breast density is conventionally
estimated by using the area of the mammogram that
appears to be ‘‘glandular’’ and dividing by the total
breast area. This is performed either visually, using
classifications such as the Wolfe system [3], or using a
computer-based thresholding technique [1]. Each of
these classifications results in a measure of breast
density which is a strong determinant of breast cancer
risk.
Several recognized breast cancer risk factors are
positively related to area measures of breast density,
including height [4, 5], parity [6] and age at first birth [7].
There are two notable exceptions to this, namely age and
body weight, both of which are positively related to
breast cancer risk, but inversely related to breast density
[5, 8]. An inverse association between body mass index
(BMI) and percent breast density is inevitable, since
inherent in the definition of these measures is the
assumption that fatty tissue is non-dense, and women
with a high BMI have higher amounts of fat in the
breasts. In addition to this discrepancy, there are other
concerns associated with the visual methods of assessing
breast density. These include: (i) the subjectivity of visual
measures; (ii) variations in visual density with breast
compression and X-ray exposure; (iii) consideration of
area measures of the breast, despite its three-dimensional
structure.
Due to these concerns regarding area-based measures,
recent research has been directed towards volume-based
measurements, which try to model the volume of
glandular tissue. We describe here initial results from
the Standard Mammogram Form (SMF2) technology,
version 2.2, a fully automated objective measurement
tool to estimate the volume of glandular tissue in the
breast from a mammogram [9, 10]. The SMF algorithm
explicitly considers breast compression, exposure and
tube voltage, and computes two volumetric measures of
breast density, (i) the absolute volume (cm3) of the breast
that is dense (SMF volume) and (ii) the percentage of the
volume of the breast which is dense (SMF%). SMF is
different from other volumetric research methods in that
it incorporates a full physics model rather than using
step-wedges in each image [11, 12].
Materials and methods
The data presented were obtained from women who
are part of the Glasgow Alumni Cohort, which has been
described in detail elsewhere [13]. The included women
were students who were registered at the University of
Glasgow during the period 1948–1968 and who attended
an annual medical examination at the Student Health
Service. Surviving cohort members were contacted by
postal questionnaire in 2001. Those women who
replied to the questionnaire and were living in
Scotland were asked to give consent for access to

378 The British Journal of Radiology, May 2006

Volumetric breast density using SMF
screening mammograms taken under the Scottish Breast
Screening Programme (SBSP) between 1989 and 2002.
Women were informed that their mammograms would
be digitized.
All screening mammograms for each woman were
retrieved from the eight breast screening centres and
were digitized on site with a Canon FS300 digitizer
scanner at a resolution of 100 mm with 8-bit precision by
a single radiographer. Data on exposure and tube voltage
were collected. Both the postal questionnaire survey and
the acquisition of digital mammograms received ethical
approval from the Multi-centre Research Ethics
Committee (Scotland).
For the visual assignment of density categories,
scanned images were displayed at 300 mm resolution
on a flat-panel display system. At this resolution, the
images were about the same size as a mammogram film.
All images were displayed to appear as if viewed on a
light-box. No other adjustment or image post-processing
was applied during the reading period. We have
previously reported on the similarity in density mea-
sures obtained when these assessments are made from
the digitized image compared with the original film [14].
Visual density measures were made by one radiologist
experienced in density assessment (RW) using a six-
point categorical scale of the percentage of the breast
area that appeared dense. The categories were: 0%, 1–
10%, 11–24%, 25–49%, 50–74% and >75% and are
referred to in this paper as the six category classification
(SCC). These scales were chosen to make our work
comparable with that of other researchers [15]. All
mammograms for each woman were presented consecu-
tively to the radiologist.
The volume of the mammogram that appeared dense
was estimated using SMF, v2.2. This is a computer
algorithm which models the image formation process to
decompose the breast into fatty and non-fatty tissue. It
achieves this through estimating the thickness of dense
tissue in each column of tissue between each pixel on the
projected image and the X-ray source. The image is then
standardized to remove the dependence on the para-
meters which were used to form the image. The input
parameters required are side (left or right), view
(craniocaudal (CC) or mediolateral oblique (MLO),
current time product (mAs), tube voltage (kVp) and
filter and target materials. The algorithm is fully
automated, only requiring user intervention if there has
been a data entry error, for example if a right sided
mammogram was entered as left sided, since this causes
the breast segmentation algorithm to fail. Detailed
explanations of the physics behind the model have been
published previously [9, 10].
Statistical analyses
Descriptive analyses were performed on all mammo-
grams of all women. Mann-Whitney tests were used to
test the differences in density measures obtained from
left and right mammograms, and from different views
(CC and MLO). Because of the log-normal distribution of
the data, both SMF and SMF% were log transformed
prior to analysis, and the natural log of these measures
was used in the regression models. The new variables are
The British Journal of Radiology, May 2006
referred to as ln(SMF volume) and ln(SMF%). Random
effects linear regression models were used to estimate
associations between age and SMF density measures,
taking into account the clustered nature of the data
(several mammograms per woman).
Further analysis was based on paired mammograms
(i.e. left vs right, CC vs MLO) taken on the same day. For
these, the first visit per woman was used, since it is at
this visit that both views have been routinely performed
since 1994, before two-view mammography became
routine at all visits by 2003. Because of the continuous
nature of the SMF data (both volume and percentage),
Bland-Altman plots [16] of the natural log of the SMF
measures were used instead of kappa statistics to assess
the agreement between paired data. These plot the mean
difference between pairs (expected to be zero) against the
average value (ln(SMF volume) or ln(SMF%)) of that
pair.
Results
There were 3566 women in the original Glasgow
Alumni Cohort, of whom 2169 (61%) were sent a postal
questionnaire in 2001. These were the women who could
be traced through the National Health Service Central
Register and were still alive. The response rate was 59%
(n51285). Of the respondents, 935 women (73%) were
still living in Scotland. 277 of these women (30%) had
never had a screening mammogram, and two women
refused access to their films.
The SMF algorithm was run on all 3968 mammograms
belonging to 649 of the remaining 656 women (films of
seven women were omitted inadvertently). The pro-
gramme failed on one image and produced a result
classified as ‘‘not excellent’’ for 29 (0.7%) further images.
23 (3.5%) further women (122 images) were excluded as
they reported having had breast cancer in the 2001
questionnaire. Analyses are based on the remaining 626
women with 3816 mammograms.
The median age at first breast screening was 53.6 years
(range 40.0–71.5 years). Eight women were over 65 years
at the time of their first mammogram. The median
(interquartile range (IQR)) of the absolute volume of SMF
was 66 cm3 (48–98 cm3), and of SMF% was 23.4% (18.6–
29.7%), see Figure 1.
There was a non-linear relationship between the total
volume of the breast and the volume of the breast which
was dense, see Figure 2. In women with small overall
breast volume, the absolute volume of SMF dense tissue
was low, but the percentage of the volume of the breast
which was dense was variable. In women with larger
breast volume, the absolute volume of SMF dense tissue
was variable, whereas the percentage of the volume of
the breast which was dense tended to be smaller.
Assessment of the association between the two SMF
measures and the SCC system showed no relationship
between absolute SMF volume and SCC, but a strong
positive association between SMF% and SCC (Table 1).
Each category increase in SCC was associated with a
0.04% (95%CI: 0.03–0.05) higher ln(SMF%). There were
small but significant associations between each of
the two SMF measures with age at mammography. The
regression coefficient per year older, based on the
379
 M Jeffreys, R Warren, R Highnam and G Davey Smith

Figure 1. Distribution of the volume of dense breast tissue Figure 2. Relationship between total breast volume and
in 3816 mammograms. standard mammogram form (SMF) volume/SMF% in 3816
mammograms.
ln(absolute SMF volume) was 20.008 cm3 (95%CI:
20.010 to 20.005). For ln(SMF%), the regression coeffi- Discussion
cient was 20.021% (20.023 to 20.020).
This study describes our initial experiences of using a
Comparisons of median SMF values stratified by
new method to assess mammographic breast density, in
mammogram view (CC vs MLO) indicated that the
which we estimate the proportion of the volume of the
SMF algorithm consistently produced lower SMF
breast that is dense and the absolute volume of dense
volume for CC compared with MLO views (Table 2).
tissue. This new method is fully automated, and
The CC/MLO differences for SMF% were smaller.
produced a useable density value for over 99% of
Comparing left and right sides, there was a small
images. We have shown that the percentage measure
difference in the median absolute SMF volume, being
slightly larger for left than right breasts, but no difference
for SMF%.
Table 1. Association between area percent density and
Correlations between the paired mammograms taken standard mammogram form (SMF) density in mammograms
on the same day were high. For mammograms taken on of 649 women
the same day, the left–right correlation was high, r50.92
(p,0.001) for ln(SMF volume) and r50.85 (p,0.001) for SCC n Absolute SMF SMF%
volume (cm3)
ln(SMF%). These correlations were equally high for CC
and MLO views. Bland-Altman plots (Figure 3) showed median IQR median IQR
that although left/right agreement was good for the 0% 359 69 53 to 103 16 14.1 to 18.1
majority of women, there were some women whose 1–10% 542 67 49 to 95 17.6 15.7 to 20.4
values lie well outside the mean reference range (¡2 11–24% 649 64 46 to 85 19.6 17.7 to 22.1
standard deviations of the mean difference between the 24–49% 822 69 51 to 99 23.7 21.1 to 26.8
left/right measures). The mean difference (right minus 50–74% 753 71 50 to 107 29.2 25.4 to 33.1
left) of the ln(SMF volume) was 20.027 cm3 (95% >75% 625 58 40 to 93 34.4 29.1 to 40.4
confidence interval (CI) 20.044 to 20.009) and of SCC (Six category classification) is based on a visual assess-
ln(SMF%) was 0.005% (95% confidence interval ment of the percentage of the area that appears dense.
20.008% to 0.019%). IQR (interquartile range).

380 The British Journal of Radiology, May 2006

Volumetric breast density using SMF

Table 2. Association between view, side and standard cancer in several studies [1], SMF has not yet been
mammogram form (SMF) density in mammograms of 649 validated directly against the incidence of breast cancer.
women Combining a biological understanding of breast density
with knowledge of the physics of mammogram acquisi-
View n Absolute SMF SMF %
volume (cm3) tion, we believe that this is likely to be as strong a
predictor of breast cancer as the more conventional
Median IQR Median IQR methods of mammographic density assessment. Since
CC 625 58 42–94 25.4% 20.3–32 the estimation of density using this method is entirely
MLO 3191 67 49–98 23.0% 18.3–29 objective, we expect that the non-differential misclassi-
p-value ,0.001 0.098 fication during visual estimation, which attenuates
Side associations between density and breast cancer, will not
Left 1921 67 49–98 23.3% 18.7–29.5 occur. Associations between SMF and breast cancer may
Right 1895 65 47–97 23.4% 18.6–29.8 therefore be of greater magnitude than the four to six
p-value ,0.001 0.69 fold risks reported between extreme categories of
IQR, interquartile range. visually estimated or computer-assisted methods [1].
A limitation of all currently proposed volumetric
is closely associated with frequently used visual systems, including SMF, is that they fail to acknowledge
assessments of density. We had no gold standard against that the non-fatty component consists of groups of
which to compare the absolute volume of density, but fibrous, glandular and other tissues together. It has been
the good agreement between left and right SMF suggested that the (possibly unconscious) acknowledg-
estimates provides an assurance regarding the internal ment of these structures by radiologists when assigning
consistency of the method. the visual percent density may explain the power of
Unlike the visual and semi-automated systems to these assessments in determining breast cancer risk.
assess breast density, which have been related to breast Comparison of visual area-based density measures have
shown that correlations between these and MR images
are high [17], and of similar magnitude to the associa-
tions which we report. Although correlation coefficients
are not recommended as measures of assessing agree-
ment [16, 18], we present these results as a comparison
with those reported in previous studies. For example, in
a small sample of pre-menopausal women, Pearson
correlation coefficients were between 0.86 and 0.96 [19],
very similar to the results we present. Recent data
suggest that the association between breast density and
breast cancer risk is similar in pre- and post-menopausal
women [20].
We found that the SMF algorithm appears to estimate
higher SMF% values in CC compared with MLO
mammograms, but lower SMF volume. This may be
because the MLO, but not the CC, view includes the
axillary tail, which is primarily composed of fat.
Furthermore, the MLO image captures all the tissue at
the very back of the breast, so has a higher SMF volume,
but since that tissue is often mostly fatty, the SMF% for
CC is higher.
If SMF can be proven in validated studies to be a
useful marker of breast cancer risk, its potential is
enormous. Unlike some methods of measuring volu-
metric breast density [11, 12], SMF does not require step
wedges to be included during the mammogram acquisi-
tion. We have shown that it can be used with historical
mammograms, which is a strong feature for use in
epidemiological studies. Although SMF does currently
require calibration data such as mAs and kVp to be
known, work is well underway to remove this require-
ment (RH, personal communication, 2005). Development
that would enable volume measurements of this kind
without the need for calibration data would enable wider
Figure 3. Bland-Altman plots showing agreement between
use for epidemiology in a multicentre setting.
left/right pairs of mammograms, 626 women. Note: Bland-
Altman plots of the difference against the mean of left and Despite continuing reported associations between risk
right measures, with horizontal lines showing the mean factors, breast density and breast cancer [2, 20], there is a
difference of 0 and limits of agreement (¡2 standard need for such work to be refined. First, we propose that
deviations). Both standard mammogram form (SMF) and the relative amounts of dense and non-dense tissues
SMF% have been log transformed for these analyses. should be considered as two separate outcomes, with

The British Journal of Radiology, May 2006 381


both routinely reported. This has been previously
suggested [5], but not adhered to in the majority of
published studies. This will allow better modelling of
associations between risk factors and density, without
the potential confounding influence of measures of body
fat on the results. We have found that BMI is positively
related to SMF volume [21], as would be expected if SMF
volume is thought of as a proxy marker for breast cancer
risk.
Second, investigations into the biological mechanisms
underlying associations between breast density and
breast cancer risk are needed. Some work in this area
has begun. For example, insulin-like growth factor (IGF)
and its main binding protein IGFBP-3 have both been
related to breast density [22, 23].
In summary, our findings suggest that the novel
technique of estimating the volume of dense breast
tissue, which involves modelling mammographic breast
density using a fully automated system, has the potential
to be a marker of breast cancer risk available for use in
large epidemiological studies. Work is underway to
investigate whether SMF can predict breast cancer risk.
Acknowledgments
We are grateful to the women who participated in this
study for allowing access to their mammograms, to Pat
Forrest who performed the digitization and for the
assistance of all the Scottish Breast Screening Centres.
This work was undertaken while Dr Jeffreys (née
Okasha) was employed at the University of Bristol. Dr
Highnam is employed by Mirada Solutions, where he
developed the SMF algorithm. We are grateful for the
financial support provided by Breast Cancer Campaign,
Breast Cancer Research Trust and World Cancer
Research Fund International. The Centre for Public
Health Research is supported by a programme grant
from the Health Research Council of New Zealand.
References
1. Boyd N, Lockwood G, Byng J, Tritchler D, Yaffe M.
Mammographic densities and breast cancer risk. Cancer
Epidemiol Biomarkers Prev 1998;7:1133–44.
2. Heine JJ, Malhotra P. Mammographic tissue, breast cancer
risk, serial image analysis, and digital mammography. Part
1. Tissue and related risk factors. Acad Radiol
2002;9:298–316.
3. Wolfe JN. Breast patterns as an index of risk for developing
breast cancer. AJR Am J Roentgenol 1976;126:1130–7.
4. Gram I, Funkhouser E, Tabar L. Anthropometric indices in
relation to mammographic patterns among peri-menopau-
sal women. Int J Cancer 1997;73:323–6.
5. Boyd NF, Lockwood GA, Byng JW, Little LE, Yaffe MJ,
Tritchler DL. The relationship of anthropometric measures
to radiological features of the breast in premenopausal
women. Br J Cancer 1998;78:1233–8.
6. Vachon CM, Kuni CC, Anderson K, Anderson VE, Sellers
TA. Association of mammographically defined percent
382
M Jeffreys, R Warren, R Highnam and G Davey Smith
breast density with epidemiologic risk factors for breast
cancer (United States). Cancer Causes Control
2000;11:653–62.
7. El-Bastawissi AY, White E, Mandelson MT, Taplin SH.
Reproductive and hormonal factors associated with mam-
mographic breast density by age (United States). Cancer
Causes Control 2000;11:955–63.
8. Salminen T, Hakama M, Heikkila M, Saarenmaa I.
Favorable change in mammographic parenchymal
patterns and breast cancer risk factors. Int J Cancer
1998;78:410–4.
9. Highnam R, Brady M. Mammographic image analysis, 1
edn. Kluwer Academic Publishers, 1999.
10. Marias K, Behrenbruch C, Highnam R, Parbhoo S, Seifalian
A, Brady M. A mammographic image analysis method to
detect and measure changes in breast density. Eur J Radiol
2004;52:276–82.
11. Hufton A, Astley S, Marchant T, Patel H. A method for the
quantification of dense breast tissue from digitised mam-
mograms. Proceedings of the IWDM 2004. 2005 (In press).
12. Pawluczyk O, Augustine BJ, Yaffe MJ, Rico D, Yang J,
Mawdsley GE, et al. A volumetric method for estimation of
breast density on digitized screen-film mammograms. Med
Phys 2003;30:352–64.
13. McCarron P, Davey Smith G, Okasha M, McEwen J. Life
course exposure and later disease: a follow-up study based
on medical examinations carried out in Glasgow University
(1948–68). Public Health 1999;113:265–71.
14. Jeffreys M, Warren RM, Davey Smith G, Gunnell D. Breast
density: agreement of measures from film and digital
image. Br J Radiol 2003;76:561–3.
15. Byng J, Boyd N, Fishell E, Jong R, Yaffe M. The quantitative
analysis of mammographic densities. Phys Med Biol
1994;39:1629–38.
16. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986;1:307–10.
17. Wei J, Chan HP, Helvie MA, Roubidoux MA, Sahiner B,
Hadjiiski LM, et al. Correlation between mammographic
density and volumetric fibroglandular tissue estimated on
breast MR images. Med Phys 2004;31:933–42.
18. Silman A, Macfarlane G. Epidemiological studies: a
practical guide, 2 edn. Cambridge: Cambridge University
Press, 2002.
19. Byng JW, Boyd NF, Little L, Lockwood G, Fishell E, Jong
RA, et al. Symmetry of projection in the quantitative
analysis of mammographic images. Eur J Cancer Prev
1996;5:319–27.
20. Vacek PM, Geller BM. A prospective study of breast
cancer risk using routine mammographic breast density
measurements. Cancer Epidemiol Biomarkers Prev
2004;13:715–22.
21. Jeffreys M, Warren R, Gunnell D, McCarron P, Highnam R,
Davey Smith G. Body mass index in young adulthood and
breast cancer risk (abstract). Australasian Epidemiologist
2003;10:17.
22. Guo YP, Martin LJ, Hanna W, Banerjee D, Miller N, Fishell
E, et al. Growth factors and stromal matrix proteins
associated with mammographic densities. Cancer
Epidemiol Biomarkers Prev 2001;10:243–8.
23. Maskarinec G, Williams AE, Kaaks R. A cross-sectional
investigation of breast density and insulin-like growth
factor I. Int J Cancer 2003;107:991–6.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 383–388

Occupational radiation doses in interventional cardiology: a

15-year follow-up
1,2 1 1,2 3
E VAÑO, PhD, L GONZALEZ, PhD, J M FERNANDEZ, BSc, F ALFONSO, PhD, MD and 3C MACAYA,
PhD, MD

1
Department of Radiology, Complutense University Medical School 28040 Madrid, Spain, 2San
Carlos University Hospital, Medical Physics Service and 3Cardiovascular Institute, 28040 Madrid,
Spain

ABSTRACT. This report describes occupational radiation doses of interventional

cardiologists over 15 years and assesses action undertaken to optimize radiation
protection. Personal dosimetry records of nine staff cardiologists and eight
interventional cardiology fellows were recorded using personal dosemeters worn over
and under their lead aprons. The hospital in which this study was conducted currently
performs 5000 cardiology procedures per year. The hospital has improved its facilities
since 1989, when it had two old-fashioned theatres, to include four rooms with more
advanced and safer equipment. Intensive radiation protection training was also
implemented since 1989. Initially, some individual dose values in the range of 100–
300 mSv month21, which risked exceeding some regulatory dose limits, were measured
over the lead apron. Several doses in the range of 5–11 mSv month21 were recorded
under the apron (mean510.2 mSv year21). During the last 5 years of the study, after the
implementation of the radiation protection actions and a programme of patient-dose
optimization, the mean dose under the apron was reduced to 1.2 mSv year21. Current
mean occupational doses recorded under the lead apron are 14% of those recorded
Received 11 May 2005
during 1989–1992 and those recorded over the apron are 14-fold less than those Revised 1 August 2005
recorded during 1989–1992. The regulatory dose limits and the threshold for lens Accepted 1 September
injuries might have been exceeded if radiation protection facilities had not been used 2005
systematically. The most effective actions involved in reducing the radiation risk were
DOI: 10.1259/bjr/26829723
training in radiation protection, a programme of patient-dose reduction and the
systematic use of radiation protection facilities, specifically ceiling-suspended ’ 2006 The British Institute of
protective screens. Radiology

Radiation exposure is a significant concern for inter-
ventional cardiologists (ICs) because workloads and the
complexity of procedures have increased over the past
few years without a corresponding increase in the
number of specialists [1]. Although reduced scatter
radiation in catheterization laboratories compared with
that in old X-ray system laboratories, improved radio-
logical protection facilities, and better, more inclusive
radiation protection training for ICs have substantially
reduced the risk of radiation exposure, the complexity
and number of procedures have increased. Therefore,
interventional cardiology is recognized as a high-radia-
tion-risk practice [1–3], and evaluation and follow-up of
occupational doses should be considered an important
part of quality assurance (QA) programmes.
Address correspondence to: Prof. Luciano Gonzalez.
This study was partially funded by the European Commission 5th
Framework Programme, Contract DIMOND FIGM-CT-2000-00061,
the Spanish Department for Science and Technology (project
BFI2003-09434) and the Spanish Nuclear Safety Council.
Validation of some results with TLD chips was carried out with
experimental equipment partially funded with EC FEDER
resources.
Several aspects of radiation safety in the practice of
cardiology have been addressed by the American
College of Cardiology in a consensus document [4].
The UNSCEAR 2000 report [5] states that fluoroscopic
procedures are by far the largest source of occupational
exposure in medicine. Cardiac catheterization, in parti-
cular, can represent a major source of exposure. A study
performed in the UK [6] indicated that ICs receive a
mean annual dose of 0.4 mSv, twice that received by
radiologists and many times that received by nurses and
technicians.
There are substantial differences in occupational doses
between cardiac laboratories [7–10]. This is caused by
differences in X-ray systems (old film-based systems
versus digital units) and their particular settings, levels
of training in radiation protection, frequency of use of
radiation protection facilities and personal dosemeters,
and workloads of specialists.
Renaud et al [11] described a 5-year follow-up of the
radiation doses received by the in-room personnel of
three cardiac catheterization laboratories and concluded
that some workers may have exceeded the occupational
limit for the lens of the eye. Lens injuries have been

The British Journal of Radiology, May 2006 383


reported for several interventional radiology suites in
which radiation protection conditions were not appro-
priate to the level of risk [12].
This report describes occupational radiation doses
from interventional cardiology in a university hospital
over a period of 15 years and the actions that were taken
to optimize radiation protection. Data were gathered
from a dosemeter worn on the trunk of the body under
the apron and a dosemeter worn outside the apron, as
recommended by the International Commission on
Radiological Protection (ICRP) [1].
Methods and materials
Follow-up of IC’s personal dosimetry records was
performed in a university hospital currently performing
more than 5000 procedures per year in four catheteriza-
tion laboratories with nine staff cardiologists and eight
fellows. In 1989, this interventional cardiology service
used two old-fashioned X-ray units. In 1994, a Philips
Optimus M-200 Poly C X-ray unit (Philips, Best, The
Netherlands), installed in 1988, was upgraded and an old
CGR unit was exchanged for a Philips Integris HM-3000.
In 2000, two new Philips Integris H-5000 units were
installed. All systems now have protective screens
suspended from the ceiling. This radiation protection
tool, which had previously not been installed in one of
the rooms, was not used regularly by all specialists until
they were made aware of its importance. In addition,
lead aprons, thyroid protectors and lead glasses were
also available and are used routinely at present (with a
few exceptions).
Two personal dosemeters with thermoluminescent
dosimetry chips, as recommended by the radiation
protection service of the hospital, were used for occupa-
tional dosimetry: one was worn on the trunk of the body
under the apron and the other was worn outside the
apron at the level of the collar or the left shoulder. A
dosemeter under the apron provides an estimate of the
dose to the organs of the shielded region. A dosemeter
worn outside the apron supplies an estimate of the dose
to the organs of the head and neck, including the thyroid
and lenses of the eyes (if unshielded), but greatly
overestimates the doses to organs of the trunk. Results
obtained from both dosemeters were used to estimate the
occupational effective dose as recommended by the
NCRP [13] and ICRP [1]. Dosemeters were read monthly
by a public dosimetry service accredited and audited by
the National Regulatory Authority.
Before 1992, training in radiation protection for ICs was
scant, if performed at all. Subsequently, a radiation
protection training programme was initiated in accordance
with national regulations [14]. Of the staff cardiologists
working in the centre, 90% attended the courses and were
accredited in radiation protection, as required by the
National Regulatory Authority. Some new cardiologists,
especially fellows, did not attend the courses. New
regulations in force since 1999 [15] require a second level
of radiation protection training for interventionalists,
which includes training in radiation protection of patients
and QA, as recommended by the ICRP [1]. Training in
radiation protection of patients is also required by
European Directive 43/97/EURATOM [16].
384
E Vaño, L Gonzalez, J M Fernandez et al
An interactive CD-ROM, co-sponsored by the
European Commission [17], is used to provide radiation
protection training for residents and fellows who
commence work in interventional suites during the
intervals between radiation-protection training courses.
A copy of this CD-ROM is given to all new doctors on
commencement of duty at the hospital’s interventional
cardiology service. In addition, refresher sessions on
radiation protection are presented periodically.
Detailed analysis of personal dosimetry records of IC
personnel is conducted every month. This is followed by
individual interviews with persons exposed to monthly
doses greater than 1.0 mSv under the apron (1/20 of the
annual effective dose limit) or greater than 7.5 mSv over
the apron (1/20 of the annual lens dose limit). In addition,
a progressive audit programme was implemented to
detect high patient doses, facilitate clinical follow-up in
cases of likely skin radiation injury and to implement
corrective action when necessary. Since 1999, a national
standard [15] stipulates that patient doses in interventional
procedures must be estimated and recorded. Because this
patient-dose audit has reduced patient doses, occupational
doses have also been reduced [2].
Personal dosimetry services typically provide monthly
estimates of Hp(10) (the dose equivalent in soft tissue at
10 mm depth), which is usually compared with the
annual limit of effective dose and with the eye lens limit
[18], and Hp(0.07) (the dose equivalent in soft tissue at
0.07 mm depth) [18]. Usually, no significant differences
between values are found in cardiac catheterization
suites. The values reported in this paper are for estimates
of Hp(10) obtained from personal dosimetry readings.
The effective dose, E, can be estimated [13] from the
dosemeter values for Hw (under the apron at the waist,
although this position is not critical) and Hn (above the
apron at the neck) from the equation:
E~0:5Hw z0:025Hn
NCRP report 122 [13] contains specific recommenda-
tions for calculating the effective dose when protective
aprons are worn during diagnostic and interventional
medical procedures involving fluoroscopy. In addition to
the above formula, it states that the effective dose can be
estimated as Hn/21 if only one dosemeter is worn on the
neck outside the apron.
Results
The data from occupational dosimetry were allocated
to one of three periods for purposes of analysis.
First period (1989–1992): investigation of high dose
values and implementation of a customized
radiation protection programme
Table 1 shows the findings from this period. Most
values were in the range of 100–300 mSv month21, but in
one case a dose of 1600 mSv month21 was recorded by
the left shoulder dosemeter outside the lead apron.
Values in the range of 5–11 mSv month21 were recorded
The British Journal of Radiology, May 2006
Occupational radiation doses in IC
Table 1. Individual monthly high values of personal dose equivalent Hp(10) and total Hp(10) values under apron during the year
(except for cases indicated in the footnotes). Capital letters and numbers in the staff column are an internal code allowing
traceability of the reported data
Staff member Year
I1 Senior cardiologist 1989
B1 Senior cardiologist 1989
F1 Senior cardiologist 1989
G1 Senior cardiologist 1990
R1 Senior cardiologist 1990
G1 Senior cardiologist 1991
G1 Senior cardiologist 1992
A1 Senior cardiologist 1992
F2 Senior cardiologist 1992
R1 Fellow cardiologist 1992
B1 Fellow cardiologist 1992
C1 Resident 1992
a
Values over apron not available. Incorrect use of the dosemeter cannot be excluded.
b
Value under apron during the month receiving 1640 mSv over the apron. 47 mSv are 2.9% of the dose over the apron.
Abnormal irradiation of the over-apron dosemeter was not demonstrated.
c
Only some months.
under the apron. An initial complete evaluation of the
radiation protection conditions of the catheterization
laboratories was done, after which follow-up of abnor-
mal values was investigated and corrective actions
proposed. Consequently, the occupational medical ser-
vice of the hospital advised some staff to abstain from
catheterization duties for several months. The National
Regulatory Authority was informed of these actions.
Lens injuries would have occurred in those situations if
the corrective actions had not been put into practice
immediately.
Second period (1993–1998): consolidation of the
radiation protection programme
Training courses in radiation protection and seminars
with ICs (including fellows) were commenced, new X-
ray systems with radiation protection facilities were
installed, and a formal programme of quality control
(QC) and strategies to reduce patient and staff doses
were launched. Maximum monthly dose values (over the
apron) ranged from 7 mSv to 10 mSv, with the exception
of a new fellow, for whom high readings of up to
28 mSv month21 were recorded on two occasions. The
highest yearly Hp(10) values under the apron were
between 2 mSv and 3 mSv.
Third period (1999–2004): implementation of
occupational radiation protection in the QA
programme
During this period, the frequency of the X-ray system
QC programme increased from once yearly to two or
three times per year. The old CGR X-ray system was
removed in 1999. Full characterization was done by
measuring patient entrance dose, image quality and
scatter radiation levels for all fluoroscopy and cine
modes. Closer contact with the maintenance engineers
was established to customize the operation modes to
fulfil the image quality requirements of the cardiologists
while keeping doses as low as possible. Since this
The British Journal of Radiology, May 2006
Max. mSv/month Total Hp(10) (under apron) (mSv)
51 (over apron) 7.4
8 (under apron) 27.8a
4.6 (under apron) 12.3
62 (over apron) 5.2
65 (over apron) 9.2
346 (over apron) 27
180 (over apron) 4.2
155 (over apron) 7.1
54 (over apron) 23.7
1640 (over apron) 47b
185 (over apron) 3.1c
179 (over apron) 11
initiative, patient dose values were measured, recorded
in a database and analysed periodically.
Since 2000, the MARTIR training CD-ROM [17] has been
distributed to new personnel joining the interventional
cardiology service, and radiation protection refresher
seminars are held two or three times per year. Individual
real-time occupational dosimetry has also been implemen-
ted for some procedures. Electronic dosemeters (Unfors
EED-30; www.unfors.se) measure the dose accumulated
by the specialist throughout a procedure and the max-
imum dose rate, which provides information about the
correct use of the protective screen.
Maximum values recorded by dosemeters placed over
the apron were lower during the third period than
during the second period and ranged between
3 mSv month21 and 4 mSv month21. The maximum
dose under the apron was generally 2 mSv month21, but
some abnormally high values were recorded for specia-
lists doing electrophysiology cardiac procedures (in
service since 2000). A maximum over-the-apron dose of
26 mSv month21 was recorded for one specialist.
The workloads during the three periods were similar:
five to six procedures per day and room, shared between
one to three cardiologists. Some of the fellows stayed at
the hospital for short periods and often performed many
procedures per day to improve their skills. Typical
workloads were two to four procedures per day for staff
and three to six procedures per day for fellows. Table 2
shows monthly doses before, during and after radiation
protection training. Mean and median doses decreased
significantly after the training courses.
Unpaired t-test analysis revealed statistically signifi-
cant differences between means for doses before and
after the training periods. In two-tailed tests, p-values
were less than 0.05 (p50.01 for 1996 vs 1991; p50.02 for
1995 vs 1992). Table 3 presents the annual dose values for
the three periods. Only personal dosimetry records
comprising all the monthly dose values were used.
Data in which background dose values were recorded by
the over-apron dosemeter of specialists who had a
substantial workload were excluded from the analysis,
as this indicated that the personal dosemeter had not
been used. Between 20% and 30% of the cardiologists
385

Table 2. Relevant dose values (in mSv/month) under the lead apron, before, during and after the training courses on RP for
numbers of IC specialists indicated
Year Sample Range
1991 8 1.9–26.5
1992 11 0.9–24.2
1993 (training) 7 1.0–4.4
1994 (training) 12 0.6–13.0
1995 10 0.7–4.1
1996 13 0.4–5.8
neglected to send their personal dosemeters to the
dosimetry service for processing every month.
Differences between under-apron doses during 1989–
1992 and the other two periods were statistically
significant (p,0.01), and a more significant difference
was noted for values over the apron (p,0.004).
Table 4 presents estimates of the transmitted fraction
of energy across different lead aprons with thickness
equivalents in the range of 0.25–0.5 mm lead. The IPEM
software application [19] for spectra from 70 kVp to
90 kVp was used for calculations.
The real spectra of scattered radiation in the catheteri-
zation rooms are difficult to determine. However, the X-
ray beam used for interventional cardiology in our
laboratories (with the Philips Integris systems) typically
ranges between 80 kVp and 110 kVp. Thus, the energy
degradation in the scattering process would yield dose-
transmitted fractions of between 3.3% and 8.3%
for 0.25 mm lead aprons, between 1.5% and 4.9% for
0.35 mm lead aprons, and between 0.5% and 2.4% for
0.5 mm lead. Thus, a dose under the apron of between
0.5% and 8.3% of the values measured over the apron
was considered compatible with the personal protection
used and was regarded as a good indicator of proper use
of personal dosemeters. The same criterion has been
used to reject unreliable data, and the values from
Tables 2 and 3 are fully compatible.
In summary, the radiation protection programme
during the 15-year period reduced the effective dose to
cardiologists by one order of magnitude, avoiding cases
of high individual doses. The real mean effective dose for
cardiologists in our centre during the last 4 years of our
study was 1.2 mSv year21, which is compatible with
results recently reported by Delichas et al [20] (1.2–
2.7 mSv/procedure, a dose of 0.7–1.5 mSv year21 for a
workload of 50 procedures per month).
Discussion
Several questions arise from the results presented in this
paper. First, it should be determined if the high doses
Table 3. Mean values (and standard deviation) in mSv/year of occupational doses of cardiologists during the periods referred
to. The percentage of dose under apron in relation to the dose over apron is indicated in the Hp(10) ‘‘under apron’’ column
between brackets
Number of Period Hp(10)
reliable data
over apron under apron
15 1989–1992 259¡249 10.2¡8.6 (3.9%)
24 1993–1998 31¡15 1.7¡1.1 (5.5%)
11 1999–2004 18¡7 1.4¡0.4 (7.7%)
386
E Vaño, L Gonzalez, J M Fernandez et al
Mean¡SD Median
9.0¡9.3 5.1
7.4¡8.5 3.7
1.9¡1.0 1.6
3.0¡3.3 1.6
1.8¡1.2 1.3
1.5¡1.6 0.9
reported during the years 1989–1992 are real dose values
or incidental readings caused by inappropriate use of the
dosemeters. In fact, the bulk of the results in Table 1 should
correspond to real dose values received by the cardiolo-
gists during a period in which there was no culture of
safety: ceiling-suspended screens were absent or unused,
the X-ray systems were used in relatively high-dose
fluoroscopy modes and film cine acquisition was done at
25 frames s21. The high dose values shown in Table 1
cannot be considered a consequence of the incorrect use of
the dosemeters. All abnormal doses were reported to the
doctors wearing the dosemeters and investigated with
them, and no reason was found to suggest that incidental
dosemeter irradiation occurred.
For the 1640 mSv measured at the left shoulder of a
visiting cardiologist in 1 month, it was not possible to
prove any abnormal dosemeter irradiation. The dose
measured by the dosemeter worn under the apron was
42 mSv in that month, 2.8% of the dose over the apron.
This figure is compatible with the transmitted fraction
across the lead apron (Table 4). Moreover, experimental
measurements in one of the cardiology rooms used by a
fellow simulating clinical conditions produced doses in
good agreement with the dosemeter readings, taking into
account the presumed work rate, fluoroscopy time and
frame rate per procedure, and the mean scatter dose rate
for a non-pulsed fluoroscopy mode.
Distance is an important factor that could increase (or
decrease) the scatter dose rate. A distance of 65 cm
between the cardiologist and the isocentre has been
supposed, but a variation of 15 cm nearer to the patient
could increase the occupational dose by 70%. In addition,
considering that the protective screens—typically
equivalent to a shielding of 0.5–1.0 mm lead—can reduce
the scatter dose by a factor of 100 if properly used,
differences in the reported occupational doses in the
scientific literature of two orders of magnitude measured
over the lead apron are not surprising. In fact, Pratt and
Shaw showed that the relationships between the cardio-
logist’s eye dose and factors such as the dose efficiency of
the X-ray equipment, scattered-dose rates, examination
Effective dose Effective dose (NCRP,
(NCRP, using two using over-apron
dosemeters) dosemeter)
11.6 12.3
1.6 1.5
1.2 0.86
The British Journal of Radiology, May 2006
Occupational radiation doses in IC
Table 4. Protection of different lead aprons for X-ray beams
filtered with 3 mm Al and generated at the kVp indicated
kVp Protective apron Fraction of energy
mm lead equivalent transmitted (%)
90 0.25 8.3
90 0.35 4.9
90 0.50 2.4
80 0.25 5.7
80 0.35 3.0
80 0.50 1.3
70 0.25 3.3
70 0.35 1.5
70 0.50 0.5
protocols and workload are complex and vary from
centre to centre [21].
Data considered reliable are scarce in Tables 1 and 2
because between 1989 and 1996, a significant number of
cardiologists did not use the personal dosemeters during
all procedures or overlooked the established procedure
of sending the dosemeters to the medical physics service
monthly. Compliance with the radiation badge policies is
one of the main problems in many interventional
cardiology services. Reported occupational dose values
are often surprisingly low and the reason is not a high
level of radiation protection, but a lack of use of personal
dosemeters. McCormick et al [10] reported that after a
mandatory radiation protection training programme,
compliance with the radiation badge policy was only
36% in 1999, reaching 67% in 2000 and 77% in 2001 for
physicians and nurse clinicians. Therefore, confidence in
the mean dose values determined by the regional
dosimetry services, and sometimes by the regulatory
bodies, to assign occupational doses is open to discus-
sion, as stated by UNSCEAR [5].
Another important point for improving the occupa-
tional dosemetry data is reporting dose results from
dosemeters over and under the lead apron [1, 13] and
combining them to calculate a more realistic effective
dose. The over-apron dosemeter provides very useful
information on the risk of lens injuries in interventional
suites. The data in Table 3 show that differences up to
15% with the conventional under-apron dose approach
can be found when considering the proposed NCRP
formula [13].
Finally, one may wonder if occupational dose values
as high as those reported for the period 1989–1993 could
be reached with new X-ray systems and radiation
protection facilities. Fortunately, the probability is low.
Modern interventional cardiology X-ray systems have
significantly decreased the radiation level for the patient
and, as a consequence, the scatter radiation level. In
addition, radiation protection facilities [8, 22], especially
ceiling-suspended screens, are in common use, access to
advice of medical physics experts is more frequent and
dose parameters are included in QA programmes,
among other positive factors. However, the workload
of cardiologists and the complexity of procedures are
increasing, and thus vigilance should be maintained. In
the new X-ray laboratories, high levels of scattered dose
are still measured and sometimes a surprising lack of
training in radiation protection is the cause of avoidable
and unjustified occupational over-irradiations.
The British Journal of Radiology, May 2006
Conclusions
Occupational doses measured on specialists who are
routinely using their personal dosemeters show that the
radiation protection level has significantly improved in
the last decade. A reduction in the effective dose by a
factor of 10 has been achieved. The most successful
action to reduce occupational doses has been training in
radiation protection. The use of ceiling-suspended
protective screens in a systematic way by the cardiolo-
gists and the programme of patient dose reduction were
important complementary actions. New X-ray equip-
ment also contributed to further dose reductions, but its
relative impact cannot be distinguished from the training
effect in this study because of their interdependence.
Another significant conclusion is that mean values of the
occupational doses in catheterization laboratories could
provide an incorrect estimate of the real radiological risk
if some specialists are not using their personal dose-
meters on a regular basis.
Acknowledgments
The authors thank Mercedes Lago for her help in
gathering dosimetry data.
References
1. ICRP Publication 85. Avoidance of radiation injuries from
medical interventional procedures. Ann ICRP. Oxford, UK:
Pergamon, Elsevier Science Ltd, 2000;30(2).
2. Vano E. Radiation exposure to cardiologists: how it could
be reduced. Heart 2003;89:1123–4.
3. Finkelstein MM. Is brain cancer an occupational disease of
cardiologists? Can J Cardiol 1998;14:1385–8.
4. American College of Cardiology. Radiation safety in the
practice of cardiology. ACC expert consensus document.
J Am Coll Cardiol 1998;31:892–913.
5. United Nations Scientific Committee on the Effects of
Atomic Radiation. Sources and effects of ionizing radiation.
UNSCEAR 2000 report to the General Assembly, with
scientific annexes. Annex E. Occupational radiation expo-
sures. (Available at www.unscear.org). New York: United
Nations, 2000.
6. Hughes JS, O’Riordan MC. Radiation exposure of the UK
population—1993 review. NRPB-R263 (1993). Quoted in
UNSCEAR 2000 report.
7. Tsapaki V, Kottou S, Vano E, Komppa T, Padovani R,
Dowling A, et al. Occupational dose constraints in inter-
ventional cardiology procedures: the DIMOND approach.
Phys Med Biol 2004;49:997–1005.
8. Kuon E, Schmitt M, Dahm JB. Significant reduction of
radiation exposure to operator and staff during cardiac
interventions by analysis of radiation leakage and
improved lead shielding. Am J Cardiol 2002;89:44–9.
9. Vano E, Gonzalez L, Guibelalde E, Fernandez JM, Ten JI.
Radiation exposure to medical staff in interventional and
cardiac radiology. Br J Radiol 1998;71:954–60.
10. McCormick VA, Schultz CC, Hollingsworth-Schuler V,
Campbell JM, O’Neill WW, Ramos R. Reducing radiation
dose in the cardiac catheterization laboratory by design
alterations and staff education. Am J Cardiol 2002;90:903–5.
11. Renaud L. A 5-year follow-up of the radiation exposure to
in-room personnel during cardiac catheterization. Health
Physics 1992;62:10–5.
12. Vano E, Gonzalez L, Beneytez F, Moreno F. Lens injuries
induced by occupational exposure in non-optimized
387

interventional radiology laboratories. Br J Radiol
1998;71:728–33.
13. NCRP report 122. Use of personal monitors to estimate
effective dose equivalent and effective dose to workers for
external exposure to low-LET radiation. Bethesda, MD:
NRCP, 1995.
14. Royal Decree 1891/1991, from the Industry and Energy
Department, on installation and utilisation of X-ray devices
for medical diagnostic purposes. In: Spanish State Official
Bulletin January 3rd 1992: 138–48 [In Spanish].
15. Royal Decree 1976/1999, from the Health and Consumer
Affairs Department, establishing quality criteria in radio-
diagnostics. In: Spanish State Official Bulletin of January
29th 1999: 45891–900 [In Spanish].
16. Council Directive 97/43 Euratom, on health protection of
individuals against the dangers of ionizing radiation in
relation to medical exposure, and repealing Directive 84/
466 Euratom. Official Journal of the European Communities
No L 180, 9th July 1997:22–27.
17. MARTIR (Multimedia and Audiovisual Radiation
Protection Training in Interventional Radiology). CD-ROM.
Radiation Protection 119, European Commission. Directorate
388
E Vaño, L Gonzalez, J M Fernandez et al
General Environment, Nuclear Safety and Civil Protection.
Luxembourg, 2002 (available for free from the Publication
Department of the European Commission, Office for Official
Publications of the European Communities, Luxembourg,
e-mail: tren-radprot@cec.eu.int).
18. ICRU report 51. Quantities and units in radiation protection
dosimetry. Bethesda, MD: International Commission on
Radiation Units and Measurements, 1993.
19. IPEM report 78. The Institute of Physics and Engineering in
Medicine. Catalogue of Diagnostic X-ray Spectra and other
data. Cranley K, Gilmore BJ, Fogarty GWA, Desponds L.
Software prepared by Reilly AJ and Sutton D. 1997.
20. Delichas M, Psarrakos K, Molyvda-Athanassopoulou E,
Giannoglou G, Sioundas A, Hatziioannou K, et al. Radiation
exposure to cardiologists performing interventional cardio-
logy procedures. Eur J Radiol 2003;48:268–73.
21. Pratt TA, Shaw AJ. Factors affecting the radiation dose to
the lens of the eye during cardiac catheterization proce-
dures. Br J Radiol 1993;66:346–50.
22. Kuon E, Birkel J, Schmitt M, Dahm JB. Radiation exposure
benefit of a lead cap in invasive cardiology. Heart
2003;89:1205–10.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 389–400

A four-dimensional computer simulation model of the in vivo

response to radiotherapy of glioblastoma multiforme: studies on
the effect of clonogenic cell density
1 1 1
G S STAMATAKOS, PhD, V P ANTIPAS, PhD, N K UZUNOGLU, PhD and 2R G DALE, PhD

1
In Silico Oncology Group, Microwave and Fibre Optics Laboratory, School of Electrical and
Computer Engineering, National Technical University of Athens, 9 Iroon Polytechniou St., GR 157 80
Greece and 2Department of Radiation Physics and Radiobiology, Hammersmith Hospitals NHS Trust
and Faculty of Medicine, Imperial College, Charing Cross Hospital, Fulham Palace Road, London W6
8RF, UK

ABSTRACT. Tumours behave as complex, self-organizing, opportunistic dynamic

systems. In an attempt to better understand and describe the highly complicated
tumour behaviour, a novel four-dimensional simulation model of in vivo tumour
growth and response to radiotherapy has been developed. This paper presents the
latest improvements to the model as well as a parametric validation of it. Improvements
include an advanced algorithm leading to conformal tumour shrinkage, a quantitative
consideration of the influence of oxygenation on radiosensitivity and a more realistic,
imaging based description of the neovasculature distribution. The tumours selected for
the validation of the model are a wild type and a mutated p53 gene glioblastomas
multiforme. According to the model predictions, a whole tumour with larger cell cycle
duration tends to repopulate more slowly. A lower oxygen enhancement ratio value
leads to a more radiosensitive whole tumour. Higher clonogenic cell density (CCD)
produces a higher number of proliferating tumour cells and, therefore, a more difficult
tumour to treat. Simulation predictions agree at least semi-quantitatively with clinical
experience, and particularly with the outcome of the Radiation Therapy Oncology
Group (RTOG) Study 83-02. It is stressed that the model allows a quantitative study of Received 7 June 2004
the interrelationship between the competing influences in a complex, dynamic tumour Revised 8 August 2005
environment. Therefore, the model can already be useful as an educational tool with Accepted 30 August 2005
which to study, understand and demonstrate the role of various parameters in tumour
DOI: 10.1259/bjr/30604050
growth and response to irradiation. A long term quantitative clinical adaptation and
validation of the model aiming at its integration into the treatment planning ’ 2006 The British Institute of
procedure is in progress. Radiology

The rapid growth and resilience of tumours make it
difficult to believe that they behave as random, dis-
organized and diffuse cell masses and suggests instead
that they are emerging, opportunistic systems [1, 2]. If
this hypothesis holds true, the growing tumour and not
only the single cell [3] must be investigated and treated
as a self-organizing complex dynamic system. This
cannot be done with currently available simple in vitro/
in vivo models or common mathematical approaches.
Therefore, there is a need for novel computational
models to simulate the complexity of solid tumour
growth and invasion, combining a range of disciplines
including medical, biological, engineering and statistical
physics research [1, 2]. Recent research efforts have
focused on modelling of tumour response to various
therapeutic modalities but, unfortunately, current mod-
els of dynamic processes need substantial improvement
due to the complexity of the problem and the paucity of
large series’ of clinical data.
This paragraph provides a brief outline to several of
the concepts and earlier research efforts. Duechting et al
[4, 5] introduced a simulation model which concerns
only the in vitro case or the early avascular stages of
small in vivo tumours and is based on a consideration of
the distinct phases of the cell cycle. Kocher et al [6, 7]
developed a simulation model of the development of a
tumour spheroid and its response to radiosurgery.
However the detailed geometry of the clinical tumour
as depicted by imaging data has not been considered in
this model. Instead, an equivalent spherical tumour is
considered in place of the generally arbitrarily shaped
actual tumour. Additionally, detailed cell cycle phase
biology (phases G1, S, G2, M) has not been taken into
account, with grouping of the cells into only proliferating
and dormant classes being considered instead. None of
the above mentioned models have been applied to large
clinical tumours of varied geometrical shapes, and none
of them simulates conformal shrinkage for an arbitrarily
shaped clinical tumour undergoing treatment. In the
tumour growth models presented by Kansal et al [1, 2], a
discretising grid is used in which each geometrical cell is
able to contain a large number of biological cells, but the

The British Journal of Radiology, May 2006 389


grid has not been used to discretise clinical tumours of
arbitrary shape. Neither is the response of the tumour to
irradiation addressed in this model. Swanson et al [8, 10]
and Mandonnet et al [9] have developed clinically
significant spatiotemporally models of tumour growth
and invasion concerning glioblastoma multiforme
(GBM). Nevertheless, although growth and invasion
constitute fundamental phenomena related to GBM
treatment optimization, the investigators have not
focused on the radiobiological mechanisms underlying
radiotherapy. Byrne et al and Alarcon et al [11–14] have
developed mathematical models of avascular tumour
growth and angiogenesis evolution pertinent mainly to
the initial stages of tumour development. Although
valuable insight can be gained using such models,
extension to clinical voluminous tumours is not an a
priori manageable task.
An effort to overcome these shortcomings has pre-
viously been made by our group through the develop-
ment of a four-dimensional patient-specific in vivo
simulation model [15–17]. All parameters used in the
model have already been defined and can be determined
(in principle) experimentally or clinically. Therefore, use
of new mathematically dictated parameters of ambig-
uous physical meaning is avoided. Furthermore, the
authors believe that the development of an experimental
and clinical biology based model provides substantial
insight into the interdependence of the mechanisms
involved, even if some parameters cannot currently be
accurately quantified for individual patients.
This work presents the latest advances and improve-
ments of a four-dimensional, patient individualized, in
vivo simulation model of tumour response to radio-
therapy and discusses attempts at parametric and clinical
validation. To this end, the outcome of a pertinent
clinical study is exploited. It is also noted that an
eventual combination of the model with the approaches
of the previously mentioned research groups might be
worth a careful investigation. In such a case, a more
scientifically complete and clinically meaningful picture
of the various aspects of tumour modelling might
emerge. It is suggested that the work presented might
also be considered a step towards shaping an emerging
analytical-computational discipline of ‘‘In Silico
Radiation Oncology’’.
Tumour cell distribution
In the following, a brief outline of the model’s
construction is given.
1. A discretizing cubic mesh is superimposed upon a
three-dimensional reconstruction of the tumour,
including its necrotic region and the surrounding
anatomical features, based on the imaging data.
2. Within each geometrical cell of the mesh, a number of
equivalence classes (compartments) of the contained
biological cells are defined based on their distribution
over the various phases within or out of the cell cycle.
Sufficient registers are used in order to store the
current state of each equivalence class (e.g. time spent
in phase G1, etc.).
3. The mesh is scanned every hour.
390
G S Stamatakos, V P Antipas, N K Uzunoglu and R G Dale
4. The basic biological ‘‘laws’’ (metabolic activity,
cell cycling, mechanical restrictions, cell survival
probability following irradiation with dose D, etc.)
are applied on each geometrical cell at each scanning.
5. A spatial and functional restructuring of the tumour
may take place during each scanning as e.g. new cells
may be produced (leading to differential tumour
growth) or as existing cells may disappear (leading to
differential tumour shrinkage).
6. The simulation predictions can be two- or three-
dimensionally visualized at any simulated instant of
interest.
The basic philosophy of our approach can also be found
(http://www.in-silico-oncology.iccs.ntua.gr/) at [17].
Recent improvements
Vessel network, neoangiogenesis and oxygen
supply based on imaging data
The observation that angiogenesis occurs around
tumours was made nearly 100 years ago [18–21]. The
hypothesis that tumours produce a diffusible ‘‘angio-
genic’’ substance was put forward in 1968 [22, 23].
Mammalian cells require oxygen and nutrients for their
survival, and functional cells must therefore be located
within a distance of 100–200 mm from the nearest capillary
blood vessels, which is the diffusion limit for oxygen.
Vessels in an embryo are assembled from endothelial
precursors and this primitive network subsequently
expands by sprouting (angiogenesis) or intussusception,
in which interstitial tissue columns are inserted into the
lumen of pre-existing vessels and partition the vessel
lumen [24]. Tumour vessels develop by sprouting or
intussusception from pre-existing vessels. Circulating
endothelial precursors, shed from the vessel wall or
mobilized from the bone marrow, can also contribute to
tumour angiogenesis [25, 26]. Tumour cells can also grow
around an existing vessel to form a perivascular cuff.
Without blood vessels, tumours cannot grow beyond a
critical size or metastasise to another organ [27].
In contrast to normal vessels, tumour vasculature has
the following distinct characteristics:
1. Tumour vessel ultrastructure is abnormal [27]
2. Tumour vessel ultrastructure is highly disorganized
[27]
3. Tumour vessels are tortuous and dilated, with uneven
diameter, excessive branching and shunts [27]
4. Tumour blood flow is chaotic and variable [27, 28],
and leads to hypoxic and acidic regions in tumours
[27, 29]
5. Vessel walls have numerous ‘‘openings’’, widened
interendothelial junctions, and a discontinuous or
absent basement membrane [27]
6. Tumour vessels are ‘‘leaky’’ and have tremendous
heterogeneity in leakiness over space and time [27, 30,
31]
Definition of the imaging based tumour layers dictates
the number and the metabolic state of the individual
The British Journal of Radiology, May 2006
Simulation of in vivo glioblastoma multiforme response to radiotherapy
biological cells included within each layer. During the
simulation process, and in the case of tumour growth,
the normal tissue capillaries are shifted away and
tumour capillaries are generated in their place [27, 32].
Consequently, the new tumour cells are assumed to be
sufficiently oxygenated and able to divide.
A ‘‘proliferating layer’’ is assumed to exist between the
external surface of the gross tumour volume and a
hypothetical surface (HYP) enclosing its necrotic kernel
and lying 1.5 mm further out. The tumour volume
contained between HYP and the surface of the necrotic
region has been assumed to contain large numbers of
dormant G0 cells; therefore this is called the ‘‘G0 cell
layer’’. This layer contains a substantial number of
dormant cells around the necrotic area of the tumour
as it appears on the imaging data. We have assumed that
the clonogenic cell density (CCD) in the ‘‘proliferating
layer’’ is two times the CCD in the G0 cell layer. CCD in
the necrotic or dead cell layer of the tumour has also
been assumed to be one fifth of the CCD in the
proliferating cell layer.
Consideration of oxic and hypoxic cells
The resistance of cells that are hypoxic at the time of
therapy will influence the efficacy of treatment with
radiation, chemotherapy and combined modality regi-
mens. Tumour cell response to ionizing radiation is
strongly dependent upon oxygen, any given dose killing
substantially fewer hypoxic than oxic cells. The radiation
dose that allows a particular level of survival tends to
increase by the same factor at all levels of survival when
oxygen is removed. This allows calculation of oxygen
enhancement ratio (OER) for the same level of biological
effect. For most cells, OER for X-rays is around 3.0 [6, 7,
33]. Some researchers [33–35] report that OER reduces
for radiation doses to 3.0 Gy or less. In practical terms,
within a tumour microenvironment the oxic cells are
those which proliferate, whereas the hypoxic cells are
dormant or G0 cells.
Extensive work has been done to measure hypoxia in
human brain tumours (especially for gliomas) [36–41].
Different LQ a and b parameter values for the oxic (G1, S,
G2, M) and hypoxic (G0) cells are considered. The
interrelation between the hypoxic and oxic LQ para-
meters is given by the following expressions [7]:
aoxic
ahypoxic ~ ð1Þ
OER
boxic
bhypoxic ~ ð2Þ
OER2
Nygren and Ahnstrom [42] suggest that OER can
range from 2.0 to 3.0, and Palcic et al [34], Stuschke et al
[43] and Speke and Hill [44] reported a value for OER of
2.3, 2.7, 2.75, respectively.
It should be noted that hypoxic cells in the clinical setting
can become oxic again when either new microvasculature
vessels have emerged in their vicinity, or when the space
between them and the nearest blood vessels has been
cleared of other cells already killed by irradiation.
The British Journal of Radiology, May 2006
An advanced algorithm leading to conformal
tumour shrinkage
The tumour shrinkage process usually tends to behave
as a conformal contraction [45]. The biological rationale
for the cells to be ‘‘pulled’’ towards the centre-of-mass of
the tumour is that the surrounding normal tissues exert a
rather uniform pressure upon the tumour in such a way
that the brain tends to recover its (physiological) normal
shape (homeostasis). Nevertheless, deviations of this rule
due to local inhomogeneities are to be expected.
In order to satisfactorily simulate this process in
conjunction with the rest of the simulation strategy, the
‘‘centre-of-mass algorithm’’ (CMA) is introduced. Its
primary mission is to ‘‘pull’’ individual tumour cells
towards the centre-of-mass of the entire tumour. It is
emphasised that the term ‘‘centre-of-mass’’ is not used in
a strict context as it refers to a tumour of uniform mass
density. The model does not consider the eventual
appearance and behaviour of new tumour foci out of
the space occupied by the primary tumour due to
infiltration of adjacent tissues.
Any geometrical cell which is in two-dimensional side
contact with at least one other geometrical cell of the
main tumour mass is considered to belong to the main
cohesive tumour mass. Let us suppose that the distance
from the centre of an isolated geometrical cell C2 to the
nearest geometrical cell occupied by the main tumour
mass is larger than one geometrical cell. The content of
the isolated geometrical cell tends to move towards the
centre of the main cohesive tumour mass. Only shifts
along the x+ or x- or y- or y+ or z- or z+ directions are
allowed. During the application of CMA, the distance
from the centre of C2 to the centre of the main tumour
along each of the six possible directions is calculated.
Subsequently, the least distance is selected. In cases
where more than one distance has the same least value,
the direction is selected randomly. After completion of
the next discretization mesh scan, C2 will be closer to the
tumour centre-of-mass and after a number of scans, C2
will be connected with the nearest cell of the main
tumour mass.
A more realistic approach to tumour cell
distribution
A random number generator is used to produce a
uniform cell distribution over the individual time units
constituting each phase within or out of the cell cycle (e.g.
S, G0), this being a more realistic approach. This is
because cells within the same cell cycle phase are not
considered to be synchronized during the initialization
process.
Parameter value selection: the paradigm of in
vivo glioblastoma multiforme
The selected paradigm is a recently irradiated glio-
blastoma multiforme tumour. An oncology specialist
delineates the gadolinium enhanced T1 weighted MRI
imaging-based apparent boundary of the tumour as well
as the boundary of its necrotic region. As a first
391

approach, the neovasculature field is assumed to coin-
cide with the area of the tumour where pronounced
metabolism is apparent on the imaging data. A cube
defining the anatomical region of interest is super-
imposed on the three-dimensionally reconstructed
tumour and surrounding anatomical features. The cube
includes a volume discretizing mesh. The dimensions of
each geometrical cell of the mesh, considered to be able
to accommodate 106 biological cells [46] (NBC5106), are
1 mm61 mm61 mm.
A number of researchers have focused on the
measurement of the duration of cell cycle (Tc), especially
for gliomas. Hoshino et al [47] have reported a mean Tc
of 57 h. Crafts et al [48] suggested that Tc can range from
2 days to 3 days. Hoshino and Wilson [49] have
mentioned a Tc of 75.6 h. Pertuiset et al [50] have found
an average value of 1–2.5 days. The Tc values considered
in this paper have been 24 h, 48 h and 72 h. Salmon et al
[51] have suggested that proliferating tumour cells
would spend their time in the various cell cycle phases
as follows: fraction of time spent in G1: TG150.4 TC, S:
TS50.39 TC, G2: TG250.19 TC and M: TM50.02 TC.
After irradiation, most often the reproductively dead
cells will continue to cycle for (usually) 1–3 divisions
before their ultimate (biological) death. In the model
developed by our group, reproductively dead cells are
assumed to undergo two mitoses before their biological
death [17]. Reproductively dead cells and their offspring
which are still cycling are considered proliferating until
their ultimate biological death because they are detect-
able through imaging modalities. This point emphasises
the conceptual differences that may arise between an
engineering and a medical physics/clinical approach to
the same biological phenomenon. Exploratory simula-
tion runs have shown that if still-cycling reproductively
(but not yet biologically) dead cells are added to the
unaffected proliferating cells, an increase in the number
of the latter by up to a factor of 10 (1 log) is to be
expected during a typical radiotherapy course.
A standard fractionation scheme (2 Gy once a day,
5 days per week, 60 Gy in total) has been simulated. The
LQ model parameters of the tumour have been assumed
to be aoxic50.17 Gy21, boxic50.02 Gy22, and ahypoxic5
(0.17/OER) Gy21, bhypoxic5(0.02/OER2) Gy22 for a GBM
with known mt p53 gene [52] and aoxic50.6 Gy21,
boxic50.06 Gy22, ahypoxic5(0.6/OER) Gy21, bhypoxic5
(0.06/OER2) Gy22 for a GBM with known wild type
(wt) p53 gene [7, 53]. Both the aoxic and boxic parameters
are assumed to remain constant throughout the cell
cycle. For visualization purposes, cells are ‘‘painted as
dead’’ during the time interval between a lethal cell hit
and necrosis or apoptosis.
For the specific type of poorly differentiated tumour
under consideration, and for simplification reasons, all
non-clonogenic cells have been considered to be necrotic
and sterile cells have not been taken into account. The
contribution of the living non-clonogenic cells will be
considered in a future version of the model.
A typical tumour volume of 20 cm3 contains 4–5%
clonogenic tumour cells [54], i.e. the typical CCD is
around 107 cm23 [55]. Most calculations of the biological
effect of radiation on tumours assume that the CCD is
uniform. But in practice, tumours will almost certainly
have a non-uniform CCD [56]. This factor has not been
392
G S Stamatakos, V P Antipas, N K Uzunoglu and R G Dale
modelled except by Brahme and Agren [57] and Webb
and Nahum [56]. At present, as there is paucity of
experimental data in vivo, a range of reasonable
values has been assumed. Webb and Nahum [56] have
reported that the variation of CCD across tumours is a
very important factor, especially for brain tumours.
Hence the CCD has been assumed to be
16104 cells mm23, 26104 cells mm23, 36104 cells mm23,
46104 cells mm23.
Within each geometrical cell of the discretizing mesh,
the initial distribution of the clonogenic cells throughout
the cell cycle, the G0 and the necrotic phases depends on
the layer of the tumour to which the geometrical cell
under consideration belongs [17].
The cell loss factor [58] has been taken to be 0.3 as cell
death products are removed from brain with substantial
difficulty and it has been expressed as the sum of the cell
loss factor due to necrosis (0.2) and the cell loss factor
due to apoptosis (0.1).
It is noted that estimates of the percentage of
proliferating, dormant and dead cells in the various
tumour layers are included in Stamatakos et al [17] and
are based on a rather semi-quantitative representation of
the cycling status of tumour cells depending on the
imaging-based layer in which the considered tumour
cells lie. For example, proliferating cells included in the
dark tumour areas that appear on gadolinium enhanced
T1 weighted MRI slices are expected to be scant.
Concerning cell density, the standard assumption of
106 cells mm23 [46] has been made.
Validation and results
In order to clinically evaluate the simulation model,
several arm simulations of the Radiation Therapy
Oncology Group (RTOG) study 83-02 [18] have been
performed. The GBM imaging data considered through-
out the whole paper have been used as the spatial basis
for performing in silico experiments. The following
typical parameter values have been adopted: OER53.0,
clonogenic cell density 516CCD516104 cells mm23,
cell cycle duration Tc530 h and the LQ parameters have
been assumed to match the values of a GBM with mt p53
gene as previously mentioned.
Figure 1a shows the total number of proliferating and
dormant tumour cells as a function of time for the
hyperfractionated (1.2 Gy twice daily to the dose of
81.6 Gy, ‘‘HF-81.6’’) and accelerated hyperfractionated
(1.6 Gy twice daily to the dose of 54.4 Gy, ‘‘AHF-54.4’’)
radiotherapy schedules. All schemes considered in this
paper start on the first day of the radiotherapy course.
HF-81.6 is completed on day 46 after initiation of
treatment whereas AHF-54.4 is completed on day 23.
Figure 1b depicts the total number of proliferating and
dormant tumour cells as a function of time for the
hyperfractionated (1.2 Gy twice daily to the dose of
76.8 Gy, ‘‘HF-76.8’’) and accelerated hyperfractionated
(1.6 Gy twice daily to the dose of 48 Gy, ‘‘AHF-48’’)
radiotherapy schedules. Both irradiation schedules start
on the first day of the first week of treatment. HF-76.8 is
completed on day 44 after initiation of treatment whereas
AHF-48 is completed on day 19. According to the graphs,
before completion of the AHF course, cell kill due to
The British Journal of Radiology, May 2006
Simulation of in vivo glioblastoma multiforme response to radiotherapy

Figure 1. (a) Total number of proliferating and dormant tumour cells as a function of time for the hyperfractionated (1.2 Gy
twice daily, 5 days per week to the dose of 81.6 Gy, ‘‘HF-81.6’’) and accelerated hyperfractionated (1.6 Gy twice daily, 5 days per
week to the dose of 54.4 Gy, ‘‘AHF-54.4’’) radiotherapy schedules. HF-81.6 is completed on day 46 after initiation of treatment
whereas AHF-54.4 is completed on day 23. In all fractionation schedules considered in this paper, no radiation is administered on
Saturdays or Sundays. (b) Total number of proliferating and dormant tumour cells as a function of time for the
hyperfractionated (1.2 Gy twice daily, 5 days per week to the dose of 76.8 Gy, ‘‘HF-76.8’’) and accelerated hyperfractionated
(1.6 Gy twice daily, 5 days per week to the dose of 48 Gy, ‘‘AHF-48’’) radiotherapy schedules. Both irradiation schedules start on
the first day of the first week of treatment. HF-76.8 is completed on day 44 after initiation of treatment whereas AHF-48 is
completed on day 19.

AHF irradiation is more pronounced than cell kill
induced by the HF scheme. This can be explained by
the fact that a higher total dose has been administered to
the tumour by the AHF scheme whereas, for the period
under consideration, both schemes are characterized by
the same time intervals between consecutive sessions. In
cases where not all living cells have been killed by AHF
irradiation, tumour repopulation is considerable so that,
by the time the HF scheme is completed, living tumour
cells and their progeny which have escaped AHF
irradiation outnumber tumour cells which have escaped
HF irradiation. Improved tumour control following HF
irradiation in comparison with the AHF scheme is in
agreement with the conclusions of the clinical trial RTOG
83-02.
Figure 2a shows the total number of proliferating and
dormant tumour cells as a function of time for the
hyperfractionated (1.2 Gy twice daily to the dose of
72 Gy, ‘‘HF-72’’) and accelerated hyperfractionated
(1.6 Gy twice daily to the dose of 48 Gy, ‘‘AHF-48’’)
radiotherapy schedules. HF-72 is completed on day 40
after initiation of treatment whereas AHF-48 is com-
pleted on day 19. Figure 2b shows the total number of
proliferating and dormant tumour cells as a function of
time for the hyperfractionated (1.2 Gy twice daily to
the dose of 64.8 Gy, ‘‘HF-64.8’’) and accelerated hyper-
fractionated (1.6 Gy twice daily to the dose of 48 Gy,
‘‘AHF-48’’) radiotherapy schedules. HF-64.8 is com-
pleted on day 37 after initiation of treatment whereas
AHF-48 is completed on day 19. Both irradiation

Figure 2. (a) Total number of proliferating and dormant tumour cells as a function of time for the hyperfractionated (1.2 Gy
twice daily, 5 days per week to the dose of 72 Gy, ‘‘HF-72’’) and accelerated hyperfractionated (1.6 Gy twice daily, 5 days per
week to the dose of 48 Gy, ‘‘AHF-48’’) radiotherapy schedules. HF-72 is completed on day 40 after initiation of treatment
whereas AHF-48 is completed on day 19. (b) Total number of proliferating and dormant tumour cells as a function of time for
the hyperfractionated (1.2 Gy twice daily, 5 days per week to the dose of 64.8 Gy, ‘‘HF-64.8’’) and accelerated hyperfractionated
(1.6 Gy twice daily, 5 days per week to the dose of 48 Gy, ‘‘AHF-48’’) radiotherapy schedules. HF-64.8 is completed on day 37
after initiation of treatment whereas AHF-48 is completed on day 19. Both irradiation schedules start on the first day of the first
week of treatment.

The British Journal of Radiology, May 2006 393


schedules start on the first day of the first week of
treatment. According to the graphs, before completion of
the AHF course, cell kill due to AHF irradiation is more
pronounced than cell kill induced by the HF scheme.
This can be explained by the fact that higher total dose
has been administered to the tumour by the AHF scheme
with the same fractionation as the one followed by the
HF scheme for the period under consideration. In cases
where not all living cells have been killed by AHF
irradiation, tumour repopulation is considerable so that,
by the time the HF scheme is completed, living tumour
cells and their progeny which have escaped AHF
irradiation outnumber tumour cells which have escaped
HF irradiation.
Figure 3 depicts the total number of tumour cells
(proliferating, dormant and dead cells) as a function of
time for the hyperfractionated (1.2 Gy twice daily to the
dose of 76.8 Gy, ‘‘HF-76.8’’) and accelerated hyperfrac-
tionated (1.6 Gy twice daily to the dose of 48 Gy, ‘‘AHF-
48’’) radiotherapy schedules. Irradiation starts on the
first day of the first week. HF-76.8 is completed on day 44
after initiation of treatment whereas AHF-48 is com-
pleted on day 19. Both irradiation schedules start on the
first day of the first week of treatment. At the end of
week 10, repopulation following AHF-48 treatment is
more pronounced than repopulation following HF-76.8,
in accordance with Figure 1b.
All of the above mentioned observations are in
agreement with the statistically significant results of the
RTOG 83-02 study [18] which refer to the survival for the
following radiotherapy fractionations: high dose hyper-
fractionation (HF, total doses: 76.8 Gy and 81.6 Gy), low
dose hyperfractionation (HF, total doses: 64.8 Gy and
72.0 Gy) and accelerated hyperfractionation (AHF, total
doses: 48.0 Gy and 54.4 Gy). It should be noted that
carmustine (BCNU), which was also administered to all
patients included in that study, was assumed not to
significantly modify the effectiveness relations among
the different radiotherapeutic schemes considered. The
Figure 3. Total number of tumour cells (proliferating,
dormant and dead cells) as a function of time for the
hyperfractionated (1.2 Gy twice daily, 5 days per week to the
dose of 76.8 Gy, ‘‘HF-76.8’’) and accelerated hyperfractio-
nated (1.6 Gy twice daily, 5 days per week to the dose of
48 Gy, ‘‘AHF-48’’) radiotherapy schedules. Irradiation starts
on the first day of the first week. HF-76.8 is completed on day
44 after initiation of treatment whereas AHF-48 is completed
on day 19. Both irradiation schedules start on the first day of
the first week of treatment.
394
G S Stamatakos, V P Antipas, N K Uzunoglu and R G Dale
reason for this is that the same chemotherapy adminis-
tration schedule was followed for all patients regardless
of the specific radiotherapeutic scheme administered to
each individual patient. All patients received BCNU
80 mg m22, intravenously, on days 1, 2 and 3 of the first
week of radiotherapy and subsequently on 3 consecutive
days every 8 weeks for a period of 1 year, to a maximum
dose of 1440 mg m22. Only limited modifications of the
schedule based on age and not on the radiotherapy
scheme administered were made during the trial.
Agreement with the results of the RTOG 83-02 study
implies that the simulation model has successfully
captured and integrated the critical biological aspects
that determine the clinical outcome. Furthermore, it
strengthens the proposition that advanced cancer inte-
grative (radio)biology might explain and predict the
therapy outcome, despite the complexity of the clinical
setting.
In the following, a parametric analysis is carried out in
order to study the effect of the clonogenic cell density as
well as that of the OER and cell cycle duration. The
model code has been executed for a simulated period of
up to 6 weeks, an interval which normally covers the
treatment period of the radiotherapy course and may
extend to some days after its completion. The cell cycle
duration has been assumed to be 48 h, OER equal to 3.0,
CCD equal to 16104 cells mm23 and the LQ parameters
have been assumed to match the values of a GBM with
known mt p53 gene, unless otherwise stated.
The simulation results of Figure 4 demonstrate the
ability of the algorithm to effectively simulate the
tumour response to a standard irradiation scheme
under different values of CCD (16104 cells mm23,
26104 cells mm23, 36104 cells mm23, 46104 cells
23
mm ). Higher values of CCD not only produce a higher
number of proliferating tumour cells, but also affect the
entire tumour composition. Such behaviour is in accor-
dance with the previously described interdependence
among the various metabolic layers of the tumour
(dormant, dead, proliferating). Figure 5 provides a two-
dimensional visualization of the simulated response of a
clinical glioblastoma multiforme tumour to the standard
fractionation scheme for different CCD values. At the
end of day 3, the tumour with 16CCD appears to be
more radiosensitive compared with tumours with
26CCD, 36CCD and 46CCD as its population of
proliferating cells is lower compared with the others
(Figure 5a–d). At the end of day 5, all tumours are
strongly affected by radiation treatment, whereas the
highest number of proliferating cells is contained in
tumour with 46CCD (Figure 5h). At the beginning of
the first day of the second week (day 8), newly-produced
proliferating cells are present in sufficient numbers to be
apparent in Figure 5k,l. The population of the proliferat-
ing cells is larger in Figure 5l compared with Figure 5k
while the majority of the slice of tumour Figure 5l is
painted as ‘‘proliferating’’. For this specific case, the LQ
model parameters of the tumour are in accordance with a
GBM with known mt p53 gene.
The simulation results in Figure 6 demonstrate the
ability of the algorithm to effectively simulate the tumour
response to a standard irradiation scheme for different
OER values (1.0, 2.0, 3.0). Higher OER values describe a
more radioresistant tumour.
The British Journal of Radiology, May 2006
Simulation of in vivo glioblastoma multiforme response to radiotherapy

Figure 4. Simulation predictions of (a) the number of proliferating and (b) the total number of proliferating, dormant and dead
mt p53 tumour cells in the case of the standard fractionation scheme (2 Gy once a day, 5 days per week, 60 Gy in total) for
different clonogenic cells densities (CCD5104 cells mm23). Irradiation schedule starts on the first day of the first week of
treatment. It should be stressed that reproductively dead cells and their offspring that are still cycling are considered
proliferating until their ultimate biological death. The periodicity noticed on all graphs reflects the weekly irradiation
periodicity. It is noted that no irradiation takes place during the weekend.

Figure 7 depicts the corresponding simulation results
of a clinical glioblastoma tumour to the standard
fractionation scheme for different Tc values. For the
short period simulated (10 days) the tumour with
Tc548 h appears to be more difficult to treat than the
one with Tc560 h. It can nevertheless be noted that
during most of the working days the tumour with
Tc548 h shows a better response to irradiation whereas
the tumour with Tc560 h shows a slower response to
radiation due to its long Tc.
Figure 8 is a three-dimensional reconstruction of the
simulated response of a GBM with known mt p53 gene
for Tc524 h (Figure 8a) and Tc572 h (Figure 8b) at the
end of day 8. Figure 8a contains few geometrical cells
which have been painted as ‘‘proliferating’’, in contrast
to Figure 8b in which all geometrical cells have been
painted as ‘‘dead’’. During the interval in which the
tumour is not irradiated (days 6 and 7), the proliferating
tumour cells with Tc524 h multiply prominently faster
than in the case of the tumour with Tc572 h.
Concerning the model sensitivity issue, Figures 4, 6
and 7 can also be used in order to quantitatively
visualize the effect of perturbing critical model para-
meters on the simulation predictions. For example,
according to Figure 4a, on the 30th day after initiation
of the radiotherapy course the number of proliferating
tumour cells for the case of CCD equal to 46CCD
(CCD5104 cells mm23) is as high as 270% of the number
of proliferating cells for the case of 16CCD on the same
day. According to Figure 6a, on the 30th day after
initiation of the radiotherapy course the number of
proliferating tumour cells for the case of OER53.0 is as
high as 127% of the number of proliferating cells for the
case of OER51.0 on the same day. According to
Figure 7a, on the 30th day after initiation of the radio-
therapy course the number of proliferating tumour cells
for the case of Tc548 h is as high as 339% of the number
of proliferating cells for the case of Tc560 h on the same
day. Concerning estimation of confidence intervals it is
noted that, due to the limited information contained in
the RTOG 83-02 clinical study data and the small number
of complete sets of clinical data available by our group,
no strict confidence limits can be estimated as yet. It is
pointed out that cancer is inherently a partly stochastic
phenomenon and, therefore, an accurate estimation of
the confidence limits of its predicted in vivo behaviour is
not always feasible. It is also noted that, although the
simulation model predicts the obvious, it provides a
means for quantitatively studying the response of real
clinical malignant tumours to radiotherapeutic schemes.
Discussion and conclusion
The results presented are in agreement with both
qualitative clinical experience and the outcome of the
RTOG 83-02 clinical study. Predictably, a whole tumour
with shorter Tc tends to repopulate faster and therefore is
more difficult to treat.
Greater CCD produces a higher number of proliferat-
ing tumour cells, and therefore a tumour which is more
difficult to treat, and eventually produces a lower
tumour control probability (TCP) [59]. Webb and
Nahum [56] report that the TCP is a complicated
function of the variation in both dose and CCD and
TCP will depend on the assumption made about CCD. It
is pointed out that reproductively dead cells can cycle
only for a very limited number of divisions (one to three)
before they and their progeny die biologically. Therefore,
re-growth of reproductively dead cells or reproductively
‘‘killed’’ clonogenic cells is unimportant for clinical
outcome [60].
Although simulation prediction curves look quite
similar (Figures 4, 6 and 7), a closer observation reveals
several differences among them. Random numbers have
actually been used, but the fact that all biological
parameters for all curves in Figure 4 are the same except
for the CCD has led to a pretty similar (analogous)
behaviour of the composite biological system.
Furthermore, a careful and successful application of the

The British Journal of Radiology, May 2006 395

 G S Stamatakos, V P Antipas, N K Uzunoglu and R G Dale

Figure 5. Two-dimensional visualization of the simulated response of a radiosensitive clinical glioblastoma multiforme tumour
to the standard fractionation scheme, for a range of clonogenic cell density (CCD). The figure shows a centrally located
horizontal slice of a tumour with clonogenic cell density equal to 16CCD(5104 cells mm23) (a) 3 fictitious days after the
beginning of the radiotherapy course (e) 5 fictitious days after the beginning of the radiotherapy course, (i) 8 fictitious days
after the beginning of the radiotherapy course. A centrally located horizontal slice of a tumour with clonogenic cell density
equal to 26CCD (b) 3 fictitious days after the beginning of the radiotherapy course (f) 5 fictitious days after the beginning of
the radiotherapy course (j) 8 fictitious days after the beginning of the radiotherapy course. A centrally located horizontal slice of
a tumour with clonogenic cell density equal to 36CCD, (c) 3 fictitious days after the beginning of the radiotherapy course, (g) 5
fictitious days after the beginning of the radiotherapy course, (k) 8 fictitious days after the beginning of the radiotherapy
course. A centrally located horizontal slice of a tumour with clonogenic cell density equal to 46CCD, (d) 3 fictitious days after
the beginning of the radiotherapy course, (h) 5 fictitious days after the beginning of the radiotherapy course, (l) 8 fictitious days
after the beginning of the radiotherapy course. For this specific case, the LQ model parameters of the tumour are in accordance
with a GBM cell line with known mt p53 gene. Irradiation schedule starts on the first day of the first week of treatment. Colour
Code: dark grey: proliferating cell layer, light grey: dormant cell layer (G0), white: dead cell layer. The colouring criterion
‘‘99.8%’’ used to visualize the predictions has been defined as follows. ‘‘For a geometrical cell of the discretising mesh, if
the percentage of dead cells is lower than 99.8% then {if percentage of proliferating cells . percentage of G0 cells, then paint
the geometrical cell dark grey (proliferating cell layer)}, else paint the geometrical cell light grey (G0 cell layer) else paint the
geometrical cell white (dead cell layer)’’.

396 The British Journal of Radiology, May 2006

Simulation of in vivo glioblastoma multiforme response to radiotherapy

Figure 6. Simulation predictions of the number of (a) proliferating and (b) dormant mt p53 tumour cells in the case of standard
fractionation for different OER (1.0, 2.0, 3.0) values. It is stressed that reproductively dead cells and their offspring that are still
cycling are considered proliferating until their ultimate biological death. Irradiation schedule starts on the first day of the first
week of treatment.

Figure 7. Simulation predictions of the number of (a) proliferating and (b) dormant mt p53 tumour cells in the case of standard
fractionation for different Tc (cell cycle time) values. It is stressed that reproductively dead cells and their offspring that are still
cycling are considered proliferating until their ultimate biological death. Irradiation schedule starts on the first day of the first
week of treatment.

Monte Carlo technique has led to a remarkably stable
numerical behaviour of the simulation model.
It has been experimentally demonstrated that most
solid tumours contain resistant hypoxic cells, with
estimates of the hypoxic fractions ranging from below
1% to well over 50% of the total viable cell population
[33]. Additionally, evidence that hypoxia exists in human
tumours to a degree that can influence radiation
response comes from those clinical trials in which some
form of hypoxia modification has been attempted and
found to improve tumour response [61]. Knisely and
Rockwell [62] have reported that the resistance of
gliomas to treatment with radiation and antineoplastic
drugs may result in part from the effects of the extensive
and severe hypoxia that is present in such tumours. They
have emphasised that the brain tumours contain exten-
sive regions in which the tumour cells are subjected to an
unphysiological degree of hypoxia, this being involved in
the evolution of cells in low-grade malignancies to the
resistant, aggressive phenotype characteristic of glioblas-
tomas. Furthermore, the results of the parametric OER
study agree with data presented by Horsman and
Overgaard [33]. It is stressed that although the model
confirms the obvious, it allows a quantitative study of the
inter-relationship between the competing influences in a
complex, dynamic tumour environment. A possible future
refinement of the model would include a detailed
description of the modulation of cell cycling by oxygen
tension by taking into account the latest pertinent
experimental observations. It is also noted that, according
to our model, intrinsic sensitivity, at least as modified, by
OER, is an important determinant of the poor outcome of
glioblastoma multiforme irradiation, which appears to be
in disagreement with Taghian et al [63].

The British Journal of Radiology, May 2006 397


The CMA tumour shrinkage algorithm applied in this
paper is more realistic than the one previously described
by Stamatakos et al [17], as CMA is able to substantially
conserve the tumour shape (conformal shrinkage) in
accordance with Perez and Brady [45]. Additionally, the
imaging based neovasculature distribution has been in
agreement with Horsman and Overgaard [33]. Finally,
the uniform cell distribution algorithm produces a
realistic initial cell phase distribution. The predictions
of all indicative simulations performed agree at least
qualitatively with the clinical experience and with the
data presented by Duechting et al [1, 5], Kocher et al [6,
7] and Horsman and Overgaard [33].
It is stressed that the model presented addresses the
imageable gross tumour. In order to take into account the
brain micrometastases (diffuse invasion), an approach
similar to the one developed by Swanson et al [8, 10] and
Mandonnet et al [9] or Kansal et al [1, 2] must be
considered. It is also noted that the end points of most
clinical trials have not been the volumetric and/or
metabolic response of tumours, but rather the overall
response to treatment such as survival and tumour
relapse interval. Consequently, currently available clini-
cal data can be exploited for the validation of the models
only in a rough (approximate) way. This implies that a
better survival following treatment scheme X compared
with survival following treatment scheme Y can be
roughly used as an indicator of better tumour control
achievable with scheme X. Obviously this can be the case
if radiation toxicity is within acceptable limits for both
treatment schemes. Concerning fluctuations of radio-
sensitivity throughout the cell cycle, differing radio-
sensitivity of cells in the various phases of the cell cycle
has been considered and successfully simulated by our
group in the case of small in vitro or in vivo pre-
angiogenetic/avascular tumour spheroids [17].
Nevertheless, although easily includable, this variation
has not been addressed in the present version of the
in vivo gross tumour model in order to keep computer
memory and execution time requirements as low as
possible.
Modelling of the irradiation effects on the surrounding
normal tissues in vivo, modelling of the tumour response
to chemotherapy in vivo and further enhancement of our
models with more genetic data are currently under way.
Systematic comparison with clinical data is expected to
lead to more clinically adapted parameter values. The
clinical validation procedure is in progress and involves
comparison of the model predictions with pertinent
clinical data before, during and after the radiotherapy
course. The easily adjustable, modular simulation model
‘‘follows’’ the clinical practice and activates a self-
optimization procedure.
It should also be noted that very rare deviations from
the poor standard prognosis of glioblastoma multiforme
do exist. Nevertheless, the aim of the simulation model is
to predict the most likely time course of the treatment
response and, consequently, it is unlikely that an extreme
scenario would be predicted. Execution of the computer
code on supercomputer systems where a more dense
discretizing mesh could be considered might refine the
model’s prediction accuracy.
The validation process in conjunction with generic
parameter estimation techniques (neural networks
398
G S Stamatakos, V P Antipas, N K Uzunoglu and R G Dale
technique, taboo searching, etc.) can be used to achieve
better estimates of the input parameters. Agreement with
clinical observations strengthens the applicability of the
model to real situations. An integrated and patient-
individualized decision support and spatiotemporal
treatment planning system is expected to emerge after
completion of the necessary clinical adaptation and
validation processes. Such a system could also serve as
an educational platform for professionals and patients by
means of virtual reality demonstrations of the likely
natural development and treatment responsiveness of
specific cancers so that all groups might positively
contribute to the discussion about treatment procedure.
Use of genotyping data might enhance the potential of
the model, as more accurate estimates of the patient’s
individualized linear quadratic radiobiological para-
meters might be achieved through the use of molecular
interaction networks.
A more quantitative validation can be achieved using
the patient data to be collected and applying multiple
parameter adaptation methods such as genetic algo-
rithms or neural networks. The imaging (e.g. MRI, PET,
etc.), histopathological and, eventually, genetic data are
introduced into the simulation model and a candidate
radiotherapeutic or chemotherapeutic scheme is defined.
The output of the simulation run, which is the
prediction of the tumour and the most affected normal
tissue response to the treatment scheme, is then
evaluated by the supervising doctor. If a further scheme
is to be tested in silico, the simulation run is repeated
with the same imaging and radiobiological/pharmaco-
dynamic data as previously.
In the end, the modelling platform might serve as a
generic ‘‘decision-support system’’. In this way, the
medical doctor might make his or her final decision on
the selection of the most promising therapeutic scheme
by taking into account both the predicted outcomes of all
simulated regimens as well as his or her own medical
knowledge and expertise. This computational platform
does not therefore intend to replace the medical doctor’s
input, but to add the possibility of investigating the
impact of specific treatment-induced perturbations.
Acknowledgments
The project has been supported by the Hellenic
Ministry of Education and Religious Affairs under the
program ‘‘Irakleitos: Fellowships for Research of the
National Technical University of Athens’’. The project
has been co-funded by the European Social Fund (75%)
and National Resources (25%).
References
1. Kansal AR, Torquato S, Harsh IV GR, Chiocca EA,
Deisboeck TS. Simulated brain tumour growth dynamics
using a three-dimensional cellular automaton. J Theor Biol
2000;203:367–82.
2. Kansal AR, Torquato S, Harsh IV GR, Chiocca EA,
Deisboeck TS. Cellular automaton of idealized brain
tumour growth dynamics. BioSystems 2000;55:119–27.
3. Kraus M, Wolf B. Emergence of self-organization in tumour
cells: relevance for diagnosis and therapy. Tumour Biol
1993;14:338–53.
The British Journal of Radiology, May 2006
Simulation of in vivo glioblastoma multiforme response to radiotherapy

Figure 8. Three-dimensional visualization of the simulated response of a radiosensitive clinical glioblastoma multiforme tumour
to the standard fractionation scheme, for (a) Tc524 h and (b) Tc572 h at the end of day 8. It is pointed out that total tumour
cells include all morphologically existing cells, living (proliferating and quiescent) and dead (but not yet lysed or fragmented)
alike. Irradiation schedule starts on the first day of the first week of treatment. Colour Code: red: proliferating cell layer, green:
G0 layer, blue: dead cell layer. The colouring criterion ‘‘99.8%’’ was used to visualize the predictions (Figure 5).

4. Duechting W, Ulmer W, Lehrig R, Ginsberg T, Dedeleit E. 16. Stamatakos G, Zacharaki E, Makropoulou M,

Computer simulation and modelling of tumour spheroid
growth and their relevance for optimization of fractionated
radiotherapy. Strahlenther Onkol 1992;168:354–60.
5. Duechting W, Ginsberg T, Ulmer W. Modelling of radio-
genic responses induced by fractionated irradiation in
malignant and normal tissue. Stem Cells 1995;13(S1):301–6.
6. Kocher M, Treuer H. Reoxygenation of hypoxic cells by
tumour shrinkage during irradiation. A computer simula-
tion. Strahlentherapie und Onkologie 1995;171:219–30.
7. Kocher M, Treuer H, Voges J, Hoevels M, Sturm V, Muller
RP. Computer simulation of cytoxic and vascular effects of
radiosurgery in solid and necrotic brain metastasis.
Radiother Oncol 2000;54:149–56.
8. Swanson KR, Alvord Jr EC, Murray JD. Virtual brain
tumours (gliomas) enhance the reality of medical imaging
and highlight inadequacies of current therapy. Br J Cancer
2002;86:14–8.
9. Mandonnet E, Delattre JY, Tanguy ML, Swanson KR,
Carpentier AF, Duffau H, et al. Continuous growth of
mean tumor diameter in a subset of grade II gliomas. Ann
Neurol 2003;53:524–8.
10. Swanson KR, Bridge C, Murray JD, Alvord Jr EC. Virtual
and real brain tumors: using mathematical modelling to
quantify glioma growth and invasion. J Neurol Sci
2003;216:1–10.
11. Byrne H, Matthews P. Asymetric growth of avascular solid
tumours: exploiting symmetries. IMA J Mathematics
Applied in Medicine and Biology 2002;19:1–29.
12. Alarcon T, Byrne HM, Maini PK. A cellular automaton
model for tumour growth in inhomogeneous environment.
J Theor Biol 2003;225:257–74.
13. Alarcon T, Byrne HM, Maini PK. Towards whole-organ
modelling of tumour growth. Prog Biophys Mol Biol
2004;85:451–72.
14. Alarcon T Byrne HM, Maini PK. A mathematical model of
the effects of hypoxia on the cell-cycle of normal and cancer
cells. J Theor Biol 2004;229:395–411.
15. Stamatakos G, Dionysiou D, Nikita K, Zamboglou N, Baltas
D, Pissakas G, et al. In vivo tumour growth and response to
radiation therapy: a novel algorithmic description. Int J
Radiat Oncol Biol Phys 2001;51 Suppl. 1:240.
Mouravliansky N, Marsh A, Nikita K, et al. Modelling
tumour growth and irradiation response in vitro - a
combination of high-performance computing and web
based technologies including VRML visualization. IEEE
Trans Inform Tech Biomed 2001;5:279–89.
17. Stamatakos GS, Dionysiou DD, Zacharaki EI,
Mouravliansky NA, Nikita KS, Uzunoglu NK. In Silico
radiation oncology: combining novel simulation algorithms
with current visualization techniques. Proceedings of the
IEEE 2002;90:(11).
18. Werner-Wasik M, Scott CB, Nelson DF, Gaspar LE, Murray
KJ, Fischbach JA, et al. Final report of a phase I/II trial of
hyperfractionated and accelerated hyperfractionated radia-
tion therapy with carmustine for adults with supratentorial
malignant gliomas. Radiation Oncology Therapy Group
Study 83–02. Cancer 1996;77:1535–43.
19. Goldman E. The growth of malignant disease in man and
the lower animals with special reference to the vascular
system. Lancet 1907;2:1236–40.
20. Ide AG, Baker NH, Warren SL. Vascularization of the
Brown-Pearce rabbit epithelioma transplant as seen in the
transparent ear chamber. Am J Radiol 1939;42:891–9.
21. Algire GH, Chalkley HW. Vascular reactions of normal and
malignant tissues in vivo. I. Vascular reactions of mice to
wounds and to normal and neoplastic transplants. J Natl
Cancer Inst USA 1945;6:73–85.
22. Greenblatt M, Shubik P. Tumour angiogenesis: transfilter
diffusion studies in the hamster by the transparant chamber
technique. J Natl Cancer Inst 1968;41:111–24.
23. Ehrmann RL, Knoth M. Choriocarcinoma: transfilter stimu-
lation of vasoproliferation in the hamster cheek pouch
studied by light and electron microscopy. J Natl Cancer Inst
1968;41:1329–41.
24. Patan S, Munn LL, Jain RK. Intussusceptive microvascular
growth in a human colon adenocarcinoma xenograft: a
novel mechanism of tumour angiogenesis. Microvasc Res
1996;51:260–72.
25. Asahara T, Kalka C, Isner JM. Stem cell therapy and gene
transfer for regeneration. Gene Ther 2000;7:451–7.
26. Rafii S. Circulating endothelial precursors: mystery, reality,
and promise. J Clin Invest 2000;105:17–9.

The British Journal of Radiology, May 2006 399


27. Carmeliet P, Jain RK. Angiogenesis in cancer and other
diseases. Nature 2002;407:249–59.
28. Baish JW, Jain RF. Fractals and cancer. Cancer Res
2000;60:3683–8.
29. Helmlinger G Yuan F Dellian M, Jain RK. Interstitial pH
and pO2 gradients in solid tumours in vivo: high-resolution
measurements reveal a lack of correlation. Nat Med
1997;3:177–82.
30. Hobbs SK, Monsky WL, Yuan F, Roberts WG, Griffith L,
Torchilin VP, et al. Regulation of transport pathways in
tumour vessels: role of tumour type and microenvironment.
Proc Natl Acad Sci USA 1998;95:4607–12.
31. Hashizume H, Baluk P, Morikawa S, McLean JW, Thurston
G, Roberge S, et al. Openings between defective endothelial
cells explain tumour vessel leakiness. Am J Pathol
2000;156:1363–80.
32. Michelson S, Leith JT. Positive feedback and angiogenesis
in tumour growth control. Bull Math Biol 1997;59:233–54.
33. Horsman MR, Overgaard J. The oxygen effect and tumour
microenvironment. In: Steel GG, editor. Basic Clinical
Radiobiology. 3rd edition, London UK: Hodder Arnold,
2002:158–168.
34. Palcic B, Skarsgard LD. Reduced oxygen enhancement ratio
at low doses of ionizing radiation. Radiat Res
1984;100:328–39.
35. Revesz L, Palcic B. Radiation dose dependence of the
sensitization by oxygen and oxygen mimic sensitizers. Acta
Radiol Oncol 1985;24:209–17.
36. Beppu T, Kamada K, Yoshida Y, Arai H, Ogasawara K,
Ogawa A. Change of oxygen pressure in glioblastoma
tissue under various conditions. J Neurooncol
2002;58:47–52.
37. Potapov AA, Rapman I, Liass FM, Manevich AZ,
Iarmonenko SP. Oxygenation of glioblastomas and normal
brain tissue during radiation therapy. Med Radiol (Mosk)
1983;28:21–4.
38. Groshar D, McEwan AJ, Parliament MB, Rtasun RC,
Golberg LE, Hoskinson M, et al. Imaging tumour hypoxia
and tumour perfusion. J Nucl Med 1993;34:885–8.
39. Yetkin FZ Mendelsohn D. Hypoxia imaging in brain
tumours. Neuroimaging Clin N Am 2002;12:537–52.
40. Valk PE, Mathis CA, Prados MD, Gilbert JC, Budinger TF.
Hypoxia in human gliomas: demonstration by PET with
fluorine-18-fluoromisonidazole. J Nucl Med 1992;33:2133–7.
41. Leenders KL. PET: blood flow and oxygen consumption in
brain tumours. J Neurooncol 1994;22:269–73.
42. Nygren J, Ahnstrom G. The oxygen effect in permeabilized
and histone-depleted cells: an enhanced OER for DNA
double-strand breaks compared to single-strand breaks is
abolished by soluble scavengers. Int J Radiat Biol
1997;72:163–70.
43. Stuschke M, Budach V, Budach W, Feldmann HJ, Sack H.
Radioresponsiveness sublethal damage repair and stem cell
rate in spheroids from three human tumour lines: compa-
rison with xenograft data. Int J Radiat Oncol Biol Phys
1992;24:119–26.
44. Speke A, Hill R. the effects of clamping and reoxygenation
of repopulation during fractionated irradiation. Int J Radiat
Biol 1995;31:857–63.
45. Perez C, Brady L, Roti Roti J. Overview. In: Perez C, Brady
L, editors. Principles and practice of radiation oncology.
Philadelphia, PA: Lippincott-Raven, 1998:10.
400
G S Stamatakos, V P Antipas, N K Uzunoglu and R G Dale
46. Nahum A, Sanchez-Nieto B. Tumour control probability
modelling: basic principles and applications in treatment
planning. Phys Med 2001;Suppl. 2:13–23.
47. Hoshino T, Wilson CB, Rosenblum ML, Barker MJ.
Chemotherapeutic implications of growth fraction and cell
cycle time in glioblastomas. Neurosurg 1975;43:127–35.
48. Crafts DC, Hoshino T, Wilson CB. Current status of
population kinetics in gliomas. Bull Cancer 1977;64:115–24.
49. Hoshino T, Wilson CB. Cell kinetic analyses of human
malignant brain tumours (gliomas). Cancer 1979;44:956–62.
50. Pertuiset B, Dougherty D, Cromeyer C, Hoshino T, Berger
M, Rosenblum MLJ. Stem cell studies of human malignant
brain tumours. Part 2: Proliferation kinetics of brain-tumour
cells in vitro in early-passage cultures. Neurosurg
1985;63:426–32.
51. Salmon ES, Sartorelli AC. Cancer chemotherapy. In:
Katzung BG, editor. Basic and clinical pharmacology.
USA: Lange Medical Books/McGraw-Hill, 2001:926.
52. Haas-Kogan DA, Yount G, Haas M, Levi D, Kogan SS, Hu
L, et al. p53-dependent G1 arrest and p53 independent
apoptosis influence the radiobiologic response of glioblas-
toma. Int J Radiat Oncol Biol Phys 1995;36:95–103.
53. Jones B, Bleasdale C. Influence of tumour regression and
clonogen repopulation on tumour control by brachytherapy.
In: Baier K, Baltas D, editors. Modelling in clinical radio-
biology. Freiburg, Germany: Albert-Ludwigs-University,
Freiburg Oncology Series, Monograph 1997;2:116–26.
54. Antipas V, Dale RG, Coles IP. A theoretical investigation
into the role of tumour radiosensitivity, clonogen repopula-
tion, tumour shrinkage and radionuclide RBE in permanent
brachytherapy implants of 125-I and 103-Pd. Phys Med Biol
2001;46:2557–69.
55. Webb S. Optimum parameters in a model for tumour-
control probability including interpatient heterogeneity.
Phys Med Biol 1994;39:1895–914.
56. Webb S, Nahum AE. A model for calculating tumour
control probability in radiotherapy including the effects of
inhomogeneous distributions of dose and clonogenic cell
density. Phys Med Biol 1993;38:653–66.
57. Brahme A, Ågren AK. Optimal dose distribution for
eradication of heterogeneous tumours. Acta Radiol Oncol
1987;26:377–85.
58. Begg AC, Steel GG. Cell proliferation and growth rate of
tumour. In: Steel GG, editor. Basic Clinical Radiobiology.
London, UK: Hodder Arnold 2002:8–22.
59. Ebert MA, Hoban W. Some characteristics of tumour
control probability for heterogeneous tumours. Phys Med
Biol 1996;41:2125–33.
60. Steel GG. Cell survival as a determinant of tumour
response. In: Steel GG, editor. Basic Clinical Radiobiology.
London, UK: Hodder Arnold 2002:52–63.
61. Horsman MR, Overgaard J. Overcoming tumour radio-
resistance resulting from hypoxia. In: Steel GG, editor. Basic
Clinical Radiobiology. London, UK: Hodder Arnold
2002:169–181.
62. Knisely JP, Rockwell S. Importance of hypoxia in the
biology and treatment of brain tumours. Neuroimaging
Clin N Am 2002;12:525–36.
63. Taghian A, Ramsay J, Allalunis-Turner J, Budach W,
Gioioso D, Pardo F, et al. Intrinsic radiation sensistivity
may not be the major determinant of the poor clinical
outcome of glioblastoma multiforme. Int J Radiat Oncol Biol
Phys 1993;25:243–9.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 401–408

A quantitative study of IMRT delivery effects in commercial

planning systems for the case of oesophagus and prostate
tumours
J SECO, PhD, C H CLARK, PhD, P M EVANS, DPhil and S WEBB, PhD, DIC

Joint Department of Physics, Institute of Cancer Research and Royal Marsden NHS Foundation
Trust, Downs Road, Sutton, Surrey SM2 5PT, UK

ABSTRACT. This study focuses on understanding the impact of intensity-modulated

radiotherapy (IMRT) delivery effects when applied to plans generated by commercial
treatment-planning systems such as Pinnacle (ADAC Laboratories Inc.) and CadPlan/
Helios (Varian Medical Systems). These commercial planning systems have had several
version upgrades (with improvements in the optimization algorithm), but the IMRT
delivery effects have not been incorporated into the optimization process. IMRT
delivery effects include head-scatter fluence from IMRT fields, transmission through
leaves and the effect of the rounded shape of the leaf ends. They are usually accounted
for after optimization when leaf sequencing the ‘‘optimal’’ fluence profiles, to derive
the delivered fluence profile. The study was divided into two main parts: (a) analysing
the dose distribution within the planning-target volume (PTV), produced by each of the
commercial treatment-planning systems, after the delivered fluence had been
renormalized to deliver the correct dose to the PTV; and (b) studying the impact of the
IMRT delivery technique on the surrounding critical organs such as the spinal cord,
lungs, rectum, bladder etc. The study was performed for tumours of (i) the oesophagus
and (ii) the prostate and pelvic nodes. An oesophagus case was planned with the
Pinnacle planning system for IMRT delivery, via multiple-static fields (MSF) and
compensators, using the Elekta SL25 with a multileaf collimator (MLC) component. A
prostate and pelvic nodes IMRT plan was performed with the Cadplan/Helios system for
a dynamic delivery (DMLC) using the Varian 120-leaf Millennium MLC. In these
commercial planning systems, since IMRT delivery effects are not included into the
optimization process, fluence renormalization is required such that the median
delivered PTV dose equals the initial prescribed PTV dose. In preparing the optimum
fluence profile for delivery, the PTV dose has been ‘‘smeared’’ by the IMRT delivery
techniques. In the case of the oesophagus, the critical organ, spinal cord, received a
greater dose than initially planned, due to the delivery effects. The increase in the
spinal cord dose is of the order of 2–3 Gy. In the case of the prostate and pelvic nodes,
the IMRT delivery effects led to an increase of approximately 2 Gy in the dose delivered
Received 9 February 2005
to the secondary PTV, the pelvic nodes. In addition to this, the small bowel, rectum and Revised 7 September 2005
bladder received an increased dose of the order of 2–3 Gy to 50% of their total volume. Accepted 7 September
IMRT delivery techniques strongly influence the delivered dose distributions for the 2005
oesophagus and prostate/pelvic nodes tumour sites and these effects are not yet
DOI: 10.1259/bjr/91588055
accounted for in the Pinnacle and the CadPlan/Helios planning systems. Currently, they
must be taken into account during the optimization stage by altering the dose limits ’ 2006 The British Institute of
accepted during optimization so that the final (sequenced) dose is within the constraints. Radiology

Inverse IMRT treatment-planning optimization is
usually performed based on a set of physical and/or
biological constraints, independent of the delivery
method. Leaf sequencing then creates a set of MLC leaf
positions/motions to generate an actual delivered
fluence which matches the optimal inverse fluence as
closely as possible. When converting the optimal plan, as
calculated by the planning system, into a deliverable
The work is supported by Cancer Research UK (under reference
grant SP 2313|0201) and the Institute of Cancer Research.
sequence of leaf positions (multiple-static fields, MSF or
dynamic delivery, DMLC), the dose delivered to the
planning target volume (PTV) and the organs at risk
(OAR) will change depending on the MLC leaf char-
acteristics taken into account during the leaf sequencing
[1, 2]. In the present study, a quantitative analysis of the
impact of IMRT delivery effects for 3 or 5 fields,
uniformly distributed around the patient, for multiple-
static fields (MSF), dynamic MLC or compensator
delivery techniques has been investigated, for the case
of two commercial planning systems and for two
anatomical sites.

The British Journal of Radiology, May 2006 401


Method
Overview of IMRT delivery techniques: MSF, DMLC
or Compensator
In converting the optimal fluence plan from the
planning system into a deliverable sequence of leaf
positions (MSF or DMLC) or compensator thickness,
different sequencing algorithms are used. To calculate
the leaf sequences for the MSF technique the ADAC
(now Philips) Pinnacle planning system uses the K-
means clustering algorithm [3]. This breaks the optimal
fluence into smaller groups or clusters of equal value.
For the compensator, the density of the compensation
material, width, height, resolution for the compensator
and depth of the plane for the compensator shape
optimization are defined. The energy fluence is then
iteratively attenuated by the corresponding thickness of
the modifier at each optimization fluence point. A
‘‘granular compensator’’ was selected with spatial
resolution of 0.05 cm and material density of
4.9 g cm23 (approximate density of standard steel
granulate compensator material). The source-to-
compensator distance was 56.6 cm and the compensator
was allowed a maximum thickness of 10 cm.
In the case of the DMLC technique, the optimal
fluences are converted to the actual fluences (Varian
terms) using the leaf motion calculator (LMC), which
designs the leaf motion patterns. The LMC takes into
account the various MLC parameters such as maximum
leaf span (i.e. the physical length of the leaf), leaf speed,
transmission, rounded-end effects and minimum leaf
gaps. Since the X and Y jaws do not move during beam-
on and hence cannot follow the trailing leaf, the
maximum leaf span will determine how many MLC
carriage positions (or static overlapping segments) will
be required to deliver the fluence for a given field width
(X jaws). The field is split into multiple overlapping
fields of the appropriate number of carriage or jaw
positions. Although the leaf motions are not fully
synchronised, the time of travel across the field is the
same for all leaf pairs, which helps to reduce the tongue-
and-groove effect [4].
Patient setup and treatment objectives
Oesophagus tumour site and Pinnacle planning
system
A patient with oesophageal carcinoma was planned
with an Elekta SL25 linac using 3- or 5-field IMRT plans
(gantry angles: 0 ˚ (anterior field), 120 ˚, 240 ˚ and 0 ˚, 72 ˚,
144 ˚, 216 ˚ and 288 ˚), respectively, using P3IMRT
(Pinnacle, ADAC version 6.0 g). In version 6.0 g, IMRT
delivery effects have not been included in the optimiza-
tion process.
The clinical target volumes (CTV), spinal cord and
lung parenchyma were outlined on each image. The CTV
region included both the oesophagus tumour and
adjacent lymph nodes. The PTV region was generated
by adding a three-dimensional margin of 15 mm to the
CTV to account for movement and target definition
uncertainties. The goal of the plan was to deliver 55 Gy
402
J Seco, C H Clark, P M Evans and S Webb
to the PTV, while maintaining the spinal cord dose at less
than 45 Gy and minimizing the dose to the lungs. The
spinal cord dose constraint of 45 Gy is conservative.
Martel et al [5] and Emami et al [6] have shown that, in
head and neck cancers, the tolerance dose for the spinal
cord is around 50 Gy, with a 5% chance of a complication
occurring in 5 years. None of the patients treated by
Martel et al [5] developed radiation myelitis, with the
spinal cord receiving doses up to 50 Gy.
In addition to the spinal-cord dose constraint, the
volume of lung irradiated to 18 Gy has been used as a
planning constraint at the Royal Marsden NHS
Foundation Trust. In the present study, no more than
20% of the lung could receive more than 18 Gy (V18,
20%). The prediction of lung complications at the
treatment-planning stage is not straightforward. There
is no consensus on which dosimetric parameter should
be used to reflect the clinical incidence of pneumonitis;
however the volume of lung receiving 18 Gy was chosen
[7]. Cardiac radiation toxicity for carcinoma of the
oesophagus is not a major clinical concern because of
the small number of long-term survivors. Therefore, the
heart was not included as one of the OARs.
Prostate and pelvic nodes tumour site and Helios/
CadPlan planning system
The patient with prostate carcinoma was planned with
a 5-beam technique using CadPlan version 6.3.5, with the
Helios inverse-planning module. Gantry angles of 180 ˚
(posterior), 270 ˚ (right lateral), 325 ˚ (right anterior
oblique), 35 ˚ (left anterior oblique), 100 ˚ (left posterior
oblique) were chosen such that the beams were
approximately equispaced and were not opposing.
These beam angles were found to provide good bowel
sparing. The treatment was designed to deliver a dose of
70 Gy to the prostate and 50 Gy to the seminal vesicles
and pelvic nodes. The prostate CTV was considered to be
the entire visible prostate and was grown to a PTV with a
1 cm margin. However, if the overlap between the PTV
and rectum was large, then the posterior margin was
reduced to 8 mm. The nodal CTV was expanded to a
PTV with a uniform 5 mm margin. The dose constraints
used as goals for the prostate and pelvic node treatment
were given by Clark et al [8].
For a prostate and pelvic node treatment with five
gantry angles, typical beam lengths were 16–18 cm and
beam widths were 10–18 cm. Typical monitor units
(MUs) were 95 (for a section of the field) and 135
(for a maximum-width single field). The prescribed dose
was 2.0 Gy per fraction to the median of the prostate
PTV [8].
Clinical impact of IMRT delivery effects
An ‘‘optimum’’ IMRT plan was obtained from each of
the commercial planning systems used. The optimum
IMRT fluence profiles were then leaf-sequenced in order
to generate the leaf positions or compensator thicknesses,
to allow the delivery of the planned profile. A step-and-
shoot K-means clustering algorithm from Pinnacle or the
dynamic leaf motion calculators from CadPlan/Helios
was used to leaf-sequence the IMRT profiles.
The British Journal of Radiology, May 2006
Quantitative study of IMRT delivery effects
Each delivered IMRT plan would deliver a slightly
different dose to the PTV, from that planned. The
delivered IMRT fluence maps were then renormalized
by the user such as to attain the prescribed dose to the
median of the PTV, i.e. 55 Gy for the oesophagus tumour
and 70 Gy for the prostate tumour. In the case of the
oesophagus tumour site, the impact of clustering the
IMRT fluence profiles was studied, where each fluence
map was divided into equal fluence levels, using an error
tolerance method equivalent to that described in Bär et al
[9]. The Pinnacle sequencing software allows the error
tolerance, i.e. the maximum difference between the
optimal and sequenced fluence maps, to be specified.
The effect of this was studied by specifying 2%, 5% and
10% tolerance levels.
In the case of the CadPlan/Helios planning system,
only dose–volume values were compared before and
after sequencing, for the secondary PTV and critical
organs (rectum, bladder, etc.), for three prostate patients
with nodal involvement because it was not possible to
retrieve dose or fluence information before leaf sequen-
cing, only dose–volume values. The dose–volume values
compared correspond to the dose delivered to 90%, 75%,
50%, 25% and 10% of the volume of the bladder, rectum
and small bowel and 95%, 75%, 50%, 25% and 5% of the
volume of the PTV (prostate), right and left nodes.
Results and discussion
Oesophagus tumour site and Pinnacle planning
system
Clinical impact of IMRT delivery effects on PTV
An ‘‘optimum’’ IMRT plan was obtained with the
Pinnacle planning system for a prescribed dose to the
PTV region of 55 Gy. The IMRT profiles were delivered
Figure 1. Dose–volume histograms (DVHs) of the optimum dose (OPT_PTV) or delivered dose to the planning target volume
(PTV) by (a) 3 and (b) 5 intensity-modulated radiotherapy (IMRT) plans. The delivery techniques represented are the (i) multiple
static fields (MSF) with 2%/compensator (red), 5% (blue) and 10% (green) error tolerance for the 3 beam and 5% (blue) and 10%
(green) for the 5 beam and the (ii) compensator (orange) for the 5 beam case.
The British Journal of Radiology, May 2006
using either the MSF or the compensator delivery
techniques. In Figure 1, the DVH obtained for the
delivered plans in case of 3 and 5 beams are compared
with the optimal plan for the PTV (vertical axis starts at
45 Gy). In delivering the optimum fluence profile, the
PTV dose has been ‘‘smeared’’ by both the MSF and the
compensator techniques.
In the case of the 3-beam delivery, the error tolerance
of 2% has produced the dose distribution that better
approximates the planned optimum (OPT_PTV). This is
a consequence of the greater number of fluence levels
allowed by the 2% than by the 5% and 10% cases (the
compensator was not shown for the 3-beam case since
the result was equal to the PTV_2% case).
In the case of the 5-beam plan, the delivered plans for
the MSF for 5% and 10% error tolerance and the
compensator are shown to ‘‘smear’’ the dose distribution
planned for the PTV. The 2% is not shown because the
system would ‘‘crash’’, destroying any data previously
calculated, possibly due to hardware and memory
limitations. The compensator has produced the best
delivered plan, of all the delivery methods represented
for 5 beams.
In Figure 2, the dose distribution obtained on a CT
slice of the patient for the 3-beam case, obtained for
(a) the optimum plan, and MSF delivery with (b) 2% and
(c) 5% error tolerance is shown. In Figure 2a, the 98%
isodose curve conforms closely to the PTV region
represented by the dark red circle. After delivery with
the MSF technique, the high 98% isodose covers part of
the adjacent lungs. In addition to this, the 80% isodose
has significantly changed from the planned distribution
(green curve in Figure 2a) to the delivered (green curve
in Figure 2b,c). The lungs have received increased dose
due to the inclusion of head-scatter and transmission.
The spinal cord is also receiving an increased dose as
may be observed comparing the 50% isodose (blue)
curves in Figure 2.
403

(a)
(b)
(c)
Figure 2. A standard oesophageal dose distribution
obtained from the (a) optimal and actual fluence delivered
with MSF for (b) 2% and (c) 5% error tolerance. The 98%,
80%, 50% and 20% isodose levels are shown, respectively, as
red, green, blue and yellow and the planning target volume
(PTV) is thick red.
404
J Seco, C H Clark, P M Evans and S Webb
Impact of IMRT delivery effects on OARs
The dose–volume histogram (DVH) for the spinal cord
is presented in Figure 3 for both the 3- and 5-beam cases.
In the case of the 3 beam, the spinal cord has received
significantly more dose than initially planned, due to
delivery effects. The increase in the maximum dose
delivered to the spinal cord is of the order of 2–3 Gy, in
addition to an overall increase of 5–10 Gy in the dose
delivered to 50% of the total volume of the spinal cord. In
the case of the 5 beams, a similar increase in the spinal
cord dose is observed.
In Figure 4, the DVH obtained for the left and right
lungs are presented. The maximum dose delivered to the
lungs increases, respectively, by 1 Gy and 3 Gy for the 3
and 5 beams. For the case of 5 beams, the dose delivered
to 50% of the lung volume varies dramatically with the
tolerance error used to segment modulated beam
profiles. However, for the three beams, there is no
significant difference between the tolerance errors in
terms of the delivered dose to the lungs by the MSF
technique. In the case of the 5 beams, the delivered doses
to the lungs by the MSF (5% and 10%) or the
compensator are distinctly different. As shown in these
results, the IMRT delivery techniques strongly influence
the delivered dose distributions to the tumour site
(oesophagus) and the surrounding OARs (spinal cord
and lungs). The overall increase in 1–3 Gy in the dose
delivered to the spinal cord and lungs corresponds to a
fractional increase of approximately 1.8–5.4% of the total
planned PTV dose of 55 Gy. This increase is a conse-
quence of the additional head scatter and transmission
fluence that is given to the patient after delivering the
optimum IMRT profile. The additional head scatter and
transmission fluence due to these MLC fields is not
accounted for in the optimization.
This increase in the delivered IMRT fluence due to
head scatter and transmission is consistent with that
observed by Seco et al [1] when delivering optimum
profiles for prostate tumours with the ELEKTA SL25
MLC. In Seco et al [1], the head scatter and transmission
contribution for the MSF technique led to an increase of
3% in the delivered IMRT fluence to the prostate tumour
and surrounding OARs.
Prostate and pelvic nodes and Helios planning
system
An ‘‘optimum’’ IMRT plan was obtained with the
Helios planning system with a prescribed dose to the
PTV region of 70 Gy to the primary prostate tumour and
48 Gy to the nodes. The ‘‘prescribed’’ dose to the nodes
was set at 48 Gy (not 50 Gy) at the planning stage, in
order to compensate for the increase in the nodal dose
after leaf-sequencing the optimum fluence. Therefore,
the final dose delivered to the nodes will be approxi-
mately equal to 50 Gy. The optimum profiles were then
leaf-sequenced in order to generate the leaf positions for
the DMLC delivery. The plan was then calculated to
allow the evaluation of the delivered profile and dose
distribution. Each delivered IMRT plan would yield a
different dose to the primary PTV (c.f. that determined
by the optimization), with subsequent increase in dose to
The British Journal of Radiology, May 2006
Quantitative study of IMRT delivery effects
Figure 3. Dose–volume histograms (DVHs) obtained for the spinal cord in the case of (a) the 3 and (b) the 5 beam plans. The
delivery techniques represented are the (i) multiple static fields (MSF) with 2% (red), 5% (blue) and 10% (green) error tolerance
for the 3 beam and 5% (blue) and 10% (green) for the 5 beam and the (ii) compensator (orange) for the 5 beam case.
the secondary PTV (nodes) and OARs (bladder, small
bowel and rectum).
The mean dose (over the group of three prostate
patients) delivered to the volumes of interest ‘‘before’’
and ‘‘after’’ leaf-sequencing are presented (Table 1), for
the various dose/volume points studied (the variation in
the delivered dose, between the ‘‘after’’ and ‘‘before’’
case, is also given after each volume of interest in
brackets). In addition to this, a typical DVH for a single
patient is also presented in Figure 5 to illustrate the effect
of leaf-sequencing on the final patient DVH. In the case
of the OARs there is an increase of up to 3.32 Gy in the
delivered dose to 50% of the total volume of the small
bowel (c.f. Table 1 and Figure 5). The bladder and
rectum are subject to 2.15 Gy and 3.03 Gy more dose,
respectively, delivered to 50% of their total volume. The
majority of the OARs volumes (bladder, small bowel and
rectum) have received significantly more dose delivered
to 90% of their volume. In the case of the rectum, this
increase was the largest, being in the order of 5 Gy to
90% of its total volume. In the case of the bladder and
small bowel, the increases in the dose delivered to 90% of
the total volume were 3.78 Gy and 3.06 Gy, respectively.
The greater increase in dose delivered to the OARs as
compared with the increase to the PTVs is due to the
extra transmission and scatter delivered to the OAR
during the times when they are shielded by the MLCs.
The proportion of time when the OARs are covered by
the MLCs is greater than for the targets and therefore the
increase in dose to the OARs is also greater.
For the secondary PTV (right and left nodes), the leaf-
sequencing of an IMRT ‘‘optimum’’ fluence profile (to be
delivered by an MLC) leads to an average increase of,
approximately, 2 Gy to 50% of the volume of the nodes,
while the dose to the prostate tumour is not significantly
altered. In addition, the dose to 90% of the nodal
volumes is increased by 3 Gy. The increase is due
mainly to head scatter and transmission radiation
associated with the MLC delivery equipment and that
The British Journal of Radiology, May 2006
is not accounted for by the Helios inverse planning
module.
In Van Esch et al [10], Clark et al [11] and Hong et al
[12] the transmission fluence for MLC leaves measured
contributed an additional 1.7%, approximately, to the
delivered fluence from Varian Linacs using the dynamic
mode of delivery. The increase of approximately 2 Gy
in the dose delivered to 50% volume of the pelvic
nodes (Table 1) is mainly constituted of approximately
1.7% of transmission fluence (leading to an extra 0.85 Gy
in deposited dose) and 2.3% of additional head scatter
fluence (leading to an extra 1.15 Gy in deposited
dose). The overall increase in the delivered dose is
observed for all the volumes of interest: bladder, small
bowel, rectum and nodes but not the primary PTV/
prostate, to which the dose is renormalized ‘‘after’’
leaf-sequencing.
Conclusions
The impact of IMRT delivery effects on two commer-
cial treatment-planning systems: Pinnacle planning
system (ADAC Laboratories Inc.) and CadPlan/Helios
(Varian Medical Systems) planning system, was evalu-
ated. The study was performed separately for the
oesophagus and the prostate (with pelvic nodes) tumour
sites, where multiple static fields and compensator
delivery techniques were evaluated for the first tumour
site and DMLC was evaluated for the second tumour
site.
In the case of the oesophagus tumour site and using
the Pinnacle planning system, the IMRT delivery effects
were shown to produce a ‘‘smearing’’ of the dose to the
PTV (oesophagus) and an increase of 2–3 Gy in the dose
delivered to the spinal cord. For the prostate (and pelvic
nodes) tumour site planned with the CadPlan/Helios
(Varian Medical Systems) system, an increase in
delivered dose of approximately 3 Gy was observed for
405
 J Seco, C H Clark, P M Evans and S Webb

Figure 4. Dose–volume histograms (DVHs) obtained for the left and right lungs in the case of the 3 and 5 beams intensity-
modulated radiotherapy (IMRT) plan.

Table 1. The mean dose (over a group of three patients with prostate and nodal planning target volumes (PTVs)) delivered to
the volumes of interest before and after leaf-sequencing. The difference in delivered dose, to the volumes of interest, between
before and after leaf-sequencing is also presented in brackets in Table 1B
(A) Mean dose ‘‘before’’ leaf sequencing

Dose delivered to volumes of interest (Gy)

Vol. Bladder Bowel Rectum PTV prostate PTV right node PTV left node

5% 71.00 53.33 65.33

10% 69.73 41.17 68.00
25% 62.37 31.50 51.00 50.17 50.50
50% 48.83 24.17 40.83 70.00 48.30 48.30
75% 39.70 13.67 36.10 46.43 46.50
90% 32.07 2.67 31.17
95% 67.33 43.67 43.83

406 The British Journal of Radiology, May 2006

Quantitative study of IMRT delivery effects

(B) Mean dose ‘‘after’’ leaf sequencing

Dose delivered to volumes of interest (Gy)

Vol. Bladder Bowel Rectum PTV prostate PTV right node PTV left node

5% 72.28 (+1.28) 53.62 (+0.29) 63.14 (22.20)

10% 68.24 (21.49) 41.80 (+0.63) 66.16 (21.84)
25% 61.85 (20.52) 33.74 (+2.24) 52.46 (+1.46) 51.62 (+1.45) 51.60 (+1.10)
50% 50.98 (+2.15) 27.49 (+3.32) 43.86 (+3.03) 69.77 (20.23) 50.34 (+2.04) 49.98 (+1.68)
75% 42.59 (+2.89) 15.24 (+1.58) 39.34 (+3.24) 48.90 (+2.47) 48.56 (+2.06)
90% 35.85 (+3.78) 5.73 (+3.06) 36.23 (+5.07)
95% 66.36 (20.98) 46.76 (+3.10) 46.48 (+2.65)

Figure 5. A dose–volume histogram (DVH) of an example patient with prostate carcinoma and nodal involvement (with right
(Node_RT) and left (Node_LT) nodes considered separately). The organs at risk considered are the bladder, small bowel and
rectum. The dose values are presented for the two cases: (i) before (BEF) shown as points and (ii) after (continuous lines) leaf
sequencing with the planning target volume (PTV) dose being normalized to 70 Gy. Arrows indicate the increase in the dose
delivered to the volume of interest due to leaf-sequencing effects.

50% of the total volume of the OARs: bladder, rectum References

and small bowel. In addition to this, an increase of
approximately 2 Gy was observed in the dose delivered
to the 50% of the total volume of the pelvic nodes, the
secondary PTV. The results obtained show that if
delivery effects are not accounted for at the planning/
optimization stage, then an increase in delivered dose to
the volumes of interest may be expected, after leaf
sequencing the optimum IMRT profiles. Currently, they
must be taken into account during the optimization stage
by altering the dose limits accepted during optimization
so that the final (sequenced) dose is within the desired
constraints.
1. Seco J, Evans P, Webb S. Analysis of the effects of the
delivery technique on an IMRT plan: comparison for
multiple static field, dynamic and NOMOS MIMiC collima-
tion. Phys Med Biol 2001;46:3073–87.
2. Seco J. Comparison of the efficacy of intensity modulated
radiotherapy delivered by competing technologies. PhD
thesis, University of London, 2002.
3. Hartigan JA. Clustering algorithms. New York, NY: Wiley,
1975.
4. Essers M, de Langen M, Dirkx MLP, Heijmen BJM.
Commissioning of a commercially available system for
intensity-modulated radiotherapy dose delivery with

The British Journal of Radiology, May 2006 407

 J Seco, C H Clark, P M Evans and S Webb

dynamic multileaf collimation. Radiother Oncol 9. Bär W, Alber M, Nusslin F. A variable fluence step
2001;60:215–24. clustering and segmentation algorithm for step and shoot
5. Martel MK, Eisbruch A, Lawrence TS, et al. Spinal cord IMRT. Phys Med Biol 2001;46:1997–2008.
dose from standard head and neck irradiation: implications 10. Van Esch A, Bohsung J, Sorvari P, et al. Acceptance tests
for three-dimensional treatment planning. Radiother Oncol and quality control (QC) procedures for the clinical
1997;47:185–9. implementation of intensity modulated radiotherapy
6. Emami B, Lyman J, Brown A, et al. Tolerance of normal (IMRT) using inverse planning and the sliding window
tissue to therapeutic irradiation. Int J Radiat Oncol Biol technique: experience from five radiotherapy departments.
Phys 1991;21:109–22. Radiother Oncol 2002;65:53–70.
7. Martel MK, Ten Haken RK, Hazuka MB, et al. Dose-volume 11. Clark CH, Bidmead AM, Mubata CD, et al. Intensity-
histogram and 3-D treatment planning evaluation of modulated radiotherapy improves target coverage, spinal
patients with pneumonitis. Int J Radiat Oncol Biol Phys cord sparing and allows dose escalation in patients with
1994;28:575–81. locally advanced cancer of the larynx. Radiother Oncol
8. Clark CH, Mubata CD, Meehan CA, et al. IMRT clinical 2004;79:189–98.
implementation: prostate and pelvic node irradiation using 12. Hong L, Alektiar K, Chui C, et al. IMRT of large fields:
Helios and a 120-leaf MLC. J Appl Clin Med Phys whole-abdomen irradiation. Int J Radiat Oncol Biol Phys
2002;3:273–84. 2002;54:278–89.

408 The British Journal of Radiology, May 2006

The British Journal of Radiology, 79 (2006), 409–414

Effect of ticlopidine in the prevention of radiation enteropathy

1
S AKYUREK, MD, 2L ATAHAN, MD, 2M CENGIZ, MD, 3C SOKMENSUER, MD, 4I HABERAL, MD, 2F YILDIZ,
1
MD and C ONAL, MD

1
Department of Radiation Oncology, Ankara University School of Medicine, Ankara, Departments
of 2Radiation Oncology and 3Pathology, Hacettepe University School of Medicine, Ankara and
4
ENT and Head and Neck Surgery Department, Health Ministry Ankara Education and Training
Hospital, Ankara, Turkey

ABSTRACT. Impairment of vascular function is considered to play an important role in

chronic radiation enteropathy. In this experimental study, the role of ticlopidine, an
inhibitor of ADP-induced platelet aggregation, was investigated in radiation
enteropathy. 80 male Wistar albino rats, each weighing 170–200 g, were divided into
four groups: (a) radiation alone (n520); (b) radiotherapy plus ticlopidine (n520);
(c) ticlopidine control (n520) and (d) control (n520). Both radiation groups received
19 Gy radiation to the exteriorized intestinal segments in a single fraction. Ticlopidine
or vehicle was administered 12 h after radiotherapy and continued for 1 month. Rats
from every group were euthanized randomly at intervals of 6 weeks from 2 weeks to 26
weeks. Histopathological radiation injury was assessed using radiation injury scoring
(RIS). Radiation with ticlopidine or radiation alone groups showed significant RIS
deterioration compared with controls in all time points studied. Comparison of median
RIS of radiotherapy and radiotherapy+ticlopidine groups at the 2nd, 14th and 26th
weeks yielded statistically significant RIS in favour of radiotherapy+ticlopidine group
(p50.05). However, these differences were less pronounced at the 8th and 20th week
(p50.07). Both radiation groups had poor weight gain when compared with control Received 17 March 2005
and ticlopidine groups. The weight gain in radiotherapy+ticlopidine group was Revised 22 June 2005
significantly superior to only radiation group between 10th and 20th weeks (p50.05). Accepted 30 August 2005
This study showed that inhibition of platelet aggregation with ticlopidine might be
DOI: 10.1259/bjr/16265085
useful in radiation enteropathy. However, the precise role of antiaggregant therapies
on radiation enteropathy should be comprehensively studied before clinical ’ 2006 The British Institute of
consideration. Radiology

Normal tissue damage is the main dose-limiting factor
in clinical radiotherapy. The small intestine is one of the
most radiosensitive organs and unavoidably included in
most abdominal or pelvic radiotherapy fields [1]. Despite
efforts to keep doses below tolerance doses, severe
chronic radiation enteropathy (CRE) occurs in about
10–15% of patients and has an adverse effect on the
quality of life of long term survivors [2, 3]. To date there
has been neither an effective method of preventing
radiation-induced intestinal injury nor the treatment
once it has occurred.
Intestinal radiation injury (radiation enteropathy) is
classified as either acute or chronic. Acute radiation
enteropathy is usually a self-limiting condition, which is
the result of the cell kill of the rapidly renewing
epithelial cells of the mucosa leading to a temporary
breakdown of the mucosal barrier and inflammation.
CRE, on the other hand, is associated with high rates of
morbidity and mortality [3, 4], and the underlying
molecular mechanisms have not been clearly identified
yet. The characteristic histopathological findings after
radiotherapy in acute phase are reported to take the form
of endothelial cell swelling, increased permeability and
interstitial fibrin deposition, and development of
platelet-fibrin thrombi in submucosal capillaries and
microvasculature has been accused of being the main
target for radiation damage. As the injury progress to
chronic radiation enteropathy, the prominent histopatho-
logical changes appear to be intestinal wall fibrosis and
vascular sclerosis [5–8].
It has previously been reported that CRE is associated
with significant upregulation of the transforming growth
factor b (TGF-b) and fibrogenic cytokine, and down-
regulation of endothelial cell surface protein, thrombomo-
dulin (TM) [9–12]. Thrombomodulin plays a key role in
protein C anticoagulant pathway. Reduced endothelial TM
levels and increased prothrombotic properties of endothe-
lium seem to contribute to hypercoagulation with
increased fibrin formation, tissue factor activity and platelet
aggregation [8, 13–15]. Platelets are the most important
sources of fibrogenic factors and cytokines, such as TGF-b
which is critically involved in radiation enteropathy [16,
17]. Adenosine 59-dipohosphate (ADP), on the other hand,
is a potent and specific mediator of platelet aggregation

The British Journal of Radiology, May 2006 409


and activation that is released by damaged endothelial
cells, red blood cells and activated platelets [18].
There are some encouraging data in the literature which
show that anticoagulation therapy, such as heparin,
warfarin, or acetyl salicylic acid (ASA) may be beneficial
in the treatment of radiation injury [19–21]. Ticlopidine is a
powerful platelet aggregation inhibitor recently used with
encouraging success in human clinical trials. It has been
reported to decrease recurrence and severity of cerebro-
vascular and cardiovascular accidents, peripheral arterial
disease, and to decrease the occlusion rates of coronary
saphenos vein bypass grafts [22]. Ticlopidine achieves
antiplatelet efficacy by blocking activation of platelets by
ADP. It selectively and irreversibly inhibits the binding of
ADP to its receptor on platelets, thereby effecting ADP-
dependent activation of the glycoprotein IIb/IIIa complex,
the major receptor for fibrinogen present on the platelet
surface [23]. The study reported here was designed to
determine whether ticlopidine has ability to prevent or
retard the development of radiation induced enteropathy
if it is used prophylactically.
Methods and materials
Animals
80 male Wistar rats, weighing 170–200 g (aged 8–12
weeks) at the time of irradiation, were used in this
experimental study. They were housed with five to a
cage and had free access to water and commercial food
pellets. All animals were kept under the same experi-
mental conditions.
Irradiation
Before irradiation, all rats including the controls were
anaesthetized with 80 mg kg21 ketamin hydrochloride
(Ketalar; Parke-Davis, Morris Plains, NJ) and 10 mg kg21
xylazin (Rompun; Milles Lab, Shawnee, KS). A surgical
procedure which has been described by Hauer-Jensen
et al [24] was performed to all rats, including sham-
irradiated animals, before radiotherapy. Briefly, a 1.5–
2 cm midline incision was made and 10 cm mid-small
intestinal segment, 30–40 cm proximal to the caecum,
was exteriorized and marked with silk ligatures in the
adjacent mesentery. Rats were placed in a specially
designed box, with the exteriorized intestinal segment
lying on gauze compress soaked in 0.9% saline, and care
was taken to avoid tension of the mesentery during
irradiation. Only the exteriorized intestinal segment was
irradiated and the rest of animal was protected by a 5 mm
lead shield. The intestine was again placed into the
abdomen at the end of the procedure.
The intestinal segment was irradiated with 6 MeV
electron beams using a Philips SL-25 linear accelerator
(Philips, Best, The Netherlands). Before irradiation,
animals were randomly assigned to one of the four
groups. A single fraction of 19 Gy was delivered through
a 6 cm66 cm anterior portal in group 1 (RT, n520).
Same dose was given in group 2 (RT+T, n520) with
100 mg kg21 daily oral ticlopidine HCl (Ticlid-Sanofi) for
1 month starting from 12 h after irradiation. Similar
410
S Akyurek, L Atahan, M Cengiz et al
procedures except irradiation were performed in 2 age-
matched control groups, a ticlopidine control group (T,
n520) and no-drug control group (C, n520).
All animals were followed up weekly for weight loss
and for complications. Experiments were performed in
accordance with the national regulations for animal
experimentation, and experimental protocols were
approved by the local animal ethics committee before
the start of studies.
Morphological assessment
Four animals from each group were randomly selected
and euthanized by bleeding under ether anaesthesia at
intervals of 6 weeks from 2 (early inflammatory phase) to
26 weeks (chronic, fibrotic phase). Specimens suturated
before irradiation procedure from irradiated and sham-
irradiated animals were excised and fixed in 4%
formaldehyde solution, dehydrated and embedded in
paraffin. After that cross sections and longitudinal
sections of the intestine were performed then
stained with haematoxylin-eosin for histopathological
evaluation.
Each specimen was examined with light microscopy
and histopathological changes were scored by a single
pathologist (CS) blinded to the study. Scoring (RIS) was
done by recording the macroscopical and 7 histopatho-
logical changes: mucosal ulcerations (0–2), epithelial
atypia (0–3), thickening of subserosa (0–3), vascular
sclerosis (0–3), intestinal wall fibrosis (0–3), ileitis cystica
profunda (0–3), and lymphatic congestion (0–1) within
the irradiated segment. The RIS has been shown to be a
reliable indicator of the severity of intestinal radiation
injury [3, 24–28]. Total RIS was calculated by adding the
scores of the individual parameters.
Statistical methods
The Mann-Whitney U-test was used to compare
median RIS values among the treatment groups.
Student t-test was used to compare mean weights¡
standard errors. A significance level of p,0.05 was
selected as the minimal level for significance.
Results
All animals were followed-up to 6 months after
treatment. All but eight rats survived during this period.
Five deaths (14th , 28th, 49th, 126th, 161st days) from RT
group and three deaths (14th, 42nd, 147th days ) from
RT+T group were found dead during the follow-up time.
Autopsies revealed intestinal perforation in three rats,
obstruction in two rats and haemoperitoneum for the
rest. There is no difference regarding to haemoperito-
neum between RT and RT+T groups.
Weight changes
Time-related weight changes for each group are shown
in Figure 1. Both radiation groups had significantly poor
The British Journal of Radiology, May 2006
The prevention of radiation enteropathy with ticlopidine
Figure 1. Weight changes as a function of time after
irradiation.
weight gain when compared with C and T groups. There
was no statistically significant difference between irra-
diated groups during the first 20 weeks, but a significant
difference (p,0.05) in favour of RT+T group was observed
between 10th and 20th weeks.
Morphological changes
There were no significant changes in control groups
except minor histopathological alterations, which were
considered to be due to ageing. The median RIS value, as
a function of time, is given in Figure 2 and Table 1.
At 2 weeks after irradiation, the RT group showed
deterioration in the crypt morphology, increase in
epithelial proliferation, superficial and deep ulceration,
and serosal thickening. Total RIS score was assessed as
11 in this group. The RT+T group on the other hand only
showed minimal crypt epithelial changes and subserosal
thickening due to increased amount of fibrin deposits
and oedema. Median RIS score was 6 and the difference
was statistically significant between these 2 groups
(p,0.05).
At 8 weeks after irradiation, macroscopic thickening of
the bowel wall was apparent in the RT group and
Figure 2. Median radiation injury scores as a function of
time after irradiation.
The British Journal of Radiology, May 2006
prominent vascular hyalinization in addition to histo-
pathological changes in the crypt morphology was
recorded. Median RIS score was 11. For the RT+T group,
macroscopic thickening of the bowel wall was similar to
the RT only group, whereas light microscopic evaluation
revealed only superficial mucosal ulcerations and serosal
thickening. Median RIS was 7, which was not statistically
significant, although there was a trend in favour of RT+T
(p50.07).
At the 14th week after irradiation, prominent bowel
wall thickening, stenosis and adhesions were recorded
macroscopically for RT only group. Severe mucosal
ulcerations, vascular sclerosis, hyalinization and appar-
ent serosal thickening were observed by light micro-
scopy and the median RIS was 14. RT+T group, on the
other hand, showed only wall thickening with superficial
ulceration, minimal serosal thickening and increase in
vascular hyalinization at this week. Median RIS was 6,
which was significantly better than the RT only group
(p,0.05).
At the 20th week, histopathological appearance in the
RT group was not remarkably different from the 14th
week of evaluation. The RT+T group on the other hand
showed some improvement on the mucosal crypt
morphology, although vascular hyalinization and sclero-
sis continued. There was a significant difference both
macroscopically and microscopically between the two
groups at the 26th week of evaluation. Apparent wall
thickening, adhesions and stenosis with similar light
microscopic findings to the previous weeks were
recorded for the RT group. However, although minimal,
some improvement was shown compared with the 20th
week for the RT+T group (Figures 3–6).
Discussion
The present study demonstrates that ticlopidine, an
inhibitor of ADP-induced platelet aggregation, has an
ability to reduce radiation induced damage in intestine
both in early and late phases in the animal model.
It is important to remember that the animal model
does not exactly reflect its human counterpart. Radiation
damage in different animal models can be quite different
in terms of severity and the presentation of the damage.
However, basic physiopathological chain of events and
the final appearance looks similar [29, 30]. The histo-
pathological appearance of chronic enteropathy in
patients is described as atropic mucosa, thick and fibrotic
submucosa and subserosa [31], which were similar to
animal model findings. We therefore assume that animal
model might be taken as representative of human
gastrointestinal toxicity.
Radiation induced morphological changes in the
intestine were reported to begin just after irradiation.
Progressive crypt shrinkage during the first 24–36 h after
irradiation was shown by Potten et al [32]. Recently,
radiation-induced apoptosis of endothelial cells were
reported to play an essential role in the pathogenesis of
acute gastrointestinal syndrome [33, 34] and it was
prevented when endothelial cell apoptosis was inhibited
[34]. Although vascular damage was not prominent at
2 weeks after irradiation in our study, vascular hyalini-
zation and sclerosis was apparent at 8 weeks in the RT
411
 S Akyurek, L Atahan, M Cengiz et al

Table 1. Median RIS values stratified by groups

2nd week 8th week 14th week 20th week 26th week

RT 10 (9–12) 11 (10–12) 13 (12–14) 13 (12–14) 12 (11–13)

RT+T 6 (6–7) 8 (7–9) 8 (4–10) 9 (8–9) 6 (6–6)
T 0 (0–0) 0 (0–0) 1 (0–1) 0 (0–0) 1 (0–2)
C 0 (0–0) 0 (0–0) 0 (0–0) 0 (0–0) 0 (0–0)
RT, radiotherapy alone group; RT+T, radiotherapy plus ticlopidine group; T, ticlopidine control group; C, control group

only group. Although endothelial cell injury was not

assessed separately, we think that vascular damage at 8
weeks supports the endothelial cell damage in the early
phase after irradiation. Our findings are consistent with
those of Jensen et al, who reported that vascular changes
were almost absent 2 weeks after irradiation, increased
during the following 8 weeks and stabilized 8 weeks
after irradiation [35].

Figure 5. Macroscopic appearance of the radiotherapy (RT)

Figure 3. Radiotherapy (RT) alone group. Specimen exhibits
extensive mucosal ulceration and vascular sclerosis, 26 weeks
after irradiation (vs, vascular sclerosis; ul, ulcer; s, serosa).
alone group 26 weeks after irradiation demonstrates
apparent bowel wall thickening, adhesion and stenosis.
Arrows show suturated specimen (vs, vascular sclerosis; ul,
ulcer; s, serosa).

Figure 6. Macroscopic appearance of the radiotherapy+ti-

Figure 4. Radiotherapy+ticlopidine (RT+T) group. Specimen clopidine (RT+T) group 26 weeks after irradiation demon-
exhibits mild thickening of muscularis mucosa, 26 weeks strates mild bowel wall thickening. Arrows show suturated
after irradiation (vs, vascular sclerosis; m, muscularis mucosa). specimen (vs, vascular sclerosis; m, muscularis mucosa).

412 The British Journal of Radiology, May 2006

The prevention of radiation enteropathy with ticlopidine
Radiation induced damage to the vasculature has long
been considered an important pathophysiological
mechanism of CRE [36]. The predominant damage
after irradiation has been reported to be in the
microvasculature causing endothelial cell swelling,
detachment of the endothelium from underlying matrix,
increased capillary permeability, progressive loss of
endothelial cells, and platelet adhesion and aggregation
[37, 38]. It has also been shown that radiation enhances
von Willebrand factor release, down-regulation of
endothelial TM and prostacyclin, which promote pro-
thrombotic properties. Some encouraging results were
reported, that administration of anticoagulation therapy
such as heparin, warfarin, or ASA may be effective
against radiation injury in some organs [19–21]. Glantz et
al studied heparin and warfarin in radiation induced
nervous system injury [20] and reported that anti-
coagulation therapy may reverse small vessel endothelial
injury, which is the fundamental lesion of radiation
necrosis, and produce clinical improvement in some
patients. Furthermore, Ludgate et al demonstrated
antiplatelet properties of ASA in radiation enteropathy,
whereas 5-amino salicylic acid, a potent anti-inflamma-
tory agent with minimal effect on platelet aggregation,
was found to be ineffective and possibly even harmful in
radiation enteropathy [39, 40]. These observations sup-
port the hypothesis that decreasing the platelet adhe-
siveness by antiaggregant therapy may produce the
decrease of microvascular thrombosis and improve the
microcirculation, finally decreasing late damage.
Antiplatelet agents in the prevention of CRE, on the
other hand, have been studied in the literature with
promising results [22, 23]. The role of these agents in
prevention are assumed to be directed on their activities
against thrombocytes, since a large number of studies
have shown that these drugs selectively inhibit ADP-
induced platelet activation by antagonizing the binding
of ADP to platelet receptors [23]. ADP elicits various
effects on platelets after binding to the specific receptor
for ADP, such as inhibition of adenyl cyclase, mobiliza-
tion of calcium from internal stores and expression of
fibrinogen receptor [23, 41, 42].
In a study by Wang et al, short-term clopidogrel (a
recently developed analogue of ticlopidine), administra-
tion starting 2 days before until 10 days after irradiation,
produced protection against early and, to a lesser extent,
delayed radiation enteropathy [43]. In this particular
report, the authors suggested that temporary inhibition
of platelet aggregation is not sufficient to permanently
interrupt the fibrogenic cycle responsible for intestinal
fibrosis, and that prolonged treatment or combination
therapies may be necessary. Contemporary to this study
we administered ticlopidine for 1 month. Median RIS
values in our study showed statistically significant
differences between RT and RT+T groups in the 2nd,
14th and 26th weeks of evaluation, which leads us to
think that the drug not only attenuates the early, but also
the late effects of irradiation. Platelets are the first
cellular elements that involved and initiate the haemo-
static and inflammatory responses and release a number
of proinflammatory and fibrinogenic mediators, and the
main effect of antiaggregants in prevention are assumed
to be directed by their antiplatelet activity. In recent
years, however, it was shown that ticlopidine and
The British Journal of Radiology, May 2006
clopidogrel increase endothelial nitric oxide and prosta-
cyclin production [42, 44–46] and modulate the contrac-
tile response of vascular smooth muscles [46]. So it may
be rational to think that the main effect is not only due to
antiaggregant property but also positive modulation of
endothelial cell function.
We used ticlopidine as a daily dose of 100 mg kg21
based on the preliminary experiments, in which
Yamamoto et al studied antithrombotic effect of ticlopi-
dine on He-Ne laser induced thrombus formation in rat
mesenteric microvessels [47]. In that experimental study,
Yamamoto et al reported that at a dose of 100 mg kg21,
ticlopidine inhibited thrombus formation both in arter-
ioles and venules and appeared to be more potent than
ASA. Our results confirm the efficacy of this dose in
terms of preventing radiation induced enteropathy.
Conclusion
This study supports that ticlopidine can ameliorate
early and chronic radiation enteropathy. The effect of
antiaggregant therapies on radiation enteropathy should
be comprehensively studied in phase II and III studies
before clinical considerations.
References
1. Berthrong M, Fajardo LF. Radiation injury in surgical
pathology. Part II. Alimentary tract. Am J Surg Pathol
1981;5:153–78.
2. Eifel PJ, Levenback C, Wharton JD, Oswald MJ. Time course
and incidence of late complications in patients treated with
radiation therapy for FIGO stage IB carcinoma of the
uterine cervix. Int J Radiat Oncol Biol Phys
1995;32:1289–300.
3. Hauer-Jensen M, Wang J, Denham JW. Bowel injury:
current and evolving management strategies. Semin
Radiat Oncol 2003;13:357–71.
4. Awward H. Late reacting tissues. In: Radiobiological and
physiological perspectives. Dordrecht, The Netherlands:
Kluver Academic Publishers 1990;395–420.
5. Eldor A, Vlodavsky I, HyAm E, Atzmon R, Fuks Z. The
effect of radiation on prostacyclin (PGI2) production by
cultured endothelial cells. Prostaglandins 1983;25:263–79.
6. Ts’ao C, Ward WF. Acute radiation effects on the
content and release of plasminogen activator activity in
cultured aortic endothelial cells. Radiat Res
1985;101:394–401.
7. Verheij M, Dewit LG, Boomgaard MN, Brinkman HJ, van
Mourik JA. Ionizing radiation enhances platelet adhesion to
the extracellular matrix of human endothelial cells by an
increase in the release of von Willebrand factor. Radiat Res
1994;137:202–7.
8. Zhou Q, Zhao Y, Li P, Bai X, Ruan C. Thrombomodulin as a
marker of radiation-induced endothelial cell injury. Radiat
Res 1992;131:285–9.
9. Langberg CW, Hauer-Jensen M, Sung CC, Kane CJ.
Expression of fibrogenic cytokines in rat small intestine
after fractionated irradiation. Radiother Oncol
1994;32:29–36.
10. Canney PA, Dean S. Transforming growth factor beta: a
promotor of late connective tissue injury following radio-
therapy? Br J Radiol 1990;63:620–3.
11. Richter KK, Langberg CW, Sung CC, Hauer-Jensen M.
Increased transforming growth factor beta (TGF-beta)
immunoreactivity is independently associated with chronic
413

injury in both consequential and primary radiation entero-
pathy. Int J Radiat Oncol Biol Phys 1997;39:187–95.
12. Richter KK, Langberg CW, Sung CC, Hauer-Jensen M.
Association of transforming growth factor beta (TGF-beta)
immunoreactivity with specific histopathologic lesions in
subacute and chronic experimental radiation enteropathy.
Radiother Oncol 1996;39:243–51.
13. Richter KK, Fink LM, Hughes BM, Sung CC, Hauer-Jensen
M. Is the loss of endothelial thrombomodulin involved in
the mechanism of chronicity in late radiation enteropathy?
Radiother Oncol 1997;44:65–71.
14. Dittman WA, Majerus PW. Structure and function of
thrombomodulin: a natural anticoagulant. Blood
1990;75:329–36.
15. Ohji T, Urano H, Shirahata A, Yamagishi M, Higashi K,
Gotoh S, et al. Transforming growth factor beta 1 and beta 2
induce down-modulation of thrombomodulin in human
umbilical vein endothelial cells. Thromb Haemost
1995;73:812–8.
16. Fava RA, Casey TT, Wilcox J, Pelton RW, Moses HL,
Nanney LB. Synthesis of transforming growth factor-beta 1
by megakaryocytes and its localization to megakaryocyte
and platelet alpha-granules. Blood 1990;76:1946–55.
17. Zheng H, Wang J, Koteliansky VE, Gotwals PJ, Hauer-
Jensen M. Recombinant soluble transforming growth factor
beta type II receptor ameliorates radiation enteropathy in
mice. Gastroenterology 2000;119:1286–96.
18. Mills DC. ADP receptors on platelets. Thromb Haemost
1996;76:835–56.
19. Hornsey S, Myers R, Jenkinson T. The reduction of
radiation damage to the spinal cord by post-irradiation
administration of vasoactive drugs. Int J Radiat Oncol Biol
Phys 1990;18:1437–42.
20. Glantz MJ, Burger PC, Friedman AH, Radtke RA, Massey
EW, Schold SC Jr. Treatment of radiation-induced nervous
system injury with heparin and warfarin. Neurology
1994;44:2020–7.
21. Mennie AT, Dalley V. Aspirin in radiation-induced diar-
rhoea. Lancet 1973;19:1131.
22. Saltiel E, Ward A. Ticlopidine. A review of its pharmaco-
dynamic and pharmacokinetic properties, and therapeutic
efficacy in platelet-dependent disease states. Drugs
1987;34:222–62.
23. Sharis PJ, Cannon CP, Loscalzo J. The antiplatelet effects of
ticlopidine and clopidogrel. Ann Intern Med
1998;129:394–405.
24. Hauer-Jensen M, Sauer T, Devik F, Nygaard K. Late
changes following single dose roentgen irradiation of rat
small intestine. Acta Radiol Oncol 1983;22:299–303.
25. Hauer-Jensen M, Poulakos L, Osborne JW. Intestinal
complications following accelerated fractionated x-irradia-
tion. An experimental study in the rat. Acta Oncol
1990;29:229–34.
26. Langberg CW, Sauer T, Reitan JB, Hauer-Jensen M.
Influence of fractionation schedule on development of
intestinal complications following localized irradiation.
An experimental study in the rat. Acta Oncol 1994;33:403–8.
27. Langberg CW, Sauer T, Reitan JB, Hauer-Jensen M.
Tolerance of rat small intestine to localized single dose
and fractionated irradiation. Acta Oncol 1992;31:781–7.
28. Langberg CW, Sauer T, Reitan JB, Hauer-Jensen M.
Relationship between intestinal fibrosis and histopathologic
and morphometric changes in consequential and late
radiation enteropathy. Acta Oncol 1996;35:81–7.
414
S Akyurek, L Atahan, M Cengiz et al
29. Skwarchuk MW, Travis EL. Murine strain differences in the
volume effects and incidence of radiation-induced colo-
rectal obstruction. Int J Radiat Oncol Biol Phys
1998;41:889–95.
30. Skwarchuk MW, Travis EL. Changes in histology and
fibrogenic cytokines in irradiated colorectum of two murine
strains. Int J Radiat Oncol Biol Phys 1998;42:169–78.
31. Hauer-Jensen M, Fink ML, Wang J. Radiation injury and the
protein C pathway. Crit Care Med 2004;32:325–30.
32. Potten CS, Owen G, Roberts SA. The temporal and spatial
changes in cell proliferation within the irradiated crypts of the
murine small intestine. Int J Radiat Biol 1990;57:185–99.
33. Maj JG, Paris F, Haimovitz-Friedman A, Venkatraman E,
Kolesnick R, Fuks Z. Microvascular function regulates
intestinal crypt response to radiation. Cancer Res
2003;63:4338–41.
34. Paris F, Fuks Z, Kang A, Capodieci P, Juan G, Ehleiter D, et
al. Endothelial apoptosis as the primary lesion initiating
intestinal radiation damage in mice. Science 2001;293:293–7.
35. Hauer Jensen M, Sauer T, Devik F, Nygaard K. Effects of
dose fractionation on late roentgen radiation damage of rat
small intestine. Acta Radiol Oncol 1983;22:381–4.
36. Fajardo LF. The complexity of endothelial cells. A review.
Am J Clin Pathol 1989;92:241–50.
37. Rubin DB, Drab EA, Ts’ao CH, Gardner D, Ward WF.
Prostacyclin synthesis in irradiated endothelial cells cul-
tured from bovine aorta. J Appl Physiol 1985;58:592–7.
38. Verheij ML, Dewit, van Mourik JA. Radiation-induced von
Willebrand factor release. Blood 1997;90:2109–10.
39. Baughan CA, Canney PA, Buchanan RB, Pickering RM. A
randomized trial to assess the efficacy of 5-aminosalicylic
acid for the prevention of radiation enteritis. Clin Oncol
1993;5:19–24.
40. Ludgate CM. Preliminary report: acetylsalicylic acid ther-
apy in the treatment of complications following abdominal
radiation. J Can Assoc Radiol 1985;36:138–40.
41. Sasaki Y, Ishii I, Giddings JC, Yamamoto J. Protective effects
of ticlopidine and aspirin, administered alone and in
combination, on thrombus formation in rat cerebral vessels.
Haemostasis 1996;26:150–6.
42. Jakubowski A, Chlopicki S, Olszanecki R, Jawien J,
Lomnicka M, Dupin JP, et al. Endothelial action of
thienopyridines and thienopyrimidinones in the isolated
guinea pig heart. Prostaglandins Leukot Essent Fatty Acids
2005;72:139–45.
43. Wang J, Albertson CM, Zheng H, Fink LM, Herbert JM,
Hauer-Jensen M. Short-term inhibition of ADP-induced
platelet aggregation by clopidogrel ameliorates radiation-
induced toxicity in rat small intestine. Thromb Haemost
2002;87:122–8.
44. Arrebola F, Zabiti S, Canizares FJ, Cubero MA, Crespo PV,
Fernandez-Segura E. Changes in intracellular sodium,
chlorine, and potassium concentrations in staurosporine-
induced apoptosis. J Cell Physiol 2005;204:500–7.
45. de Lorgeril M, Bordet JC, Salen P, Durbin S, Defreyn G,
Delaye J, et al. Ticlopidine increases nitric oxide generation
in heart-transplant recipients: a possible novel property of
ticlopidine. J Cardiovasc Pharmacol 1998;32:225–30.
46. Yang LH, Hoppensteadt D, Fareed J. Modulation of
vasoconstriction by clopidogrel and ticlopidine. Thromb
Res 1998;92:83–9.
47. Yamamoto J, Ishii I, Sasaki Y, Nagamatsu Y, Matsuda T,
Ando E. Antithrombotic effect of ticlopidine on He-Ne
laser-induced thrombus formation in rat mesenteric micro-
vessels. Haemostasis 1992;22:147–52.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 415–418

Radioprotective effects of hesperidin against gamma irradiation

in mouse bone marrow cells
1
S J HOSSEINIMEHR, PhD and 2A NEMATI, MSc

1
Department of Medicinal Chemistry, Faculty of Pharmacy, Mazandarn University of Medical
Sciences, Sari and 2Novin Medical Radiation Institute, Tehran, Iran

ABSTRACT. The radioprotective effects of hesperidin (HES), a flavonone glucoside, were

investigated by using the micronucleus test for anticlastogenic and cell proliferation
activity. A single intraperitoneal (ip) administration of hesperidin at doses of
10 mg kg21, 20 mg kg21, 40 mg kg21, 80 mg kg21 and 160 mg kg21 45 min prior to
gamma irradiation (2 Gy) reduced the frequencies of micronuleated polychromatic
erythrocytes (MnPCEs). All five doses of HES significantly reduced the frequencies of
MnPCEs and increased PCE/PCE+NCE ratio in mice bone marrow compared with non-
drug-treated irradiated control (p,0.0001). There was a drug dose–response effect of
Received 22 April 2005
HES in reducing MnPCE and increasing the PCE/PCE+NCE ratio in bone marrow cells. The Revised 20 September
maximum reduction in MnPCE was observed in mice treated with HES at a dose of 2005
80 mg kg21. The total MnPCE values were 2.85 fold less in the 80 mg kg21 HES group Accepted 12 October 2005
after being exposed to 2 Gy of c-rays than those in the respective irradiated control.
DOI: 10.1259/bjr/40692384
Our study demonstrates that hesperidin has powerful protective effects on the
radiation-induced DNA damage and on the decline in cell proliferation in mouse bone ’ 2006 The British Institute of
marrow. Radiology

Ionizing radiation generates reactive oxygen species in frequency of micronuclated polychromatic erythrocytes
the cells. These free radicals can induce damage to (MnPCEs) in mouse bone marrow exposed to gamma
critical macromolecules such as DNA. The cellular DNA rays.
damage leads to mutation and cancer [1]. High levels of
gamma irradiation can induce mortality in mammals.
With respect to radiation damage to humans, it is
important to protect biological systems from radiation- Materials and methods
induced genotoxicity or lethality. The main
radioprotective class is thiol synthetic compounds such Animals
as amifostine. Amifostine is a powerful radioprotective
Male NMRI mice weighing 25¡3 g were
agent compared with other agents, but this drug is
purchased from the Razi Institute (Tehran, Iran). Mice
limited in the use in clinical practice due to side effects
were housed in accordance to the principles outlined
and toxicity [2–4]. The search for less-toxic radiation
in ‘‘The Guide for The Care and Use of Laboratory
protectors has spurred interest in the development of
Animals’’ prepared by Tehran University of Medical
natural products. Recently, we reported that citrus
Sciences in the university animal house, and given
extract protects mouse bone marrow cells against gamma
standard mouse pellets and water ad libitum. All
irradiation. The citrus extracts contained high amounts
animals were kept under controlled lighting
of flavonoids [5]. Flavonoids are a family of polyphenolic
conditions (light:dark, 12:12 h) and temperature
compounds found in fruits and vegetables. Flavonoids
(22¡1 ˚C).
have wide biological properties including antibacterial,
antiviral, anticancer, immunostimulant and antioxidant
effects [6]. Hesperidin (HES) is a flavonone glycoside,
belonging to the flavonoid family. This natural product Chemicals and treatment
is found in citrus species. HES was reported to have
many biological effects including anti-inflammatory, HES was from Aldrich (USA). HES was dissolved
antimicrobial, anticarcinogenic and antioxidant effects, in phosphate buffered saline (pH 7.6). Mice were
and decreasing capillary fragility [7]. HES, in combina- injected intraperitoneally for all experiments. For
tion with a flavon called diosmin is used as DaflonH selection of the optimum dose of HES for radioprotec-
(Servier, France) to treat chronic venous insufficiency in tion, five doses (10 mg kg21, 20 mg kg21, 40 mg kg21,
Europe [8]. Other biological effects of HES are 80 mg kg21, 160 mg kg21) of HES were administrated
unknown. In continuation of this line of investigation, to the experimental animals 45 min before gamma
the in vivo radioprotective activity of HES was investi- irradiation. The control animals received the same
gated by using gamma rays as an oxidative DNA volume of normal saline or buffer phosphate. Five
damaging agent, and evaluating any reduction in the mice were used for each treatment group.

The British Journal of Radiology, May 2006 415

 S J Hosseinimehr and A Nemati

Irradiation Statistical analysis

Whole-body irradiation was performed with a cobalt-
60 c-radiation source (Teratron 780, Canada). Mice were
placed in ventilated PlexiglasH cages and irradiated in
groups of five mice, simultaneously. The source-to-skin
distance was 80 cm with a dose rate of 1.03 Gy min21 at
room temperature (23¡2 ˚C). The mice were irradiated
with a total dose of 2 Gy c-rays.
Micronucleus assay
The mouse bone marrow micronucleus test was
carried out according to the method described by
Schmid for evaluation of the chromosomal damage in
experimental animals [5, 9]. The animals were eutha-
nized by cervical dislocation 24 h after irradiation. The
bone marrow from both femurs was flushed in the form
of a fine suspension into a centrifuge tube containing
fetal calf serum (FCS). The cells were dispersed by gentle
pipetting and collected by centrifuge at 2000 rpm for
5 min at 4 ˚C. The cell pellet was resuspended in a drop of
FCS and bone marrow smears were prepared. The slides
were coded to avoid observer bias. After 24 h air-drying,
the smears were stained with May-Grunwald/Giemsa,
as described by Schmid. With this method, polychro-
matic erythrocytes (PCEs) stain reddish-blue and nor-
mochromatic erythrocytes (NCEs) stain orange, while
nuclear material is dark purple. For each experimental
point, five mice were used and a total of 5000 PCEs were
scored for each experimental point to determine the
percentage of micronucleated polychromatic erythro-
cytes (MnPCEs), micronucleated normochromatic ery-
throcytes (MnNCEs) and ratio of PCE to (PCE + NCE).
The ratio of PCE to (PCE + NCE) was determined
for each experimental group to assess radiation
effects with or without HES on bone marrow prolifera-
tion [10].
The data are presented as mean¡standard deviation
(SD). One-way analysis of variance (ANOVA) analysis
and Tukey’s HSD test were used for multiple compar-
isons of data.
Results
The effect of gamma irradiation with or without HES
on the induction of MnPCEs and the PCE/PCE+NCE
ratio in bone marrow cells, 24 h after c-irradiation, is
shown in Table 1. The frequency of micronuclei was
increased in all groups of mice irradiated with 2 Gy c-
irradiation compared with the control treated with
normal saline or phosphate buffer (p,0.0001). The
frequencies of MnPCE found in the HES treated groups
were significantly lower than that of the group treated
with radiation alone. The total MnPCE values were 1.15,
1.36, 1.66, 2.85 and 1.8 fold less in the 10 mg kg21,
20 mg kg21, 40 mg kg21, 80 mg kg21 and 160 mg kg21
HES group after being exposed to 2 Gy of c-rays,
respectively, than those in the respective irradiated
control (Figure 1). All five doses were effective in
significantly reducing (p,0.0001) the frequency of
MnPCE induced by 2 Gy irradiation and there was a
significant difference between the effects of various
doses of HES. There was a drug dose–response effect
of HES in the reduction of MnPCE in bone marrow cells.
The maximum reduced MnPCE was observed in mice
treated with HES at a dose of 80 mg kg21 (Figure 1). The
frequency of MnPCE in the latter group was 5.01¡0.32%,
much lower than in the group receiving radiation alone
(14.29¡0.5). With a further increase in the HES dose to
160 mg kg21, there was a reduced effect of HES on the
frequency of MnPCE induced by c-irradiation. The ratio
of PCE/PCE+NCE reduced significantly after exposure
to 2 Gy of c-irradiation (p,0.0001). Determination of

Table 1. Effects of hesperidin (HES) on the formation of radiation-induced micronulei polychromatic erythrocytes (PCEs) and
normochromatic erythrocytes (NCEs) and the ratio of PCE/PCE+NCE in mice bone marrow exposed to 2 Gy c-irradiation
Treatment MnPCE/PCE (%) MnNCE/NCE (%) PCE/PCE+NCE (%)

Control 0.236¡0.108 0.700¡0.300 62.872¡2.139

Normal saline 0.482¡0.239 0.718¡0.285 62.716¡3.160
Phosphate buffer 0.674¡0.295 1.43¡0.319 58.744¡2.643
Irradiation 14.292¡0.503 3.200¡0.661 39.010¡1.537
10 mg kg21 HES 12.422¡0.674 2.861¡0.440 42.194¡1.063
+irradiation
20 mg kg21 HES 10.943¡0.496 3.644¡0.760 44.800¡1.659
+irradiation
40 mg kg21 HES 8.604¡0.768 3.248¡1.096 50.046¡1.302
+irradiation
80 mg kg21 HES 5.012¡0.317 2.71¡0.549 54.658¡1.633
+irradiation
160 mg kg21 HES 7.957¡0.537 3.187¡0.489 49.974¡1.674
+irradiation
40 mg kg21 HES 0.755¡0.208 0.654¡0.234 57.338¡1.878
80 mg kg21 HES 0.786¡0.266 0.838¡0.229 55.796¡1.152
160 mg kg21 HES 2.405¡0.492 1.575¡0.338 49.910¡1.668
MnPCE, micronucleated polychromatic erythrocyte; MnNCE, micronucleated normochromatic erythrocyte.

416 The British Journal of Radiology, May 2006

Radioprotective effects of hesperidin
Figure 1. Effect of various doses of hesperidin (HES) on the
frequency of micronucleated polychromatic erythrocytes
(MnPCEs) in the bone marrow of mice exposed to c-radiation
(R) at a dose of 2 Gy.
ratio of PCE/PCE+NCE in the gamma irradiated mice
showed a pronounced cytotoxic effect of radiation on
bone marrow proliferation. Treatment of mice with HES
arrested the radiation-induced decline in the PCE/
PCE+NCE ratio (Table 1), and this increase in the PCE/
PCE+NCE ratio in the HES+irradiated group (at doses
20 mg kg21, 40 mg kg21, 80 mg kg21, 160 mg kg21) was
higher than that of the irradiated-alone group (p,0.001).
The highest PCE/PCE+NCE ratio was observed in HES
treated mice with 80 mg kg21 before c-irradiation
(Figure 2). There was no significant difference between
phosphate buffer-control and HES treated mice at this
dose before c-rays. In this study, HES did not indicate
any genotoxic and toxic effects at 40 mg kg21 and
80 mg kg21, but genotoxicity was observed at
160 mg kg21 compared with phosphate buffer (p,0.001).
Discussion
We previously reported that citrus extracts could pro-
tect mice bone marrow cells against gamma irradiation
and cyclophosphamide, when injected prior to exposure
and cyclophosphamide treatment. These citrus extracts
contain high levels of flavonoids [5, 11]. The results of
this study demonstrated the protective effects of HES, a
flavonone, against genotoxicity and toxicity induced by
Figure 2. Effects of various doses of hesperidin (HES) on the
radiation induced polychromatic erythrocyte (PCEs)/PCE+
normochromatic erythrocytes (NCEs) ratio in the bone
marrow of mice exposed to c-radiation (R) at dose 2 Gy.
The British Journal of Radiology, May 2006
c-irradiation in mouse bone marrow cells. However,
synthetic compounds, mainly thiol compounds, have
good radioprotective effects, but they are limited in their
use by side effects. Natural compounds, including
flavonoids, may play a role in scavenging free radicals,
such as hydroxyl radicals generated by c-rays in cells.
Ionizing radiation generates free radical damage in DNA
and induces genotoxic effects and death in the cells [1,
12]. There is a possibility that pre-treatment with
flavonoids could induce protection against oxidative
stress. Orientin and Vicenin, two flavonoids, protect
mice against chromosomal aberration induced by c-
irradiation when administrated before 2 Gy c-rays [13].
HES is the predominant flavonoid in lemons and
oranges. The peel and membranous parts of these fruits
have the highest HES concentrations. Sweet or navel
oranges (Citrus sinensis) and lemons (Citrus limon) are the
richest dietary sources of HES [14]. In this study, HES
had a dose-dependent protective effect on reducing
MnPCEs induced by gamma irradiation. The greatest
protective effect was observed at a dose of 80 mg kg21.
HES did not show any genotoxic or toxic effects at doses
up to 80 mg kg21 in mice bone marrow cells. Treating
mice with HES at dose 80 mg kg21 before exposure to
2 Gy radiation reduced the frequency of MnPCEs almost
2.85 fold. The percentage PCE/PCE+NCE ratio declined
in irradiated mice, since this ratio gives a direct index of
cell division. HES protected mice against radiation-
induced decline in cell proliferation, as evidenced by
the increased PCE/PCE+NCE ratio. The molecular
mechanism of the radioprotective effects of HES is not
clear. It has been reported that flavonoids have anti-
oxidant and chelating properties. These polyphenols are
excellent scavengers of free radicals due to the high
reactivity of their hydroxyl substituents [12]. Free-radical
scavenging is apparently responsible for the inhibitory
effect of flavonoids such as rutin, morin, quercetin and
genestin on the clastogenic activity induced by c-
irradiation in mice [15]. Among naturally occurring
flavonoids, HES has been pharmacologically evaluated
as a potential anticarcinogenic agent because of its
antioxidant activity [8, 16]. Other biological effects
include immune-modulation, treatment of venous insuf-
ficiency and scavenging of peroxynitrite as a reactive
oxidant [7, 17]. HES has also protected against photo-
induced breakage of DNA [18]. It is possible that HES
protects bone marrow cells with its antioxidant activity.
However, the antioxidant capacity of HES is not as high
as that of other flavonoids such as quercetin and
myricetin [19, 20]. Other mechanisms probably contri-
bute to its radioprotective effects. Thus, further experi-
ments are needed to explain the molecular mechanism of
HESs protective effects.
In conclusion, our results demonstrate that HES gives
significant protection to mice bone marrow against the
clastogenic effects of gamma irradiation.
Acknowledgments
We would like to thank Mr Mahmoudzadeh, Dr
Zahmatkesh, Dr Akhlaghpour and Dr Abassi from the
Novin Medical Radiation Institute for their assistance.
417

References
1. Reily PA. Free radicals in biology: oxidative stress and the
effect of ionizing radiation. Int J Radiat Biol 1994;65:27–33.
2. Hosseinimehr SJ, Shafiee A, Mozdarani H, Akhlagpour S.
Radioprotective effects of 2-iminothiazolidine derivatives
against lethal dose of gamma radiation in mice. J Radiat Res
2001;42:401–8.
3. Hosseinimehr SJ, Shafiee A, Mozdarani H, Akhlagpour S,
Froughizadeh M. Radioprotective effects of 2-imino-
3{(chromone-2-yl) carbonyl} thiazolidine against gamma
irradiation in mice. J Radiat Res 2002;43:293–300.
4. Turrisi AT, Glover DG, Hurwitz S, et al. The final reports of
the phase I trial of single dose WR-2721, s-2-(3-aminoprpy-
lamino)ethyl phosphorothioic acid. Cancer Treat Rep
1986;70:1389–93.
5. Hosseinimehr SJ, Tavakoli H, Pourheidari GR, Sobhani A,
Shafiee A. Radioprotective effects of citrus extract against
gamma irradiation in mouse bone marrow cell. J Radiat Res
2003;44:237–41.
6. Tiwari AK. Imbalance in antioxidant defense and human
diseases: multiple approach of natural antioxidants ther-
apy. Curr Sci 2001;81:1179–87.
7. Garg A, Garg S, Zaneveled JD, Singla AK. Chemistry and
pharmacology of the citrus bioflavonoid hesperidin.
Phytotherapy Res 2001;15:655–69.
8. Yang M, Tanaka T, Hirose Y, Deguchi T, Mori H, Kawada
Y. Chemopreventive effects of diosmin and hesperidin on
N-butyl-N-(4-hydroxybutyl)nitrosamine-induced urinary-
bladder carcinogenesis in male ICR mice. Int J Cancer
1997;73:719–24.
9. Schmid W. The micronucleus test. Mutat Res 1975;31:9–15.
10. Mozdarani H, Gharabi A. Radioprotective effects of
cimetidine in mouse bone marrow cells exposed to c-rays
418
S J Hosseinimehr and A Nemati
as assayed by micronucleus test. Int J Radiat Biol
1993;64:189–94.
11. Hosseinimehr SJ, Karami M. Citrus extract modulates
genotoxicity induced by cyclophosphamide in mice bone
marrow cells. J Pharm Pharmacol 2005;75:505–9.
12. Pietta PG. Flavonoids as antioxidants. J Nat Prod
2000;63:1035–42.
13. Devi PU, Bisht KS, Vinitha M. A comparative study of
radioprotection by ocimum flavonoids and synthetic
aminothiol protectors. Br J Radiol 1998;71:782–4.
14. Tomas-Barberan F, Clifford MN. Flavanones, chalcones and
dihydrochalcones-nature, occurrence and dietary burden. J
Sci Food Agric 2000;80:1073–80.
15. Shimoi K, Masuda S, Furugori M, Esaki S, Kinae N.
Radioprotective effect on antioxidative flavonoids in c-ray
irradiated mice. Carcinogenesis 1994;15:2669–72.
16. Tanaka T, Kohno H, Murakami M, Shimada R, Kagami S,
Sumida T, et al. Suppression of azoxymethane-induced
colon carcinogenesis in male F344 rats by mandarin juice in
b-cryptoxanthin and hesperidin. Int J Cancer
2000;88:146–50.
17. Kim JY, Jung KJ, Choi JS, Chung HY. Hesperetin: a potent
antioxidant against peroxynitrite. Free Radic Res
2004;38:761–9.
18. Yoshikawa Y, Suzuki M, Yamada N, Yoshikawa K. Double-
strand break of giant DNA: protection by glucosyl-
hesperidin as evidenced through direct observation on
individual DNA molecules. FEBS Lett 2004;566:39–42.
19. Rice-Evans CA, Miller NJ. Antioxidant activities of flavo-
noids as bioactive components of food. Biochem Soc Trans
1996;24:790–5.
20. Bonina F, Lanza M, Montenegro L, Puglisi C, Tomaino A.
Flavonoids as potential protective agents against photo-
oxidative skin damage. Int J Pharm 1996;145:87–94.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 419–424

Characteristics and predictive factors of early-onset diarrhoea

during pelvic irradiation
1,2 1 1
E Y HUANG, MD, C J WANG, MD, H C HSU, MD and 1L M SUN, MD, MS

1
Department of Radiation Oncology, Kaohsiung Chang Gung Memorial Hospital, 123 Ta-Pei Road,
Niao-Sung Hsiang, Kaohsiung Hsien, Taiwan and 2School of Traditional Chinese Medicine, Chang
Gung University, Taiwan

ABSTRACT. This study reported characteristics and predictive factors of early-onset

diarrhoea in patients receiving pelvic irradiation. We retrospectively reviewed 229
patients undergoing radiotherapy alone for cervical or uterine cancer. Oral barium was
taken 90 min before simulation. According to contrast medium within small intestine
only or colon in simulation films, we categorised patients as normal and rapid transit
groups. Small or large volume of small-bowel was also evaluated according to barium
distribution of simulation films. Whole-pelvic irradiation (39.6–45 Gy/22–25 fractions)
was delivered to all patients initially. We recorded the onset of diarrhoea during pelvic
irradiation. The rates of early-onset diarrhoea (,10 Gy) were compared between these
two groups. The incidence of diarrhoea before 10 Gy was 7% and 17% (p50.138) in
patients with normal and rapid transit, respectively. In multivariate analysis, interaction
among rapid transit, prior abdomen operation and large small-bowel volume
(p50.019) were noted for early-onset diarrhoea. Further subgroup analysis revealed
that rapid transit (p50.046) was a significant factor in patients with both prior
abdominal operation and large small-bowel volume. The incidence of early-onset
diarrhoea was as high as 40% in this particular group. Patients experiencing early-onset
diarrhoea had a higher incidence of moderate to severe diarrhoea (65%) than those Received 4 May 2005
without early-onset diarrhoea (23%) (p,0.001). In multivariate analysis, early-onset Revised 8 August 2005
diarrhoea was the only factor of moderate to severe diarrhoea (p50.001). In Accepted 17 August 2005
conclusion, rapid small-bowel transit may be predisposed to early-onset diarrhoea
DOI: 10.1259/bjr/51376226
during pelvic radiotherapy in patients with both prior abdominal operations and large
small-bowel volume. Early-onset diarrhoea is considered as a predictive factor of ’ 2006 The British Institute of
diarrhoea of a higher grade. Radiology

Radiation-induced acute diarrhoea is common during
abdominal and pelvic irradiation. It results from a
variety of different pathophysiological mechanisms,
including malabsorption of bile salts and lactose,
imbalances in local bacterial flora and changes in the
intestinal patterns of motility [1, 2]. Although villus
atrophy [3] of the small bowel is a key factor, the cause of
diarrhoea is not only pathological, but also a functional
change of the intestine. Changes in gastrointestinal (GI)
motility during fractionated irradiation precede the
appearance of histopathological lesions in the GI tract,
and the symptoms experienced particularly within the
first week of radiotherapy are directly related to changes
in bowel motility [4].
The onset of diarrhoea during pelvic radiotherapy
usually takes place in the second to third week of
radiotherapy with conventional fractionation. However,
unusually and inexplicably early onset of diarrhoea was
noted in a few patients. Morphological change of small
bowel can not explain the phenomenon. It is known that
laxatives such as castor oil can rapidly induce diarrhoea by
changing the small bowel motility [5]. Similarly, we
hypothesize that patients with rapid small-bowel transit
can be predisposed to early-onset diarrhoea during pelvic
radiotherapy. Interestingly, we used multivariate analysis
and noted interaction among rapid small-bowel transit,
prior abdominal operation and large small-bowel volume
within pelvis. The result is that it is worth investigating the
mechanism of early-onset diarrhoea in the future.
Materials and methods
Patients’ characteristics
Between January 1996 and January 2004, patients with
histologically proven cervix or uterine malignancies were
reviewed. The selection criteria of patients were four-field
whole-pelvic irradiation and oral barium contrast taken
90 min before simulation [6]. The exclusion criteria of
patients were: (1) receiving antispasmodic drugs, laxatives
or chemotherapy before or during the first week of
radiotherapy; (2) recent history of gastroenteritis; (3)
history of diabetes mellitus; or (4) history of pelvis
irradiation. Characteristics of 229 patients and treatment
are shown in Table 1. Orthogonal films including anterior–
posterior (AP) and lateral view were taken during
simulation. In the AP view, at source–axis distance
(SAD) 100 cm, x-axis size is the width (median: 15.5; range
13–18 cm) and y-axis size is the cephalocaudal distance

The British Journal of Radiology, May 2006 419

 E Y Huang, C J Wang, H C Hsu and L M Sun

Table 1. Characteristics of patients (n5229) conditions: (1) CTC Grade 2 or greater diarrhoea; (2) no
improvement of initial diarrhoea with minor antispas-
Parameters Number modic drug and change to major antispasmodic medica-
Age tion; (3) watery diarrhoea. The dose of onset of diarrhoea
,62 years 112 was recorded for actuarial analysis.
> 62 years 117
Diagnosis
Cervical cancer
Stage IB–IIA 76 Definition of small-bowel movement and volume
Stage IIB–IIIA 73 distribution
Stage IIIB–IVB 27
When barium filled the small intestine only, we defined
Post-operative 14
Recurrent 21 this as normal bowel transit. If barium could be noted
Endometrial cancer within the colon, rapid bowel transit was categorised. In
Stage I 4 order to evaluate small-bowel volume/distribution within
Stage II 3 the whole pelvis target, we used a simple volumetric
Stage III 6 method. We scored barium-filled amount of small bowel
No operation 2 within the radiation fields in orthogonal films. If caudal
Uterine sarcoma 3 small-bowel was below inferior aspect of sacroiliac (SI)
Prior abdomen operation joint and dorsal small-bowel was behind vertical line of
Yes 83
sacral promontory, large-volume distribution was defined.
No 146
Small-bowel volume Otherwise, the distribution was classified as small volume.
Small volume 138 Small or large volume presents corresponding upper/
Large volume 91 anterior or lower/posterior position of small bowel within
Small-bowel transit pelvis. The early-onset of diarrhoea was defined as
Normal 211 diarrhoea before 10 Gy.
Rapid 18

(median: 17; range 15–21 cm). In the lateral view, at SAD
100 cm, z-axis size is the ventrodorsal distance (median:
13; range 11–18 cm). The upper margin of AP film was
located at L4–5. The lateral margin was 1–1.5 cm beyond
the widest part of the pelvic brim. For lateral film, the
anterior margin was anterior part pubic symphysis. The
posterior margin is behind S2–3 at least.
Before radiotherapy, an interview of each patient was
performed for information of pelvic radiotherapy. We
had explained the possible toxicities of radiotherapy. We
informed each patient that she could ask for medication
at the onset of symptoms. After the beginning of
radiotherapy, we inspected patients weekly if no
progression of symptoms was noted, especially during
the first 1–2 weeks of radiotherapy. When symptoms
attacked or progressed, immediate medication was
prescribed to patients according to the bowel habits
and stool consistency recorded in the charts. If patients
had no improvement of initial diarrhoea with minor
antispasmodic drug such as mebeverine, we could
change to major antispasmodic medication such as
loperamide and the onset of moderate to severe
diarrhoea was recorded in the chart. Whole-pelvic
irradiation (39.6–45 Gy) of four-field technique was
initially administered with 10 MV or 15 MV photons
through supine position. The daily fraction dose was
1.8 Gy, five fractions weekly. We reviewed the onset of
diarrhoea in chart record. The severity of diarrhoea
could be measured according to common toxicity criteria
(CTC) [7]. Grade 1 is an increase of 2–3 stools per day
over pre-treatment. Grade 2 is an increase of 4–6 stools
per day, or nocturnal stools. Grade 3 is an increase of >7
stools per day or incontinence; or need for parenteral
support for dehydration. Grade 4 is physiological
consequences requiring intensive care; or haemody-
namic collapse. It was considered moderate to
severe diarrhoea when patients had any of the following
Statistics
Univariate analysis of early-onset and moderate to
severe diarrhoea rate was performed using Chi-square/
Fisher’s exact test. For multivariate analysis of early-
onset and moderate to severe diarrhoea, we used a
logistic regression model with stepwise forward proce-
dure. Age, rapid transit, small-bowel volume, and prior
abdominal surgery were evaluated as categorical data. A
p-value of less than 0.05 was considered to be statistically
significant. The risk of diarrhoea or large-volume
distribution was presented as odds ratio (OR) with 95%
confidence interval (CI). Data processing and statistics
were carried out on a personal computer using the
software SPSS 12.0 for Windows (SPSS Inc., Chicago, IL).
Results
Early-onset diarrhoea between normal and rapid
small bowel transit groups
There were 17 patients (7.4%) with diarrhoea before
10 Gy. No moderate to severe diarrhoea was noted
before 10 Gy. The result of univariate analysis is shown
in Table 2. The incidence of early-onset diarrhoea was
7% and 17% in patients with normal and rapid small
bowel transit (p50.138), respectively. The corresponding
rates were 4% and 40% (p50.068) in patients with both
large volume and prior abdominal operation.
Multivariate analysis of early-onset diarrhoea
among patients during pelvic irradiation
We considered the possible predisposed factors for the
analysis of early-onset diarrhoea. The results are shown

420 The British Journal of Radiology, May 2006

Early-onset diarrhoea after pelvic irradiation

Table 2. Univariate analysis of factors associated with early-onset and moderate to severe diarrhoea during pelvic irradiation
Parameters Early-onset p-value Moderate to severe p-value

Age > 62 years (yes vs no) 7% vs 8% 0.730 24% vs 28% 0.517

Prior abdominal operation (yes vs no) 8% vs 7% 0.660 33% vs 22% 0.078
Large small-bowel volume (yes vs no) 10% vs 6% 0.248 29% vs 24% 0.430
Rapid small-bowel transit (yes vs no) 17% vs 7% 0.138 28% vs 26% 0.785
Prior abdomen operation and large volume (both vs 10% vs 7% 0.459 38% vs 24% 0.109
another 3 groups)
Prior abdomen operation and rapid transit (both vs 29% vs 7% 0.087 43% vs 25% 0.378
another 3 groups)
Rapid transit and large volume (both vs another 3 18% vs 7% 0.192 27% vs 26% 1.000
groups)
Prior abdomen operation, large volume, and rapid 40% vs 7% 0.045* 40% vs 25% 0.605
transit (all vs another 7 groups)
Early-onset diarrhoea (yes vs no) 2 2 65% vs 23% ,0.001*
*Statistically significant.

in Table 3. If we did not consider interaction, no Discussion

significant factors were noted. Interestingly, we noted
interaction among operation, large volume and rapid
transit (OR: 9.29, 95% CI: 1.44–59.93) (p50.019).
Subgroup analysis revealed that rapid small-bowel
transit was a significant factor of early-onset diarrhoea
in patients with simultaneous large volume and prior
abdominal operation (OR: 15.33, 95% CI: 1.05–224.78)
(p50.046), but not another combinations.
Early-onset diarrhoea predicts moderate to severe
diarrhoea among patients during pelvic irradiation
Incidence of moderate to severe diarrhoea was 65%,
and 23% in patients with and without early-onset
diarrhoea (p,0.001), respectively. In multivariate analy-
sis, early-onset diarrhoea was the only factor of moderate
to severe diarrhoea (p50.001) (OR: 6.26, 95% CI: 2.20–
17.82) while age, small-bowel volume, prior abdominal
operation, and small-bowel transit were adjusted.
One of the most serious complications of abdominal or
pelvic radiotherapy is bowel damage, leading to nausea,
vomiting, diarrhoea and abdominal cramps during or
shortly after therapy. These symptoms are probably due
to a rapid modification of the intestinal motility and to
the structural alteration of the intestinal mucosa (cell loss
and altered crypt integrity) [2]. Clinically, fractionated
radiotherapy of the pelvis usually results in diarrhoea
while dose is around 10–20 Gy. The mechanism is villus
atrophy [3] caused by impaired crypt proliferation.
According to our results, the incidence of early-onset
diarrhoea was about 7%. It is difficult to explain
diarrhoea before 10 Gy with the model of cell loss and
altered crypt integrity because structural alteration of the
intestinal mucosa may not appear before 10 Gy. Hence,
small-bowel hypermotility may be considered for the
mechanism of early-onset diarrhoea.
The radiation-induced change of small-bowel motility
had been reported in some investigations. Yeoh et al
conducted a series of human and animal studies about

Table 3. Multivariate analysis of factors associated with early-onset and moderate to severe diarrhoea during pelvic irradiation
Parameters Early-onset Moderate to severe
a
p-value p-value p-value p-valuea

Age > 62 years 0.730 0.668 0.562 0.562

Prior abdominal operation 0.660 0.900 0.087 0.087
Large small-bowel volume 0.248 0.495 0.606 0.606
Rapid small-bowel transit 0.119 0.882 0.846 0.846
Interaction between operation and 2 0.600 2 0.137
large volume
Interaction between operation and rapid 2 0.704 2 0.609
transit
Interaction between rapid transit and 2 0.506 2 0.810
large volume
Interaction between operation, large 2 0.019*b 2 0.978
volume, and rapid transit
Early-onset diarrhoea 2 2 0.001*c 0.001*c
*Statistically significant.
a
Consideration of interaction.
b
Odds ratio (95% confidence interval)59.29 (1.44–59.93).
c
Odds ratio (95% confidence interval)56.26 (2.20–17.82).

The British Journal of Radiology, May 2006 421


small-bowel motility during or after pelvis/abdomen
irradiation. During radiotherapy, increased stool fre-
quency (p,0.001) was associated with more rapid small-
intestinal (p,0.01) and whole-gut (p,0.05) transit [8]. 1–6
years after radiotherapy, small intestinal transit (p,0.05)
was inversely related to stool frequency [9]. In an animal
study, there was an initial increase in frequency followed
by a non-significant reduction, but not in the amplitude
of ileal pressure waves following fractionated abdominal
irradiation (22.5 Gy in 9 fractions at 3 fractions/week).
There was no relationship between motility and histol-
ogy [10]. Erickson et al used an animal model with the
same fractionation that dramatically increased the
frequency of giant migrating contractions (GMCS) of
the small intestine, with the incidence peaking after the
second dose. The increased frequency of GMCS asso-
ciated with abdominal cramps and diarrhoea occurred as
early as a few hours after the first radiation fraction, and
returned to normal within days of cessation of radiation.
Altogether, hypermotility of small bowel is associated
with diarrhoea during fractionated irradiation.
Our multivariate analysis revealed interaction among
rapid transit, prior abdominal operation and large small-
bowel volume in pelvis. Subgroup analysis showed that
rapid transit was a significant factor of early-onset
diarrhoea in patients with both prior abdominal opera-
tion and large volume. Prior abdominal operation [11–
13] and small-bowel volume [11–15] are associated with
radiation-induced enteropathy. Our previous study
noted volume effect of small bowel for acute diarrhoea
in patients with prior abdominal operation [16]. When
patients have both factors, the possibility of large full-
dose volume is high. If small-bowel transit is rapid
before radiotherapy, there may be hypermotility within a
large volume of the small bowel irradiated. While water
can not be absorbed in this segment of small bowel,
early-onset diarrhoea may appear. Jankovic et al noted
that X-radiation produced acute sensitization (2 h) of rat
GI tract to acetylcholine and histamine [17]. Hence,
radiation-induced hypermotility effect of neurotransmit-
ters may be enhanced in patients with rapid transit.
The effect of abdominal operation in small bowel
distribution is important and associated with enterotoxi-
city. If small bowel is located in posterior pelvis (i.e.
large-volume distribution), patients with prior abdom-
inal operation may have a fixed loop of small bowel that
could be repeatedly exposed by fractionated radio-
therapy. Gallagher et al [11] reported that pelvic small-
bowel volume was larger after pelvic surgery, especially
abdominoperineal resection. In post-operative patients
with rectosigmoidal and endometrial cancers, 26–65% of
fixed small bowel was noted in the posterior pelvis [14,
18, 19]. Operation could exacerbate radiation-induced
acute diarrhoea in rectal cancer because small bowel
could be fixed in posterior pelvis by operation [12, 20].
We could calculate comparable large-volume distribu-
tion (35%) in our study. Limited movement of small
bowel within posterior pelvis caused by operation may
result in repeated large-volume and full-dose exposure
of small bowel. If these patients have rapid small-bowel
transit before radiotherapy, hypermotility of large
volume can result in early-onset diarrhoea.
Some methods of measurement for transit time of
small intestine were reported. Radioisotopes allow
422
E Y Huang, C J Wang, H C Hsu and L M Sun
accurate quantitation of the pattern and effectiveness of
the transit of chyme through the small and large
intestines [21, 22]. Birkebaek et al measured healthy
adult test subjects using a scintigraphic method with
111
In marked single unit tablets. The small intestine
transit time was 5 h (median) with an interquartile range
of 4–7 h [23]. Benmair et al used a magnetic method for
the measurement of small intestinal transit time in a
group of 20 normal subjects [24]. The mean transit time
was 157.5¡63.9 min. Pia de la Maza et al used hydrogen
breath test [25] and noted mean intestinal orocecal transit
time decreased from 1.7¡0.9 h to 1.36¡0.9 h after
5 weeks of radiation (p,0.05) [26]. The variations may
depend on the methodology. Simulation at 1.5 h after
barium intake seems to be applicable for mimicking
transit time in our study. Although oral contrast is not
the optimal method for transit time of small bowel, we
could still note increased incidence of early-onset
diarrhoea in patients with short transit time.
The severity of early-onset diarrhoea was mild in our
results. The role of early-onset diarrhoea in moderate to
severe diarrhoea was tested in our study. We noted that
early-onset of diarrhoea was predisposed to moderate to
severe diarrhoea in both univariate and multivariate
analysis. Patients experiencing early-onset of diarrhoea
may be radiosensitive. Hence, more accumulative
damage and more severe diarrhoea developed in those
who received further fractionated radiotherapy after
early-onset diarrhoea. The result suggests that more
severe diarrhoea may appear if early onset diarrhoea is
noted.
There are some flaws in the present study. The small-
bowel volume and transit were evaluated by orthogonal
simulation films. Although they are not standard
methods, they are acceptable for prediction. Because
diabetes mellitus [27–31] and chemotherapy [12, 32]
could influence diarrhoea, we excluded patients with
these factors and sample size was decreased.
Furthermore, the incidences of rapid transit and early-
onset diarrhoea were low. Hence, a large sample size is
needed to present statistical differences. Regardless, the
role of small-bowel transit was established while inter-
action was considered for multivariate analysis. This is
the first clinical result investigated about the association
between small-bowel transit and early-onset diarrhoea.
Prospective study with more sensitive evaluation for
small-bowel volume and transit is encouraged.
Although we focused on acute diarrhoea during whole
pelvis irradiation, late bowel complications were also
important because quality of life was impaired [33, 34].
Andreyev et al stated that most studies have used
inadequate assessments of GI toxicity [35]. The popula-
tion of our patients included those receiving operations
or not. Hence, the radiation dose, field size, and boost
technique varied after 39.6–45 Gy whole pelvis irradia-
tions. These factors limited effective assessment of
influence of early-onset diarrhoea in late bowel compli-
cations. However, the prospective study using more
comprehensive records of symptoms, such as diary and
questionnaire, is encouraged. Furthermore, our previous
study used early-onset diarrhoea as part of severity of
diarrhoea and showed correlation between acute and
late toxicity [36]. If early-onset diarrhoea appears,
the patient may have more severe diarrhoea during
The British Journal of Radiology, May 2006
Early-onset diarrhoea after pelvic irradiation
subsequent irradiation and increased incidence of late
bowel complications. More supportive care and dose/
target volume modification to avoid late complications
may be considered. In conclusion, early-onset diarrhoea
may be associated with rapid small-bowel transit in
patients with simultaneous prior abdominal operation
and large volume of small-bowel irradiated. Patients
with early-onset diarrhoea are predisposed to more
severe diarrhoea during subsequent pelvic radiotherapy.
References
1. Classen J, Belka C, Paulsen F, Budach W, Hoffmann W,
Bamberg M. Radiation-induced gastrointestinal toxicity.
Pathophysiology, approaches to treatment and prophylaxis.
Strahlenther Onkol 1998;174 Suppl. 3:82–4.
2. Somosy Z, Horvath G, Telbisz A, Rez G, Palfia Z.
Morphological aspects of ionizing radiation response of
small intestine. Micron 2002;33:167–78.
3. Carr KE, Hume SP, Ettarh R, et al. Radiation-induced
changes to epithelial and non-epithelial tissue. In: Dubois
A, King GL, Livergood DR, editors. Radiation and gastro-
intestinal tract. Boca Raton, FL: CRC Press, 1994:113.
4. Erickson BA, Otterson MF, Moulder JE, Sarna SK. Altered
motility causes the early gastrointestinal toxicity of irradia-
tion. Int J Radiat Oncol Biol Phys 1994;28:905–12.
5. Mathias JR, Martin JL, Burns TW, Carlson GM, Shields RP.
Ricinoleic acid effect on the electrical activity of the small
intestine in rabbits. J Clin Invest 1978;61:640–4.
6. Huh SJ, Lim DH, Ahn YC, Kim DY, Kim MK, Wu HG, et al.
Effect of customized small bowel displacement system in
pelvic irradiation. Int J Radiat Oncol Biol Phys 1998;40:623–7.
7. Trotti A, Byhardt R, Stetz J, Gwede C, Corn B, Fu K, et al.
Common toxicity criteria: version 2.0. An improved
reference for grading the acute effects of cancer treatment:
impact on radiotherapy. Int J Radiat Oncol Biol Phys
2000;47:13–47.
8. Yeoh E, Horowitz M, Russo A, Muecke T, Robb T, Maddox
A, et al. Effect of pelvic irradiation on gastrointestinal
function: a prospective longitudinal study. Am J Med
1993;95:397–406.
9. Yeoh E, Horowitz M, Russo A, Muecke T, Ahmad A, Robb
T, et al. A retrospective study of the effects of pelvic
irradiation for carcinoma of the cervix on gastrointestinal
function. Int J Radiat Oncol Biol Phys 1993;26:229–37.
10. Fraser R, Frisby C, Schirmer M, Blackshaw A, Langman J,
Yeoh E, et al. Effects of fractionated abdominal irradiation
on small intestinal motility--studies in a novel in vitro
animal model. Acta Oncol 1997;36:705–10.
11. Gallagher MJ, Brereton HD, Rostock RA, Zero JM, Zekoski
DA, Poyss LF, et al. A prospective study of treatment
techniques to minimize the volume of pelvic small bowel
with reduction of acute and late effects associated with
pelvic irradiation. Int J Radiat Oncol Biol Phys
1986;12:1565–73.
12. Minsky BD, Conti JA, Huang Y, Knopf K. Relationship of
acute gastrointestinal toxicity and the volume of irradiated
small bowel in patients receiving combined modality
therapy for rectal cancer. J Clin Oncol 1995;13:1409–16.
13. Nuyttens JJ, Robertson JM, Yan D, Martinez A. The position
and volume of the small bowel during adjuvant radiation
therapy for rectal cancer. Int J Radiat Oncol Biol Phys
2001;51:1271–80.
14. Capirci C, Polico C, Mandoliti G. Dislocation of small bowel
volume within box pelvic treatment fields, using new ‘‘up
down table’’ device. Int J Radiat Oncol Biol Phys 2001;51:
465–73.
15. Baglan KL, Frazier RC, Yan D, Huang RR, Martinez AA,
Robertson JM. The dose-volume relationship of acute small
The British Journal of Radiology, May 2006
bowel toxicity from concurrent 5-FU-based chemotherapy
and radiation therapy for rectal cancer. Int J Radiat Oncol
Biol Phys 2002;52:176–83.
16. Huang EY, Hsu HC, Yang KD, Lin H, Wang FS, Sun LM,
et al. Acute diarrhea during pelvic irradiation: is small-
bowel volume effect different in gynecologic patients with
prior abdomen operation or not? Gynecol Oncol 2005;97:
116–23.
17. Jankovic SM, Matovic M, Milaranovic D, Igrutinovic I.
Sensitization of rat gastrointestinal tract to acetylcholine
and histamine produced by X-radiation. Acta Physiol Hung
1997;85:215–30.
18. Brierley JD, Cummings BJ, Wong CS, McLean M, Cashell A,
Manter S. The variation of small bowel volume within the
pelvis before and during adjuvant radiation for rectal
cancer. Radiother Oncol 1994;31:110–6.
19. Green N. The avoidance of small intestine injury in
gynecologic cancer. Int J Radiat Oncol Biol Phys 1983;9:
1385–90.
20. Rodel C, Fietkau R, Keilholz L, Grabenbauer GG, Kessler H,
Martus P, et al. The acute toxicity of the simultaneous
radiochemotherapy of rectal carcinoma. Strahlenther Onkol
1997;173:415–21.
21. Kamm MA. The small intestine and colon: scintigraphic
quantitation of motility in health and disease. Eur J Nucl
Med 1992;19:902–12.
22. von der Ohe MR, Camilleri M. Measurement of small bowel
and colonic transit: indications and methods. Mayo Clin
Proc 1992;67:1169–79.
23. Birkebaek NH, Memmert K, Mortensen J, Dirksen H,
Christensen MF. Fractional gastrointestinal transit time:
intra- and interindividual variation. Nucl Med Commun
1990;11:247–52.
24. Benmair Y, Fischel B, Frei EH, Gilat T. Evaluation of a
magnetic method for the measurement of small intestinal
transit time. Am J Gastroenterol 1977;68:470–5.
25. Pereira SP, Khin-Maung U, Bolin TD, Duncombe VM,
Nyunt-Nyunt-Wai, Myo-Khin, et al. A pattern of breath
hydrogen excretion suggesting small bowel bacterial over-
growth in Burmese village children. Pediatr Gastroenterol
Nutr 1991;13:32–8.
26. Pia de la Maza M, Gotteland M, Ramirez C, Araya M, Yudin
T, Bunout D, et al. Acute nutritional and intestinal changes
after pelvic radiation. J Am Coll Nutr 2001;20:637–42.
27. Cuoco L, Montalto M, Jorizzo RA, Santarelli L, Arancio F,
Cammarota G, et al. Eradication of small intestinal bacterial
overgrowth and oro-cecal transit in diabetics. Hepato-
gastroenterology 2002;49:1582–6.
28. Abrahamsson H. Gastrointestinal motility disorders in
patients with diabetes mellitus. J Intern Med 1995;237:
403–9.
29. Bjornsson ES, Urbanavicius V, Eliasson B, Attvall S, Smith
U, Abrahamsson H. Effects of hyperglycemia on interdi-
gestive gastrointestinal motility in humans. Scand J
Gastroenterol 1994;29:1096–104.
30. Valdovinos MA, Camilleri M, Zimmerman BR. Chronic
diarrhea in diabetes mellitus: mechanisms and an approach
to diagnosis and treatment. Mayo Clin Proc 1993;68:
691–702.
31. Iber FL, Parveen S, Vandrunen M, Sood KB, Reza F,
Serlovsky R, et al. Relation of symptoms to impaired
stomach, small bowel, and colon motility in long-standing
diabetes. Dig Dis Sci 1993;38:45–50.
32. Miller RC, Martenson JA, Sargent DJ, Kahn MJ, Krook JE.
Acute treatment-related diarrhea during postoperative
adjuvant therapy for high-risk rectal carcinoma. Int J
Radiat Oncol Biol Phys 1998;41:593–8.
33. Gami B, Harrington K, Blake P, Dearnaley D, Tait D, Davies
J, et al. How patients manage gastrointestinal symptoms
423

after pelvic radiotherapy. Aliment Pharmacol Ther
2003;18:987–94.
34. McGough C, Baldwin C, Frost G, Andreyev HJ. Role of
nutritional intervention in patients treated with radio-
therapy for pelvic malignancy. Br J Cancer 2004;90:2278–87.
35. Andreyev HJ, Vlavianos P, Blake P, Dearnaley D, Norman
AR, Tait D. Gastrointestinal symptoms after pelvic
424
E Y Huang, C J Wang, H C Hsu and L M Sun
radiotherapy: role for the gastroenterologist? Int J Radiat
Oncol Biol Phys 2005;62:1464–71.
36. Wang CJ, Leung SW, Chen HC, Sun LM, Fang FM, Huang
EY, et al. The correlation of acute toxicity and late rectal
injury in radiotherapy for cervical carcinoma: evidence
suggestive of consequential late effect (CQLE). Int J Radiat
Oncol Biol Phys 1998;40:85–91.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 425–431

Optical density variations in CT films and their effect on image

quality
1 2 3
I A TSALAFOUTAS, PhD, G V PAPOUTSIS, RT, P N MANIATIS, MD and 4K A GOGOS, MSc

1
Medical Physics Unit, ‘Konstantopoulio - Agia Olga’ Hospital, 3-5 Agias Olgas, Nea Ionia, 142 33,
Athens, 2Field Service Engineering Department, General Electric Medical Systems, 156 Cyprou
Avenue and 91 Konstantinoupoleos Str, Argyroupolis, 164 51, Athens, 3Computed Tomography
Department, ‘Konstantopoulio - Agia Olga’ Hospital, 3-5 Agias Olgas, Nea Ionia, 142 33, Athens
and 4Nuclear Medicine Department, ‘Hygeia’ Hospital, 4 Erythrou Stavrou, 151 23, Maroussi,
Athens, Greece

ABSTRACT. It was recently reported that optical density (OD) variations were observed
in CT films printed with a laser camera, depending on the printing format and the
frame position within the film. The purpose of the present study was to investigate if
these variations are common to both laser and dry-film printers and if the different OD
settings along with day-to-day and frame-to-frame variations may affect the image
quality. Eight laser and five dry-film printers installed at 12 different CT facilities were
tested. For each one, the SMPTE test pattern was printed on all frames of a film using
the same printing format. The ODs of the 0%, 10%, 40% and 70% patches of the 11-
step greyscale of the SMPTE patterns were measured with a densitometer in all frames,
while all films were examined on a viewing box to assess subjectively the image quality
by visual inspection of the test pattern. A wide range of OD settings and variations
were recorded. Frame-to-frame variations in the same film of up to 0.19, 0.15 and 0.21
OD, were observed for contrast index (CI, the OD difference of patches 10% and 70%),
speed index (SI, the OD of patch 40%) and maximum OD (ODmax, the OD of patch 0%),
respectively. The variations were not always of the same magnitude, nor always
followed the same pattern, even for printers of the same model. Considering all films
and frames, the CI ranged from 1.26 to 1.74, the SI from 0.68 to 1.43 and the ODmax
Received 4 May 2005
from 2.5 to 3.11 OD, well beyond the proposed settings and tolerances of 1.55¡0.15, Revised 9 September 2005
1.15¡0.1 and 2.45¡0.1 given in the literature for CI, SI and ODmax, respectively. Despite Accepted 13 September
these large differences, the various problems that were identified in image quality from 2005
the visual inspection of the films could not be directly attributed to OD settings, as films
DOI: 10.1259/bjr/28579947
with similar CI, SI and ODmax presented quite different image quality levels. Therefore,
for routine quality control, thorough visual inspection of the SMPTE test pattern ’ 2006 The British Institute of
provides all the necessary information about the imaging chain status. Radiology

CT was first introduced into clinical practice in the
early 1970s and since then has become one of the most
useful diagnostic tools. The technology of CT scanners is
continuously and rapidly evolving while significant
developments have also occurred in CT film printers,
as laser and dry printers superseded CRT based multi-
format cameras (hard-copy cameras).
Nowadays, in CT departments, diagnosis is more and
more frequently performed on workstation monitors.
However, film still remains an important means for
diagnosing, storing and transmitting CT images. Indeed,
whilst monitors and films are both currently used for the
interpretation of CT images, in many cases diagnosis is
based solely on film reading. Since for viewing a CT
image, the window width (WW) and window level (WL)
are adjusted so that the structures of interest are best
imaged on the monitor, it is consequently important to
ensure that the image characteristics are maintained
when printed on film.
In order to test medical imaging systems and hardcopy
devices, the Society of Motion Picture and Television
Engineers (SMPTE) formed a committee that developed
and tested a number of patterns, until they agreed on the
final version of the SMPTE monochrome test pattern that
provides both qualitative and quantitative information
[1]. The characteristics of SMPTE pattern, its use for
acceptance and quality control (QC) purposes and
relevant technical details on the design of film printers
are summarized in the report of the American
Association of Physicists in Medicine (AAPM)
Diagnostic X-ray Imaging Committee Task Group No 1
[2].
In the AAPM report [2], reference values are given for
video and laser cameras concerning the optical densities
(OD) that certain patches of the 11-step greyscale of the
SMPTE pattern should have when printed on film, along
with the OD variation tolerances for the QC of these
systems. However, in a recent study [3] it has been

The British Journal of Radiology, May 2006 425

 I A Tsalafoutas, G V Papoutsis, P N Maniatis and K A Gogos
reported that the ODs of these patches vary with printing
format and with frame position within the same film.
In the present study, the frame-to-frame OD variations
within the same film were further investigated for a number
of laser and dry printers, to identify if these variations are
dependent on the type of film printer and whether being
combined with different OD settings and day-to-day
variations can affect the image quality in any way.
Materials and methods
In this study, 12 different CT facilities were included;
7 equipped with a laser and 4 with a dry-film printing
device, while in one facility two film printers (one laser
and one dry) were interfaced to the same CT scanner. All
CT scanners were manufactured by the same company
(GE Medical Systems, Milwaukee, WI), while film
printers were manufactured by Agfa (Agfa-Gevaert
A.G., München, Germany) and Kodak (Eastman Kodak
Company, New York, NY).
In each facility, the SMPTE test pattern was viewed on
the CT monitor, setting the WW to 100 and the WL to 0
(except for the Hispeed models where the WL was set to
1024 automatically when the SMPTE was imaged) [2].
The SMPTE test pattern is shown in Figure 1 and a
detailed description of its characteristics is given in the
AAPM report [2]. According to the suggestions of the
AAPM report, the brightness and contrast of the TV
monitor can be considered adequately adjusted when the
5% and 95% average picture level patches inset,
respectively, in the 0% and 100% patches (located at
the two ends of the 11-step greyscale) and the low and
high contrast bar-patterns (located at the centre and the
426
four corners of the SMPTE test pattern) are clearly
discernible. In all TV monitors examined, these criteria
were fulfilled and the adjustment of contrast or bright-
ness was not considered necessary.
Subsequently, and without changing the window
settings, the SMPTE pattern was printed so as to cover
the 20 frames of a film (5 rows by 4 columns). The OD
values of the 0%, 10%, 40% and 70% patches of the 11-
step greyscale SMPTE patterns of all frames were
measured with a calibrated optical densitometer (RMI
331 Densitometer; X-Rite, Grandville, MI). The OD of
patch 40% is defined as the speed index (SI), the OD
difference of patches 10% and 70% as the contrast index
(CI), while ODmax is the OD of patch 0% [2].
In two CT facilities, the above procedure was repeated
a few months after the first test in order to identify
possible day-to-day variations. In one CT facility, the test
procedure was carried out three times with the same
laser camera and three different film-types (after apply-
ing the necessary sensitometric adjustments), while for
one film type the test was also repeated a few months
after the first test. In the latter facility, in order to adjust
the photographic processor after film changes, the
procedure described in the user manual of the laser
camera was carried out. The laser camera produced a 21-
step greyscale setting the target density at the value of 3.3
OD, as proposed by the manufacturer. The maximum
OD (the OD of the 21st step) was measured with the
optical densitometer and was manually fed into the
camera. This procedure was repeated until a film with
maximum OD within ¡0.05 of the target density was
produced. Next the ODs of all steps were measured and
manually fed into the laser camera for updating the look-
up tables [2].
Figure 1. The SMPTE test pattern.
The British Journal of Radiology, May 2006
Optical density variations in CT films and image quality
In addition to the OD measurements, all the films were
observed on the same viewing box to assess the
qualitative characteristics of the printed SMPTE patterns.
Two observers graded in consensus the images for
contrast (in both low and high ODs) and resolution (in
both low and high contrast) using an arbitrary five-point
scale ranging from 1 (not discernible) to 5 (perfectly
discernible) and commented on the presence of any
type of artefacts. For contrast grading, the visibility of the
5% and 95% insets to the 0% and 100% patches,
respectively, was assessed. For low contrast resolution
grading, the visibility of the low contrast bar-patterns
was assessed; similarly, for high contrast resolution
grading, the high contrast bar-patterns located
beside the respective low contrast bar-patterns were
evaluated.
Results
Overall, 18 films acquired from 13 different printers
were examined. The mean values and the maximum
variation (max2min) of the CI, SI and ODmax recorded in
the 20 frames of each film are given in Table 1, along
with the results of the qualitative assessment. The tests
with decimal numbers indicate those tests made on the
same film printer, in order to investigate the impact of
film changes and day-to-day variations on frame-to-
frame OD variations and image quality.
Most of the observed OD variation patterns are
given in Figures 2–6. The variation patterns depicted in
Figure 2 refer to the same laser printer (Scopix LR 3300)
for 4 tests made with 3 different films. The correspond-
ing sensitometric curves (used for updating the look-
up tables) are depicted in Figure 7. In Figure 3, the OD
variation patterns of 6 laser printers of the same model
(Scopix LR 3300) are given (for printers 1 and 2, the
results from the tests that exhibited the largest variations
are depicted). In Figure 4, the OD variation patterns of 3
different laser printers (Scopix LR 3300, Scopix LR 5200,
Ektascan 2180) are given (for Scopix LR 3300, the results
from the test that exhibited the largest variations is
depicted). In Figure 5, the OD variation patterns of 5 dry
printers (three DryView 8100, one DryView 8700 and one
Drystar 3000) are given. Finally, in Figure 6, the OD
variation patterns of the laser and the dry printers that
exhibited the largest frame-to-frame variations (one
Scopix LR 3300 and one DryView 8100, respectively)
are given for comparison.
The OD variation patterns of CI, SI and ODmax given in
Figure 2 present some similarities. All these parameters
obtain their maximum values in the upper left frame and
then, when moving to the right, decrease to a minimum
(at the 3rd frame) and then increase again. This pattern is
repeated for each row and in some cases a slight decrease
from the first to the last row is also observed. The mean
values of CI, SI and ODmax are quite similar even if the
respective sensitometric curves are quite different.
However, the maximum variations in CI and ODmax
ranged from 0.07 to 0.15 and 0.06 to 0.16, respectively,
with no obvious correlation with the sensitometric
curves. It is important to note that when the film was
changed from Agfa to Ferrania, the sensitometric
curve (shown in Figure 7 with the dashed line) was
The British Journal of Radiology, May 2006
significantly faster than that of Agfa, obtaining a
maximum OD of 3.72 instead of the 3.3 target OD.
After the look-up tables were updated, the sensitometric
curve of the Ferrania film was found to be significantly
slower than that of Agfa film. However, the CI, SI and
ODmax were almost equal to those obtained with the
Agfa film.
It can also be deduced from the tests repeated with the
same film (tests 1.3–1.4, 2.1–2.2, 8.1–8.2) that day-to-day
variations did not significantly change the OD variation
pattern. However, differences in the maximum frame-to-
frame OD variations and the mean values of CI, SI and
ODmax were observed, as in the case of the Ektascan 2180
(tests 8.1–8.2) where the mean value of SI changed
between tests from 1.02 to 1.39.
In Figure 3, it is apparent that not all Scopix LR3300
presented the same OD variation pattern or the same
maximum variations in CI, SI and ODmax. In Figure 4,
the more extreme pattern observed for Scopix LR3300 is
compared with the less extreme patterns observed for
Scopix LR5200 and Ektascan 2180 laser printers. In
Figure 5, one can appreciate the variety of OD variation
patterns observed for the various dry printers tested.
Finally, in Figure 6 it is obvious that the variation
patterns of the laser and dry printers that exhibited the
largest CI, SI and ODmax frame-to-frame variations are
completely different.
As far as the qualitative assessment of images is
concerned, films with reduced contrast and
resolution, and various artefacts were identified. These
problems were common to all film frames and no
frame-to-frame variations in image quality were
observed, with one exception discussed in the end of
this section.
No problems were observed in the contrast in the high
ODs, as the 5% inset patch was always perfectly (grade 5)
or well (grade 4) discernable from the 0% background.
However, for low ODs there were three films (tests no. 5,
2.1 and 4) where the 95% inset patch was not discernable
at all (grade 1) or was poorly discernable (grade 3) from
the 100% background. While this could be attributed to
the very low SI of all three films, for the film from test
no. 2.2, which also exhibited a very low SI, the respective
contrast grade was 4.
On the other hand, the deterioration of low and high
contrast resolution observed in some of the films could
not be associated with OD settings. While in some cases a
reduced resolution could be attributed to other printer
problems causing geometrical distortion and back-
ground mottle (as in tests no. 2.1, 2.2, 9 and 11),
a reduced low contrast resolution was also observed in
the films from tests no. 8.1 and 8.2 with no obvious
reason.
Finally, it is worth mentioning that the only exception
in the image quality homogeneity among frames was the
film from test no. 4. In this film, alternating bands of
white-black and white-grey line pairs in the high contrast
bar patterns (an artefact we termed ‘‘banding’’) were
apparent in all frames. However, this artefact was more
intense in the outer left frames (the first frame of each
row) and specifically in the upper left and lower left high
contrast bar patterns. This was indeed the only case
where a definite variation of image quality among
frames was observed.
427
428

Table 1. Results of the optical density (OD) measurements and visual assessment of image quality in the 18 films studied. The mean values of contrast index (CI), speed index (SI)
and maximum OD (ODmax), and the maximum frame-to-frame differences (max2min) observed are shown. Mean values outside the ranges (1.55¡0.15, 1.15¡0.1 and 2.45¡0.1 for
CI, SI and ODmax, respectively) proposed for the quality control (QC) of a laser camera [2], are given in italics. The maximum frame-to-frame differences observed for laser (tests 8.1–
8.2) and dry printers (tests 9–13) are given in bold. The contrast grading in the high and low ODs (around the ODs of inset patches 5% and 95%, respectively) and the low contrast
(LC) and high contrast (HC) resolution grading are shown, as well as comments on the presence of artefacts
Test CT Cameraa Filmb CI max2min SI max2min ODmax max2min Contr. Contr. LC HC Comments
5% 95% Res. Res.

1.1 Prospeed Scopix LR3300 LT 2B daylight 1.53 0.15 1.05 0.08 2.87 0.16 5 5 4 5 –
1.2 Prospeed Scopix LR3300 Lifeimager HNB 1.56 0.07 1.06 0.08 2.90 0.09 5 5 4 5 –
1.3 Prospeed Scopix LR3300 KAP LP 670 A 1.65 0.09 1.07 0.09 2.91 0.06 5 5 5 4 ‘‘Banding’’c
1.4 Prospeed Scopix LR3300 KAP LP 670 A 1.58 0.1 1.08 0.06 2.90 0.12 5 5 5 4 –
2.1 Sytec Scopix LR3300 LT 2B daylight 1.29 0.08 0.86 0.04 2.59 0.1 5 3 2 3 Large GDd + ‘‘banding’’
2.2 Sytec Scopix LR3300 LT 2B daylight 1.36 0.08 0.88 0.04 2.68 0.1 5 4 2 3 Large GD + ‘‘banding’’
3 Hispeed Scopix LR3300 LT 2B daylight 1.58 0.19 1.19 0.15 2.99 0.21 5 4 5 4 ‘‘Banding’’
4 Prospeed Scopix LR3300 LT 2B daylight 1.36 0.17 0.78 0.07 2.59 0.19 5 3 4 5 Variable ‘‘banding’’, ‘‘over-
shoot’’e
5 Prospeed Scopix LR3300 LT 2B daylight 1.44 0.16 0.69 0.05 2.57 0.18 5 1 4 5 –
6 Hi-Speed Scopix LR3300 LT2B daylight 1.67 0.10 1.31 0.04 2.93 0.10 4 5 4 4 –
7 Prospeed Scopix LR5200 LT 2B daylight 1.69 0.08 1.18 0.04 3.02 0.1 4 5 4 5 Large ‘‘overshoot’’
8.1 Prospeed Ektascan 2180 EHG-10 1.43 0.1 1.02 0.04 2.85 0.18 5 5 2 4 –

I A Tsalafoutas, G V Papoutsis, P N Maniatis and K A Gogos

8.2 Prospeed Ektascan 2180 EHG-10 1.46 0.09 1.39 0.07 2.89 0.1 5 5 3 4 –
9 Prospeed DryView 8100 DVB 1.49 0.13 1.24 0.1 2.82 0.2 4 4 1 4 Bgndg mottle (horizontal
white stripes), ‘‘banding’’
10 Hi-Speed DryView 8100 DVB 1.63 0.06 1.23 0.1 3.01 0.08 4 5 4 4 ‘‘Banding’’
11f Prospeed DryView 8100 DVB 1.48 0.05 1.16 0.04 2.72 0.06 5 5 3 4 Bgnd mottle (thin vertical
white and grey lines)
12 Prospeed DryView 8700 DVB 1.65 0.07 1.02 0.1 2.66 0.07 4 4 4 5 Large ‘‘overshoot’’
The British Journal of Radiology, May 2006

13 Hispeed Drystar 3000 DT1B Dry 1.53 0.12 1.35 0.07 2.87 0.12 4 5 4 4 ‘‘Banding’’
a
Scopix LR 3300, Scopix LR 5200 and Drystar 3000 printers are manufactured by Agfa, while Ektascan 2180, DryView 8100 and DryView 8700 are manufactured by Kodak.
b
LT2B daylight and DT1B Dry films are manufactured by Agfa (Agfa–Gevaert N.V., Mortsel, Belgium), EHG-10 and DVB by Kodak (Eastman Kodak Company, New York, NY),
Lifeimager HNB by Ferrania (Ferrania S.p.A, Ferrania (SV), Italy) and KAP LP 670 A by Konica (Konica Corporation, Tokyo, Japan).
c
’’Banding’’ stands for the presence of alternating bands of white-black and white-grey line pairs in the high contrast bar patterns in horizontal, vertical or both directions.
d
GD stands for geometrical distortion (straight lines appear as curves).
e
In films with ‘‘overshoot,’’ the high contrast bar-patterns extend beyond their limits casting their shadow within the adjacent areas.
f
Test 11 was carried out in the same CT facility as tests 8.1 and 8.2.
g
Bgnd is an abbreviation for ‘‘Background’’ of the SMPTE test pattern where the OD is equal to that of the 50% patch.
Optical density variations in CT films and image quality
Figure 2. The variation of the contrast index (CI), speed
index (SI) and maximum optical density (ODmax) values with
frame numbers for tests 1.1 (Agfa film), 1.2 (Ferrania film),
1.3 and 1.4 (Konica film) performed on different days with
the same laser printer (Scopix LR3300, Agfa) at the same CT
installation. The frames have been numbered starting from
the upper left corner, and vertical lines indicate the start of
each of the 5 rows with 4 frames each.
Discussion
The AAPM report [2] gives mean values and accep-
table tolerances of CI, SI and ODmax as 1.55¡0.15,
1.15¡0.1 and 2.45¡0.1, respectively. These values,
however, refer to the periodic quality control (QC) of a
laser camera where the same format and frame position
are monitored using the Kodak SO-497 film. These serve
as a reference for the accepted OD settings and day-to-
day variations but are not applicable to frame-to-frame
variations. For the latter variations, the AAPM report [2]
recommends that OD differences of ¡0.05 or slightly
wider within one sheet of film may be accepted.
However, the report does not state whether exceeding
the proposed tolerances could have an effect on image
quality. Furthermore, as has been noted in a previous
study [3], limits have not been set for frame-to-frame
variations indicating a certain malfunction of the film
printer.
The reason for frame-to-frame variations has not yet
been identified. While some possible reasons have
been proposed by Agfa technicians concerning the
The British Journal of Radiology, May 2006
Figure 3. The variation of the contrast index (CI), speed
index (SI) and maximum optical density (ODmax) values with
frame numbers for tests performed at different CT installa-
tions which all had the same laser printer model
(Scopix LR3300, Agfa) and used the same film (LT 2B daylight,
Agfa).
results of our previous paper [3], in view of the
variations also observed in dry printers, where different
technologies are used for film exposing and processing,
none seem very convincing. The other image quality
problems identified in the films studied could be
attributed to a number of factors (e.g. the geometrical
distortion is most probably due to a worn polygon
mirror) and can usually be identified and repaired by the
camera manufacturer technicians, who usually utilize
other types of test patterns that the cameras are able to
produce. As written in the AAPM report [2], the SMPTE
test pattern can be used to identify image quality
problems, but not to trace the exact origin of the problem
within the imaging chain. This requires further investi-
gation by the camera (and perhaps the CT) technicians
and it is outside the scope of this paper. Whilst we are
eager to study the above issues further, this is not
currently possible because of the lack of detailed
information on design and sources of problems provided
by the companies concerned.
The mean CI values given in Table 1 (except for tests
2.1, 2.2 and 4) are within the reference range. However,
many of the SI and all the ODmax values given in Table 1
are outside the range of values given by the AAPM
429
 I A Tsalafoutas, G V Papoutsis, P N Maniatis and K A Gogos
Figure 4. The variation of the contrast index (CI), speed
index (SI) and maximum optical density (ODmax) values with
frame numbers for tests performed at different CT installa-
tions, using three laser printer models (3: Scopix LR3300-
Agfa; 7.1: Scopix LR5200-Agfa; 8.1: Ektascan 2180-Kodak).
report [2]. Thus, the printers and films investigated in
our study exhibit a larger variability in performance. The
maximum differences in CI, SI and ODmax observed in
the present study due to frame-to-frame variations
reached 0.19 OD, 0.15 OD and 0.21 OD, respectively
(film no. 3). Thus, if both day-to-day and frame-to-frame
variations are taken into account, the CI, SI and ODmax
on any frame for a specific film printer could well vary
within ¡0.25, ¡0.2 and ¡0.2, respectively, from the
central OD settings. It must be also noted that consider-
ing all the frames of all films, the CI ranged from 1.26 to
1.74 (0.48 OD difference), the SI from 0.68 to 1.43 (0.75
OD difference) and the ODmax from 2.5 to 3.11 (0.61 OD
difference). Despite these large OD variations, image
quality as assessed by visual inspection of the SMPTE
test pattern could not be correlated with the OD values.
The results of this study confirmed that frame-to-
frame OD variations were common among the different
printers tested, though were not always of the same
magnitude or pattern. However, an OD variation large
enough to reduce image quality only for certain frames
of the same film was not identified. Therefore, whatever
the frame-to-frame differences between ODs, if the
SMPTE test pattern characteristics remain clearly
and equally discernable in all frames during visual
430
Figure 5. The variation of the contrast index (CI), speed
index (SI) and maximum optical density (ODmax) values with
frame numbers for tests performed at different CT installa-
tions, using different dry printer models (9,10,11: DryView
8100-Kodak; 12: DryView 8700-Kodak; 13: Drystar 3000-
Agfa).
inspection, it should be safe to assume that the quality of
printed clinical CT images remains adequate for diag-
nosis.
In conclusion, we propose that for a certain CT
facility, once the reference values for the OD of all the
patches in the 11-step grey scale of the SMPTE test
pattern and the CI, SI and ODmax have been determined,
as well as the frame-to-frame OD variations (at least for
the most commonly used printing format), periodic
visual inspection of the SMPTE test pattern on the TV
monitor and on all the frames of a film should be
adequate for routine QC purposes. However, when
from the visual inspection a loss of image quality is
suspected, measurements of the ODs in a film fully
covered with the SMPTE test pattern should be carried
out to determine whether the camera should be re-
adjusted or repaired.
References
1. Specifications for medical diagnostic imaging test pattern for
television monitors and hard-copy recording cameras.
SMPTE Recommended Practice 1986, RP 133-1986. SMPTE
Journal 1986;95:693–5.
The British Journal of Radiology, May 2006
Optical density variations in CT films and image quality
Figure 6. The variation of the contrast index (CI), speed
index (SI) and maximum optical density (ODmax) values for
the laser and dry printers that exhibited the largest frame-to-
frame variations (3: Scopix LR3300-Agfa; 9: DryView 8100-
Kodak).
The British Journal of Radiology, May 2006
Figure 7. The sensitometric curves used for updating the
look-up tables before acquiring the SMPTE films (tests 1.1 to
1.4). The dashed line is the sensitometric curve that resulted
when the film changed from Agfa to Ferrania before
updating the look-up tables. The latter curve is included in
the graph to demonstrate the influence of look-up tables on
the sensitometric characteristics of CT films.
2. Gray JE, Anderson WF, Shaw CC, Shepard J, Zeremba LA,
Lin PP. Multiformat video and laser cameras: history, design
considerations, acceptance testing and quality control.
Report of AAPM Diagnostic X-ray Imaging Committee
Task Group No1. Med Phys 1993;20:427–38.
3. Tsalafoutas IA, Tsapaki V, Koulentianos E, Triantopoulou C.
Quality control of a laser camera with the SMPTE test
pattern: optical density variations with printing format and
frame position. Br J Radiol 2004;77:52–6.
431
 The British Journal of Radiology, 79 (2006), 432–436

SHORT COMMUNICATION

Assessment of environmental disturbances to the static magnetic

field in magnetic resonance installations
M A SCHMIDT, PhD

Department of Medical Physics, St George’s Hospital, Blackshaw Road, London SW17 0QT, UK

ABSTRACT. The static magnetic field of MRI scanners can be affected by environmental
factors. Magnetic resonance spectroscopy and functional imaging with single-shot
echo-planar imaging (EPI) are particularly vulnerable to the movement of lifts, vehicles,
trains and other large metallic masses in the vicinity. This work investigates the
sensitivity of two different imaging techniques to assess disturbances of the static
magnetic field: (i) phase changes in gradient-echo images of a uniform test object; and
(ii) image displacement along the phase encoding direction in single-shot EPI images.
For the latter a hexane sample was used, and the separation between CH2 and CH3
signals was taken as a reference. Both techniques were evaluated in a site known to be
free of any significant environmental disturbances and validated by inducing a
magnetic field disturbance. Both techniques provide valuable information in
acceptance tests, allowing MRI users to evaluate and manage the environmental Received 19 April 2005
conditions surrounding a scanner. The single-shot EPI technique was found to be highly Revised 26 October 2005
sensitive, being expected to detect magnetic field fluctuations down to 0.005 parts per Accepted 26 October 2005
million (ppm). The phase images method was found to be less sensitive (0.02 ppm) but
DOI: 10.1259/bjr/76396327
is more easily available. The single-shot EPI technique was used in acceptance tests and
environmental disturbances to the magnetic field of the order of 0.04 ppm were ’ 2006 The British Institute of
measured at the isocentre on two separate occasions. Radiology

MRI scanners are known to be affected by external
environmental factors. Examples are movement of lifts,
vehicles, trains and other large metallic objects, often
having some ferromagnetic content, which disturb the
static magnetic field surrounding the magnet isocentre.
Functional imaging and other techniques based on echo-
planar imaging (EPI) are particularly vulnerable: Durand
et al [1] have shown that changes in currents associated
with a rail line caused significant deterioration on
functional brain studies at 1.5 T. Peaks in magnetic
resonance spectra can also suffer considerable broaden-
ing, reducing magnetic resonance spectroscopy (MRS)
sensitivity.
MRI equipment suppliers advise on site planning and
on minimum distances between the magnet isocentre
and various sources of disturbance to the magnetic field.
Some MRI scanners are placed relatively close to possible
sources of disturbance, and checking for any environ-
mental effects then becomes a necessary part of the
acceptance test. It may also be necessary to verify the
stability of the magnetic field to aid the management of
the environment surrounding an MRI system throughout
its life-span.
At any given point within the magnet bore, a transient
environmental disturbance to the static magnetic field
will be perceived as a fluctuation of the Larmor
frequency. A second effect is a transient reduction of
the field homogeneity within a given volume. MRI
scanners always adjust the central frequency prior to
imaging, but the timescale of this automatic measure-
ment is invariably slow, not allowing the user to measure
a transient magnetic field fluctuation reliably. A loca-
lized measurement of the magnetic field fluctuation
could be made with a small sample, simply by
repeatedly acquiring a free induction decay (FID) signal.
However, in practice, this is extremely difficult to
undertake in acceptance tests, as the required pulse
sequence is often unavailable.
This work investigates the sensitivity of widely
available imaging techniques to evaluate disturbances
of the static magnetic field. Current shimming standards
allow the routine acquisition of brain 1H spectra with
2 Hz water line width at 1.5 T, and thus measurement
techniques must be sensitive to magnetic field variations
at least of the order of 0.01 parts per million (ppm). To be
truly useful in acceptance testing, a technique for the
measurement of changes to the static magnetic field must
not depend on any detailed knowledge of the pulse
sequences involved or other information to be provided
by the manufacturer. Also, it must not depend on
specific software packages, which may not be available.
It is also desirable to produce an objective record of any
transient changes in the static magnetic field intensity.
The latter is particularly relevant for disturbances which
are not very frequent, but are expected to be significant
(an infrequent train service in the vicinity, for example).

432 The British Journal of Radiology, May 2006

Short communication: Assessment of disturbances to the magnetic field in MR installations
Methods
Two separate imaging techniques were considered for
the quantification of magnetic field disturbances caused
by environmental factors, and their sensitivity was
evaluated. This work was undertaken at the Radiology
Department, St George’s Hospital (GE Signa/Echo-Speed
gradients, Milwaukee, WI). A 1.5 T MRI scanner has been
in operation on this site for 10 years and the site is known
to be free of any significant disturbances to the static
magnetic field due to the surrounding environment.
Phase images method
Changes in the magnetic field value are demonstrated
as phase changes on gradient-echo images, since the
term c.TE.b(t) is added to the image phase in the
presence of magnetic field fluctuations (c is the gyro-
magnetic ratio, TE is the echo-time and b(t) is an
unknown magnetic field fluctuation). This technique
can be used in any scanner, but some of the main
manufacturers require a research agreement or a service
password to produce phase, real and imaginary images.
Manufacturers also use different approaches for the
scaling of the phase images, sometimes using a threshold
from the magnitude images to avoid displaying the
random background phase. When the phase scaling for a
given scanner is not known, the phase must be calculated
from real and imaginary images. Another solution to
eliminate any ambiguity is to impose a known shift to the
central frequency, as this will cause a known phase
change, and acquire two images (before and after shift)
for reference.
A self-loaded uniform cylindrical test object with
19 cm diameter was employed (NiCl solution,
1.66 g l21, pH 4). A relatively short T1 was preferred to
maximize the signal-to-noise ratio (SNR) in rapid
imaging techniques. The phase change associated with
field fluctuations can be maximized by increasing TE,
but this slows down the measurement and leads to
poorer SNR. As a compromise, a series of 100 fast spoiled
gradient-echo images were acquired with the following
parameters: TE53.3 ms (shortest TE for a symmetrical
echo), repetition time (TR)57.3 ms, 24 cm field of view
(FOV), 10 mm slice thickness, 2566128 data acquisition
matrix and ¡32.25 kHz receiver bandwidth. These
parameters produce a 2p phase shift for a 4.4 ppm shift
on the Larmor frequency.
The first phase image acquired was taken as a
reference image, and all subsequent phase images were
subtracted from the reference image. A region of interest
(ROI) was chosen at the isocentre and the ROI mean
phase value was computed over the resulting series of
subtracted phase images. The range of ROI phase values
was used to estimate the sensitivity of the method.
Single-shot echo-planar imaging (EPI) method
Magnetic field fluctuations cause image shifts along the
phase encoding (PE) direction in a series of single-shot
spin-echo EPI images. The magnitude of the image shifts
depends on the PE bandwidth and, as a consequence, on
The British Journal of Radiology, May 2006
the interecho delay. However, details of the EPI pulse
sequence may not be known at the time the acceptance
test is performed, since manufacturers differ on the
amount of pulse sequence information provided as
standard. Again, a known shift in the central frequency
will produce a reference displacement, which can be used
to calibrate any other images where a displacement along
the PE direction occurs. Another solution is to use a
sample which contains more than one chemical shift, and
use the known separation of the two peaks as a reference.
The use of the single-shot EPI pulse sequence is wide-
spread, but it is not necessarily available in every scanner
as it is usually sold as a separate software package.
An 8 mm diameter test tube containing hexane (C6H14)
was employed, and the known chemical shift difference
between the CH2 and CH3 peaks (0.39 ppm) was used as
a reference. Both peaks have similar intensity and are
close enough to produce two clear images if the central
frequency is chosen as the midpoint between the two
peaks. A series of 100 single-shot spin-echo EPI images
was acquired at the rate of one image per second with
the following parameters: TE5100 ms, bandwidth
¡31.25 kHz, slice thickness 10 mm, FOV 24 cm624 cm,
producing a 2566256 image using a half-Fourier
acquisition. These parameters were chosen to make the
technique as sensitive as possible by increasing the PE
bandwidth. The two locations of the sample tube in the
images were determined by calculating their centre of
gravity, after thresholding. The distance between the two
images corresponding to the CH2 and CH3 peaks was
also calculated to provide a reference.
Both techniques described above can generate at least
one image per second, as this rate was considered the
minimum requirement to monitor magnetic field dis-
turbances associated with traffic, trains and lift move-
ment. Images can be acquired at a higher rate by
reducing the data matrix size along the PE direction.
When it was possible to control the potential source of
magnetic field disturbance, this control was exercised (by
controlling lift movement, for example). The images
were processed with in-house software written in IDL
(IDL 6.1; Research Systems Inc., Boulder, CO). Both
techniques were validated prior to use in acceptance
testing by bringing a strongly paramagnetic solution of
dimeglumine gadopentetate (concentrated Magnevist,
0.5 mol l21; Schering, Berlin, Germany) closer to the
measurement point and then removing it, creating a
magnetic field disturbance.
After validation of the imaging techniques, the
stability of the magnetic field was investigated at two
other 1.5 T MRI scanner sites as a part of acceptance
testing in new installations (Philips Intera/Explorer
gradients, Eindhoven, Netherlands and Siemens
Symphony/Sprint gradients, Erlangen, Germany).
Results
Sensitivity assessment
Phase images method
Figure 1a shows the first phase image of the test-
object, taken as a reference image, and Figure 1b shows
433
 M A Schmidt

the difference between one of the phase images in the average phase over the small central ROI indicated
series and the reference image. The maximum phase (Figure 1b), the range of variation was up to 0.02 ppm
change for any pixel over the object within the series of (0.003 ppm standard deviation). This method is therefore
100 images was under 0.25 ppm. Considering the not expected to be sensitive to field fluctuations below

Figure 1. (a) Reference phase image and (b) subtracted phase image of cylindrical uniform test object, with region of interest
(ROI) used for measurements indicated. The variation of the average ROI phase value over a series of subtracted phase images is
shown in (c), both in the absence of any magnetic field disturbance, and in the presence of a magnetic field disturbance caused
by bringing a strongly paramagnetic solution to the test object vicinity and then removing it.

434 The British Journal of Radiology, May 2006

Short communication: Assessment of disturbances to the magnetic field in MR installations

0.02 ppm. When a strongly paramagnetic solution is

brought closer to the test object, the ROI phase value
changes as shown in Figure 1c.

EPI method
Figure 2a–g shows sections of several single-shot EPI
images of the hexane sample. CH2 and CH3 peaks
produce separate images of the sample tube, of similar
intensities. The image shift along the phase encoding
direction can be seen from Figure 2d–f, as those were
acquired when a strongly paramagnetic solution was
brought close to the hexane sample, and then removed.
In the absence of any magnetic field fluctuations, the
position of the geometric centre of each tube image
changes by no more than 0.005 ppm (Figure 2h). The
standard deviation is as low as 0.0015 ppm. Figure 2i
shows how the position of the tube centre changes, for
both CH2 and CH3, when the magnetic field is disturbed
by moving a tube containing a strongly paramagnetic
solution in the vicinity. Despite some progressive
blurring associated with the central frequency offset,
the distance between the two peaks is kept constant in
the range of the measurement. Therefore the EPI
method is expected to be sensitive to magnetic field
variations of the order of 0.01 ppm, as required for
acceptance testing.

Acceptance testing
The EPI method with the hexane sample was
employed in two separate acceptance tests where the
site conditions were to be evaluated. Both sites were
equipped to perform functional brain studies with single
shot-EPI, and one of them was equipped with a spectro-
scopy package. Magnetic field fluctuations of approxi-
mately 0.04 ppm were detected at the magnet isocentre
during both tests. At one site the variation was
associated with the movement of heavy equipment
(image intensifier) along a shielded wall in the
corridor adjacent to the MRI scanner room. A sign was
placed to prevent that area from being used in a way
that could disturb MRI data acquisition. At the other site,
the variation was associated with car parking on
the pavement close to the external wall on the side of
the magnet room. The erection of a barrier was
suggested.

Discussion and conclusions

The single-shot EPI technique is the most sensitive of
the imaging techniques discussed as it is expected to
detect magnetic field fluctuations down to 0.005 ppm.
However, single-shot EPI may not be available in every
system, and it may not always be possible to decrease the
PE bandwidth to the same level of sensitivity used in
testing on the GE scanner as those parameters are clearly
undesirable for clinical applications. The toxicity of
hexane must also be mentioned, as it needs to be
transported to the site of the acceptance test with due
care.
The ‘‘phase images’’ method is less sensitive but more
easily available, and deserves further investigation and
Figure 2. (a–g) Sections of several single-shot echo-planar
imaging (EPI) images of hexane, showing the image shift
along the phase encoding (PE) direction associated with a
magnetic field disturbance. (d–f) A paramagnetic solution is
brought to the vicinity of the test object, and (g) then
removed. In the absence of any magnetic field disturbances
any measured image shifts are under 0.005 ppm, as shown in
(h) for the image associated with CH3. (i) The separation
between the CH3 and CH2 images remains constant even
when there is movement of a paramagnetic solution in the
vicinity of the isocentre.

The British Journal of Radiology, May 2006 435


optimization. It could be made more sensitive with a
higher SNR and using longer echo-times. If higher
sensitivity is achieved, the method could be used in
large uniform test objects (spherical or cylindrical) to
detect disturbance to the magnetic field homogeneity
over larger volumes.
Both imaging techniques discussed could be further
optimized, trading spatial resolution for a higher time
resolution. However, standard MRI pulse sequences
often impose restrictions to the minimum size of the
data acquisition matrix.
The interpretation of the results obtained with both
methods assumes that an actual drift of the Larmor
frequency is the only cause of the measured phase shifts
and offsets. However, these effects can easily be caused
by a number of different sources related to the scanner
hardware (electronic components warming up, for
example). For a scanner operating within specification,
hardware-related phase drifts and offsets should be of
very small magnitude, if detectable at all. Because of the
possibility of simple hardware malfunction, it is essential
to associate the measurements obtained with specific
environmental sources and to perform reproducible
measurements.
436
M A Schmidt
Both techniques discussed allow MRI users to evaluate
the environmental conditions surrounding an MRI
scanner, and can provide the basis for informed
discussion on the management of the area surrounding
it. The final decision on how much disturbance to the
static magnetic field can be tolerated is complex and
depends not only on the nature of the work to be
performed in a particular scanner, but also on the pattern
of the local magnetic field disturbance and its associated
frequency. The single-shot EPI method, in particular,
was proven to be highly sensitive and is strongly
recommended for acceptance testing.
Acknowledgment
The author would like to thank Dr Franklyn Howe for
very helpful discussions on MR spectroscopy.
References
1. Durand E, van de Moortele P-F, Pachot-Clouard M, Le Bihan
D. Artifact due to Bo fluctuations in fMRI: correction using
the k-space central line. Magn Reson Med 2001;46:198–201.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 437–440

SHORT COMMUNICATION

An investigation of search pattern extent in the threshold

contrast detection task
C J KOTRE, PhD

Regional Medical Physics Department, Newcastle General Hospital, Newcastle-upon-Tyne NE4 6BE,
UK

ABSTRACT. The threshold contrast-detail diameter test is used as a semi-quantitative

measure of image quality in radiology. This observation task is called ‘‘signal known
exactly/background known exactly’’ because the signals are usually low contrast disks in
known positions, and the background is uniform except for noise fluctuations. The
performance of the observer undertaking this task can to some extent be predicted
from knowledge of the noise power in the image background, and adoption of the
assumption that the noise is sampled through an aperture of the same area as the test
feature being observed. In order to extend this approach to optimization of clinical
images, the effect of the cluttered anatomical background on the detection task must Received 7 September
2005
be quantified. To study the effect on detection of nearby structure, a series of contrast- Revised 2 November 2005
detail tests was carried out using a progressively restricted background area of Gaussian Accepted 14 November
noise, and a range of object diameters. It was found that the observer’s ability to detect 2005
low contrast objects is progressively reduced as the area of the search area is reduced,
DOI: 10.1259/bjr/13489819
the difficulty of the task increasing rapidly as the diameter of the restricted search area
falls to less than twice that of the target disk. The results suggest the presence of a ’ 2006 The British Institute of
search pattern that scales in proportion with the size of the test feature. Radiology

The threshold contrast-detail diameter test has been
used for many years as a standard semi-quantitative test
of image quality especially in fluoroscopy, radiography
and mammography [1–3]. In this technique, the observer
is shown a set of disks imaged against a uniform
background. The disks vary both in contrast and
diameter and are laid out in a fixed pattern within the
test object. This type of stimulus is known as a ‘‘signal
known exactly/background known exactly’’ (SKE/BKE)
task. For each disk diameter the observer works along a
sequence of disks of decreasing contrast and decides
which is the last visible disk. The contrast of the last
visible disk is defined as the threshold contrast at that
disk diameter. The process is repeated over a range of
disk diameters and the results can be plotted as a
contrast-detail curve, which marks the transition from
what combination of size and contrast can be seen in the
image and what cannot be seen because it is masked by
noise.
The most widely accepted model of the behaviour of
the human visual system in undertaking the SKE/BKE
contrast-detail task is that the threshold contrast
obtained is equal to the noise observed in the back-
ground, multiplied by a constant, the threshold signal-to-
noise ratio, the value of which is thought to be around 2.5
[4]. The noise perceived by the observer is taken to be the
standard deviation of a number of luminance samples of
the background, where each sample is the mean
luminance within an aperture of the same area as the
disk under observation [5, 6]. This model broadly fits
experimental results when modified to take account of
further limitations to detection at small disk sizes due to
the finite point spread function of the eye, and at large
disk sizes due to the maximum extent of the photo-
receptor field over which summation of the noise sample
may occur [4, 6].
Although the SKE/BKE contrast-detail task has been
shown to be very useful in comparing the performance
of radiological image receptors, where the noise in the
uniform background is principally that from X-ray
quantum fluctuations plus components of system noise,
such as electronic amplification noise and the fixed noise
pattern of phosphor screens, there are difficulties in
extending the results to the prediction of detectability in
clinical images. The background in clinical images
consists of a wide variety of object shapes, sizes and
contrasts all of which tend to inhibit the detection of the
diagnostic signal required in any given examination. The
presence of the background can be thought of as a noise
source, sometimes termed structure noise or clutter, and
it is the removal of this noise component that allows
small contrast signals to be detected in digital subtraction
angiography. Although meaningful results can be
obtained in observational studies of contrast-detail test
objects overlying a complex anatomical type background
[7], it is difficult to extend this to the general problem of
dose optimization of clinical images. In order to under-
take optimization it will be necessary to be able to
quantify background structure as a noise source, and to
be able to predict its effect on observer performance. One

The British Journal of Radiology, May 2006 437


means of achieving this may be to design algorithms to
sample the radiological image in a similar way to a
human observer, so that the detectability of a low
contrast object can be analysed on a point-by-point basis
against the background of real clinical images. Although
the present work is based on the contrast-detail test,
other approaches to quantifying signal detection in
realistic radiological backgrounds have been explored,
notably alternative forced choice methods [8–10] and
receiver operating characteristic studies [11, 12].
As a starting point for the design of algorithms to
mimic an observer undertaking a contrast-detail task, it
is necessary to extend the knowledge of the human
(a)
(c)
Figure 1. Examples of test images. (a) The 16 pixel diameter disks with no search restriction (other than the limits of the image
frame). (b,c) The same diameter with Ds/Dd53 and 1.5, respectively.
438
C J Kotre
observer noise sampling to investigate not only the area
of the sampling aperture, but the spatial distribution of
the samples and the effect on observer performance of
the proximity of adjacent structures.
Experimental observations
A series of static contrast-detail test patterns were
presented to six experienced observers on a monochrome
TV display monitor. The display was first calibrated in
terms of luminance using a Hagner S2 photometer
(Hagner AB, Solna, Sweden) whose response is matched
(b)
The British Journal of Radiology, May 2006
Short communication: Search pattern extent for threshold contrast tests
to that of the human eye, and the monitor settings were
then left fixed for the duration of the experiment. The
photometer was calibrated using a source of known
luminous intensity in terms of illuminance, from which
the luminance calibration of its internal detector was
derived using the relationship between luminance of a
surface and illuminance at a point for a Lambert source
given by Guibelalde et al [13].
Each test pattern consisted of a 464 matrix of 16
circular test features decreasing in contrast in steps of
0.84 (reciprocal cube-root of 2). Computer-generated
Gaussian noise was added to the whole image sufficient
to produce a threshold cut-off within the contrast range
displayed. The test images consisted of 5126512 8-bit
pixels displayed in a 20 cm616 cm format at a viewing
distance of approximately 50 cm, although the observer
was not restricted in viewing distance. The observations
were carried out with room lights dimmed, and
observers were allowed to score disks as ‘‘half seen’’
where detection was uncertain, following common
practice in scoring contrast-detail tests in radiological
image assessment.
Four detail diameters were used (32, 16, 8 and 4 pixels,
equivalent to 11.3 mm, 5.6 mm, 2.8 mm and 1.4 mm,
respectively) and each of these was displayed within a
larger circular background area of mid-grey with noise
added to both disk and background. Outside the area of
mid-grey, the screen was blanked to black with no noise
to limit the search area of the observer. The combinations
of disk diameter (Dd) and search area diameter (Ds) were
varied to give no search restriction, then Ds/Dd54, 3, 2,
1.5, 1.25 and 1.125 for each of the four disk diameters.
The test patterns were presented in random order with
no time restriction on the observation. Three example
test patterns are reproduced in Figure 1. Figure 1a shows
the 16 pixel diameter disks with no search restriction
(other than the limits of the image frame), Figure 1b,c
shows the same diameter with Ds/Dd53 and 1.5,
respectively.
Figure 2. Contrast-detail plots of the results for the unrest-
ricted search (#), and Ds/Dd53 (%), 1.5 (n) and 1.125 (,).
The error bar illustrates a typical¡1 standard error on the
mean value of threshold contrast for the experiment.
The British Journal of Radiology, May 2006
Figure 3. The ratio of threshold contrast for the limited
search, normalized to that of the unlimited search, plotted
N
against Ds/Dd, for all four disk diameters: 1.4 mm ( ), 2.8 mm
(&), 5.6 mm (m) and 11.3 mm (.).
At the conclusion of all observations, the recorded
scores were converted to luminance threshold contrasts
(CT) and averaged across observers.
Results
Figure 2 shows conventional contrast-detail plots of
some of the results with quadratic curve fits on
logarithmic axes. To avoid obscuring the figure, only
plots for the unrestricted search and Ds/Dd53, 1.5 and
1.125 are shown. As Ds/Dd is reduced, the threshold
contrasts increase as expected. The error bar illustrates a
typical¡1 standard error of the mean value of threshold
contrast for the experiment. Standard errors for each
point were calculated, but are not shown to avoid
obscuring the figure.
Figure 3 shows the ratio of threshold contrast for the
restricted search, normalized to that of the unrestricted
search, plotted against Ds/Dd, for all four disk diameters.
Although the values are subject to considerable uncer-
tainties, the trend with Ds/Dd appears to be similar for
all diameters, suggesting the presence of a search pattern
that scales directly with the size of the object being
observed.
Taking the similarity of the results of Figure 3 as
sufficient to support the hypothesis that the search
pattern scales in proportion to the size of the object,
Figure 4 shows the data from Figure 3 averaged across
all disk diameters. It is notable that the mean ratio of
threshold contrasts already has value greater than unity
(although unity is within the error bar) at Ds/Dd54,
indicating that some interference with the observer’s
search pattern is already present with this background.
The ratio increases quickly for values of Ds/Dd , 2, as
the visual task changes from finding a disk on a
background to detecting the difference between the
intensity of the disk and the narrow band of noisy
background surrounding it. This finding is in agreement
439

Figure 4. The ratio of threshold contrast for the limited
search, normalized to that of the unlimited search, plotted
against Ds/Dd for the data from Figure 3 averaged across all
disk diameters. The error bars show¡1 standard error on the
mean.
with previous results observed for a design of mammo-
graphic contrast-detail phantom, where a marker circle
at 2.7 times the diameter of the test feature was found to
reduce detection, but marker circles at greater relative
diameters (.6) had little effect [14].
Conclusions
It had been shown that for the SKE/BKE contrast-
detail detection task undertaken against a background of
Gaussian noise, the observer’s ability to detect low
contrast objects is progressively reduced as the area of
the search area around the object is reduced, the
difficulty of the task increasing rapidly as the diameter
of the restricted search area falls to less than twice that of
the target disk. The results further suggest the presence
of a search pattern that scales in proportion with the size
of the test feature. These results are relevant to the design
of contrast-detail test phantoms for radiological imaging
440
C J Kotre
systems, and also to the broader investigation of the
effect of background anatomical structure on the detec-
tion of diagnostic features in radiology.
References
1. Hay GA, Clarke OF, Coleman NJ, Cowen AR. A set of X-ray
test objects for quality control in television fluoroscopy. Br J
Radiol 1985;58:335–44.
2. Bijkerk KR, Lindeijer JM, Thijssen MAO. The CDMAM
phantom: a contrast-detail phantom specifically for mam-
mography. Radiology 1993;185(P):395.
3. Institute of Physics and Engineering in Medicine.
Recommended standards for the routine performance
testing of diagnostic X-ray systems. (IPEM Report 77)
York: IPEM, 1997.
4. Schnitzler AD. Image-detector model and parameters of the
human visual system. J Opt Soc Am 1973;63:1357–68.
5. Rose A. The sensitivity performance of the human eye on
an absolute scale. J Opt Soc Am 1948;38:196–208.
6. Chesters MS, Hay GA. Quantitative relation between
detectability and noise power. Phys Med Biol 1983;28:
1113–25.
7. Kotre CJ. The effect of background structure on the
detection of low contrast objects in mammography. Br J
Radiol 1998;71:1162–7.
8. Ohara K, Doi K, Metz CE, Giger ML. Investigation of basic
imaging properties in digital radiography. 13. Effect of
simple structured noise on the detectability of simulated
stenotic lesions. Med Phys 1989;16:14–21.
9. Bochud F, Valley J-F, Verdun FR. Estimation of the noisy
component of anatomical backgrounds. Med Phys 1999;26:
1365–70.
10. Burgess AE, Jacobson FL, Judy PF. Human observer
detection experiments with mammograms and power-law
noise. Med Phys 2001;28:419–37.
11. Kundel HL, Nodine CF, Thickman D, Carmody D, Toto L.
Nodule detection with and without a chest image. Invest
Radiol 1985;20:94–9.
12. Reverez G. Conspicuity and uncertainty in the radiographic
detection of lesions. Radiology 1985;154:625–8.
13. Guibelalde E, Vano E, Llorca AL. Quality assurance of
viewing boxes: proposal for establishing minimum require-
ments and results from a Spanish quality control pro-
gramme. Br J Radiol 1990;63:564–7.
14. Kotre CJ, Porter DJT. A printed image quality test phantom
for mammography. Br J Radiol 2005;78:746–8.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 441–444

SHORT COMMUNICATION

Imaging pulmonary embolism in pregnancy: what is the most

appropriate imaging protocol?
S MATTHEWS, MBBS, MRCP, FRCR

Radiology Department, Northern General Hospital, Herries Road, Sheffield, South Yorkshire
S5 7AU, UK

ABSTRACT. Pulmonary embolism is the leading cause of death in pregnancy. Despite

the difficulties in clinical diagnosis and the concerns regarding radiation of the fetus,
the British Thoracic Society guidelines for imaging pulmonary embolism do not
specifically address the issue of imaging for pulmonary embolism in this group. This
communication discusses the difficulties of diagnosis and imaging pulmonary embolism
in pregnancy and proposes a suitable imaging protocol. Clinical exclusion of patients
from further imaging is recommended if the patient has a low pre-test probability of
pulmonary embolism and a normal d-dimer. It is advised that all remaining patients Received 1 August 2005
undergo bilateral leg Doppler assessment. If this test is positive, the patient should be Revised 17 November 2005
treated for pulmonary embolism; if negative, all patients should be referred for CT Accepted 19 January 2006
pulmonary angiography. Ideally, informed consent should be obtained prior to CT
DOI: 10.1259/bjr/15144573
scanning. All neonates exposed to iodinated contrast in utero should have their thyroid
function tested in the first week of life due to the theoretical risk of contrast induced ’ 2006 The British Institute of
hypothyroidism. Radiology

Pulmonary embolism (PE) is the leading cause of
maternal death [1]. The rate of PE in pregnancy is five
times greater than that for non-pregnant women of the
same age and is about 1 in 1500 deliveries; the risks are
even higher in the puerperium. The clinical diagnosis of
PE is difficult in the general population, but is further
complicated in pregnancy as some of the clinical
symptoms of PE can be normal/expected symptoms of
pregnancy. Precise PE diagnosis in pregnancy is vital to
prevent unnecessary treatment of PE as treatment is
associated with side effects for both the mother and
fetus. Accurate imaging is essential, but there are
frequently anxieties relating to fetal radiation exposure
during diagnostic procedures. Despite the complexities
of the clinical scenario, the British Thoracic Society (BTS)
guidelines on the management of PE do not address the
issue of imaging for PE in the pregnant patient [2].
There are an array of clinical, biochemical and
radiological tests available for the investigation of PE,
some of which have not been validated in pregnancy. In
general terms, the BTS guidelines recommend clinical
assessment of the pre-test probability of PE for each
patient with d-dimer assessment for patients with a low
or intermediate pre-test probability followed by either a
ventilation/perfusion scan or CT pulmonary angiogram
(CTPA; depending on the availability of nuclear medi-
cine scans locally and the presence of a chest X-ray
abnormality). The Wells criteria are the most frequently
used tool for assessing the clinical probability of PE.
However, pregnant patients were excluded from the
analysis group for validation of the criteria [3]. The
d-dimer is known to increase in pregnancy. The d-dimer
is usually normal in the first trimester of uncomplicated
pregnancy, starts to rise during the second trimester
and returns to baseline levels at 4–6 weeks post partum
[4–6]. The d-dimer is not affected by bleeding, breast
feeding or heparin prophylaxis, but is elevated in
association with many pregnancy-related complications
such as pre-eclampsia [5, 7]. Pregnant women with on-
going thrombosis have been shown to develop a
significant rise in d-dimer [8].
It is estimated that 70% of patients with a proven PE
have proximal deep venous thrombosis (DVT).
However, the proportion in pregnant women with PE
is unknown. Therefore, Doppler ultrasound of the leg
veins is recommended for the investigation of PE if the
patient has symptoms and signs suggestive of DVT [2].
The diagnosis of DVT in pregnant women can be
problematic. There is increased lower extremity vein
diameter and decreased flow secondary to hormonal
effects and the direct compressive effect of the enlarged
uterus on pelvic veins, hence the legs are frequently
swollen in the absence of DVT [9]. There is an increased
risk of iliac vein thrombosis, which is not routinely
assessed by leg Doppler studies. In addition, the
accuracy of Doppler ultrasound, including iliac vein
assessment, has not been validated in pregnant patients.
However, Doppler ultrasound does not involve ionizing
radiation and, as there are concerns about the radiation
exposure to mother and fetus, bilateral leg Doppler has
been proposed as the initial investigation of suspected
PE in pregnancy [9]. This would conform to the general
principle of maintaining doses ‘‘as low as reasonably
achievable’’ whilst still offering a valuable test that may

The British Journal of Radiology, May 2006 441


preclude the need for further investigations associated
with a significant radiation dose [10].
Ventilation/perfusion (V/Q) imaging is well estab-
lished for imaging PE and, in a survey relating to
imaging practice for the investigation of PE in pregnancy
in 1998, was the most frequently employed test for this
sub group [11]. In pregnant women the radiation dose
can be minimized by using a half-dose perfusion scan
and only proceeding to ventilation imaging if a defect is
identified on the perfusion scan [12]. However, for the
general population 50–70% of V/Q scans are indetermi-
nate. For pregnant women the proportion of patients
falling into each of the reporting categories (high
probability, normal and indeterminate) is different:
fewer pregnant women have high probability scans (less
than 5%) and many more have a normal scan (75%) [13–
15]. Therefore, only 20% of this patient group have
indeterminate scans. This change in distribution of
patients within the probability groups is thought to
reflect the younger average age and reduced presence of
co-morbidities compared with the general population.
PE can be confidently excluded with a normal V/Q scan,
but this test throws up a relatively high proportion of
indeterminate results and, in the high probability group,
up to 20% of the patients may not have PE. A small study
involving 113 de novo cases of potential PE showed that
withholding anticoagulation in pregnant women with
normal or indeterminate scans is probably safe [13].
However, larger studies are required to confirm this
finding, especially as the incidence of PE in the
proportion of pregnant women investigated is so low.
It has been suggested that CTPA has a greater
discriminatory power than V/Q scanning with a low
pre-test probability, but that CTPA and V/Q scanning
have a similar discriminatory power in those with a high
pre-test probability [16].
CTPA is now a well-validated investigation with a
sensitivity and specificity between 94% and 100% [17,
18]. The negative predictive value of a normal CTPA is
over 99% [19]. The clinical validity of a CT scan to rule
out PE is similar to that reported for conventional
pulmonary angiography [19]. Anticoagulants can be
safely withheld if the CTPA is negative for PE. CTPA
is advantageous as the emboli are directly visualized
(unlike for V/Q scanning) and alternative causes for the
patient’s symptoms may be diagnosed. However, there
are concerns regarding the radiation dose received from
CT scanning, particularly to the fetus. Recent studies
have shown that the fetal radiation exposure for CTPA
varies from 3.3 mGy to 130.0 mGy; the dose increasing
during each trimester as the fetus enlarges and
approaches the imaged area in the thorax [20].
However, the estimated fetal radiation dose for V/Q
scanning is estimated as 100–370 mGy, i.e. the dose may
be more than 3 times greater than for CTPA. In addition,
CTPA has a superior sensitivity and specificity for PE
[21].
All radiation to the fetus carries a potential risk. This
risk must be balanced against the risk to the mother/
fetus if PE is not diagnosed or treated and against the
risk of treatment of non-confirmed PE. Everyone is
exposed to radiation all the time from the atmosphere,
ground and from ingested food and drink. The average
‘‘background radiation’’ for an individual in the UK is
442
S Matthews
2.7 mSv per year, which equates to about 1000 mGy for a
fetus in utero for 9 months [22]. The worst estimated
absorbed dose for the fetus in the third trimester
undergoing CT pulmonary angiography is 130 mGy, i.e.
approximately 7 times less than the natural background
radiation. Central nervous system malformations can be
associated with excess radiation [23]. A threshold fetal
dose greater than 100 000–200 000 mGy is required to
cause such a problem. A dose of 100 mGy to the fetus is
associated with an excess death from cancer up to the
age of 15 years of 1 in 300 000. This can be put into
perspective by comparison with everyday activities and
their relative annual risk of death in the UK, e.g. smoking
10 cigarettes each day has a risk of 1 in 200 and relative
risk of death from uncomplicated pregnancy for the
mother is 1 in 170 000 [24]. The annual risk of death in
the UK for all cancer is 1 in 400 and for death from all
causes at the age of 40 years 1 in 700 [24]. Therapeutic
termination would not be considered for a fetal dose
below 100 000 mGy. However, it is recommended that
the patient is given more detailed information regarding
the risks of radiation for procedures where the fetal dose
is expected to be greater than 1000 mGy [23]. Therefore,
although the radiation risk to the fetus cannot be
ignored, the risk is very low. The risk of fetal death is
much greater if the mother has untreated PE.
In addition to the radiation risk to the fetus, the breast
radiation dose from CTPA must be considered. The
female breast is extremely radiosensitive, and it has been
shown that a sufficiently large radiation dose can cause
breast cancer [25]. The exposure of the immature breast
during early development and around the time of
menarche carries a higher risk than at other times of a
woman’s life. There is little evidence that radiation
exposure after 45 years of age increases the incidence of
breast cancer. The effect of radiation on the breast in
pregnancy is unclear. However, there has been shown to
be a relatively linear relationship between radiation dose
and subsequent breast cancer, although this relationship
does not extend into the highest radiation dose expo-
sures used for radiotherapy. The vast majority of people
exposed to radiation do not develop a cancer related to
that exposure. The 25 000 female atomic bomb survivors
in Japan have been followed for over 50 years, but only
173 breast cancer deaths have occurred of which 41 were
attributed to the radiation received in 1945.
A radiation dose of 100 cGy is associated with an
increased risk of breast cancer of 40% in young Western
women [25]. This is the same risk a woman experiences
secondary to several common conditions/lifestyle
choices, e.g. never being pregnant, menarche before the
age of 11 years or a late menopause. Epidemiological
studies have not detected a significantly increased risk of
breast cancer below a dose of 20 cGy. Parker et al have
specifically investigated female breast radiation expo-
sure during CTPA and calculated an effective minimum
dose of 20 mGy (2 cGy) [26]. This dose concurs with
other data estimating a dose between 20 mGy and
50 mGy (2–5 cGy) for a standard chest CT scan [27].
These estimates are significantly below the level of
20 cGy, below which no effect on the breast can be
demonstrated, but significantly higher than the
estimated breast radiation dose of 0.28 mGy associated
with ventilation/perfusion scanning [28]. Although this
The British Journal of Radiology, May 2006
Short communication: Imaging pulmonary embolism in pregnancy
radiation exposure is associated with an immeasurable
low malignancy risk, this exposure should not be
ignored and the development of breast shields that
may reduce this dose by up to 73% may be considered in
the future [29].
A 2003 survey of members of the Society of Thoracic
Radiology found that 53% of responding radiologists
would use CTPA as a first line investigation for
excluding PE in pregnant patients, but only 60% of
radiologists obtained informed consent from any preg-
nant patient undergoing CTPA, only 16% of departments
had a written policy for this group and only 40%
modified the imaging protocol in an attempt to reduce
radiation dose [30].
The risks of iodinated contrast media on the fetus have
not been fully investigated. Many pregnant patients have
received intravenous contrast for the investigation of
other problems such as renal colic. However, there are no
reports in the literature of any ill effects, despite the
theoretical risk of contrast induced hypothyroidism.
In summary, d-dimer concentration is known to rise in
normal pregnancy, but a normal value is clinically
helpful. Doppler ultrasound of both legs has a low pick
up rate, but does not involve ionizing radiation. If DVT is
found, the patient should be treated as for PE. This test
requires further validation in pregnant women.
Ventilation perfusion imaging is associated with a higher
radiation dose to the fetus, but lower radiation dose to
the mother/breast than CTPA. The majority of patients
have a low or indeterminate probability of PE from the
V/Q scan and it is probably safe to withhold anti-
coagulants in these groups, but this has not been
validated in a large controlled study. V/Q scanning
has a lower discriminatory threshold for PE than CTPA
in patients with a low or intermediate pre-test clinical
probability. The mother receives a higher radiation dose
from CTPA, especially to the breast, but the fetal dose is
lower than V/Q imaging. However, all quoted radiation
doses to breast or fetus are below the thresholds
estimated to be associated with any significant risk.
CTPA has a greater sensitivity and specificity than V/Q
scanning and is able to diagnose alternative causes for
the patients’ symptoms in cases where PE is absent.
However, the risk of iodinated contrast to the fetus is not
known.
Multiple clinicians may interpret these same facts and
derive differing imaging protocols based on their
perceived significance for each factor discussed. The
following imaging protocol is proposed to optimize the
accuracy of PE diagnosis whilst minimizing radiation
dose to the fetus: All patients should be clinically
assessed by a senior clinician and a pre-test probability
assigned. Patients in the first two trimesters of pregnancy
should have the d-dimer measured, provided there is no
specific contraindication. A normal d-dimer and low pre-
test probability of PE can be used to exclude PE in the
pregnant patient. Exclusion of patients from further
imaging on the basis of intermediate pre-test probability
and normal d-dimer needs further investigation. All
pregnant patients requiring further investigation for PE
should proceed to bilateral leg Doppler ultrasound. If
this test is positive for venous thromboembolism, the
patient should be treated for PE. If this test is negative,
all patients should proceed to CTPA. Informed consent
The British Journal of Radiology, May 2006
should be sought from the pregnant patient prior to
CTPA. The patient should be given simple information
explaining the risks of fetal and maternal radiation and
risks to the mother and child of failing to accurately
diagnose PE. The CTPA protocol should be modified to
minimize the radiation dose; in particular the length of
the thorax along the z-axis should be reduced. It is
recommended that the baby has thyroid function testing
within the week of birth due to the theoretical risk of
contrast induced hypothyroidism [31].
Further research is advised to validate the above
protocol and to clarify the in utero affects of iodinated
contrast on the neonate.
References
1. Pabinger I, Grafenhofer H. Thrombosis during pregnancy:
risk factors, diagnosis and treatment. Pathophysiol
Haemost Thromb 2002;32:322–4.
2. British Thoracic Society guidelines for the management of
suspected acute pulmonary embolism. Thorax 2003;58:
470–84.
3. Wells PS, Anderson DR, Rodger M, Stiell I, Dreyer JF,
Barnes D, et al. Excluding pulmonary embolism at the
bedside without diagnostic imaging: management of
patients with suspected pulmonary embolism presenting
to the emergency department by using a simple clinical
model and d-dimer. Ann Int Med 2001;135:98–107.
4. Kline JA, Williams GW, Hernandez-Nino J. D-dimer
concentrations in normal pregnancy: new diagnostic
thresholds needed. Clin Chem 2005;51:825–9.
5. Epiney M, Boehlen F, Boulvain M, Reber G, Antonelli E,
Morales M, et al. D-dimer levels during delivery and the
postpartum. J Thromb Haemost 2005;3:268–71.
6. Frascalanci I, Comeglio P, Liotta AA, Cellai AP, Fedi S,
Parretti E, et al. D-dimer concentrations during normal
pregnancy, as measured by ELISA. Thromb Res 1995;78:
399–405.
7. Nolan TE, Smith RP, Devoe LD. Maternal plasma D-dimer
levels in normal and complicated pregnancies. Obstet
Gynecol 1993;81:235–8.
8. Bombelli T, Raddatz-Mueller P, Fehr J. Coagulation activa-
tion markers do not correlate with the clinical risk of
thrombosis in pregnant women. Am J Obstet Gynecol
2001;184:382–9.
9. Chan W-S, Ginsberg JS. Diagnosis of deep vein thrombosis
and pulmonary embolism in pregnancy. Thrombosis Res
2002;107:85–91.
10. Statutory Instruments. Ionising Radiation (Medical
Exposure) Regulations 2000. London: HMSO, 2000, No.
1059.
11. Boiselle PM, Reddy SS, Villas PA, Liu A, Seibyl JP.
Pulmonary embolus in pregnant patients: survey of
ventilation-perfusion imaging policies and practice.
Radiology 1998;207:201–6.
12. Balan KK, Critchley M, Vedavathy KK, Smith ML,
Vinjamuri S. The value of ventilation-perfusion imaging
in pregnancy. Br J Radiol 1997;70:338–40.
13. Chan WS, Ray JG, Murray S, Coady GE, Coates G, Ginsberg
JS. Suspected pulmonary embolism in pregnancy: clinical
presentation, results of lung scanning, and subsequent
maternal and pediatric outcomes. Arch Intern Med
2002;162:1170–5.
14. Hull RD, Hirsh J, Carter CJ, Raskob GE, Gill GJ, Jay RM,
et al. Diagnostic value of ventilation-perfusion lung scan-
ning in patients with suspected pulmonary embolism.
Chest 1985;88:819–28.
15. Value of ventilation/perfusion scan in acute pulmonary
embolism. Results of the prospective investigation of
443

pulmonary embolism diagnosis (PIOPED). The PIOPED
investigators. JAMA 1990;263:2753–9.
16. Hayashino Y, Gotto M, Noguchi Y, Fukui T. Ventilation-
perfusion scanning and helical CT in suspected pulmonary
embolism: Meta analysis of diagnostic performance.
Radiology 2005;234:740–8.
17. Blachere H, Latrabe V, Montaudon M, Valli N, Couffinhal T,
Raherisson C, et al. Pulmonary embolism revealed on helical
CT angiography: comparison with ventilation-perfusion radio-
nuclide lung scanning. AJR Am J Roentgenol 2000;174:1041–7.
18. Remy-Jardin M, Remy J, Baghaie F, Fribourg M, Artaud D,
Duhamel A. Clinical value in the thin collimation diagnostic
workup of pulmonary embolism. AJR Am J Roentgenol
2000;175:407–11.
19. Quiroz R, Kucher N, Zou KH, Kipfmueller F, Costello P,
Goldhaber SZ, et al. Clinical validity of a negative computed
tomography scan in patients with suspected pulmonary
embolism: a systematic review. JAMA 2005;293:2012–7.
20. Winer-Muram HT, Boone JM, Brown HL, Jennings SG,
Mabie WC, Lombardo GT. Pulmonary embolism in
pregnant patients: fetal radiation dose with helical CT.
Radiology 2002;224:487–92.
21. Bates SM, Ginsberg JS. How we manage venous throm-
boembolism in pregnancy. Blood 2002;100:3470–8.
22. Watson SJ, Jones AL, Oatway WB, Hughes JS. Health
Protection Agency. Ionising radiation exposure of the UK
population: 2005 review. HPA – RPD – 001.
444
S Matthews
23. Pregnancy and medical radiation: International
Commission on Radiological Protection. Ann ICRP 2000;30:
1–43.
24. Living with radiation. (Fifth edition). Chilton: NRPB, 1998.
25. Ionising radiation and breast cancer risk. Sprecher Institute
for Comparative Cancer Research. Fact Sheet 52, October
2004.
26. Parker MS, Hui FK, Camacho MA, Chung JK, Broga DW,
Sethi NN. Female breast radiation exposure during CT
pulmonary angiography. AJR Am J Roentgenol 2005;185:
1228–33.
27. Rehani MM. CT: caution on radiation dose. Ind J Radiol
Imag 2000;10:19–20.
28. Cook JV. UK agency reports slight increase in radiation
exposure. BMJ 2005;330:1229.
29. Ocker J. New Breast Shield reduces radiation to female
breast during chest CT by up to 73%. Medical News Today,
16 May 2005.
30. Schuster ME, Fishman, JE, Copeland JF, Hatabu H, Boiselle
PM. Pulmonary embolism in pregnant patients: a survey of
practices and policies for CT pulmonary angiography. AJR
Am J Roentgenol 2003;181:1495–8.
31. Webb JA, Thomsen HS, Morcos SK, and members of the
Contrast Media Safety Committee of European Society of
Urogenital Radiology. The use of iodinated and gadolinium
contrast media for MRI during pregnancy and lactation.
Eur Rad 2005;15:1234–40.
The British Journal of Radiology, May 2006
The British Journal of Radiology, 79 (2006), 445–447

CASE REPORT

Primary subcutaneous sacrococcygeal ependymoma: a case

report and review of the literature
Y T MA, MRCP, P RAMACHANDRA, FRCR and D SPOONER, FRCR

The Cancer Centre, Queen Elizabeth Hospital, Edgbaston, Birmingham B15 2TH, UK

ABSTRACT. Extraspinal ependymomas are rare. The majority occur in the

sacrococcygeal region. The subcutaneous variety accounts for approximately two thirds
of cases, which are commonly misdiagnosed as a pilonidal cyst or sinus. Treatment is
complete surgical resection. The role of coccygectomy is controversial. Adjuvant Received 5 May 2004
radiotherapy is of benefit to those with an incompletely excised tumour. Up to 20% Revised 13 May 2005
metastasise, chiefly to the inguinal lymph glands, but pulmonary metastases are also Accepted 19 May 2005
reported. Palliative chemotherapy has not been shown to be of any benefit. Long term
DOI: 10.1259/bjr/61959899
follow-up is important as metastases can occur up to 20 years after initial presentation.
We report a 37-year-old woman with a subcutaneous sacrococcygeal ependymoma ’ 2006 The British Institute of
with iliac lymph nodal metastasis at presentation. Radiology

The ependymal cells form an epithelial sheet and line pelvis demonstrated a residual lobulated soft tissue
the ventricles in the brain and the central canal of the mass, about 3 cm in diameter, lying anterior and to the
spinal cord. Ependymomas are tumours derived from right of the tip of the coccyx. A second smaller lesion was
ependymal cells and are usually confined to the central also visible between the right internal and external iliac
nervous system, occurring predominantly in children arteries (Figure 1). CT scanning of the thorax and
and young adults. They account for only 6% of abdomen demonstrated multiple liver cysts, but no
intracranial gliomas, but in the spinal cord they are the evidence of lung or liver metastases. Isotope bone
most common tumours of glial origin, accounting for scanning revealed no evidence of skeletal metastases.
over 60%, mostly in the region of the cauda equina and
the filum terminale [1].
Rarely, ependymomas occur outside of the central
nervous system [1]. The majority occur in the sacrococ-
cygeal subcutaneous tissue or the presacral regions [2–
10]. We present a case of an ependymoma arising from
the sacrococcygeal subcutaneous tissue with iliac lymph
nodal metastasis at presentation.

Case report
A 37-year-old Caucasian lady presented with a 6-
month history of an enlarging nodule at the base of her
coccyx. She had always been aware of the presence of a
small nodule but had thought this was ‘‘normal’’ for her.
During the 6-month period the nodule had increased in
size from less than 0.5 cm to over 3 cm in diameter.
There was no history of urinary or faecal problems. She
was seen at her local hospital and a diagnosis of a
pilonidal sinus was made which was subsequently
excised. Histological examination revealed this to be a
subcutaneous myxopapillary sacrococcygeal ependy-
moma and she was referred to the regional centre for
musculoskeletal oncology.
Figure 1. MRI abdomen and pelvis. Sagittal T2 short tau
The full blood count, biochemical profile and tumour inversion recovery (STIR) sequence showing a 3 cm lobulated
markers (alpha-fetoprotein (AFP), ß-human chorionic soft tissue mass lying anterior to the tip of the coccyx, and a
gonadotrophin (HCG), carcinoembryonic antigen (CEA) second small rounded soft tissue mass superiorly in the
and CA-125) were normal. MRI of the abdomen and pelvis.

The British Journal of Radiology, May 2006 445


The residual mass was completely excised.
Intraoperatively, it was seen near the tip of the coccyx
but separate from it. Additionally, a 1 cm solid lesion
with cystic elements was found above the right common
iliac artery. This was thought to be a right iliac lymph
node and was also completely removed. The histology of
the specimens again confirmed a myxopapillary epen-
dymoma (Figure 2). The majority of the tumour cells
were positive for glial fibrillary acidic protein (GFAP)
and S-100, but epithelial and other neural and neuroen-
docrine markers were negative. The resection margins
were free of tumour cells and the smaller lesion was
confirmed to be a nodal metastasis.
She made a good post-operative recovery. No adjuvant
therapy was offered, as there was no evidence of either
macroscopic or microscopic disease. Re-staging CT
scanning of the thorax, abdomen and pelvis at 3 months
showed no residual or metastatic disease.
6 months following surgery she developed persistent
right iliac fossa pain. CT scanning of the abdomen and
pelvis now demonstrated a soft tissue mass 4.5 cm 6
4 cm 6 3.5 cm in the right pelvis and also the presence
of an enlarged right internal iliac lymph node (1.5 cm),
suggestive of recurrent disease. The right pelvic mass
and iliac lymph node were completely excised.
Histological examination of the pelvic mass revealed
reactive tissue only. Histological examination of the right
iliac lymph node confirmed metastatic myxopapillary
ependymoma. Surgery was followed with adjuvant
radiotherapy to the right pelvis (45 G in 25 daily
fractions over 5 weeks). She has made a good recovery
and is now being closely followed up in clinic.
Discussion
Primary extraspinal ependymomas are rare and usually
occur in the subcutaneous sacrococcygeal or presacral
regions. Mallory reported the first case in 1902 [5]. Since
then there have been approximately 75 reported cases of
extraspinal ependymomas, with the subcutaneous variety
accounting for about 50 cases (Table 1) [2–10]. Most of the
published cases of sacrococcygeal ependymomas have
occurred in children and young adults, although there
have been reported cases ranging from 2 months to
67 years of age [2–10]. The male to female ratio of affected
individuals is roughly equal [2, 6].
Subcutaneous sacrococcygeal ependymomas are
thought to arise from the coccygeal medullary vestige,
an ependymal lined cavity forming the remnant of the
caudal portion of the neural tube [11]. This theory is
supported by Bale who found post-coccygeal ependymal
vestiges in 10 of 15 random necropsies in infants [12].
Table 1. Comparison of subcutaneous and presacral sacrococcygeal ependymomas
Extraspinal No. of Male:female ratio
ependymomas reported cases
Subcutaneous 49 Equal
Presacral 28 Equal
446
Y T Ma, P Ramachandra and D Spooner
Presacral ependymomas are thought to arise either from
the extradural remnants of the filum terminale or as an
extension from the intradural filum terminale [2, 3]. The
frequent finding of sacral bone destruction and sacral
nerve dysfunction in patients with presacral ependymo-
mas supports this theory [4]. A germ cell origin has also
been proposed which may explain the rare occurrence of
ependymomas in the ovary and the mediastinum [2].
Most subcutaneous sacrococcygeal ependymomas are
of the myxopapillary variety [7]. The other histological
types, according to the World Health Organization
(WHO) classification, are papillary and subependymal
[2]. Characteristically, subcutaneous sacrococcygeal
ependymomas grow slowly and often reach a relatively
large size prior to presentation. The majority of patients
are asymptomatic. Consequently, these tumours are
frequently misdiagnosed clinically as a pilonidal cyst or
sinus [6, 8]. This was a feature in our case.
Ependymomas confined to the central nervous system
rarely metastasise outside of the cerebrospinal fluid (CSF)
pathway. Conversely, up to 20% of subcutaneous sacro-
coccygeal ependymomas have been reported to develop
distant metastases [2, 6, 8]. This has often been attributed
to the easier access of these tumours to lymphatic and
vascular vessels. Pulmonary, pleural, osseous, inguinal
nodal and subcutaneous metastases have been documen-
ted [2–10]. Our patient was found to have a synchronous
right iliac lymph nodal metastasis, which is an unusual
finding. Currently, no reliable clinical or histological
feature exists to help predict the development of metas-
tases. Distant metastases can occur 10–20 years after initial
presentation [6]. Consequently, patients need to be under
regular follow-up for many years.
Immunohistochemical staining helps to make the
diagnosis. Typically, there are positive reactions to glial
fibrillary acidic protein (GFAP), S-100 and vimentin [2].
This was seen in our case; the tissue reacted positively to
GFAP and S-100.
The treatment of subcutaneous sacrococcygeal epen-
dymomas remains controversial because of their rarity.
The favoured treatment is surgical excision. Complete
excision appears to be essential to prevent local
recurrence. Surgical removal of the coccyx may also be
required if the tumour is attached to this bone [2]. Some
authors have suggested routine coccygectomy as part of
the primary surgical treatment, because of an apparent
reduction in the recurrence rate. Aktuğ et al carried out a
review of all reported cases of sacrococcygeal extraspinal
ependymomas [9]. They found that local recurrence did
not occur if the coccyx was removed (four patients, mean
follow-up period of 3 years). Conversely, if the coccyx
was not excised, local recurrence occurred in 71% of
patients (5 out of 7 patients). It appears that the likely
Origin Symptoms
Coccygeal medullary vestige Asymptomatic mass
Extradural remnants of the Neurological symptoms
filum terminale, or Extension (Bowel/bladder dysfunction;
from the intradural saddle anaesthesia; paraesthesia
filum terminale and weakness of lower limbs)
The British Journal of Radiology, May 2006
Case report: Primary subcutaneous sacrococcygeal ependymoma
Figure 2. Histology of the excised mass showing plump
bipolar spindle cells and epithelial-looking cells arranged in
loose fascicles with extensive myxoid degeneration forming
pseudoacinar structures.
benefit for coccygectomy is in the group of patients with
presacral ependymomas, with involvement of the coc-
cyx. In our case, as the mass was found separate to the
coccyx, coccygectomy was therefore not performed.
Lymphadenectomy is also recommended in patients
who present concurrently with inguinal lymph nodal
metastases [8]. The role of lymphadenectomy in patients
with iliac lymph nodal metastases is unknown as it is
very rare. Although our patient had complete excision of
the ependymoma and removal of the affected iliac lymph
node, she developed local recurrence within 6 months of
surgery suggesting a possible role for regional lymph
node dissection.
Intracranial ependymomas are radiosensitive; radio-
therapy is an important local therapy [8]. However,
radiotherapy has not been found useful as primary
treatment for subcutaneous sacrococcygeal ependymo-
mas. In those with incompletely excised tumour or in the
presence of metastases, adjuvant radiotherapy is recom-
mended [2, 8]. There is no evidence that these tumours
respond to chemotherapy [2, 8].
The British Journal of Radiology, May 2006
Although extremely rare, it is important to consider
subcutaneous sacrococcygeal ependymomas in the dif-
ferential diagnosis of a sacrococcygeal mass. In the
absence of metastatic disease, complete excision of the
tumour with regional lymph node dissection will usually
result in a prolonged disease-free survival. Once metas-
tases develop, the disease tends to be slowly, but
unremittingly, progressive.
References
1. Russell DS, Rubinstein LJ. Pathology of tumours of the
nervous system (5th edn). London: Edward Arnold, 1989.
2. Johnson JM, Jessurun J, Leonard A. Sacrococcygeal epen-
dymoma: case report and review of the literature. J Pediatr
Surg 1999;34:1405–7.
3. Timmerman W, Bubrick MP. Presacral and postsacral
extraspinal ependymoma: report of a case and review of
the literature. Dis Colon Rectum 1984;27:114–9.
4. Morantz RA, Kepes JJ, Batnitzky S, Masterson BJ.
Extraspinal ependymomas: report of three cases. J
Neurosurg 1979;51:383–91.
5. Mallory FB. Three gliomata of ependymal origin: two in the
fourth ventricle, one subcutaneous over the coccyx. J Med
Res 1902;8:1–10.
6. Helwig EB, Stern JB. Subcutaneous sacrococcygeal myx-
opapillary ependymoma: a clinicopathologic study of 32
cases. Am J Clin Pathol 1984;81:156–61.
7. Maiorana A, Fante R, Fano RA. Myxopapillary ependy-
moma of the sacrococcygeal region. Pathologica
1989;81:471–6.
8. Kramer GWPM, Rutten E, Sloof J. Subcutaneous sacrococ-
cygeal ependymoma with inguinal lymph node metastasis.
J Neurosurg 1998;68:474–7.
9. Aktuğ T, Hakgüder G, Sarioğlu S, et al. Sacrococcygeal
extraspinal ependymomas: the role of coccygectomy. J
Pediatr Surg 2000;35:515–8.
10. Lynch J, Kelly N, Fitzpatrick B, Regan P. A sacrococcygeal
extraspinal ependymoma in a 67-year-old man: a case
report and review of the literature. Br J Plast Surg
2002;55:80–2.
11. Kernohan JW. Primary tumours of the spinal cord and
intradural filum terminale. In: Cytology and cellular
pathology of the nervous system. New York: Penfield W,
1932) 993–1025.
12. Bale PM. Ependymal rests and subcutaneous sacrococcy-
geal ependymoma. Pathology 1980;12:237–43.
447
’ 2006 The British Institute of Radiology The British Journal of Radiology, 79 (2006), 448

Correspondence
(The Editors do not hold themselves responsible for opinions expressed by correspondents)

Dose distribution to the mediastinum and Table 1. Dose distribution comparison for data in Cominos
et al [1] and a single patient treated by proton therapy in
heart Tzukuba
4 field X-rays 2 field protons
The Editor—Sir, (UK) (Japan)
It is encouraging to read the detailed analysis of dose
Prescribed dose (Gy) 54 72
distributions to the mediastinum and heart during radical
(30 fractions) (36 fractions)
radiotherapy for oesophageal cancer in the UK [1]. This is a Mean cardiac dose (Gy) 41.4 18.7
desirable consequence of better radiotherapy treatment Volume of heart 77 15
planning using three-dimensional (3D) techniques and receiving > 30 Gy (%)
advanced software systems. Although the authors quote Volume of heart 51 10
the Japanese experience of increased toxicity following receiving > 40 Gy (%)
chemoradiotherapy [2], there is important pioneering Volume of heart 26 8
work from Japan using protons to reduce normal tissue receiving > 45 Gy (%)
doses while delivering a high tumour target dose using
only two fields, with a respiratory gating technique linked
Thus it has been demonstrated – as proof of principle
to a pulsed output of protons.
in a small group of patients – that proton therapy can be
At The University of Tzukuba Proton Medical
successfully used in the treatment of oesophageal cancer
Research Centre, the potential indications for proton
with no evidence of late cardiac toxicity. The DVH
therapy have been extended by conducting a detailed
analysis obtained in one patient is substantially better
clinical study of 46 oesophageal cancer patients using
than with the megavoltage X-ray technique.
limited proton beam availability between 1985 and 1998.
It would be timely for the UK to acquire proton
These mostly mixed protons and X-rays studies show
therapy facilities in order to improve therapeutic dose
that high total doses can be given with long term tumour
distributions in a wide variety of anatomical sites. Japan
control at least equivalent to X-ray therapy [3]. Toxicity
– with a population of c. 160 million – will soon have 8
reduced with experience, dose adjustment, increasing
centres capable of delivering proton or ion beam radio-
the component of protons and the use of respiratory
therapy. To maintain a similar proportion, the UK would
gating. No late symptomatic complications were
need to have 3 centres.
observed in the tracheobronchial tree, heart or spinal
Yours etc.,
cord and no late grade 4 or 5 oesophageal toxicity has 1
B JONES and
been observed since 1991. Severe and persistent oeso- 2
Y AKINE
phageal ulceration in some patients had previously
occurred because of high doses per fraction (2.5–
3.7 Gy) related to limited beam availability. Since 2001, 1
University Hospital Birmingham
pure proton therapy treatments have been delivered in a Birmingham
purpose built Centre using rotating gantries and at B15 2TH, UK
fractionated doses of 1.8–2 Gy per fraction: the results of 2
Proton Medical Research Centre
these studies are awaited with interest. Software for University of Tzukuba
proton dose–volume histograms (DVH) have only been Japan
available recently at Tzukuba, so it is not possible to
retrospectively analyse the patients treated in their (Received 3 January 2006 and in revised form 18 January 2006, Accepted
publication. However, on a recently treated patient the 18 January 2006)
DVH for a two field (anterior and posterior fields) 200 DOI: 10.1259/bjr/88462532
MV proton plan has been obtained and Table 1 shows a
comparison of this plan with averaged values reported
by Cominos et al [1] for their best case scenario of a 4
References
field plan (taken from Table 2 in their paper). The patient
was a 61 year old man with a T1 N1 M0 squamous cell 1. Cominos M, Mosleh-Shirazi MA, Tait D, Henrys A, Cornes
carcinoma of the mid-thoracic oesophagus. He had P. Quantification and reduction of cardiac dose in radical
refused the offers of chemoradiotherapy and surgery. It radiotherapy for oesophageal cancer. Br J Radiol
should be noted that the UK prescribed dose is 2005;78:1069–74.
2. Satoshi I, Keiji N. Atushi O, Narikasu B. Long term toxicity
considerably lower than that used in Japan, yet the
after definitive chemoradiation of oesophageal and gastro-
cardiac doses are markedly reduced in the latter, despite oesophageal junction cancers. J Clin Oncol 2003;21:115–20.
the anterior beam field traversing through the heart. 3. Sugahara S, Tokuuye K, Okumura T, Nakahara A, Saida Y,
Furthermore, the 4 field X-ray technique will inevitably Kagei K, et al. Clinical results of proton beam therapy for
result in higher lung exposures than for the 2-field cancer of the oesophagus. Int J Radiat Oncol Biol Phys
proton technique, which were not reported. 2005;61:76–84.

448 The British Journal of Radiology, May 2006

The British Journal of Radiology, 79 (2006), 449–451

CASE OF THE MONTH

A catastrophic complication
D R VUMMIDI, MRCP, R S KASTHURI, MRCS, FRCR and R W BURY, FRCR

Department of Radiology, Blackpool Victoria Hospital and University of Manchester, Manchester,

Cheshire, UK

Received 6 April 2005

Revised 12 May 2005
Accepted 27 June 2005

DOI: 10.1259/bjr/24873219

’ 2006 The British Institute of

Radiology

Case he was in hypovolaemic shock, with a pulsatile lower

abdominal mass. Based on a clinical diagnosis of
A 61-year-old Caucasian male was admitted as an ruptured abdominal aortic aneurysm, emergency surgi-
emergency with central abdominal pain. On examination cal repair was undertaken. He was admitted to the

(a) (b)

(c) (d)

Figure 1. (a–d) Non-contrast CT brain (selected images).

The British Journal of Radiology, May 2006 449


Figure 2. Magnified section from non-contrast CT (right
frontal lobe).
intensive therapy unit following successful aneurysm
repair. Renal dysfunction and gradual obtundation were
features of a stormy post-operative period.
His biochemical profile was significantly deranged
and included elevated serum amylase.
Worsening mentation was the immediate concern and a
CT scan was organized to investigate this (Figures 1 and 2).
What are the abnormalities demonstrated on the CT
images? What is the differential diagnosis of this
appearance? What case specific diagnosis would be
appropriate?
Discussion
The CT brain demonstrates marked attenuation of the
grey/white matter differentiation. There is associated
diffuse cerebral oedema with effacement of the sulci and
basal cerebrospinal fluid (CSF) spaces. Furthermore, the
magnified images also demonstrate multiple petechial
haemorrhages (Figures 2 and 3). These are early findings
of diffuse cerebral anoxia. There is often progression to a
more diffusely low density brain, with possible ‘‘rever-
sal’’ of the grey/white matter attenuation.
At this stage, the differential for this appearance
includes profound and persistent hypotension, asphyxi-
ation and carbon monoxide fume inhalation [1].
In the given scenario, cerebral hypoperfusion would
be most appropriate from the above mentioned list of
differential diagnosis. Multiple cerebral fat emboli can
also result in generalized cerebral hypoperfusion and
result in similar appearances [1].
450
D R Vummidi, R S Kasthuri and R W Bury
Figure 3. An area of petechial haemorrhage in the cerebral
parenchyma.
Figure 4. Frozen section demonstrating arterial fat embolus
(Oil Red O stain).
The described patient unfortunately succumbed to his
illness. The post mortem examination revealed multiple
cerebral fat emboli (Figure 4), secondary to acute
necrotising pancreatitis.
Fat embolism syndrome is an uncommon but recog-
nized complication of acute necrotising pancreatitis [2].
Pathophysiologically, the chylomicrons and VLDL from
the pancreatic fat necrosis are propagated into the
vascular system resulting in cerebrovascular occlusion
and consequent infarction. Multiple such embolisation
into the intracranial vasculature results in global cerebral
hypoperfusion.
Cerebral CT scans demonstrate features as described
above, but these are not specific to fat emboli. MRI is
more sensitive in the neuroradiological diagnosis of
intracranial fat embolism [3]. MRI findings include
multiple intracranial infarcts, returning low signal on
T1 weighted and relatively high signal on T2 weighted
MRI. Similar areas are also seen to involve the corpus
callosum and the basal ganglia. More diffuse grey and
white matter abnormalities akin to CT imaging are
demonstrated in the later stages. Diffusion weighted
The British Journal of Radiology, May 2006
Case of the month: A catastrophic complication
MRI has also been reported to have role in the diagnosis
of cerebral fat emboli [4].
Whilst the cerebral imaging features are specific
to diffuse cerebral hypoperfusion, the aetiology is
broad. This case besides highlighting an unusual
cause of cerebral anoxia emphasises the importance
of clinical correlation of the radiological differential
diagnosis.
Acknowledgments
The authors acknowledge Dr Patrick Shenjere,
Specialist Registrar, Department of Histopathology,
Blackpool Victoria Hospital, Blackpool, UK.
The British Journal of Radiology, May 2006
References
1. Osborn AG, Davis WL, Jacobs J. Cerebral vasculature: normal
anatomy and pathology. In: Osborn AG, editor. Diagnostic
neuroradiology. St. Louis, MO: Mosby, 1994:355–60.
2. Bhalla A, Sachdev A, Lehl SS, Singh R, D’Cruz S. Cerebral fat
embolism as a rare possible complication of traumatic
pancreatitis. JOP 2003;4:155–7.
3. Takahashi M, Suzuki R, Osakabe Y, Asai JI, Miyo T,
Nagashima G, et al. Magnetic resonance imaging findings
in cerebral fat embolism: correlation with clinical manifesta-
tions J Trauma 1999;46:324–7.
4. Parizel PM, Demey HE, Veeckmans G, Verstreken F, Cras P,
Jorens PG, De Schepper AM. Early diagnosis of cerebral fat
embolism syndrome by diffusion weighted MRI. Stroke
2001;32:2942.
451
BJR
The British Journal
of Radiology

June 2006
Volume 79
Issue 942
June 2006, Volume 79, Issue 942

● Three-dimensional MRI of the male urethrae with implanted

artificial sphincters: initial results

● Heel bone densitometry: device specific thresholds for the

assessment of osteoporosis

● The impact of a short course of study on the performance of

radiographers when highlighting fractures on trauma
radiographs: ‘‘The Red Dot System’’

● Portal vein embolisation prior to hepatic resection for colorectal

liver metastases and the effects of periprocedure chemotherapy

● Radiation benefit and risk at the assessment stage of the UK

Breast Screening Programme

● Radiation risks for the radiologist performing transjugular

intrahepatic portosystemic shunt (TIPS)

● Organ doses from prostate radiotherapy and associated

concomitant exposures

● Intensity-modulated radiation therapy in the treatment of gastric

cancer: early clinical outcome and dosimetric comparison with
conventional techniques

● Changes in applicator positions and dose distribution

high dose rate brachytherapy fractions in cervix
patients receiving definitive radiotherapy

● Clinical and cellular ionizing radiation sensitivity in a patient with

xeroderma pigmentosum

● Survival and initial chromatid breakage in normal and tumour

cells exposed in vitro to gamma rays and carbon ions at the HIRFL

● Usefulness of mini-tracheostomy and torque controlled insertion

of applicator in fractionated endobronchial brachytherapy

● MR findings of penile lymphoma

 ● Deep tracheal laceration after balloon dilation for benign
tracheobronchial stenosis: case reports of two patients

● Correspondence

● Neck pain: an unusual presentation of a common disease

● Book reviews
The British Journal of Radiology, 79 (2006), 455–463

Three-dimensional MRI of the male urethrae with implanted

artificial sphincters: initial results
1,4
J DENG, MMed, PhD, 2M A HALL-CRAGGS, FRCR, 3M D CRAGGS, PhD, 1R RICHARDS, PhD, 3S L KNIGHT,
1 3
MSc, PhD, A D LINNEY, PhD and A R MUNDY, FRCS

1
Department of Medical Physics and Bioengineering, 2Department of Radiology and 3Institute of
Urology, 4Department of Obstetrics and Gynaecology, University College London, Gower Street,
London WC1E 6BT, UK

ABSTRACT. The aim of this study was to develop a method for simultaneous 3D
visualization of a new type of artificial urethral sphincter (AUS) and adjacent urinary
structures. Serial MR tomograms were acquired from seven men after AUS
implantation. 3D reconstruction was performed by thresholding original (positive) and
inverted (negative) image intensity and by subsequently fusing positive and negative
images. Results show that the bladder, cuff and balloons of the AUS of originally high
intensity were imaged in 3D by thresholding the positive datasets. The urethrae and
corpora cavernosa penis of originally low intensity were displayed in 3D by
thresholding the negative datasets. Fusion of the positive and negative datasets Received 1 October 2004
allowed simultaneous visualization of the AUS complex and adjacent urinary structures. Revised 19 April 2005
All the structures of interest were also clearly seen by interactive multiplanar Accepted 13 October 2005
reformatting. Coronal tomographic datasets provided better 3D and reformatted 2D
DOI: 10.1259/bjr/56511504
images than sagittal and transverse datasets. This technique offers a simple means for
evaluating the complex urethral anatomy and the AUS, and has potential for improved ’ 2006 The British Institute of
3D visualization of many other complex morphological and pathological conditions. Radiology

Artificial urinary sphincters (AUS) have been used for

many years to treat male stress urinary incontinence. The
most widely-used type is the American Medical Systems
AS800 AUS, but this is not without its problems, the most
serious one being urethral erosion caused by the constant
pressure of the inflated AUS on the urethra [1]. Post-
operative imaging is important for assessing the position,
configuration and function of an implanted AUS. The
AS800 contains radio-opaque substances to make it
visible on X-ray CT, but this exposes the patient to
ionizing radiation. In addition, the contrast agents can
degrade the device.
To reduce the erosion risk, a new type of AUS with
conditional occlusion has been developed by Craggs et al
[2]. The Craggs AUS consists of four main parts
(Figure 1). The perineal part is a cuff implanted around
the bladder neck or urethra. The next two, intrapelvic or
lower abdominal parts, are made up of two balloons,
usually placed extraperitoneally close to the bladder. The
fourth part is a scrotal pump connected to the balloons
and cuff by tubing. The whole implant is made from
silicone rubber and contains normal saline. This lends
itself to MRI, a modality using no ionizing radiation. In Figure 1. Diagram of Craggs artificial urinary sphincter.
1: urethral cuff; 2: primary reservoir/balloon; 3: additional
reservoir (stress-relief balloon); 4: scrotal pump. The design
JD is supported by an MRC Clinician Scientist Fellowship. We allows a temporarily increased intra-abdominal pressure
acknowledge some of the technical and equipment support from
(caused by a stress such as coughing) to be transmitted from
EPSRC-MRC’s MIAS IRC and RCR’s Pump Priming grant. The
prototype artificial urinary sphincters were developed in collabora- 3 into 1. This prevents stress incontinence, as well as urethral
tion with Bibby-Sterilin Limited and Isotron plc in a British erosion that could be caused by a constant high pressure on
Government Link project funded by the Department of Health the urethra even during rest when a conventional artificial
and the Department of Trade & Industry. sphincter was used (Artwork by Martin Knight).

The British Journal of Radiology, June 2006 455


addition, MRI can embrace a large body volume without
a need, unlike ultrasound, for a particular imaging
window and direct contact with the body parts being
examined. This is necessary for visualizing the Craggs
AUS because it occupies both intrapelvic and perineal
spaces, and its cuff is implanted around the delicate
urethral tissue that should not be deformed for
functional assessment. Consequently, MRI was chosen
to visualize the AUS in this study.
Ultrasound, CT and MR imaging have all been used
for visualizing the pelvis in incontinent patients [3–5].
Conventional two-dimensional (2D) imaging requires
expertise to compose a mental three-dimensional (3D)
picture from a series of 2D images. Studies have
demonstrated that 3D ultrasound [6, 7], CT [8] and
MRI [9, 10] improve comprehension of the spatially
complex morphology in the pelvis.
To the best of our knowledge, there are no previous
reports of the use of MRI for the assessment of AUS.
Initially, we used cross-sectional images, directly avail-
able from standard MR scanners, for this assessment,
and 2D MR appearances of the Craggs AUS will be
described in a separate report. In brief, it is easy to
comprehend the structures and positions of different
parts (except the connecting tubes) of the AUS and their
relationships to corresponding local body parts through
serial 2D images. However, it is difficult to form a
reliable 3D impression of all the AUS parts, the complex
connections between them and the global spatial
relationship to the relevant anatomies. This is clinically
desirable, particularly when cross-sectional imaging
planes are not perfectly aligned with ideal anatomical
axes or imaginary axes of the AUS parts.
Recently, 3D reconstruction has become available on
some more advanced MR scanners. Three 3D display
methods are often used on these systems. The first
method is multiplanar reformatting. This requires
further mental work from the operator to construct a
3D object. The second is volume 3D display, in which an
object is rendered somehow transparent so that all the
structures inside it become more or less visible, even
when views to one structure may in reality be blocked by
other structures in the front. The third, surface 3D
display, probably offers the most realistic 3D images as
structures of interest are represented as solid objects.
Surface display is often achieved by first applying a
threshold so that structures with signal intensities below
it will not be rendered. This is sufficient for some clinical
applications, such as for a CT skeletal examination; high-
intensity bone is visualized while low-intensity soft
tissues are ‘‘removed’’. However, this approach is not
sufficient for the imaging of complex biological struc-
tures where a simultaneous display of different intensity
structures is necessary. As in this study, there is a need to
show the AUS (mainly of high intensity) simultaneously
with the bladder and urethra (mainly of high and low
intensities, respectively). When a threshold appropriate
for displaying the high-intensity structures (the bladder
and AUS) is applied, the low-intensity structures (the
urethra) are not seen. If a lower threshold appropriate for
displaying the urethra is applied, irrelevant structures of
high (fat) or intermediate (other soft-tissues) intensity are
also shown, obscuring both low- and high-intensity
structures of interest. In other words, simple thresholding
456
J Deng, M A Hall-Craggs, M D Craggs et al
could not be used to present the two complexes
simultaneously and distinguishably.
In this paper, we describe the ‘‘inversion-fusion’’
method we have developed for processing sequential 2D
MR datasets of patients with implanted Craggs AUSs [2].
This facilitates the simultaneous 3D visualization of the
implant and the lower urinary tract. We then examine
whether this 3D method could bring out information
additional to 2D imaging. Finally, we discuss the feasibility
of using this relatively simple, generic methodology with
3D facilities available on advanced commercial imaging
modalities to improve 3D data reconstruction.
Materials and methods
Data acquisition
With approval from the Local Ethics Committee, nine
patients suffering persistent post-prostatectomy inconti-
nence received the Craggs AUS implantation after giving
informed consent. Excluding one patient with extensive
surgical metalwork in his pelvis, all other eight under-
went MR scans about 2 months after implantation. They
were scanned on a 1.5 T Siemens Vision system (Erlangen,
Germany) using a phase-array body coil. Images were
acquired using T2 weighted 2D Tru-FISP sequence (repeti-
tion time (TR): 6.32 ms; echo time (TE): 3 ms; field of view
(FOV): 230 mm; rectangular field of view (RFOV): 50%; flip
angle (FA): 70 ˚; time of aquisition (TA): 12–21 s) in the
coronal, transverse and sagittal planes.
In this study, the term ‘‘series’’ is being used to
describe a set of anatomically consecutive slices acquired
during one MR scan. Because of the difficulty compre-
hending global relationships between all AUS parts and
relevant urinary structures using individual series
acquired from the first MR patient, several anatomically
consecutive series were acquired from each of the remain-
ing seven patients (aged from 59 years to 76 years). This
allowed the entire AUS and the neighbouring anatomy to
be sampled by these series at regular intervals.
A total of 86 series were recorded from the 7 patients (4
to 20 from each). Each series consisted of 7 to 12
contiguous slices, all with a slice thickness of 5 mm
(minimum achievable by the scanner at the time). A
serial 3D dataset was then formed by one series or by
combining two to six consecutive series, consisting of 7
to 42 slices (Table 1).
In five patients, some serial datasets were acquired in
pairs, offset by 50% of the slice thickness (Figure 2a,b),
and each of the pairs was later used for creating an
interlaced 3D dataset (Figure 2c).
The pixel matrix in each slice was 2566256, with a
pixel size of 0.898460.8984 (mm). The voxel matrix (a 3D
dataset) was therefore 2566256 for x and y dimensions
and from 7 to 42 for z dimension. The resultant voxel size
is 0.898460.898465.0000 (mm) for serial datasets and
0.898460.898462.5 (mm) for interlaced datasets.
Data processing
The datasets were transferred to an MGI 3D
Workstation (Medical Graphics & Imaging Group,
The British Journal of Radiology, June 2006
3D MRI of male urethra and artificial sphincter

Table 1. 3D results from 35 original 3D datasets and 8 interlaced 3D datasets

Patient Sequences Slices per Series Slice Slice Orientation 3D image Notes on reconstruction
age series per 3D thickness interval quality quality
(years) dataset

59 B006-026 7 3 5 5 Coronal Adequate Big slice interval

B027-047 7 3 5 5 Coronal Adequate Big slice interval
B006-047mix 7 6 5 2.5* Coronal Good Mild motion artefact,
not affecting ROI
76 D069-077 9 1 5 5 Coronal Adequate Big slice interval
D078-085 9 1 5 5 Coronal Adequate Big slice interval
D128-135 9 1 5 5 Coronal Adequate Big slice interval
D145-154 9 1 5 5 Coronal Adequate Big slice interval
D015-041mix 9 3 5 5 Coronal Good
D101-127mix 9 3 5 5 Coronal Good
D005-014 9 1 5 5 Sagittal Adequate Big slice interval
D042-050 9 1 5 5 Sagittal Poor Too big slice interval
D050-058 9 1 5 5 Sagittal Poor Too big slice interval
D092-100 9 1 5 5 Sagittal Poor Too big slice interval
D137-145 9 1 5 5 Sagittal Poor Too big slice interval
55 H005-033mix 7 4 5 5 Coronal Good
H034-051mix 7 4 5 5 Coronal Good
H074-108mix 7 5 5 5 Coronal Good
H109-143mix 7 5 5 5 Coronal Good
H005-034int-mix 7 8 2.5 2.5* Coronal Good Mild motion artefact,
not affecting ROI
H074-109int-mix 7 10 2.5 2.5* Coronal Good Mild motion artefact,
not affecting ROI
H052-058 7 1 5 5 Sagittal Adequate Big slice interval
H144-150 7 1 5 5 Sagittal Adequate Big slice interval
60 M005-033 7 4 5 5 Coronal Good
M034-051 7 4 5 5 Coronal Good
M057-094 7 4 5 5 Coronal Good
M095-122 7 4 5 5 Coronal Good
M005-034int 7 8 2.5 2.5* Coronal Good Mild motion artefact,
not affecting ROI
M057-095int 7 8 2.5 2.5* Coronal Good Mild motion artefact,
not affecting ROI
69 S011-031mix 7 3 5 5 Coronal Good
S032-052mix 7 3 5 5 Coronal Good
S072-092mix 7 3 5 5 Coronal Good
S093-113mix 7 3 5 5 Coronal Good
S011-032int-mix 7 5 2.5 2.5* Coronal Good Mild motion artefact,
not affecting ROI
S072-093int-mix 7 5 2.5 2.5* Coronal Good Mild motion artefact,
not affecting ROI
S053-059 7 1 5 5 Sagittal Adequate Big slice interval
S114-120 7 1 5 5 Sagittal Adequate Big slice interval
64 V005-017 12 1 5 5 Coronal Good
V018-029 12 1 5 5 Sagittal Adequate Big slice interval
V030-041mix 12 1 5 5 Transverse Poor Big slice interval
67 Y005-029mix 12 2 5 5 Coronal Good
Y030-053mix 12 2 5 5 Coronal Adequate Moderate motion artefact
Y005-030int-mix 12 4 5 2.5* Coronal Adequate Moderate motion artefact
Y055-101mix 12 3 5 5 Transverse Adequate Big slice interval
*Interlaced datasets.
ROI, region of interest.

University College London, www.medphys.ucl.ac.uk/
mgi/workstat.htm), and converted into MGI format
using DispImage program [11]. Interlaced 3D datasets
were generated by interweaving and then renumber-
ing paired slices from paired serial datasets
(Figure 2c).
There were two complexes of interest: the urinary
complex consisting of the bladder, urethra and penis,
and the AUS complex consisting of the cuff, the balloons
and the tubes connecting them. On T2 weighted images,
the urinary complex mainly showed low intensity except
for the urine within the bladder and urinary tract, which

The British Journal of Radiology, June 2006 457

 J Deng, M A Hall-Craggs, M D Craggs et al

Figure 2. Schematic drawings of data

acquisition and post-processing. Pixels
(voxels) of high, intermediate and low
intensities on original (positive) images
are numbered 1, 2, and 3, respectively.
(a,b) Two series are acquired, each with a
5 mm slice interval. The second series is
scanned with a 2.5 mm offset from the
first one. (c) After interweaving the slices
of the two series, a (positive) dataset
(Slices 0–5) is created, with the slice
interval halved to 2.5 mm and the num-
ber of slices doubled. A negative dataset
(Slices 09–59) is also created, changing
pixels of low-intensity (numbered 3) into
high-intensity. (d) Two thresholds appro-
priate for displaying originally high
intensity pixels in Slices 0–5 and nega-
tively high intensity pixels in Slices 09–59,
correspondingly, remove intermediate
and low intensity pixels. Background
pixels in both positive and negative
datasets are not displayed. The remain-
ing pixels (voxels) are areas of interest,
which can then be fused into a single 3D
image (not shown here), restoring the
original spatial relationship between
pixels (voxels) 1 s and 3 s.

(a) (b)

Figure 3. One of serial slices from a coronal scan. (a) In the original (positive) sequence, the bladder, balloon and cuff of the
artificial sphincter are of high-intensity, which can be segmented by simply applying a threshold to remove low-intensity signals,
and then rendered in 3D surface display (Figure 4a). However, this will also remove the corpora cavernosa from being displayed
because their intensities are below the threshold. (b) After inversing the grey scale, the corpora become high-intensity structures
which can be displayed by simply applying another threshold (Figure 4b). In the 2D images, it is difficult to tell whether the cuff
is fully sealed around the corpora cavernosa urethrae.

was high signal. The AUS complex showed high To address this problem, a negative dataset was
intensity. As mentioned earlier, simple thresholding created by inverting the intensity of the native, positive
could not be used to present the two complexes together dataset (Figure 2c,d and Figure 3). The low-intensity
clearly. structures (urethra and penis) became high-intensity

458 The British Journal of Radiology, June 2006

3D MRI of male urethra and artificial sphincter

(a) (b)

Figure 4. 3D reconstruction with surface displays. (a) Structures above corresponding thresholds in the positive and negative
sequences are separately visualized. In order not to obscure the smaller structures (in grey), the positive sequence is placed closer
to the readers, so the bladder is placed in front of the lower abdomen. The threshold in the negative sequence is set to
intermediate low so that the separations of the three spongy structures are displayed as solid while their inner regions as
(artificially) hollow. (b) The structures from both positive and negative sequences are fused together to reveal the relationship
between the cuff and the urethra; in this case, the cuff is fully sealed (also see movies on our website for a better 3D perception).

ones in the negative dataset. The positive and negative Figure 4a). Finally, by putting the two objects into the
datasets were then combined into a new, positive+nega- same co-ordinate system, a fused image is created for
tive dataset (Figure 2c). assessing their spatial relationships (Figure 4b and
Figure 5).

3D reconstruction
This was also carried out on the MGI Workstation. The
system provides various options for the reconstruction
and manipulation of 3D datasets from serial tomo-
graphic imaging modalities [12–16]. 3D surface display
and multiplanar reformatting were used in this study to
present the structures of interest for analysis.
3D surface display for structures in the overlapping
datasets was achieved by four main steps. First, two
blocks containing structures of interest in a positive+ne-
gative sequence were defined under the guidance of
multiplanar reformatting. This avoided segmenting most
of the fat in the abdomen, skeletal muscles in the pelvic
floor and thighs (in a negative dataset which had more or
less the same intensity as that of the urethra) and the
image background area (which became the highest
intensity area in a negative dataset). Second, two
thresholds appropriate for rendering the originally
high-intensity structures (the bladder, the balloon and
the cuff) and the inversely high-intensity structures (the
corpora cavernosa) were applied to the two blocks
separately. This resulted in the structures of interest
being roughly visualized (Figure 2d). Third, some
‘‘image surgery’’ was performed to remove remaining
unwanted parts that were still above thresholds and
therefore might obscure the viewing of structures of
interest. Having done these, the two complexes could be
displayed for inspection separately (Figure 2d and
Data analysis
The level of analysis was at the dataset level (Table 1)
and between various scan settings (Table 2). The quality
of the final reconstructed 3D images in terms of
depicting the spatial relationship between the structures
of interest was scored by two observers (JD, RR) as poor,
adequate and good. ‘‘Poor’’ means that in a single fused
3D object, the spatial relationship between the structures
of interest cannot be visualized, and subsequently, the
AUS functioning position cannot be diagnostically
assessed. ‘‘Adequate’’ means that in a single fused 3D
object, the spatial relationship of the AUS cuff to the
corpora cavernosa urethra was visualized. Some other
parts of the urinary tract or of the AUS were not clearly
visualized, but this did not affect our judgement of the
completeness of the AUS cuff wrapping the urethra, the
most important diagnostic issue to be assessed. ‘‘Good’’
means that in addition to the cuff-urethral relationship,
the entire AUS complex and the entire urinary complex
were well visualized.
Results
A total of 35 datasets containing the region of interest
for 2D and 3D analysis were obtained directly or
combined from 86 series acquired. Eight interlaced

The British Journal of Radiology, June 2006 459


Figure 5. A 3D surface display showing that the cuff is not
fully wrapped around the corporus cavernosum urethrae.
Again, it is an image after fusing the cuff in the positive and
the penis in the negative sequences. The threshold in the
negative sequence is set very low so that the entire spongy
structures are displayed as solid (compare with Figure 4).
Also see a movie on our website.
datasets were further created from the 35 datasets
(Table 1). Depending on data quality and suitability, it
took about 10–30 min to process one sequence and attain
fused 3D images. The longer time was caused by
processing few, but graphically more difficult, sagittal
and transverse datasets using manual segmentation of
regions of non-interest (i.e. unwanted abdominal fat or
other tissues with complex greyscale composition, see
Discussion on Image processing). The ‘‘inversion-fusion’’
method itself, together with series-combination and slice-
interlacing, only took a few seconds.
The 3D results were listed in Table 1. The comparison
of the 3D images acquired from coronal, sagittal and
transverse base data is shown in Table 2. The coronal
datasets offered the best 3D results with 65.2% good and
34.8% adequate images. None of the sagittal and
transverse datasets produced good 3D images, but six
of 10 sagittal datasets were adequate. There were only
two transverse datasets, precluding assessment.
Interlaced datasets provided the best 3D results with
good 3D images obtained in seven (87.5%) of eight
datasets (all derived from coronal datasets). Motion
Table 2. Initial comparison of the number and percentage of useful 3D series/datasets between various scan settings and post-
processings
3D image quality Coronal series/datasets
Poor Adequate Good Subtotal Poor
Original No. (%) 0 (0.00) 8 (34.8) 15 (55.2) 23
Interlaced No. (%) 0 (0.00) 1 (12.5) 7 (87.5) 8
460
J Deng, M A Hall-Craggs, M D Craggs et al
artefact was noticeable in all eight datasets, but this
degraded the bladder more than the cuff (Figure 6).
The spatial relationships of all the structures of interest
were best assessed when the images from both positive
and negative datasets were fused. Urinary structures and
the AUS were clearly demonstrated using 3D surface
display, and this was particularly well seen when
viewed as a 3D movie (available at: www.medphys.
ucl.ac.uk/mgi/jdeng/ follow the links ‘‘Dynamic Face/
Body Parts’’ on the top panel, then ‘‘Erecting Penis’’ on
the left panel). The 3D images have depicted various cuff
conditions between the well-implanted, fully wrapped
(Figure 4) to the failed, fully open (Figure 5), and from
un-inflated to fully inflated.
2D views of the region of interest can be interactively
obtained using multiplanar reformatting under the
guidance of 3D objects (Figure 6), offering a means for
detailed examination of structures that may not be well
visualized by simple 3D reconstruction. This is particu-
larly useful for following up structures of heterogeneous
intensity such as the tubing between the balloons and
cuffs and for following the thin urine residues along the
urethra that were not acquired in single original imaging
planes. All these have been found difficult when only
viewing original 2D slices.
Discussion
2D versus 3D imaging
This initial study has shown that, by using our
‘‘inversion-fusion’’ method, 3D images simultaneously
displaying both AUS and urinary complexes can be
obtained from serial MR datasets. Although subject to
further large-scale studies, visualizing the datasets as
fused 3D objects and through reformatted 2D images
appears to have at least three advantages over viewing
them as original individual 2D slices.
First, images reconstructed by 3D surface rendering
can offer more straightforward, panoramic views of all
four main parts of the AUS and their relationships to
each other and to the urinary tract. This is important for
post-operative assessment as the four parts are
implanted in three very different anatomical places. It
is possible for an experienced radiologist to form such a
3D impression from individual slices, but it requires
greater training while still subject to human errors.
Second, multiplanar images reformatted from a 3D
dataset in any desirable orientation allow tracking of the
entire course of the tubing connection between the AUS
parts. This has proved to be very difficult when tracking
through the original 2D slices. It is crucial to check this
Sagittal series/datasets Transverse series/datasets
Adequate Good Subtotal Poor Adequate Good Subtotal
4 (40.0) 5 (50.0) 0 (0.0) 10 1 (50.0) 1 (50.0) 0 (0.0) 2
The British Journal of Radiology, June 2006
3D MRI of male urethra and artificial sphincter

(a) (b)

Figure 6. Multiplanar reformatting of the same 3D dataset in Figure 4a: Three orthogonal images are obtained by cross-
sectioning the bladder, showing the motion artefact (arrowheads) in the interlaced data, which appears more severe in the
upper side of the bladder. (b) Another three orthogonal images are obtained by cross-sectioning the cuff, showing the well-
sealed cuff, although not so apparent as in 3D images. Note the depiction of the tubing (arrows) connecting the cuff and the
balloons. Its entire course may not be visualized by 3D surface display due to its heterogeneous signal intensity, but can be
traced by interactively reformatting sequential 2D images.

connection when blockage and/or entanglement are different AUS hydraulic and other physiological circum-
suspected. stances. It may also enable real-time checking of leakage
Third, the full 3D visualization of the AUS complex during pumping [17–19].
has paved the way for its volumetric quantification. This
will be helpful for evaluation of possible AUS leakage by
comparing the given amount of saline pumped in to the Imaging resolution
increment in volume before and after the pumping. It is
The in-plane image quality obtained from this MR
impossible to work out such a change mentally unless
study is much higher than that obtained from our
there is serious leak.
unpublished 3D ultrasound study. This is because,
More generically speaking, this study has shown that
compared with ultrasound, MRI has higher intrinsic
clinically useful 3D images can be attained by combined
tissue contrast and no restriction on imaging window
use of commercially available modality and acquisition
[20, 21]. However, the out-of-plane (z-dimension) spatial
methods (MR interlaced acquisition with coronal
resolution is far from ideal. For instance, while the voxel
sequences) with relatively simple post-processing and
sizes in the x and y dimensions were equally small (about
display methods (intensity inversion and image fusion).
1 mm), the size in the z-dimension was 5 mm. To
Admittedly, 3D techniques have not been sophisticat-
address this unbalanced match, interlaced acquisitions
edly developed such that they can render texture details.
were performed in the last five patients and this halved
In this study, these features, such as the intricate
the z voxel size to 2.5 mm (Figure 2b). Thinner slices and
structures of the AUS were seen better on the 2D slices
true 3D data sets can now be acquired in our centre after
(either from original cross-sectional data or from multi-
this study. Although this will reduce the degree of
planar reformatting).
anisotropy, it is unlikely to be eliminated. In addition, no
matter how thin a slice may become available from a
Imaging modality newer scanner, radiologists will immediately expect to
see tissue details at an even smaller scale. Consequently,
Ultrasound examination has a narrow FOV for super- interlaced acquisition may still be helpful in future
ficial structures, and soft-tissue deformation results from studies.
the direct contact between transducer and the skin [15].
There is also insufficient definition of the AUS structures,
as seen in our unpublished 3D ultrasound study of some Imaging orientation
of the patients. These prevent ultrasound from being our
first choice for structural assessment. However, the latest In this study, coronal datasets provided the best 3D
real-time 3D colour Doppler ultrasound is likely to play a results (Table 2). This is because the slices are obtained
role in evaluation of penile haemodynamics under perpendicular to the long axis of the penile urethra and

The British Journal of Radiology, June 2006 461


the cuff, offering multiple serial slices through the
structures of interest. In contrast a sagittal or transverse
series acquires only three to four slices through the cuff
in the penis, simply not offering sufficient data for 3D
reconstruction. Though, recently available thinner slice
acquisition may change this.
Motion artefacts
There are several sources of movement and misregistra-
tion artefacts that cause image degradation: respiration,
peristalsis, urine and blood flow, gross body and penile
movement. Because the sequence acquisition time was up
to 21 s, periodic motion was likely to cause artefacts.
However, obvious motion artefact was mainly due to
misregistration between the added image series. This is
mainly because it is difficult to breath-hold at exactly the
same point of respiration for each image series.
Fortunately, the impact of motion artefact appeared
minimal in the region of the cuff (Figure 6), primarily
because it was extraperitoneal and stabilized by the
urethra and local ligaments. In the future, faster
acquisition and motion tracking [22] may minimize
motion artefacts.
Image processing
With a few computer mouse clicks, the interlacing and
inverting techniques have improved resolution and
speeded up segmentation of the region of interest,
respectively. Nonetheless, cumbersome manual segmen-
tation is still needed for removal of regions of non-
interest but of a wide-range of mixed signal intensities.
Clinicians are frequently caught in a dilemma in 3D
segmentation of graphically complicated structures. On
one hand, simple algorithms such as thresholding are
often unable to depict all the structures of interest that
have heterogeneous intensities (even when they are made
up of the same biological tissues), or have variable
intensity gradients between different tissue borders. This
often results in incomprehensible or even misleading 3D
images. On the other hand, complex algorithms often
demand extra expertise and are time-consuming. This
study has tried to strike a balance between the simplicity of
the image processing and the quality of the resulting
images. Some of the generic tricks developed in this study
such as ‘‘inversion-fusion’’, can be easily adapted by
commercial systems to speed up and simplify 3D
reconstruction. In fact this has been implemented in the
latest MGI 3D Workstation for commercial release. In
addition, automatic contouring with operator interactive
correction may allow rapid segmentation with sufficient
accuracy. Automatic contouring may also be used for fast
removal of extraneous, but homogeneous, tissues, such as
high-intensity fat from the abdominal wall (Figure 6),
which would otherwise obscure the region of interest in
3D images. Hence, the entire post-processing and display
time are expected to be reduced to about 10 min.
A final point to make is that for visual reality and
clarity, the ‘‘inversion-fusion’’ method was only used for
3D surface rendering in this study, but in principle, it can
also benefit volume rendering.
462
J Deng, M A Hall-Craggs, M D Craggs et al
Conclusion
This MR study has demonstrated the feasibility of
simultaneous visualization of all relevant parts of the
AUS and the lower urinary tract in three dimensions
using these novel image processing methods. This has
potential for reducing difficulty in comprehending the
spatial position, geometry and function (and malfunc-
tion) of implanted sphincters and the impact on the local
anatomy. The inversion-fusion approach can be deve-
loped as a general application for simplifying and
improving 3D visualization of many other complex
morphological and pathological conditions.
Acknowledgments
We thank the MR staff of Middlesex Hospital for data
acquisition, D Plummer and D McDonald of Medical
Physics Department for developing DispImage and other
image conversion programmes. We are also grateful to A
Todd-Pokropek for advice on MR data processing and C
H Rodeck of Obstetric Department for research planning.
References
1. Venn SN, Greenwell TJ, Mundy AR. The long-term
outcome of artificial urinary sphincters. J Urol
2000;164:702–6.
2. Craggs MD, Mundy AR, Dwyer PL, Susser J, Knight SL.
Long-term clinical outcome of a new AUS with conditional
occlusion for genuine stress incontinence: optimal cuff
pressures for continence. BJU Int 2002;90:18.
3. McFarlane IP, Foley SJ, Shah PJ. Long-term outcome of
permanent urethral stents in the treatment of detrusor-
sphincter dyssynergia. Br J Urol 1996;78:729–32.
4. Maki DD, et al. Injected periurethral collagen for post-
prostatectomy urinary incontinence: MR and CT appear-
ance. Abdom Imaging 2000;25:658–62.
5. Mostwin JL, Genadry R, Saunders R, Yang A. Stress
incontinence observed with real time sonography and
dynamic fastscan magnetic resonance imaging--insights
into pathophysiology. Scand J Urol Nephrol Suppl
2001;94–9.
6. Klein HM, Kirschner-Hermanns R, Lagunilla J, Gunther
RW. Assessment of incontinence with intraurethral US:
preliminary results. Radiology 1993;187:141–3.
7. Umek W, et al. The urethra during pelvic floor contraction:
observations on three-dimensional ultrasound. Obstet
Gynecol 2002;100:796.
8. Stenzl A, et al. 3-dimensional computerized tomography
and virtual reality of the lower urinary tract in women. Int
Urogynecol J Pelvic Floor Dysfunct 1999;10:248–53.
9. Myers RP, et al. Puboperineales: muscular boundaries of
the male urogenital hiatus in 3D from magnetic resonance
imaging. J Urol 2000;164:1412–5.
10. Ueno S, et al. Three-dimensional display of the pelvic
structure of anorectal malformations based on CT and MR
images. J Pediatr Surg 1995;30:682–6.
11. Plummer DL. A display and analysis tool for medical
images. Rivista di Neuroradiologia 1992;5:489–95.
12. Linney AD, et al. Three-dimensional visualization of data
on human anatomy: diagnosis and surgical planning.
J Audiov Media Med 1993;16:4–10.
13. Brookes JA, Deng J, Wilkinson ID, Lees WR. Three-
dimensional imaging of the postmortem fetus by MRI:
early experience. Fetal Diagn Ther 1999;14:166–71.
The British Journal of Radiology, June 2006
3D MRI of male urethra and artificial sphincter
14. Deng J, Brookes JA, Gardener JE, Rodeck CH, Lees WR.
Three-dimensional magnetic resonance imaging of the
postmortem fetal heart. Fetal Diagn Ther 1996;11:
417–21.
15. Deng J, et al. Novel technique for three-dimensional
visualisation and quantification of deformable, moving
soft-tissue body parts. Lancet 2000;356:127–31.
16. Deng J, et al. Dynamic three-dimensional colour Doppler
ultrasound of human fetal intracardiac flow. Ultrasound
Obstet Gynecol 2002;20:131–6.
17. Deng J. Terminology of three-dimensional and four-
dimensional ultrasound imaging of the fetal heart and
other moving body parts. Ultrasound Obstet Gynecol
2003;22:336–44.
18. Deng J. Minimally compressed imaging of deformable
moving body parts using dynamic 3D ultrasonography and
colour Doppler - with facial and genital parts as exemplars.
The British Journal of Radiology, June 2006
Euroson School on 3D Ultrasound Imaging. London:
Hammersmith Hospital, UK, May 4–6, 2005. p81–93.
19. Deng J, et al. Real time three-dimensional ultrasound
visualization of erection and artificial coitus. Int J Androl
2005; Online early (doi: 10.1111/j.1365-2605.2005.00617.x).
20. Coakley FV, et al. Urinary continence after radical retro-
pubic prostatectomy: relationship with membranous
urethral length on preoperative endorectal magnetic reso-
nance imaging. J Urol 2002;168:1032–5.
21. Mikuma N, Tamagawa M, Morita K, Tsukamoto T.
Magnetic resonance imaging of the male pelvic floor: the
anatomical configuration and dynamic movement in
healthy men. Neurourol Urodyn 1998;17:591–7.
22. Keegan J, Gatehouse P, Yang GZ, Firmin D. Coronary artery
motion with the respiratory cycle during breath-holding
and free-breathing: implications for slice-followed coronary
artery imaging. Magn Reson Med 2002;47:476–81.
463
 The British Journal of Radiology, 79 (2006), 464–467

Heel bone densitometry: device specific thresholds for the

assessment of osteoporosis
1 1 3
E MCCAULEY, MSc, A MACKIE, PhD, D ELLIOTT, RGN, ONC and 2A CHUCK, FRCP

1
Regional Medical Physics Department and 2Department of Rheumatology, University Hospital of
North Durham, North Road, Durham DH1 5TW and 3Durham and Chester-le-Street Primary Care
Trust, John Snow House, Durham University Science Park, County Durham DH1 3YG, UK

ABSTRACT. A diagnosis of osteoporosis is facilitated by bone mineral density (BMD)

measurement of the lumbar spine and hip using dual energy X-ray absorptiometry
(DXA), interpreted in accordance with criteria published by the World Health
Organization (WHO). The use of peripheral DXA is growing in primary care and
guidance on its use has recently been published by the National Osteoporosis Society
(NOS), recommending a triage approach using thresholds specific to each type of
peripheral device. However, no data currently exist for the Norland Apollo heel
densitometer (Cooper Surgical, Trumbull, USA). 215 women between 50 years and
75 years of age (mean age 64.6 years) referred for hip and spine BMD measurements
also had a heel BMD measurement. Device specific upper and lower thresholds were Received 30 August 2005
calculated for the Norland Apollo heel densitometer to give a 90% sensitivity and 90% Revised 14 October 2005
specificity for osteoporosis at the hip or spine. Patients with a heel T-score of above –1.2 Accepted 31 October 2005
are very likely to have normal bone density on axial densitometry, whilst patients with
DOI: 10.1259/bjr/47352663
heel T-score of below –2.2 are very likely to have osteoporosis at the hip or spine.
Patients whose measurements lie between the thresholds should be referred for axial ’ 2006 The British Institute of
DXA. Radiology

A diagnosis of osteoporosis or osteopenia can be made
using dual energy X-ray absorptiometry (DXA) of the
lumbar spine or femur. The defining criterion for the
diagnosis is the value of the T-score, which is the number
of standard deviations the measured bone mineral
density (BMD) is above or below the young adult
reference BMD. According to the World Health
Organization (WHO) criteria [1], the patient is osteo-
porotic if the T-score # –2.5 and osteopenic if the T-score
.–2.5 but # –1.
Axial densitometers, which are designed to measure
BMD at the hip and/or spine, are large, expensive and
require a permanent location, ordinarily in a secondary
care setting. Significant research into the development of
alternative methodologies has resulted in a plethora of
techniques and alternative measurement sites [2]. The
measurement of peripheral bone density is growing
outside secondary care centres, principally because of the
much lower cost of these devices compared with axial
densitometers. They are also very easy to use, requiring
minimal set up and positioning, and are relatively
portable.
Heel densitometers produce results in terms of BMD
(g cm22) and a heel T-score. Manufacturers’ data
accompanying devices often quote the WHO criteria
indicating that a heel T-score of # 22.5 would be
indicative of osteoporosis. It is well established, how-
ever, that it is inappropriate to interpret any peripheral
T-score as one would for axial densitometry [3, 4] and
WHO T-score criteria applied to heel DXA would
significantly underestimate the prevalence of osteopo-
rosis [3, 5].
In addition, it is well known that measurements of
BMD vary from scanner to scanner and between scanner
types; particularly between manufacturers because of the
variation in acquisition and analysis techniques as well
as the absence of a standard calibration technique [6–10].
The International Society for Clinical Densitometry
(ISCD) state that although peripheral DXA measure-
ments can theoretically be used to identify patients
unlikely to have osteoporosis, and also identify patients
who should be treated, this cannot be applied in clinical
practice until device-specific cut-off points are estab-
lished [11].
The National Osteoporosis Society (NOS) has recently
produced a revised position statement concerning the
recommended use of peripheral densitometry [12]. NOS
recommend a triage approach using T-score thresholds,
which are specific to each type of peripheral device.
These thresholds are defined so that patients with
osteoporosis at the hip or spine are identified with 90%
sensitivity and 90% specificity. Patients with a peripheral
T-score below the lower threshold are very likely to have
osteoporosis at the hip or spine, whereas patients with a
peripheral T-score above the upper threshold are
unlikely to have osteoporosis. Those patients who have
a heel T-score that falls between the two thresholds
require a hip and spine BMD examination for a definitive
diagnosis.
The NOS statement gives device specific threshold
data for five peripheral devices, two of which are heel
densitometers. Data in this form are not yet published for
the Norland Apollo heel densitometer (Cooper Surgical,
Trumbull, USA).

464 The British Journal of Radiology, June 2006

Heel bone densitometry

In 2001, the Durham and Chester-le-Street Primary

Care Trust (PCT) purchased a Norland Apollo DXA
scanner for the measurement of heel bone density within
the primary care setting. The device was to be operated
by the osteoporosis prevention specialist nurse within a
protocol agreed between primary and secondary care.
The aim of this study was to establish thresholds for
Apollo heel BMD T-scores for use in risk stratifying
patients within a primary care management protocol.

Method
Ethical committee approval was obtained to measure
heel BMD with the Apollo in women attending for axial
bone density measurements; all fulfilled at least one
accepted referral criterion [13]. 76 women aged between
50 years and 75 years (mean age 64.0 years) gave
informed written consent. Axial BMD was measured
for the lumbar spine, L2–4, and for the non-dominant
femur using a Lunar DPX-IQ (General Electric/Lunar
Corp., Madison, WI) densitometer. Heel BMD, of the
dominant side, was measured with a Norland Apollo
portable densitometer. Measurement of the dominant
side was acceptable since published data demonstrates Figure 1. Scatter plot of heel T-scores for the initial group of
no significant difference between BMD measurements of patients; 90% of the 57 not osteoporotic patients lie above
the heel between the right and left sides [5, 14]. the lower threshold of 22.2, whilst 90% of the 19 patients
Patients were then categorised as either not osteo- who are osteoporotic lie below the higher threshold of 0.1.
porotic (T-score . –2.5) or osteoporotic (T-score # –2.5), The limited data points in the osteoporotic group do not
dependent upon their lowest axial T-score, i.e. total allow an accurate cut off value to be selected.
femur or spine (L2–4). 23 of these patients were found to
have normal axial T-scores, 34 were osteopenic and 19
patients were osteoporotic. Threshold values were
initially calculated using this preliminary data so that
the Apollo DXA scanner could be used in clinical
practice as quickly as possible, although insufficient
patients had been examined for the data to achieve the
statistical accuracy recommended by NOS, i.e. 95%
confidence that the true sensitivity and specificity do
not fall below 80%. The initial lower threshold was a
T-score of –2.2. The initial upper threshold was 0.1,
although this value was obviously very inaccurate since
there was a large difference in the T-scores of the patients
at the cut off point and those patients immediately above
and below them, as is demonstrated in Figure 1.
Patients were seen in a primary care setting by an
osteoporosis prevention specialist nurse, who performed
heel DXA measurements on patients considered at risk
and referred those patients for axial densitometry with
heel T-scores below the upper threshold value of 0.1.
Heel and axial T-scores were recorded for every patient
referred in this way. Data for a further 139 patients
between the ages of 50 years and 75 years were recorded,
mean age 64.9 years.

Results
In total, heel and axial DXA T-scores were measured
for 215 women, mean age 64.6 years. Of these, 71 patients
were found to be osteoporotic on axial densitometry and
144 patients were not osteoporotic.
Figure 2 shows a scatter plot of heel T-scores for
patients in the two diagnostic categories. The line
Figure 2. Scatter plot of all heel T-scores for patients in the
two diagnostic categories; 90% of patients who have a T-
score above the lower threshold of 22.2 do not have
osteoporosis, whilst 90% of patients who have a T-score
below the upper T-score threshold of 21.2 are osteoporotic.
Patients with heel T-score between the two threshold values
should be referred for axial densitometry.

The British Journal of Radiology, June 2006 465


through the not osteoporotic group shows the lower
threshold, a T-score of 22.2; 90% of patients who have a
T-score above this threshold do not have osteoporosis.
The line through the osteoporotic group shows the upper
threshold, a T-score of 21.2; 90% of patients who have a
T-score below this threshold are osteoporotic.
Discussion
The information provided by the manufacturers of the
Norland Apollo heel densitometer states that a heel T-
score of # 22.5 is indicative of osteoporosis. This
assumption that the WHO criterion for axial DXA can
be applied to heel DXA has been previously demon-
strated to be incorrect. Only 15 of the 71 patients (21.1%)
who were found to be osteoporotic on axial densitometry
had heel T-scores of # 22.5.
Data were initially acquired so that a working thresh-
old value could be quickly calculated and used within
the primary care setting. These original 76 patients had
their hip, spine and heel DXA measurements all
performed on the same day. Subsequently, patients were
then referred for axial DXA on the basis of established
referral criteria or on the value of their heel DXA
measurement. These latter results will therefore intro-
duce some bias in the data, as patients who were found
to have high T-scores at the heel, above the higher
threshold value, were not referred for axial DXA, despite
the presence of other risk factors. These patients, with a
very low risk of axial osteoporosis, were therefore
excluded from the study as it was the intention of the
initial threshold values to avoid unnecessary referrals for
axial DXA. Since patients with high heel T-scores were
excluded, this bias will have the effect of lowering the
average T-score measured in the latter group of patients.
These patients represent 65% of the total patient group.
This measurement bias prevents these data from being
incorporated into the NOS published data for device
specific thresholds. Also, NOS consider only women
between the ages of 55 years and 70 years, whereas these
data have been acquired for patients between 50 years
and 75 years of age. However, the acquisition of truly
unbiased data within a more narrow age range and with
enough patients for statistical accuracy, i.e. 95% con-
fidence that the true sensitivity and specificity do not fall
below 80%, would take more than 10 months with this
department’s current workload. This estimation assumes
that only half of all patients referred are within the
required age range (based on 2004 data) and less than
one third of patients in this age group are likely to be
osteoporotic.
These biased data therefore, may be considered an
approximation of the true thresholds for the Norland
Apollo heel densitometer. The true values are likely be
slightly higher, due to the exclusion of patients with high
T-scores, therefore caution is advised with heel T-scores
just above the upper threshold of 21.2. These patients
should be considered to have an equivocal result and be
referred for axial DXA to prevent misdiagnosis. Patients
with heel T-scores just above the lower threshold (22.2)
will be referred for axial densitometry rather than be
given a diagnosis of osteoporosis from the heel DXA
alone. In this case, no patients should be misdiagnosed,
466
E McCauley, A Mackie, D Elliott and A Chuck
but the number of referrals to axial DXA will therefore be
higher than if the threshold values were unbiased.
The lower threshold of 22.2 was unchanged from that
calculated using the initial group of 57 patients. The
higher threshold was significantly lowered to 21.2 when
a total of 70 osteoporotic patients were considered,
compared with the initial value of 0.1 which was
calculated from only 19 patients. The difference in these
figures is not surprising, and is simply due to the
variability of T-scores within small numbers of patients.
Conclusions
Thresholds have been calculated for a Norland Apollo
heel densitometer; patients with a heel T-score of above
21.2 are very likely to have normal bone density on axial
densitometry, whilst patients with heel T-score of below
22.2 are very likely to have osteoporosis at the hip or
spine. Patients with heel T-scores that lie between these
two thresholds should be referred for axial densitometry.
Users of a Norland Apollo heel densitometer may wish
to implement these threshold values, with the caveat that
they are an approximation but are the best data currently
available for this scanner.
References
1. World Health Organization. Assessment of fracture risk
and its application to screening for postmenopausal
osteoporosis: Report of WHO Study Group. Technical
report series 843. Geneva, Switzerland, WHO 1994.
2. Placide J, Martens MG. Comparing screening methods for
osteoporosis. Curr Womens Health Rep 2003;3:207–10.
3. Fordham JN, Chinn DJ, Kumar N. Identification of women
with reduced bone density at the lumbar spine and femoral
neck using BMD at the os calcis. Osteoporos Int
2000;11:797–802.
4. Miller PD, Njeh CF, Jankowski LG, Lenchik L. International
Society for Clinical Densitometry Position Development
Panel and Scientific Advisory Committee. What are the
standards by which bone mass measurement at peripheral
skeletal sites should be used in the diagnosis of osteoporo-
sis? J Clin Densitom 2002;5Suppl.:S39–45.
5. Michalska D, Zikan V, Stepan J, Weichetova M, Kubova V,
Krenkova J, et al. X-ray densitometry and ultrasonography
of the heel bone – sensitivity and comparison with
densitometry of the axial skeleton. Cas Lek Cesk
2000;139:231–6.
6. Genant HK, Grampp S, Glueer CC, Faulkner KG, Jergas M,
Engelke K, et al. Universal standardisation for dual x-ray
absorptiometry: patient and phantom cross-calibration
results. J Bone Miner Res 1994;9:1503–14.
7. Pearson J, Dequeker J, Reeve J, Felsenberg D, Henley M,
Bright J, et al. Dual x-ray absorptiometry of the proximal
femur: normal European values standardized with the
European spine phantom. J Bone Miner Res 1995;10:315–24.
8. Formica CA. Standardization of BMD measurements
(editorial). Osteoporos Int 1998;8:1–3.
9. Grigorian M, Shepherd JA, Cheng XG, Njeh CF, Toschke JO,
Genant HK. Does osteoporosis classification using heel BMD
agree across manufacturers? Osteoporos Int 2002;13:613–7.
10. Faulkner KG, Roberts LA, McClung MR. Discrepancies in
normative data between Lunar and Hologic DXA systems.
Osteoporosis Int 1996;6:432–6.
11. Leib ES, Lewiecki EM, Binkley N, Hamdy RC. Official
positions of the International Society for Clinical
Densitometry. J Clin Densitom 2004;7:1–6.
The British Journal of Radiology, June 2006
Heel bone densitometry
12. Blake G, Chinn D, Steel S, Patel R, Panayiotou E, Thorpe J,
et al. The Revised NOS Position Statement on peripheral
X-ray Absorptiometry: A listing of device specific T-score
thresholds for the clinical interpretation of pDXA examina-
tions. NOS, 2005.
The British Journal of Radiology, June 2006
13. Royal College of Physicians. Osteoporosis: Clinical guide-
lines for prevention and treatment. London: RCP, 1999;7:66.
14. Williams ED, Daymond TJ. Evaluation of calcaneus bone
densitometry against hip and spine for diagnosis of
osteoporosis. Br J Radiol 2003;76:123–8.
467
 The British Journal of Radiology, 79 (2006), 468–472

The impact of a short course of study on the performance of

radiographers when highlighting fractures on trauma
radiographs: ‘‘The Red Dot System’’
S J MACKAY, TDCR, MSc, PhD

Salford University, Frederick Road Campus, Salford, Manchester M6 6PU, UK

ABSTRACT. This study was set up to determine the effect of a short course of study on
radiographers’ ability to identify fractures in a trauma context (Red Dot system). The
performance of 133 radiographers attending Red Dot courses delivered from April 1999
to September 2003 was evaluated prospectively using a selected sample of radiographs.
Assessment of fracture identification occurred before, immediately after and 6 months
after the course. The same film set was used and consisted of 30 axial and appendicular
cases, 18 with fractures and 12 normal or normal variants. Following a test for
normality of the data, a Wilcoxon Signed-Rank Test was selected and paired tests were
done between each assessment for sensitivity and specificity. There were significant
differences in sensitivity between all three assessments (p#0.05), the pre-course/post-
course and the post-course/6 month comparison being highly significant (p#0.01).
Specificity showed significant differences between the pre-course/post-course (p#0.01)
and the post-course/6 month follow-up scores (p#0.05), but no difference was found Received 8 March 2005
between the 6 months/pre-course scores. The participants improved their ability to Revised 7 October 2005
identify fractures and this appears to be as a result of the course. This improvement was Accepted 31 October 2005
not demonstrable after 6 months, although only 30% of participants took part in this
DOI: 10.1259/bjr/53513558
follow-up. Radiographers can improve their image interpretation skills from a short
course of study, but probably need continuing professional development to maintain ’ 2006 The British Institute of
these skills. Radiology

The Health Professions Council standards of profi-
ciency for radiographers [1] are designed to ensure the
safe and effective practice of radiographers in the UK.
They state that part of the role of the radiographer is to
be able to ‘‘distinguish between normal and abnormal
appearances evident on images’’ [1].
The Quality Assurance Agency benchmark statements
[2], which also describe the role of the radiographer, but
from an educational perspective, are more precise on this
matter and state that radiographers should be able to
‘‘recognize and respond appropriately to abnormal,
aberrant and pathological appearances on radiographic
images’’ [2].
In an accident and emergency setting, this practice is
termed a Red Dot system [3] and radiographers use these
skills to alert the casualty officer to fractures.
It is now over 30 years since Swinburne [4] suggested
that radiographers could be used to distinguish between
normal and abnormal radiographs, and such trauma
flagging systems were reported in 2000 to be in operation
in over 85% of accident and emergency departments in
the UK (excluding Northern Ireland) [5]. Several authors
have provided evidence of the potential of radiographers
to perform this role with training [5, 6] and without [7].
However, these evaluations have limited external valid-
ity. This is mainly due to methodological limitation.
These studies used either small numbers (n57) in one
department [6], a small non-random national sample
(n522) [5], or radiographers untrained in triaging
casualty radiographs [7]. Many studies have also
purposefully selected senior or experienced radiogra-
phers as subjects and cannot therefore provide evidence
on the performance of the range of radiographers
performing Red Dot or similar systems in the UK today.
A short course of study in image interpretation to
enable radiographers to highlight abnormal trauma
radiographs has been delivered at the author’s university
since 1999 and has attracted radiographers of varying
experience from throughout the UK. Their performance
has been evaluated in an attempt to add to the evidence
base in radiographer’s performance in Red Dot systems.
One study [4] selected senior or experienced radio-
graphers, the assumption presumably being that their
experience of image interpretation in their role as a
radiographer will have an effect on their film viewing
performance. This study will also investigate the
relationship between a radiographer’s years of experi-
ence and performance in a Red Dot scheme.
Method
Overview
The performance of radiographers attending one of
seven Red Dot courses delivered from April 1999 to
September 2003 was evaluated prospectively using a
selected sample of radiographs. They were assessed

468 The British Journal of Radiology, June 2006

Impact of a short course on radiographers’ ability to detect fractures
before the start of the course, immediately after the
course and then after 6 months of clinical practice. The
same film series was used for each evaluation.
The course
This was a 2 day course with short keynote lectures and
small group tutorials delivered by radiologists and report-
ing radiographers. Topics covered were the identification
of fractures and some non-traumatic pathologies of the
axial and appendicular skeleton along with normal and
abnormal variants. Both adult and paediatric patients were
included so as to represent the range of clinical work that
might present to an accident and emergency department.
The sample
The 133 radiographers from across the UK self-selected
for the courses. Their clinical experience ranged from
newly qualified radiographers to those with 36 years of
experience (mean512.1 years, SD510.4). 36 (27.0%) of the
sample had no experience of working within a Red Dot
system before and, of the 97 (73%) who had, the maximum
number of years was 20 years (mean53.1 years, SD53.4).
Consent was obtained from the participants for the use of
their test scores, which form part of the course feedback, in
this research and it was explained that the data would be
anonymized so no individuals could be identified.
The film assessment
The sample of radiographs were selected by a consultant
radiologist and a reporting radiographer who aimed to
make them representative of the range and quality of films
that might pass through an accident and emergency
department at a district general hospital. There were 30
cases of which 18 were pathological and 12 normal or
normal variant. The pathological films had been reported
upon by a consultant radiologist in practice and the
consultant and reporting radiographer who selected the
film were also agreed on the diagnosis. 10 radiographs
were of the upper and nine of the lower limbs, seven spine
and pelvis, and four skull and face. Three were paediatric
cases.
The assessment was undertaken at the university
under controlled conditions. Radiographs were viewed
in one session and each radiographer was allowed to
view a case for 90 s before recording one of two answers,
either normal (which included normal variants) or
abnormal. For the 6 month follow-up assessment, radio-
graphers were invited back to the university to complete
the assessment. Those who were unable to do this were
sent a CD containing the digitized film set along with
instructions regarding the conditions in which to view
the radiographs, but this could not be controlled.
Analysis
True and false, and positive and negative scores
were calculated to determine sensitivity, specificity,
The British Journal of Radiology, June 2006
and positive and negative predictive values. Accuracy
values were not used, following the advice of Maisey
and Hutton [8] who describe this as a global parameter,
‘‘…which is often misleading as a measure of test
performance and is rarely of any value’’ [8].
The sensitivity and specificity data obtained was
inspected for normality using a Kolmogorov-Smirnov
statistic [9]. This showed that the distributions were
significantly different from the normal distribution for
the pre-course and post-course (p,0.01) sensitivity data
and the pre-course (p,0.05) and post-course (p,0.01)
specificity data. No significant differences were found
for the 6 month data for either variable. Therefore data
were treated as non-parametric. Comparison was
required of pairs of scores in relation to the course, e.g.
pre/post-course; so the non-parametric Wilcoxon
Signed-Rank Test [9] was selected for analysis. All
analyses were performed using the Statistical Package
for the Social Sciences (SPSS v11.5; SPSS, Chicago, IL)
and a 0.05 significance level.
The number of years subjects had been qualified
and the number of years they had spent working within
a Red Dot scheme were correlated (Spearman’s Rho,
p# 0.05) with sensitivity and specificity values to
examine the effect of experience on image interpretation
performance.
Errors
False negative errors constitute missed fractures and
could be considered as the more important type of
detection error, which could lead to significant morbidity
and/or legal liability. Therefore, the most commonly
occurring cases of false negatives were highlighted and
described. This will help to inform course curricula
and enable radiographers to focus their professional
development.
Results
Response rate
133 (100%) radiographers completed the pre-course
assessment and 132 (99.2%) the post-course assessment.
39 (29.3%) radiographers completed the 6 month follow
up with 10 of these (25.6% of the responders at 6 months)
doing so using the digitized film set.
Sensitivity and specificity
The results are presented in Tables 1 and 2, and box
plots (using SPSS v11.5) are presented in Figures 1 and 2.
The positive and negative predictive values were not
analysed using inferential statistics and are presented for
comparative purposes only.
The median scores for sensitivity show that the
radiographers achieved a score of 78.9% initially. This
improved to 88.2% following the course, but then fell
away to less than the pre-course score after 6 months. All
differences were statistically significant.
469
 S J Mackay

Table 1. Breakdown of raw scores post-course and 6 month follow-up. Visual inspection of
the box plot in Figure 2 demonstrate that the radio-
Median Interquartile range graphers did worse immediately after the course, with
Sensitivity 25 75 more scores below the median. They then improved
slightly at 6 months but were not significantly different
Pre-course 78.9 70.6 88.2 from the pre-course score.
Post-course 88.2 82.4 94.1
6 months 76.5 64.7 88.2
Specificity 6 month follow-up
Pre-course 76.9 69.2 85.0
Further analysis was undertaken as less than one third
Post-course 76.9 62.0 85.0
6 months 76.9 69.2 86.0
of the subjects completed the 6 month follow-up. This
investigated whether this group of subjects (the com-
Positive predictive value pleters) had scored differently at other assessment times
Pre-course 86.0 81.0 93.0 to those who did not complete the assessment. This
Post-course 84.2 79.0 89.1 would help to clarify whether the completers were
6 months 84.4 79.0 92.0 representative of the group as a whole. A Wilcoxon test
found no statistically significant differences between the
Negative predictive value completers and non-completers for number of years
Pre-course 69.6 62.2 79.0 practicing radiography or the number of years in a Red
Post-course 77.8 70.3 88.0 Dot scheme. The differences in the pre- and post-test
6 months 65.7 56.0 78.0 scores for these groups are displayed in Table 3.

The median scores for specificity were the same, but

Effect of experience
the distributions varied and showed a highly statistically
significant difference between the pre-course and post- No significant correlations were found between the
course tests, and significant differences between the number of years since qualification or length of time in a

Table 2. Wilcoxon Signed-Rank Test results for sensitivity and specificity

Pre-course/post-course Post-course/6 months Pre-course/6 months

Sensitivity Median change 9.3 –11.7 –2.4

Median difference 5.9 –11.8 0
Z –7.049 –4.190 –1.984
p, 0.01 0.01 0.05
Specificity Median change 0 0 0
Median difference –7.7 7.7 0
Z –4.135 –1.983 –0.587
p, 0.01 0.05 NS
NS, not significant.

Figure 1. Sensitivity changes over

time. Box plot where the box length
is the interquartile range with the
median identified as a thick black
line traversing the box. Outliers
(ringed) are cases with values
between 1.5 and 3 box lengths from
the upper or lower edge of the box.
From SPSS v11.5.

470 The British Journal of Radiology, June 2006

Impact of a short course on radiographers’ ability to detect fractures

Figure 2. Specificity changes over

time. Box plot where the box length
is the interquartile range with the
median identified as a thick black
line traversing the box. Outliers
(ringed) are cases with values
between 1.5 and 3 box lengths from
the upper or lower edge of the box.
From SPSS v11.5.

Red Dot scheme and sensitivity or specificity scores pre- to the course, the main bulk of which shows students
course, post-course or 6 months after the course. how to recognize fractures. The students may then have
an immediate tendency to look for the abnormal rather
than the normal.
False negative errors A similar study [5] followed up subjects for 6–10 weeks
and found a decrease in sensitivity which, although not
The 10 most common false negative errors before and statistically significant, would support the trend found in
immediately after the course were ranked and collated (11 this work suggesting that improvements in performance
cases appear in column 1 as there were 2 cases of 10.9% were short lived. This suggests that radiographers need to
tied for 10th place). These key errors are where further continue to acquire the knowledge and skills necessary for
learning should be focused. These can be found in Table 4. interpreting images following any course of study. Based
on these data, they should be topping up their knowledge
and skills before a period of 6 months has elapsed.
Discussion In light of the impending mandatory continuing
professional development (CPD) requirements of the
The ability of the radiographers to identify the Health Professions Council, radiographers will need to
abnormal trauma radiographs, sensitivity, increased consider undertaking a range of CPD activities to
immediately after the course, suggesting that they had consolidate and enhance their image interpretation skills.
learned from the course. This improvement declined
over 6 months and ended up being worse than before
they started the course. Table 4. Most commonly occurring false negative errors
The ability of the radiographers to identify normality
Case description Proportion of subjects making false
on trauma radiographs, specificity, decreasd after the negative errors per case (%)
course and then improved slightly, but was then no
different from the original scores. This ‘‘post-course dip’’ Pre-course Post-course
in score was also identified by another author [5] and Skull fracture 50.0 63.0
appears to be a phenomenon inherent in the image Fracture of head of 5th 40.2 17.4
interpretation learning process. This may be due in part metatarsal with Os
Peroneum
Greenstick fracture of tibia 39.1 32.6
Table 3. Wilcoxon Signed-Rank Test results completers and
Slipped upper femoral 38.0 10.9
non-completers
epyphysis
Completers Completers Sacral fracture 27.2 20.7
pre-course/ post-course/ Fracture radius 22.8 7.6
non-completers non-completers Fracture radial head 18.5 12.0
pre-course post-course Subtle fracture base of 1st 17.4 0
Sensitivity Z –1.733 –2.646 metacarpal
p, NS 0.01 Fracture humeral head 16.3 10.9
Fracture lunate 10.9 25.0
Specificity Z –2.120 –1.169 Fracture base of 1st distal 10.9 0
p, 0.05 NS phalanx foot
Fracture zygoma 0 16.3
NS, not significant.

The British Journal of Radiology, June 2006 471


Table 5. Comparison of pre-training sensitivity and specificity levels with other authors
Sensitivity (%) Specificity (%)
Loughran [10] 62.7 97.9
McConnell [5] 92 64
Hargreaves [6] 76.2 96.4
Mackay (this study) 78.9 76.9
It is acknowledged that the small sample that
completed the 6 month assessment may not have been
representative of the group as a whole. However,
comparisons between the completers and non-comple-
ters suggest that on the measures of length of experience
since qualification and working in a Red Dot scheme
they were not significantly different. In addition, the
differences in performance between these two groups
were on the post-course sensitivity scores in which the
completers did better. This would suggest that the scores
from this study at 6 months may even have exaggerated
the performance of the whole group of radiographers.
The specificity scores between these groups were not
significantly different after the course and are therefore
likely to be representative.
Comparison with other results [5, 6, 10] shows a range
of sensitivity and specificity levels for radiographers
(Table 5). Close statistical comparison is difficult as the
measures of central tendency used are different.
However, looking across the values it would appear
that there is a need for radiographers to improve their
performance in image interpretation within Red Dot
protocols. No published performance levels were found
for radiographers working in a Red Dot scheme.
However, post-graduate reporting radiographers are
expected to achieve 90–95% sensitivity and specificity
levels [11]. If we apply this standard of performance to
Red Dot radiographers, they are clearly falling short of
this level.
One notable trend when comparing the data in this
study with two others [5, 6] is that subjects in one [5]
tended to over report (high sensitivity and low specifi-
city) and in the other [6] tended to under report (low
sensitivity and high specificity). In the current study,
relative to these other two, the subjects maintained a
more stable performance over both measures.
The areas where radiographers missed fractures have
been highlighted in this study and could be added to
those from another study [6]. These data could help
to inform ‘‘Red Dot’’ course developers and evaluators to
take a more evidence based approach to radiographer
education.
Experience did not appear to be a factor in the
performance of this group of radiographers. Further
research is needed to explore the phenomenon of
‘‘experience’’ in radiography and try to tease out the
factors that might affect the performance of radio-
graphers in image interpretation.
472
S J Mackay
Year published Measure
1996 mean
2000 median
2003 mean
2006 median
Conclusions
The radiographers did appear to learn from the short
course and improved their levels of performance in
image interpretation. These improvements were short
lived and it is recommended that regular updating is
undertaken by radiographers working within a Red Dot
scheme.
There appeared to be a reduction in performance
immediately following the image interpretation course,
which caused an increase in false positives.
Radiographers should be aware of this and attempt to
counteract the effect. Programme evaluators should
factor in this effect when planning any assessments of
image interpretation performance after a short course of
study.
The scores achieved in this evaluation fall well short of
the 90–95% sensitivity and specificity levels used in
radiographer reporting. Radiographers would benefit
from continuing to improve their image interpretation
skills throughout their working lives.
References
1. Health Professions Council. Standards of Proficiency:
Radiographers. London, UK: HPC; 2003.
2. The Quality Assurance Agency for Higher Education.
Benchmark Statement for Radiographers. QAA; 2001.
3. Berman L, de Lacey G, Twomey E, Twomey B, Welch T,
Eban R. Reducing errors in the accident and emergency
department :a simple method using radiographers. Br Med
J 1985;290:421–2.
4. Swinburn K. Pattern recognition for radiographers. Lancet
1971;1:589–90.
5. McConnell JR, Webster AJ. Improving radiographer high-
lighting of trauma films in the Accident and Emergency
department with a short course of study – an evaluation. Br
J Radiol 2000;73:608–12.
6. Hargreaves J, Mackay S. The accuracy of the red dot system:
can it improve with training? Radiography 2003;9:283–9.
7. Renwick I, Butt W, Steele B. How well can radiographers
triage x-ray films in accident and emergency departments?
Br Med J 1991;302:568–9.
8. Maisey MN, Hutton J. Guidelines for the evaluation of
radiological technologies. Report No.: 13874. London, UK:
British Institute of Radiology, 1995.
9. Pallant J. SPSS survival manual (SPSS v10). 1st edn.
Buckingham, UK: O.U. Press; 2001.
10. Loughran CF, Raynor J, Quine M, Mulley A. The red dot
system: how good is it? In: Radiology UK; May 1996.
11. Salford University. MSc Advanced Practice (Clinical
Radiographic Reporting) Programme Handbook. 2004.
The British Journal of Radiology, June 2006
The British Journal of Radiology, 79 (2006), 473–478

Portal vein embolisation prior to hepatic resection for colorectal

liver metastases and the effects of periprocedure chemotherapy
1
I K BEAL, MBBS, MRCP, FRCR, 1S ANTHONY, MBBS, MRCP, FRCR, 1A PAPADOPOULOU, MBBS, FRCR,
2
R HUTCHINS, MBBS, MS, FRCS, 2G FUSAI, FRCS, 3R BEGENT, MBBS, MRCS, LRCP, 1N DAVIES, MBBS, FRCS, FRCR,
1
J TIBBALLS, MRCP, FRCR, FRANZCR and 2B DAVIDSON, MBChB, MD, FRCS

Departments of 1Radiology 2HPB Surgery and 3Medical Oncology, Royal Free Hospital NHS Trust
and Royal Free and University College School of Medicine, Pond Street, London NW3 2QG, UK

ABSTRACT. Portal vein embolisation (PVE) is an effective method of increasing future

liver remnant (FLR) but may stimulate tumour growth. The effect of periprocedure
chemotherapy has not been established. 15 consecutive patients underwent PVE prior
to hepatic resection for colorectal liver metastases with a FLR ,30% of tumour-free
liver (TFL). Liver and tumour volumes pre-PVE and 6 weeks post-PVE were calculated by
CT or MRI volumetry and correlated with the periprocedure chemotherapy regimen.
PVE increased the FLR from 18¡5% of TFL to 27¡8% post-PVE (p,0.01). Post-PVE
chemotherapy did not prevent hypertrophy of the FLR but the volume increase with
chemotherapy (median 89 ml, range 7–149 ml) was significantly reduced (median
135 ml, range 110–254 ml without chemotherapy) (p50.016). Tumour volume (TV)
Received 26 April 2005
decreased in those receiving post-PVE chemotherapy (median TV decrease 8 ml, range Revised 2 October 2005
277 ml to +450 ml) and increased without chemotherapy (median TV increase 39 ml, Accepted 14 November
range 258 ml to +239 ml). Of the 15 patients, eight underwent resection; four were not 2005
resected due to disease progression and three due to insufficient hypertrophy of the
DOI: 10.1259/bjr/29855825
FLR. PVE increased the FLR by an average of 9% allowing resection in 50% of patients.
Periprocedure chemotherapy did not prevent but did reduce hypertrophy. A trend ’ 2006 The British Institute of
towards tumour regression was observed. Radiology

Liver resection surgery provides the main possibility
of cure in patients with colorectal liver metastases. 5-year
survival ranges from 25% to 40% [1–4]. One of the
contraindications to hepatic resection is a small future
liver remnant (FLR), which usually arises in patients
with congenitally small left lateral segments (II and III)
who require an extended right hepatectomy. The small
residual liver volume increases the risk of post-operative
liver failure. A FLR of 25–30% of non-tumour liver
volume has been recommended in patients with normal
liver function who are to undergo hepatic resection [3, 5–
8], although resections have been carried out successfully
with a FLR of ,20% [9]. Selective portal vein embolisa-
tion (PVE) can produce atrophy of the segments affected
by the cancer and compensatory hypertrophy of the
contralateral segments [6–8]. The technique was first
developed in patients with hilar cholangiocarcinoma
where extended right hepatectomy is often required [10].
Although residual liver volumes can be increased, there
remains some controversy as to whether this results in
improved post-operative outcome [11]. There are now
several reports in the literature of PVE applied to
patients with colorectal metastases [6–8, 12–14].
One of the potential difficulties with using PVE in
metastatic colorectal cancer is that tumour growth may
be promoted by PVE. Kokudo and colleagues [15]
studied patients undergoing PVE prior to hepatic
resection for colorectal liver metastases and found
that PVE increased tumour growth in the embolised
segments, possibly increasing the incidence of recurrence
following liver resection. This also risks tumours
becoming unresectable during the interval between
PVE and surgery. The growth of liver metastases in the
non-embolised segments following PVE has also been
shown to be greater than the hypertrophy rate of the
normal liver parenchyma [16]. However, in a study of
the long term survival following PVE, which included 41
patients with colorectal cancer (CRC) liver metastases,
there was no evidence to suggest that patients whose
surgery had been made possible by PVE were associated
with a poorer long term survival [7].
PVE has also been applied to patients with hepatocel-
lular carcinoma and has been associated with improved
operative outcomes as well as long term survival [17, 18].
Chemotherapy has been shown to increase the
survival of patients with unresectable liver metastases
[1]. Downstaging chemotherapy using Oxaloplatinum
and 5-Fluorouracil (5FU) allows some initially unresect-
able CRC liver metastases to be successfully resected
[19–21].
The effect of chemotherapy administered after PVE on
the FLR hypertrophy and on the tumour growth in
embolised segments, in patients with colorectal liver
metastases, has not been investigated. The aim of this
study was to assess, first, whether chemotherapy
prevents tumour progression between PVE and surgery
and, second, whether chemotherapy has a detrimental
effect on FLR hypertrophy.

The British Journal of Radiology, June 2006 473


Patients and methods
Between September 1999 and December 2002, a single
surgical firm in a specialist tertiary referral centre
performed right or extended right hepatectomy for colo-
rectal liver metastases in 57 patients. Over this time period,
15 patients who required a right or extended right
hepatectomy to resect their metastases were found to have
a calculated FLR of less than 30% in an otherwise normal
liver (n514), or less than 40% (n51) in a diseased liver
(imaging findings consistent with steatosis). These patients
underwent right PVE. Nine patients had synchronous and
six metachronous colorectal liver metastases. There were
six females and nine males, mean age 65 years (range 52–
74 years). The stage of the original tumour was Dukes’ B in
4 and Dukes’ C in 11. The mean number of metastases was
4 (range 1–11). Three of the 15 patients had undergone
resection for metastases in the left lobe of liver prior to
PVE. Three patients had minor extension of tumour into
segment IV, but no patient had discrete metastases in non-
embolised liver.
Chemotherapy
Following initial diagnosis and bowel cancer resection,
adjuvant chemotherapy had been carried out in the
referring hospital in 11 of the 15 patients (Dukes’ C n59,
Dukes’ B n52). In all cases this comprised a standard
5FU and folinic acid based protocol. In five of these
patients, this had been completed between 6 months and
18 months prior to PVE. In six patients (synchronous
metastases n55, metachronous metastases n51) che-
motherapy was continued until 1 month prior to PVE
(5FU/FA (n52) or 5FU/Oxaliplatin (n54)). In the
remaining four patients no chemotherapy was adminis-
tered following bowel cancer resection. Following PVE,
eight patients who had already received chemotherapy
underwent a further course of 5FU/FA (n51), 5FU/
FA+oxaliplatin (n55) or 5FU/FA+Irinotecan (n52).
Two patients who had not previously had chemother-
apy were given post-PVE chemotherapy with 5FU/FA
+oxaliplatinum (n51), 5FU/FA alone (n51).
Chemotherapy post-PVE was started at 2 weeks post-
embolisation and continued for 6 weeks. Three patients
who had received chemotherapy post-bowel resection or
prior to embolisation did not receive post-PVE che-
motherapy.
Two patients remained chemo-naive.
For the purpose of analysis, the patients were grouped
into post-PVE chemotherapy (n510) and no chemother-
apy post-PVE (n55).
Volumetric analysis
Liver volume measurements of the whole liver (WL),
segments I-III or I-IV (FLR) and tumour (TV) were
performed by CT (n52) (GE Medical Systems High
speed system, Milwaukee, WI) or MRI (n513) using
Philips Intera 1 T system (Philips, Best, Netherlands).
Total functioning liver (TFL) was measured by subtract-
ing TV from WL (Philips Easyvision workstation [release
4.3]) and GE Medical Systems Advantage Windows 2.0
474
I K Beal, S Anthony, A Papadopoulou et al
3D Analysis). The technique for measuring volumes by
CT has been described elsewhere [5, 7]. Briefly, axial
10 mm sections through the liver were obtained in a
single breath-hold. Using a workstation, the areas of TV,
WL and FLR were calculated by multiplying the area of
each liver image by the slice thickness. The percentage
fraction of FLR was calculated: FLR/TFL6100. The
measurements performed by MRI involved the same
process using the scanner software and a similar
technique as described by Caldwell et al [22]. An
experienced Hepatobiliary Radiologist performed all
measurements. The liver volume measurements were
repeated approximately 6 weeks post-PVE.
Portal vein embolisation (PVE)
This was performed by a percutaneous transhepatic
approach under general anaesthetic (n52) or local
anaesthesia and intravenous sedoanalgesia (n513). An
ultrasound-guided percutaneous portal vein puncture
was performed using a Neff set (William Cook Europe
A/S, Bjaeverskov, Denmark). A left-sided approach was
preferred (n59) but when not possible due to the small
size of the left lobe a right-sided approach was used
(n56). 5 Fr or 6 Fr sheaths were introduced and the
branches of the right portal vein were embolised
sequentially using 4 Fr cobra or sidewinder shaped
catheters. Embolisation was performed with alcohol
(n53) and with/without 500 mm polyvinylalcohol
(PVA) particles (William Cook Europe A/S,
Bjaeverskov, Denmark) (n52/1, respectively) or enbu-
crilate tissue adhesive (Histoacryl; B Braun Medical AG,
Emmenbrucke, Switzerland) and Lipiodol (Guerbet,
Roissy, France) at an enbucrilate:Lipiodol ratio of 1:3–
1:5 (n512). Prior to PVE and daily for 2 days post-PVE,
full blood count, clotting profile and liver function tests
were measured.
Statistics
Paired and two sample equal variance t-tests were
used for statistical analysis. Data are expressed as means
(¡ 1 SD), median values with range and 95% confidence
intervals (CIs). p,0.05 was considered significant.
Results
Procedure of portal vein embolisation (PVE)
Embolisation of the right portal vein was technically
successful in all 15 patients with no immediate complica-
tions. The first three patients were embolised with
alcohol, which produced severe right upper quadrant
pain. Subsequent embolisations were performed using
enbucrilate. Post-embolisation, most patients had mild
discomfort and low-grade pyrexia, which settled within
24 h. In one patient the 6-week follow-up MRI demon-
strated complete portal vein thrombosis. This had not
been recognized at the time of the procedure and his
clinical course post-PVE was unremarkable. A CT scan
showed evidence of enbucrilate:Lipiodol within the
The British Journal of Radiology, June 2006
Portal vein embolisation and the effects of periprocedure chemotherapy

Table 1. Effect of portal vein embolisation (PVE) on liver Table 2. The effect of chemotherapy on liver hypertrophy
volumes
Patient Chemo Pre-PVE Pre-PVE Post-PVE Post-PVE
Pre-PVE Post-PVE p regimen FLR/TFL% FLR FLR/TFL% FLR
vol. ml vol. ml
Mean WL (ml) 1912 1761
Median WL 1739 (1222–3114) 1631 (1160–3398) 0.8 Post-PVE
chemo
(range) (ml)
Mean TV (ml) 235 262 1 16.5 297 25 395
Median TV 123 (29–1147) 139 (10–1597) 0.71 2 20 237 31 320
(range) (ml) 3 15 182 25 256
Mean FLR (ml) 270 400 4 22 423 30 533
Median FLR 281 (110–558) 404 (245–653) ,0.01 5 23 303 31 418
(range) (ml) 6 24 330 31 413
Mean 18(¡5) 27(¡8) ,0.01 7 12 258 19.5 407
FLR/TFL% 8 30 558 44 653
WL, whole liver volume; TV, tumour volume; FLR, future liver 9 17.5 344 22 398
remnant volume; TFL, tumour free liver volume. 10 15 254 15 261
No post-PVE
chemo
proximal left portal vein branch. This patient was not
excluded from analysis. 11 16 293 26 404
12 19 245 33 421
13 22 281 39 535
14 16 235 21 345
Effect of portal vein embolisation (PVE) on future 15 7.5 110 17 245
liver remnant (FLR) volume
Chemo regimen, chemotherapy regimen; PVE, portal vein
Following PVE, there was no significant change in embolisation; FLR, future liver remnant; TFL, tumour free
whole liver (WL) volume (median volume pre-PVE liver.
1739 ml (range 1222–3114 ml) to a median volume
post-PVE 1631 ml (range 1160–3398 ml)).
Volumetric analysis was carried out prior to PVE and median5135 ml, range5110–254 ml) (p50.016). The dif-
was repeated at a median of 6 weeks post-PVE (range 5– ference between the means was 70 ml (CIs526–115 ml).
12 weeks). Prior to PVE, the future liver remnant (FLR) The effects of chemotherapy on FLR hypertrophy are
volume ranged from 110 ml to 558 ml (median 281 ml), illustrated by Figures 1 and 2.
which represented 18(¡5)% of the tumour free liver
(TFL). The volume was significantly increased following
PVE to a FLR range of 245–653 ml (median 404 ml) Effect of chemotherapy on tumour volume (TV)
which represented 27(¡8)% of the TFL (p50.0089). The
mean increase in parenchymal volume was 123 ml, The effect of chemotherapy on TV following PVE is
which represented a mean increase in FLR/TFL ratio of shown in Table 3.
9% (Table 1). The one patient who had no significant There was a median TV decrease of 8 ml (range –77 ml
increase in FLR post-PVE was found to have complete to +450 ml) in those receiving post-PVE chemotherapy
portal vein thrombosis. (n510) and a median TV increase of 39 ml (range –58 ml
to +239 ml) in those who did not receive post-PVE
chemotherapy (p50.476). The difference between the
means was 58 ml (CIs5–82–198 ml).
Effect of portal vein embolisation (PVE) on tumour Of the three patients who had tumour extending into
volume (TV) segment IV, two showed no significant change in TV
The patient group had a median TV of 123 ml (range
29–1147 ml) prior to PVE and 139 ml following PVE
(range 10–1597 ml). This difference was not statistically
significant (p50.71). Of the 15 patients 8 had TV increase
and 7 had a TV reduction.

Effect of chemotherapy on liver hypertrophy

(future liver remnant (FLR) volume)
The effect of chemotherapy on the hypertrophy of the
FLR following PVE is shown in Table 2. There was a
significant increase in FLR following PVE in both groups.
Patients who received chemotherapy post-PVE (n510)
had less hypertrophy (increase in FLR volume med- Figure 1. Change in FLR volume after embolisation in all
ian589 ml, range 7–149 ml) than those who received no patients. FLR, future liver remnant volume; Pre, volume prior
chemotherapy post-PVE (n55) (increase in FLR volume to PVE; Post, volume after PVE.

The British Journal of Radiology, June 2006 475

 I K Beal, S Anthony, A Papadopoulou et al

(AST) in all patients (p50.04) and bilirubin in 12 patients

(p50.0007). An increase in international normalized ratio
(INR) was seen in eight patients (p50.019) and a decrease
in platelets noted in 13 patients (p50.002). A transient
increase in white cell count (WCC) was noted in 10
patients (p50.009).

Surgery following portal vein embolisation (PVE)

Figure 2. Change in FLR volume after embolisation, accord-
ing to chemotherapy administered. FLR, future liver remnant
volume; Pre, volume prior to PVE; Post, volume after PVE;
Post-PVE Chemo, chemotherapy administered after PVE; No
post-PVE Chemo, no chemotherapy administered after PVE.
post-PVE (no post-PVE chemotherapy n51, post-PVE
chemotherapy n51) and one had an overall increase in
TV but no significant change of the TV in segment IV (no
post-PVE chemotherapy).
The effects of chemotherapy on TV are illustrated by
Figure 3.
Figures 4 and 5 demonstrate typical MRI findings pre-
PVE and post-PVE in a patient receiving post-PVE
chemotherapy with hypertrophy of the FLR and tumour
volume regression.
Haematology and biochemistry
Following PVE there were significant but transient
increases in the levels of aspartate aminotransferase
Table 3. Chemotherapy regimen and tumour volume
following PVE
Patient Chemo Pre-PVE TV ml Post-PVE TV ml
regimen
Post-PVE
chemo
Of the 15 patients who underwent PVE who would
otherwise have been deemed unresectable due to their
small-anticipated FLR, eight proceeded to a right or
extended right hepatectomy (53%). Four patients had
disease progression on their post-PVE imaging at
6 weeks, which precluded potentially curative resection
(extrahepatic disease progression, n53 and intrahepatic
and extrahepatic disease progression, n51). Of these
four, two had received post-PVE chemotherapy and two
had not received chemotherapy following embolisation.
Three patients had insufficient hypertrophy of the FLR.
Of the eight who had surgery, there were no operative or
post-operative mortalities.
Follow up
Five of the eight patients undergoing surgery remain
well and free of local or distant disease at a median of
18 months post-liver resection (range 14–26 months).
Two patients developed distant metastatic disease and
the third has undergone abdominoperineal resection for
local recurrence of a rectal cancer. The three patients who
had insufficient hypertrophy of the FLR, including the
first patient with complete portal vein thrombosis,
underwent palliative chemotherapy. The second of the
three patients had an initial FLR of just 7.5% of the TFL.
Despite a 123% increase in FLR this was still only 17% of
TFL at 6 weeks post-PVE. The third patient had an
extremely large tumour prior to embolisation, measuring
1147 ml. The percentage increase in FLR in this patient
was just 16%. All three patients were included in the
analysis.

1 236 228 Discussion

2 64 32
3 232 155 PVE is one of the major recent developments to
4 203 242 facilitate resection in patients who would otherwise be
5 46 54 contraindicated for liver surgery due to a small-antici-
6 29 10 pated residual liver volume. Patients in this study all had
7 123 116 an anticipated FLR of 30% or less of tumour-free liver
8 179 139 and were therefore deemed at risk of post-operative liver
9 1147 1597
failure.
10 29 34
Alcohol was initially used for embolisation, but due to
No post-PVE severe alcohol-induced pain requiring general anaes-
chemo
thetic, subsequent embolisations were performed using
11 727 966 enbucrilate. This was well tolerated and produced
12 90 32 results comparable with that in other series. A literature
13 66 77 review has suggested no clear advantage to any specific
14 266 486 embolic substance [5].
15 120 159 One patient in the series developed complete portal vein
Chemo regimen, chemotherapy regimen; PVE, portal vein thrombosis. In retrospect, a number of factors may have
embolisation; TV, tumour volume. contributed to this complication. Due to the small left lobe

476 The British Journal of Radiology, June 2006

Portal vein embolisation and the effects of periprocedure chemotherapy
Figure 3. Change in tumour volume after embolisation,
according to chemotherapy administered. Pre, volume prior
to PVE; Post, volume after PVE; Post-PVE Chemo, chemother-
apy administered after PVE; No post-PVE Chemo, no
chemotherapy administered after PVE.
volume, the approach was right-sided and, on completion
of embolisation, the catheter was withdrawn so no post-
procedure venogram was obtained. This does not cause
the complication, but prevented it being recognized at the
time of the procedure. Embolisation of the left portal vein
may occur if too large a volume of enbucrilate:Lipiodol is
injected, i.e. due to operator error. This is not thought to
have occurred in this case. Anatomical variants such as
portal vein trifurcation, present in this case, may predis-
pose to left portal vein embolisation.
The ability of PVE to increase the FLR was confirmed in
the present study with a mean percentage increase in FLR/
TFL ratio of 9%. This figure correlates well with the
previous reports of 8–13% [5, 7, 8]. 50% of the patients
undergoing PVE were able to proceed with potentially
curative resection. In the series reported by Elias and
colleagues, 88% of patients proceeded to resection [7]. This
difference in resection rate may relate to the centre’s
selection of patients for PVE. In the present series the mean
pre-PVE FLR/TFL ratio was 18%. This is lower than the
19–32% reported by other centres [5, 7]. Four of the
patients in the present series had a FLR/TFL ratio of 15%
or less prior to PVE and only one of these underwent
resection. If a minimum FLR of 30% is to be achieved and
the percentage increase with PVE is in the order of 10%,
then patients with an FLR of less than 20% may not benefit
from PVE. The minimal remnant volume required for safe
resection, however, remains controversial and may be
Figure 4. MRI demonstrates axial section through the liver
prior to portal vein embolisation.
The British Journal of Radiology, June 2006
Figure 5. Post-embolisation: MRI demonstrates future liver
remnant volume (FLR) hypertrophy and tumour volume
regression.
increased by any underlying parenchymal abnormality,
such as steatosis. One series concluded that the hyper-
trophy of FLR induced by PVE had no beneficial effect on
the post-operative outcome in patients with normal liver
[11]. However, in this series, patients were undergoing a
standard right hepatectomy with a satisfactory pre-
embolisation FLR/TFL volume ratio of 33¡10%. The
same study concluded that in patients with chronic liver
disease, pre-operative PVE significantly decreased post-
operative complications. Another study has confirmed an
association between a small FLR and an increase in
post-operative complications [9]. Those patients with a
standardized FLR of 20% or less had an increase in post-
operative morbidity, but not mortality. A FLR of 40% has
been recommended in patients who have previously
undergone high dose chemotherapy [12].
A time interval between PVE and hepatic resection is
necessary to allow parenchymal hypertrophy. Although
the FLR growth is maximal within the first 2 weeks [5, 8] it
continues for 4–5 weeks following PVE [5, 6, 8]. For this
reason, in the anticipation of maximum hypertrophy, this
study allowed 6 weeks between PVE and repeat volume
measurement. The delay of 12 weeks before repeat volume
measurement in one patient was due to external circum-
stances and subsequent surgical resection was successful.
This is the first study to examine the use of chemo-
therapy in patients with colorectal liver metastases under-
going pre-operative PVE. Chemotherapy has usually been
discontinued prior to PVE [7, 12] to encourage hypertrophy
of the remnant liver parenchyma. In our study, chemo-
therapy was discontinued at least 1 month prior to PVE
and it has been assumed that effects of this chemotherapy
on post-PVE liver regeneration and oncogenesis would be
minimal or insignificant. Chemotherapy in these circum-
stances is unproven. The particular chemotherapy regimen
was administered according to individual oncologist
practice. There was no significant difference in disease
stage between the two groups. Due to the small study size,
only the effects of chemotherapy post-PVE have been
analysed. In the present study FLR hypertrophy occurred
whether or not chemotherapy was given post-PVE. The
hypertrophy was, however, significantly less in those
patients who received post-PVE chemotherapy suggesting
that although chemotherapy is not contraindicated, it
477

should be considered carefully in those patients who
require a large compensatory hypertrophy following the
procedure (FLR of ,20%).
There is evidence to suggest PVE may stimulate
tumour growth in both the embolised and non-embo-
lised lobes of the liver. We have only assessed tumours in
the embolised segments. Elias et al [16] reported
increased growth of liver metastases compared with
normal liver parenchyma following PVE in five patients
with tumour in the non-embolised lobe. Kokudo et al
[15] assessed tumour growth in the embolised lobe and
found oncogenesis was promoted by PVE, with the
tumour doubling time being reduced from 92 days to
76 days. In the present study, patients undergoing
chemotherapy had no overall increase in tumour
volume, which would suggest that chemotherapy did
inhibit the tumour growth associated with PVE. Tumour
volume regression was not demonstrated by the use of
chemotherapy but this would seem unlikely over the
short period between pre-PVE staging and re-assessment
at 6 weeks post-PVE. The confidence interval for the data
is wide, suggesting that further numbers would be
helpful.
This is the first study on the effects of peri-PVE
chemotherapy and the relationship with FLR hypertrophy
and tumour volume. Further studies are required to define
the optimal chemotherapy regimen for patients with
colorectal liver metastases who are undergoing PVE.
References
1. Scheithauer W, Rosen H, Kornek GV, Sebesta C, Depisch D.
Randomised comparison of combination chemotherapy
plus supportive care with supportive care alone in patients
with metastatic colorectal cancer. BMJ 1993;306:752–5.
2. Fong Y, Salo J. Surgical therapy of hepatic colorectal
metastasis. Semin Oncol 1999;26:514–23.
3. Fusai G, Davidson BR. Management of colorectal liver
metastases. Colorectal Dis 2003;5:2–23.
4. Jaeck D, Bachellier P, Guiguet M, Boudjema K, Vaillant JC,
Balladur P, et al. Long-term survival following resection of
colorectal hepatic metastases. Association Francaise de
Chirurgie. Br J Surg 1997;84:977–80.
5. Abdalla EK, Hicks ME, Vauthey JN. Portal vein emboliza-
tion: rationale, technique and future prospects. Br J Surg
2001;88:165–75.
6. de Baere T, Roche A, Vavasseur D, Therasse E, Indushekar
S, Elias D, et al. Portal vein embolization: utility for
inducing left hepatic lobe hypertrophy before surgery.
Radiology 1993;188:73–7.
7. Elias D, Ouellet JF, de Baere T, Lasser P, Roche A.
Preoperative selective portal vein embolization before
hepatectomy for liver metastases: long-term results and
impact on survival. Surgery 2002;131:294–9.
8. Madoff DC, Hicks ME, Abdalla EK, Morris JS, Vauthey JN.
Portal vein embolization with polyvinyl alcohol particles
and coils in preparation for major liver resection for
478
I K Beal, S Anthony, A Papadopoulou et al
hepatobiliary malignancy: safety and effectiveness--study
in 26 patients. Radiology 2003;227:251–60.
9. Abdalla EK, Barnett CC, Doherty D, Curley SA, Vauthey
JN. Extended hepatectomy in patients with hepatobiliary
malignancies with and without preoperative portal vein
embolization. Arch Surg 2002;137:675–81.
10. Makuuchi M, Thai BL, Takayasu K, Takayama T, Kosuge T,
Gunven P, et al. Preoperative portal embolization to
increase safety of major hepatectomy for hilar bile duct
carcinoma: a preliminary report. Surgery 1990;107:521–7.
11. Farges O, Belghiti J, Kianmanesh R, Regimbeau JM, Santoro
R, Vilgrain V, et al. Portal vein embolization before right
hepatectomy: prospective clinical trial. Ann Surg
2003;237:208–17.
12. Azoulay D, Castaing D, Smail A, Adam R, Cailliez V,
Laurent A, et al. Resection of nonresectable liver metastases
from colorectal cancer after percutaneous portal vein
embolization. Ann Surg 2000;231:480–6.
13. de Baere T, Roche A, Elias D, Lasser P, Lagrange C, Bousson
V. Preoperative portal vein embolization for extension of
hepatectomy indications. Hepatology 1996;24:1386–91.
14. Imamura H, Shimada R, Kubota M, Matsuyama Y,
Nakayama A, Miyagawa S, et al. Preoperative portal vein
embolization: an audit of 84 patients. Hepatology
1999;29:1099–105.
15. Kokudo N, Tada K, Seki M, Ohta H, Azekura K, Ueno M,
et al. Proliferative activity of intrahepatic colorectal metas-
tases after preoperative hemihepatic portal vein emboliza-
tion. Hepatology 2001;34:267–72.
16. Elias D, de Baere T, Roche A, Leclere J, Leclere J, Lasser P.
During liver regeneration following right portal emboliza-
tion the growth rate of liver metastases is more rapid than
that of the liver parenchyma. Br J Surg 1999;86:784–8.
17. Azoulay D, Castaing D, Krissat J, Smail A, Hargreaves GM,
Lemoine A, et al. Percutaneous portal vein embolization
increases the feasibility and safety of major liver resection
for hepatocellular carcinoma in injured liver. Ann Surg
2000;232:665–72.
18. Tanaka H, Hirohashi K, Kubo S, Shuto T, Higaki I,
Kinoshita H. Preoperative portal vein embolization
improves prognosis after right hepatectomy for hepatocel-
lular carcinoma in patients with impaired hepatic function.
Br J Surg 2000;87:879–82.
19. Adam R, Avisar E, Ariche A, Giachetti S, Azoulay D,
Castaing D, et al. Five-year survival following hepatic
resection after neoadjuvant therapy for nonresectable color-
ectal [liver] metastases. Ann Surg Oncol 2001;8:347–53.
20. Bismuth H, Adam R, Levi F, Farabos C, Waechter F,
Castaing D, et al. Resection of nonresectable liver metas-
tases from colorectal cancer after neoadjuvant chemother-
apy. Ann Surg 1996;224:509–20.
21. Elias D, Lasser P, Rougier P, Ducreux M, Bognel C, Roche
A. Frequency, technical aspects, results, and indications of
major hepatectomy after prolonged intra-arterial hepatic
chemotherapy for initially unresectable hepatic tumors. J
Am Coll Surg 1995;180:213–9.
22. Caldwell SH, de Lange EE, Gaffey MJ, Sue M, Boyd JC,
Dickson RC, et al. Accuracy and significance of pretrans-
plant liver volume measured by magnetic resonance
imaging. Liver Transpl Surg 1996;2:438–42.
The British Journal of Radiology, June 2006
The British Journal of Radiology, 79 (2006), 479–482

Radiation benefit and risk at the assessment stage of the UK

Breast Screening Programme
1
J LAW, PhD and 2K FAULKNER, PhD

1
Edinburgh University Department of Medical Physics, Chancellor’s Building, Little France Crescent,
Edinburgh EH16 4SB and 2Quality Assurance Reference Centre, 9 Kingfisher Way, Silverlink Business
Park, Wallsend NE28 9ND, UK

ABSTRACT. The balance between benefit and radiation risk in a breast cancer screening
programme has received much attention at the initial screening stage. This paper
extends that attention to first stage assessment, i.e. the first stage at which women are
recalled for further investigation because of suspected lesions or other suspect film
features, and prior to any biopsy. Numbers of films, including magnification films,
taken at this stage, have been established in two UK regions by different methods.
Average total mean glandular dose has been calculated using published data of dose
per film from initial screening and a multiplying factor to allow for magnification film Received 7 April 2005
Revised 20 July 2005
dose. It is concluded that the benefit/risk ratio is considerably higher at first stage Accepted 31 August 2005
assessment than at initial screening by a factor of between 4 and 9, because of the very
much higher cancer detection rate in this well-defined sub-group of women. DOI: 10.1259/bjr/33577478
Qualitatively, this conclusion is unaffected by the quite wide variation between
’ 2006 The British Institute of
screening centres in the numbers of films taken at first stage assessment.
Radiology

Considerable attention has been given to benefit and
risk in the initial screening stage of a breast cancer
screening programme. This attention has sometimes
centred on estimations of the numbers of cancers
detected and induced as an indicator of the benefit/risk
balance. Numbers of cancers detected and induced are
not in themselves a direct measure of benefit and risk.
Not all cancers detected by screening will be cured
(though earlier detection usually gives a better prog-
nosis), and not all of those induced will be fatal. A fuller
discussion of these and related points has been given by
us elsewhere [1]. Subsequent assessment stages, after the
initial screening visit, have received very much less
attention of this kind. This may be because it is easy to
see that qualitatively the benefit/risk-balance will be
better at assessment than at initial screening, but this
point appears not to have been examined in quantitative
detail.
In the UK Breast Screening Programme, screening
rounds occur at 3-yearly intervals. Within each round,
after the initial screening visit and examination, a
proportion of the screened women are recalled for
assessment because of suspicious or doubtful features
on their films. This proportion is typically between 5%
and 10% of those screened [2]. Women recalled in this
way may have further mammograms taken in ‘‘contact’’
or conventional geometry and some will also have
magnification films. Films of either of these kinds may
be full field, i.e. cover the whole breast, or be collimated
to the region of interest. This is sometimes referred to as
first stage assessment, and that term will be used in this
paper. Most of the women requiring such assessment are
eventually given reassurance that all is well, while a
minority (say 10–20%) of these are recalled for further
assessment (second stage) which involves biopsy or
needle samples [3].
This paper will concentrate principally on the benefit/
risk balance at the first of these two stages, but some
aspects of that balance at the second stage will also be
discussed.
The ratio of cancers detected to estimates of cancers
induced at the radiation dose levels involved is not the
same as the benefit/risk ratio, but the two ratios have
been shown to be related, with the benefit/risk ratio
being probably in the broad region of half to two thirds
of the detection/induction ratio [4, 5]. It is the detection/
induction ratio which will be mainly considered in this
paper, because the numbers required can be estimated
with relative ease and reliability.
Numbers of films taken at first stage
assessment
In order to estimate the numbers of cancers detected
and induced at first stage assessment it is necessary to
know the average numbers of films of each kind that are
taken (i.e. conventional, magnification, coned or full
field), the mean dose per film, and the cancer detection
rate among women undergoing this assessment stage.
Data for the second and third of these three can be
obtained from published sources and will be discussed
later. This section will describe attempts to find out
numbers of films taken.
Initially, information was obtained from nine screen-
ing centres in the North East, Yorkshire and the Humber
regions of England by means of a questionnaire sent to
superintendent radiographers. The superintendents

The British Journal of Radiology, June 2006 479


Table 1. Average numbers of films taken per woman
attending at first stage assessment – England (North East,
Yorkshire and The Humber) (9 screening centres)
Plain contact films Magnification films
Full field Coned
0.8–1.9 1.1–2.1 0.7–1.1
These are ranges of values, derived from a questionnaire sent
to nine screening centres.
were asked to provide estimates of the numbers of films
of each kind taken per woman attending in their own
centre. In doing this, they were asked to use their own
judgement regarding current normal practice rather than
keeping detailed records for a prospective period, and to
provide estimates of both the minimum numbers of films
normally taken on all or most women and a maximum
number. It was emphasised that this maximum should
represent the upper end of the normal range and not the
maximum ever known or the worst possible case, since
the latter would tend to distort the general picture.
The results of this enquiry are given in Table 1. All the
centres concerned, except one, now take two views
(normally craniocaudal (CC) and lateral oblique (LO)) at
the initial visit on each screening round, and this will
tend to reduce the numbers of films required at
assessment. In the absence of sufficient detailed informa-
tion, it is assumed for calculation purposes that all films
taken in conventional contact geometry are full-field,
though this is almost certainly not the case. All the films
will have been taken on one breast only, unless a
suspicious feature was seen in each breast; in the latter
case, the consequent greater dose is balanced by the
corresponding increased probability of a cancer being
identified.
In a separate exercise, data on film numbers were
obtained from four Scottish Screening Centres, who
extracted numbers from their records retrospectively.
These were therefore actual numbers for a defined
period chosen essentially at random, e.g. the previous
month, and did not depend on memory or judgement.
The results obtained in this way are shown in Table 2.
This region has not yet adopted two-view screening
except for the first screening round. Consequently at first
stage assessment films are taken of the CC view on each
breast unless these have already been taken at the basic
screening visit. Occasionally the mediolateral (ML) view
is substituted for the CC view. When two-view screening
on all rounds is adopted, all these films will be taken at
Table 2. Average numbers of films taken per woman
attending at first stage assessment – Scotland (four
Screening Centres)
Centre Plain contact films: Magnification films
(all coned)
Full field Coned
A 0.34 1.01 0.36
B 2.93 0 0.30
C 1.10 0 1.09
D 0.73 0.26 0.23
Mean* 1.22 0.29 0.70
*Weighted mean, related to size of centre.
480
J Law and K Faulkner
the basic screening visit, and therefore their numbers are
excluded from the following account. All other films are
categorised as plain films with full field exposed, plain
films coned down, and magnification films, all of which
are coned down in this region. Average numbers of films
of each kind in each of four centres are given in Table 2.
It is clear that practice varies between these four
centres in some respects. For example, Centre C does
three times as many magnification films as Centres A
and B; this is a general policy in that centre, not related to
the preference of any one radiologist. In some centres,
films taken with a small compression paddle are
routinely coned down to about the area of the paddle;
in other centres the beam may not be coned in this way,
to allow the area of interest to be viewed in relation to the
architecture of the breast as a whole. Centre B follows
this latter policy so that all their plain films are assumed
to be effectively full field for dose estimation purposes.
In this region, film numbers are broadly similar to
those in Region A, but with many fewer magnification
films. Average numbers of films from both these regions
are combined in Table 3.
Results
Mean glandular breast dose at first stage
assessment
Mean breast dose in the UK Breast Screening
Programme has been estimated at a little under
2.5 mGy [6] per woman for the LO view and 2.0 mGy
for the CC view, giving a mean dose of 2.25 mGy per
view to the assessed breast. A more recent estimate [7]
suggests very slightly lower values, but the earlier higher
values are retained in this paper. Future estimates of
mean breast dose may fluctuate around these values. The
ratio of magnification film dose to contact geometry film
dose has very recently been estimated to be 2.2¡0.15,
giving magnification film doses of 5.5 mGy (LO) and
4.4 mGy (CC) [8]. Since magnification films are equally
likely to be taken on either view, the mean magnification
film dose is 5.0 mGy.
Coned films are assumed, for this calculation, to cover
approximately one third of the full breast area, on
average [9]. In practice, this proportion varies with sizes
of individual breasts and with exact sizes of defining
cones on different X-ray units, but one third is
considered to be a reasonable average value.
In general, for a suspicious feature or features in one
breast only, only one breast is imaged at assessment. All
Table 3. Average numbers of films taken per woman
attending at first stage assessment: results combined from
two UK regions given equal weight, based on two-view
initial screening
Region Plain contact films Magnification films
Full field Coned Full field Coned
A 0.8–1.9 0 1.1–2.1 0.7–1.1
B 1.22 0.29 0 0.70
Mean* 1.6 0.15 1.0 0.9
*Taking upper end of ranges for Region A.
The British Journal of Radiology, June 2006
Radiation benefit and risk
Table 4. Mean glandular dose at first stage assessment (i.e.
number of films 6 dose per film). For further calculation
these numbers are divided by 2 because only one breast is
irradiated (i.e. effective mean glandular dose to a single
breast 5.1 mGy)
Film type No. of Mean glandular dose (mGy)
films
Per film Per woman
Full contact 1.6 2.25 3.6
Coned contact 0.15 0.75 0.11
Full magnification 1.0 5.0 5.0
Coned magnification 0.9 1.7 1.5
Total (single breast) 10.2
estimates of risk factor for breast cancer induction are
based on assumptions of equal radiation exposure to
both breasts. Therefore, if the usual breast cancer
induction risk factors [1, 8] are taken, the effective breast
dose (and hence the effective risk) is half the actual dose
to the one breast (Table 4).
The mean glandular breast dose is calculated in this
way in Table 4, using film numbers from Table 3.
Because ranges of film numbers were given in Table 1,
the upper ends of those ranges have been used in this
calculation.
Cancer detection rate at first stage assessment
Typical cancer detection rates are around 6 per 1000
women who attend for basic screening, with an age-
dependant range mostly within 5–7 per 1000 [2]. The
proportion of women recalled for assessment is usually
between 5% and 10% [2], and all cancers detected are
detected within this sub-group. Thus, for a 10% assess-
ment recall rate, the detection rate at assessment is 10
times what it is at initial screening, and for a 5% recall
rate it will be 20 times that figure.
Discussion and conclusions
Benefit and risk are difficult to define, but one simple
possibility is to regard lives saved from death by breast
cancer as benefit and lives lost to radiation induced
breast cancer as risk. Cancers detected and induced are
not the same as these, but there are reasonable grounds
Figure 1. Flow chart of assessment process.
The British Journal of Radiology, June 2006
for thinking that the benefit/risk ratio is around half to
two-thirds of the detection/induction ratio [4, 5].
It has already been shown that compared with initial
screening, the cancer detection rate has increased 10
times in the sub-group recalled for assessment, if the
recall rate is 10%, or 20 times if the recall rate is only 5%.
The effective mean glandular dose per woman at first
stage assessment has been shown in Table 4 to be up to
about 5.1 mGy as against 4.5 mGy at initial screening, an
increase of 15%. If we consider only the 5.1 mGy
received at assessment, the cancer detection/induction
ratio increases by a factor of 10/1.1558.7 compared with
its value at initial screening; if the doses at each stage are
combined to give 9.6 mGy, this factor becomes 10/
2.1554.7. In both cases these factors are doubled if the
assessment recall rate is 5% rather than 10%.
Table 5 shows ratios of cancers detected/induced at
various ages for the UK Breast Screening Programme.
Column two shows values previously published [1],
based on the most recent survey of mean glandular
breast doses, and breast cancer induction risk factors
from NRPB [1, 10]. These values refer to two-view
screening. Column three shows values for first stage
assessment, assuming a 10% recall rate; if the recall rate
were 5%, these values would be twice as large. Values in
column three are greater than those in column two by the
factor of 8.7 as outlined in the previous paragraph.
Values in column four result from combining the mean
glandular dose received at the initial screening visit with
that received at first stage assessment, and are obtained
by multiplying the values in column two by the factor 4.7
from the previous paragraph. Again, the resulting
detection/induction ratios would be doubled if the
assessment recall rate were 5% rather than 10%.
Uncertainties in the numbers quoted in column two of
Table 5 have been discussed in the paper from which
these figures have been taken. They will be at least a
factor of two in each direction due to the uncertainty in
the underlying values for cancer induction per unit dose.
Uncertainties in the numbers in columns three and four
will be further increased because of the variability of
protocol and practice between screening centres at the
assessment stage. These considerations do not, however,
affect the main conclusion that the benefit/risk balance is
much more favourable at first-stage assessment than at
initial screening.
Table 5. Cancers detected/induced at first stage assessment:
two-view screening. Assumes 10% of women are assessed
(these numbers would be doubled if the assessment rate
were only 5%)
Age range (years) Basic screening* 1st stage assessment
(a) (b)
50–54 89 780 420
55–59 120 1100 560
60–64 182 1600 850
65–70 300 2600 1400
*From Law and Faulkner [1].
(a) Taking effective mean glandular dose received at 1st
stage assessment only (5.1 mGy).
(b) Taking dose received at initial screening visit (4.5 mGy)
plus dose received at 1st stage assessment (5.1 mGy), i.e.
9.6 mGy.
481

The numerical risk factors for numbers of cancers
induced per unit dose, which underlie all these ratios,
may change at some time in the future. Any such change
will not affect the relationship between ratios given in
column two and those of columns three and four in
Table 5, and thus will not affect our main qualitative
conclusion.
Thus, whatever the relation between these figures and
benefit/risk ratios, assuming that this relation is similar
at all screening stages, the benefit/risk ratio is greater at
first stage assessment than at initial screening, despite
the greater dose level, because of the greatly increased
cancer detection rate. At later stages of assessment, when
biopsy or needle-aspiration and stereoscopic X-ray pairs
may occur, the dose will be greater still, but again the
cancer detection rate will rise, typically to the region of
30–50% as against 6 per 1000 at initial screening, or 6–
12% at first stage assessment, so that again the benefit/
risk ratio will increase compared with that at earlier
screening stages.
At the initial screening stage, the female population is
assumed to be symptom-free and therefore ‘‘healthy’’.
For this reason they are not called patients. Thus, at first
stage assessment, the status of an individual may change
from being one of a selected group of the general
population to a patient, since there are now specific
grounds to suspect that disease may be present, which
must be confirmed or excluded. For patients, justification
of an examination disregards prior exposure, for what-
ever reason. It is therefore right that at this stage
particular regard should be paid to ensuring that the
balance of benefit and risk is favourable. At second stage
assessment there is a relatively high probability that
disease is present. Many other patients will routinely
receive X-ray examinations to confirm or exclude the
presence of a disease, and often where the probability of
disease being present is less than it is at this stage of
breast screening. Justification of X-ray exposure at
second stage assessment should thus certainly be made
as it is for other patients. At first stage assessment, this
point is perhaps less clear-cut, but the recalls have been
made on suspicion of disease being present at a
probability in the region of 5–10%. Many hospital X-ray
examinations will be routinely performed at a lower
probability of disease or fracture than this, and the
women concerned cannot be regarded as ‘‘healthy’’ in
the way they are at initial screening.
The collective dose to the population of screened
women is affected by the recall to assessment rate.
Women who attend for first stage assessment receive an
additional mean glandular dose of 10.2 mGy per woman
(Table 4). If this were to be averaged across the screened
population, then it would lead to an increase in the mean
glandular dose to a typical woman of 1 mGy to 5.5 mGy
assuming a recall rate of 10%. If the recall rate was 5%,
then the mean glandular dose to a typical woman would
be 5.0 mGy. However, this is not the approach we have
482
J Law and K Faulkner
chosen to adopt, as justification of an examination
disregards prior exposure.
Regardless of these considerations, once it appears that
breast cancer may be present it is essential that all
necessary X-ray exposures are made to confirm or reject
that possibility. The dose levels reported in this paper for
first stage assessment are such that justification of the
exposure on the balance of benefit and risk should be
straightforward. Thus, there should be no problem in
terms of benefit/risk balance in taking whatever films
are necessary to this end. As in any X-ray investigation
that does not imply that films should be taken without
good cause. Similar arguments should also apply at
second stage assessment. The worst possible outcome of
a screening visit is a false-negative or a false-positive
report.
Acknowledgments
We wish to thank the Superintendent Radiographers
in Breast Screening Centres throughout the North East,
Yorkshire and the Humber Region and Scotland for
answering our questions and supplying information on
screening procedures and numbers of films taken at
assessment. This research was partially supported by the
European Commission’s Radiation Protection Research
Programme (DIMOND III and SENTINEL Projects).
References
1. Law J, Faulkner K. Cancers detected and induced, and
associated risk and benefit, in a breast screening pro-
gramme. Br J Radiol 2001;74:1121–7.
2. National Health Service Breast Screening Programme
Reviews, 1994-5 to 2003-4. Statistical Bulletin, London:
Department of Health.
3. NHS BSP Annual Review 2004 Changing Lives. Sheffield
NHS BSP, 2004.
4. Young KC, Faulkner K, Wall BF, Muirhead C. Review of
Radiation risk in Breast Screening. NHS BSP Report 54;
Sheffield, NHS BSP, 2003.
5. Law J, Faulkner K. Concerning the relationship between
benefit and radiation risk, and cancers detected and
induced, in a breast screening programme. Br J Radiol
2002;75:678–84.
6. Young KC. Radiation doses in the UK trial of breast
screening in women aged 40-48 years. Br J Radiol
2002;75:362–70.
7. Young KC, Burch A, Oduko JM. Radiation doses received in
the UK Breast Screening Programme in 2001 and 2002. Br J
Radiol 2005;78:207–18.
8. Law J. Breast dose from magnification films in mammo-
graphy. Br J Radiol 2005;78:816–20.
9. Faulkner K, Bennison K. An assessment of digital stereo-
taxis in the NHSBSP. Radiat Prot Dosim 2005;117:327–9.
10. Stokell PJ, Robb JD. SPIDER-1 Software for evaluating the
detriment associated with radiation exposure. NRPB-SR261.
Chilton: NRPB, 1994.
The British Journal of Radiology, June 2006
The British Journal of Radiology, 79 (2006), 483–486

Radiation risks for the radiologist performing transjugular

intrahepatic portosystemic shunt (TIPS)
N HIDAJAT, MD, P WUST, MD, M KREUSCHNER, MD, R FELIX, MD and R-J SCHRÖDER, MD

Department of Radiology, Charité Campus Virchow-Klinikum, University Medicine Berlin,

Augustenburger Platz 1, 13353 Berlin, Germany

ABSTRACT. The aim of this study is to evaluate the radiation dose to the interventional
radiologist in transjugular intrahepatic portosystemic shunt (TIPS) concerning the risk of
cancer and deterministic radiation effects and the relation to recommended dose limits. In
18 TIPS interventions radiation doses were measured with thermoluminescence
dosemeters (TLD) fixed at the eyebrow, thyroid and hand of the radiologist without
special lead shielding of these body parts and at the chest, abdomen and testes under the
lead apron. The doses of the eye lens, thyroid gland and hand were assumed to be equal
to the corresponding surface doses. The dose at the abdomen under the lead apron was
used as an estimation of the ovarian dose. Effective dose equivalent was estimated by
Webster’s method. The estimated effective dose equivalent was 0.087 mSv and the
Received 14 March 2005
effective dose 0.110 mSv. The risk of fatal cancer was of 1026 and the risk of severe genetic Revised 7 September 2005
defect of 1027 for one single intervention. The maximum permissible number of TIPS Accepted 19 September
interventions was 181, otherwise the dose limit for effective dose would be exceeded. 2005
When the radiologist performed more than 372 TIPS procedures per year for many years,
DOI: 10.1259/bjr/67632946
the dose to the lens of the eye could exceed the threshold for cataract. If the
interventionist performs a large number of TIPS procedures in a year, the risk of fatal ’ 2006 The British Institute of
cancer and developing cataracts becomes relatively high. Radiology

A major source of radiation doses to medical personnel
are fluoroscopic procedures [1]. Interventional angio-
graphic treatments are typically associated with higher
radiation exposure than diagnostic angiography. The
transjugular intrahepatic portosystemic shunt (TIPS) has
been shown as effective in the treatment of complications
of portal hypertension such as variceal bleeding and
refractory ascites [2, 3]. Furthermore, it was shown that
TIPS is more effective in preventing variceal re-bleeding
than the competitive endoscopic method [2].
Nevertheless, TIPS has been known as an intervention
with the highest radiation exposure to patients under-
going abdominal angiographic procedures [4, 5].
In association with the high radiation exposure to the
patient undergoing TIPS, it can be assumed that the
interventional radiologist is exposed to high levels of
scattered radiation. The International Commission on
Radiological Protection [6] and the Food and Drug
Administration (FDA) [7] has estimated the radiation
risk of cancer and deterministic effects, such as cataract
of the eye lens, and recommended dose limits.
The aim of this study is to evaluate the radiation dose
to the interventional radiologist in TIPS concerning the
risk of cancer, deterministic radiation effects and the
relation to the recommended dose limits.
Methods
Radiation doses were measured in 18 TIPS procedures
(18 patients, age 54.8¡10.0 years; 15 male and 3
female) with liver cirrhosis and refractory ascites as
complications of the portal hypertension. The interven-
tion was performed with angiographic equipment
(Multidiagnost; Philips, Eindhoven, The Netherlands)
with under table tube and with pulsed fluoroscopy mode
with 12 images per second.
Thermoluminescence dosemeters (TLD; Harshaw,
Cleveland, OH) with lithium fluoride crystals
(3.2 mm63.2 mm60.9 mm) doped with Mg (LiF:Mg)
were used for dosimetry. The minimum detectable level
was 0.01 mSv. As the total time between regeneration,
exposure and next regeneration of the detectors was very
short (3 days maximum), no correction for natural
radiation (0.05 mSv month21) had to be made. The
fading rate of 5% in 12 months was also negligible.
Taking into account the individual responsivity of the
detectors to the energy of X-ray radiation and the filter
that was used (2.6 mm Al), the energy correction factor
was found to range from 1.00 for a tube voltage of 60 kV
to 1.05 at 100 kV. Reproducibility was found to be 1–2%
in doses more than 10 times greater than the detection
limit.
To measure the radiation exposure to the radiologist
(NH), one TLD was attached on both sides of the
eyebrows and the thyroid gland over the lead apron, on
the back of the hand at the base of the middle finger, on
the chest at the breastbone, the abdomen at the navel and
at the testes under the lead apron. The radiologist used a
0.35 mm lead equivalent lead apron. A lead glass with
1.0 mm lead equivalent was placed permanently in front
of the patient chest between the image intensifier and the
radiologist. Thyroid collar and lead glasses were not
worn.

The British Journal of Radiology, June 2006 483


To establish a portosystemic shunt, a puncture needle
was advanced transjugularly in a catheter through the
inferior vena cava into the right hepatic vein. An
intrahepatic branch of the portal vein was punctured
and self-expandable bare nitinol stents (Angiomed,
Karlsruhe, Germany) with a diameter of 10 mm and a
length of 4 cm or 6 cm were implanted. The end of the
stent was located within the hepatic vein. The stents and
the hepatic vein were dilated to 10 mm after deploy-
ment. Pressure measurements were performed for the
portal vein and right atrium in order to ensure that the
portosystemic pressure gradient was reduced to 10–
15 mm Hg. When TIPS placement was completed, the
guide wire and angiographic sheath were removed.
Doses of the eye lens and thyroid gland were assumed
to be equal to the corresponding measured surface doses.
The dose at the abdomen under the apron was used as an
estimation of the dose to the ovaries. The effective dose
equivalent was estimated by Webster’s method [8].
Webster developed an empirical method for estimating
effective dose equivalent according to the ICRP 26 [9].
This method is based on organ dose measurements and
effective dose calculations of Faulkner and Harrison [10]
and uses the 1977 ICRP tissue weighting factors [9] to the
under apron dose (Hu) and the over collar dose (Ho) with
the equation effective dose equivalent51.56Hu+0.046Ho.
Results
Fluoroscopy time was 77.8¡66.3 min, the dose–area
product (DAP) 446.0¡279.7 Gy cm2 (mean¡standard
deviation). The surface doses are shown in Table 1. One
surface dose at the thorax and under the apron was at the
minimum detectable level. All other surface doses were
above this level.
The doses at the eyebrows and the thyroid gland were
higher on the right side than on the left side. The left
hand was exposed more strongly than the right hand
and showed the highest dose among all body parts
monitored by TLD.
Using Webster’s method [8], the effective dose
equivalent [9] was calculated by using the values of the
over collar and under apron dosemeters. For radiologists
who do not wear a thyroid collar, the effective dose
equivalent calculated with Webster’s method is 21%
below the effective dose [6] calculated with organ dose
Table 1. Radiation (surface) doses of the interventional
radiologist as mean¡standard deviation. (In brackets dose-
meter position)
Organ/body part Right (mSv) Left (mSv)
Eye lens (eyebrow) 0.403¡0.328 0.229¡0.120
Thyroid gland (over 0.589¡0.721 0.506¡0.225
apron at collar)
Chest (under 0.041¡0.025
apron at breast bone)
Abdomen (under 0.043¡0.032
apron at waist)
Testes (under 0.057¡0.041
apron at symphysis)
Hand (back 0.549¡0.313 0.918¡0.516
of the hand)
484
N Hidajat, P Wust, R Felix et al
Table 2. Comparison of the doses of the interventional
radiologist with the recommended dose limits and the
maximum permissible numer of TIPS interventions per year
Organ/body Dose (mSv) Recommended dose Maximum
part limit [6] (mSv) permissible
number/year (n)
Eye lensa 0.403 150 372
Testes 0.057 50 877
Handb 0.918 500 544
Whole body 0.110c 20 181
a
Right side.
b
Left side.
c
Effective dose derived from effective dose equivalent (EDE).
EDE estimated according to Webster [8].
tables and depth dose charts [11]. The effective dose
would be 0.110 mSv and the maximum permissible
number of TIPS would be 181 per year (Table 2).
To meet the recommended dose limit [6], the max-
imum permissible number of TIPS interventions was
estimated to be 181, otherwise the dose limit for the
effective dose would be exceeded (Table 2). A pregnant
radiologist could perform a maximum of 23 interven-
tions after the declaration of pregnancy during the
remainder of the pregnancy (Table 3). The risk of fatal
cancer was approximately 1026 and the risk of severe
genetic defect 1027 for a single intervention (Table 4). If a
radiologist performed more than 248 TIPS per year for
many years, the radiologist could reach the threshold
for radiation-induced detectable opacity of his eye lens.
For a workload of more than 372 times a year and for
many years, he could possibly develop a cataract
(Table 5).
Discussion
In angiography, the patient is exposed by direct
radiation and the radiologist by scatter radiation.
However, the radiologist may work with and be exposed
to radiation every day for many years. In a previous,
study with a very limited number of patients undergoing
transarterial hepatic chemoembolisation, we showed that
the dose to the eye lens is a quantity that may limit the
permissible number of interventions before a dose limit
would be exceeded [12]. This study shows that it may be
deduced from one to two TIPS interventions per day, the
dose limit for the eye lens can be exceeded. It also reveals
that the effective dose can be a limiting quantity for the
number of the interventions, as one TIPS per day can
cause the cumulative effective dose to be above the limit
Table 3. Comparison of the dose to the abdomen of the
interventional radiologist with the recommended dose limit
for the conceptus of a pregnant woman and the maximum
permissible number of TIPS interventions after the declara-
tion of pregnancy during the remainder of the pregnancy
Dose (mSv) Recommended Maximum
dose limit (mSv) permissible number
(n)
Conceptus 0.043a 1 23
a
Dose to abdomen under apron
The British Journal of Radiology, June 2006
Radiation risk for radiologist in TIPS

Table 4. Risk of some stochastic damage for the radiologist, calculated according to ICRP 60 [6]
Stochastic Risk coefficient Dose (mSv)* Risk*
damage

Fatal cancer 5%/Sv 0.110¡0.080a 5.5061026¡4.0061026

Genetic 1%/Sv 0.057¡0.041b 5.7061027¡4.1061027
Defect 1%/Sv 0.043¡0.032c 4.3061027¡3.2061027
a
Effective dose derived from effective dose equivalent (EDE). EDE determined according to Webster [8].
b
Dose to testes.
c
Dose at abdomen used as dose to ovaries.
*Dose and risk presented as mean¡standard deviation.

after 1 year. In the case of pregnancy, one TIPS per day
can cause the dose limit recommended by the ICRP for
the unborn child be exceeded after 24 days.
The risks of fatal cancer derived from effective dose and
of severe genetic defect are extremely small for the
radiologist after one single TIPS. However, for high
workloads, e.g. 500 per year and over 10 years, these
stochastic risks can increase to a magnitude of 1023 to 1022.
The radiologist should be aware of the radiation
exposure to the eye lens and the associated possibility
of cataract when he performs interventions like TIPS one
to two times per day over many years. Sterility will not
occur as a deterministic radiation effect to personnel
after TIPS.
The dose differences between the right and left side of
the eye lens, thyroid gland and hands can be explained
by the position of the radiologist in relation to the
X-ray tube. While standing at the head of the patient
during the TIPS intervention, the tube is mostly located
at the liver side, i.e. right side of the patient. Thus, the
distance of the right lens and side, of the thyroid gland to
the tube is shorter than on the left side. The left hand was
used to guide and rotate the catheter and placed at the
neck, while the right hand was placed at the proximal
end of the catheter to inject contrast media into the
catheter. Thus, the left hand was closer to the tube than
the right hand. Therefore, in the interests of the
radiologist, dose measurements in interventions should
be performed at both sides when it is not clear which
side is irradiated more strongly.
Not only can there be dose differences between the
sides, but the dose distribution at the same hand may not
be uniform. Whitby and Martin [13] have found that for
most interventional radiological procedures the bases of
Table 5. Comparison between the measured doses of the
radiologist and threshold for radiation-induced deterministic
damage [6, 7]
Deterministic damage Threshold Mean dose (mSv)
(mSv year21)
Eye lensa 0.403
Detectable opacity .100 (n.248)
Visual impairment .150 (n.372)
(cataract)
Testicles 0.057
Temporary sterility 400 (n.7017)
Permanent sterility 2000 (n.35087)
Ovaries 0.043b
Sterility .200 (n.4651)
a and the radiologist to reduce the amount of scatter
Right side.
b
Dose at abdomen used as dose to ovaries.
the ring and little finger receive the highest dose. In this
study we measured the dose at the base of the middle
finger, so that the maximum dose to the hand could be
higher than we have found.
Organs or body parts located near the surface, such as
the eye lens, may be measured on the surface. For the
estimation of the ovarian dose we used the surface dose
to the abdomen. In this case, it can be argued that the
lead apron shields the individual from radiation,
preferentially at lower energies. The parts of the energy
spectrum which penetrates the lead apron are the higher
energies and these should not be absorbed significantly
by the abdomen on their way to the ovaries. Thus, the
surface dose under the apron should be a reasonable
estimation of the ovarian dose. We did not measure the
dose to the leg. A study by Whitby and Martin [14]
showed that in TIPS, the leg dose could be as much as
two to three times greater than that to the hands. In
contrast, Meier et al [5] have found that the dose to the
hand was the highest in TIPS (about 0.7 mSv at the hand
and 0.2 mSv at the leg).
Regarding radiation dose to the radiologist, there are
two particular aspects of TIPS. First, it is associated with
a long fluoroscopy time [4]. This is because the
catheterization of the hepatic vein, puncture and
catheterization of the portal vein must be guided by
fluoroscopy and is often difficult. Normally, TIPS is
performed by an experienced interventionist. Second,
many patients needing a TIPS have large amount of
ascites, which automatically leads to higher voltage and
tube current. In big departments with many different
radiological procedures, the competences of the physi-
cians are divided because no one can be really
specialized in all of these procedures. Those who are
competent for TIPS and perform about 100 of these
interventions per year accumulate more radiation dose
and may have less reserve for different interventions
than others who, for example, perform similar numbers
of angioplasties of the lower extremity.
To protect personnel from stochastic and deterministic
risks, measures of radiation protection should be used
when interventions such as TIPS are performed fre-
quently. The distance to the X-ray source should be kept
as great as possible as the dose is reduced with the
square of the distance. Lead apron, thyroid collar and
lead glasses may greatly reduce the operator radiation
exposure in cardiac interventions (to 0.8%) [15]. Thyroid
collars could reduce the effective dose by approximately
a factor of two [11]. Permanent lead shielding, e.g. a lead
plate as we have used, can be placed between the patient
radiation reaching the trunk and head of the radiologist.

The British Journal of Radiology, June 2006 485


Lead gloves have been found to be flexible and could
lead to dose reduction of about 20% [16]. For a number of
years we have used the pulsed fluoroscopy mode with 12
images per second and cannot see significant reduction
of the image quality during the intervention. Zweers et al
[17] described significant reduction of the estimated staff
effective dose in TIPS using dedicated fluoroscopy
exposure settings. Last but not least, a recently described
MR guided TIPS with use of a hybrid radiography/MR
system should facilitate the puncture of the portal vein
and be associated with less radiation exposure than a
conventional TIPS procedure [18].
References
1. National Council on Radiation Protection and Measure-
ments. Implementation of the principle of as low as reason-
ably achievable (ALARA) for medical and dental personnel.
NCRP Report no. 107. Bethesda, MD: NRCP, 1990.
2. Rössle M, Deibert MP, Haag K, et al. Randomised trial of
transjugular-intrahepatic-portosystemic shunt versus endo-
scopy plus propranolol for prevention of variceal rebleed-
ing. Lancet 1997;349:1043–9.
3. Lind CD, Malisch TW, Chong WK, Richards WO, Pinson
CW, Meranze SG, et al. Incidence of shunt occlusion or
stenosis following transjugular intrahepatic portosystemic
shunt placement. Gastroenterology 1994;106:1277–83.
4. McParland BJ. A study of patient radiation doses in
interventional radiological procedures. Br J Radiol
1998;71:175–85.
5. Meier N, Lenzen H, Sudhoff A, Fiebich M, Kötter L. X-ray
exposures of radiologists and staff in interventional radi-
ology. Radiologe 1995;35:152–5.
6. ICRP. 1990 Recommendations of the International Commis-
sion on Radiological Protection. ICRP Publication 60, Ann
ICRP, 1991.
7. Food and Drug Administration. Avoidance of serious x-ray
induced skin injuries to patients during fluoroscopically-
guided procedures. Food and Drug Administration 1994;
September 9: 1–6.
486
N Hidajat, P Wust, R Felix et al
8. Webster EW. EDE for exposure with protective aprons.
Health Phys 1989;56:568–9.
9. ICRP. Recommendations of the International Commission
on Radiological Protection. ICRP Publication 26, Ann ICRP,
1977.
10. Faulkner K, Harrison RM. Estimation of effective dose
equivalent to staff in diagnostic radiology. Phys Med Biol
1988;33:83–91.
11. Niklason LT, Marx V, Chan HP. Interventional radiologists:
occupational radiation dose and risks. Radiology 1993;187:
729–33.
12. Hidajat N, Vogl T, Biamino G, Wust P, Panzer W, Zankl M,
et al. Radiation exposure in interventional radiology as
exemplified by the chemoembolization of hepatocellular
carcinoma and laser angioplasty of the pelvic arteries.
Fortschr Röntgenstr 1996;164:249–56.
13. Whitby M, Martin CJ. A study of the distribution of dose
across the hands of interventional radiologists and cardio-
logists. Br J Radiol 2005;78:219–29.
14. Whitby M, Martin CJ. Radiation doses to the legs of
radiologists performing interventional procedures: are they
a cause for concern? Br J Radiol 2003;76:321–7.
15. Kuon E, Schmitt M, Dahm JB. Significant reduction of
radiation exposure to operator and staff during
cardiac interventions by analysis of radiation leakage
and improved lead shielding. Am J Cardiol 2002;
89:44–9.
16. Damilakis J, Koukourakis M, Hatjidakis A, Karabekios S,
Gourtsoyiannis N. Radiation exposure to the hands of
operators during angiographic procedures. Eur J Radiol
1995;21:72–5.
17. Zweers D, Geleijns J, Aarts NJ, Hardam LJ, Lameris JS,
Schultz FW, et al. Patient and staff radiation dose in
fluoroscopy-guided TIPS procedures and dose reduction,
using dedicated fluoroscopy exposure settings. Br J Radiol
1998;71:1333–4.
18. Kee ST, Ganguly A, Daniel BL, Wen Z, Butts K,
Shimikawa A, et al. MR-guided transjugular intrahepatic
portosystemic shunt creation with use of a hybrid
radiography/MR system. J Vasc Interv Radiol
2005;16:227–34.
The British Journal of Radiology, June 2006
The British Journal of Radiology, 79 (2006), 487–496

Organ doses from prostate radiotherapy and associated

concomitant exposures
1
R M HARRISON, PhD, 1M WILKINSON, DCR(T), 1A SHEMILT, BSc, 1D J RAWLINGS, MPhil, 1M MOORE and
2
A R LECOMBER, PhD

1
Regional Medical Physics Department, Newcastle General Hospital, Newcastle upon Tyne NE4 6BE
and 2Regional Medical Physics Department, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4PP,
UK

ABSTRACT. In addition to the therapeutic exposure, a course of radiotherapy will

involve the additional (concomitant) irradiation of the patient using CT, simulator or
portal imaging systems, for localization of the target volume and subsequent
verification of treatment delivery. The number of concomitant exposures is likely to
increase as the developing technical capabilities for conformal, image-guided
radiotherapy make target and critical organ definition an increasingly important aspect
of radiotherapy. Estimation of doses and risks to critical organs in the body from all
sources is thus necessary to provide the basis for adequate justification of the exposures
as required by ICRP. In this paper, doses to selected organs and tissues for which ICRP
have identified fatal cancer probabilities have been measured using a realistic
anthropomorphic phantom loaded with thermoluminescent dosemeters and irradiated
using a treatment protocol for radical radiotherapy of the prostate. Independently,
doses to the same organs and tissues have been measured from concomitant CT and
portal imaging exposures given for localization and verification purposes. Although
Received 8 July 2005
negligible in comparison with the target dose, realistic numbers of concomitant Revised 25 August 2005
exposures give a small but significant contribution to the total dose to most organs and Accepted 23 September
tissues outside the target volume. Generally, this is in the range 5–10% of the total 2005
organ dose, but can be as high as 20% for bone surfaces. These data may be used to
DOI: 10.1259/bjr/16187818
estimate concomitant doses from any combination of CT and portal imaging and may
help in the justification process, especially when additional verification exposures may ’ 2006 The British Institute of
be required during treatment. Radiology

The entire process of radiotherapy involves exposure
of the patient to various sources of ionizing radiation. In
addition to the therapeutic exposure itself, localization
and verification (concomitant) exposures may be carried
out before and during the treatment using CT scanning,
conventional simulation and megavoltage portal ima-
ging. Estimation of doses to critical organs in the body
from all sources is necessary in order to assess total
stochastic (second cancer) risks and thus provide
adequate justification of the exposures as required by
ICRP [1] and the corresponding UK legislation (The
Ionising Radiation (Medical Exposures) Regulations,
2000 (IR(ME)R). [2]). This is increasingly important
because of the concomitant exposures associated with
the development of complex conformal and image-
guided radiotherapy.
Numerous workers have reported estimates of second
cancer incidence following radiotherapy [3]. This is likely
to remain an important issue as, following improve-
ments in cancer treatment, more patients may survive
longer than the latent period for expression of the second
cancer. In addition, several new developments in radio-
therapy may lead to higher doses to organs and tissues
throughout the body. Intensity-modulated radiotherapy
(IMRT) may lead to substantial increases in whole body
doses, because of the longer beam-on times and the
consequent increased exposure to leakage radiation. The
use of an increased number of fields results in a larger
volume of tissue exposed to relatively low doses,
although the target volume may be smaller. Compared
with conventional radiotherapy, Hall and Wuu have
estimated that IMRT may almost double the incidence of
second cancers [4]. New imaging techniques are also
contributing to verification of treatment field positions
during the course of radiotherapy. Although megavol-
tage portal imaging is a long-established technique,
image-guided radiotherapy (IGRT), involving CT studies
at many fractions [5], cone beam CT using megavoltage
and kilovoltage X-ray energies [6, 7] and integrated
systems such as tomotherapy [8], are also emerging as
viable techniques for improving the accuracy of dose
delivery to the target volume and may increase organ
doses and subsequent risks. In contrast, the potential for
reduction of second cancers by using proton radio-
therapy has been investigated by Miralbell et al [9].
Previous estimates of second cancer risk have neces-
sarily made simplifying assumptions about doses and
risks. Followill et al [10] estimated a single whole body
dose at a point distant from the treatment field and
applied a risk factor for the general population. Mutic

The British Journal of Radiology, June 2006 487


and Low [11] measured doses outside the target volume
using thermoluminescent dosimetry (TLD) and a block
phantom. Verellen et al [12] used personal dosemeters to
measure the personal dose equivalent Hp(10) for six
patient treatments, related this to the effective dose and
used a nominal risk factor for the general population.
More recently, Kry et al [13] have used NCRP risk factors
[14] to compare several IMRT treatments of the prostate.
The purpose of these papers was to compare different
irradiation techniques (e.g. conventional radiotherapy
versus IMRT), thus placing the emphasis on relative,
rather than absolute estimates of risk.
In other fields, diverse exposures and irradiation
patterns are usually combined using the quantity
effective dose. However, there are several problems
associated with using this concept in the radiotherapy
context [15]. ICRP risk and tissue weighting factors refer
to the general population, whereas the age distribution of
radiotherapy patients will be skewed towards the higher
ages, implying lower risk estimates. Conversely, tissue
weighting factors refer to low doses associated with
occupational exposure and a dose and dose rate
effectiveness factor (DDREF) of 2 has been applied. It is
not clear what modifications to the DDREF should be
made for the effects of radiotherapy fractionation
schedules and cell kill on induced cancer probability at
high doses.
Thus, the estimation of second cancer induction
probability from a combination of weighted organ doses
is not trivial and no attempt has been made in this paper
to combine individual organ doses, which include a
radiotherapy component, to form an overall measure of
risk. Nevertheless, as a first step, organ doses measured
in several critical organs in a realistic anthropomorphic
phantom may be useful for justification purposes and
will help to place concomitant doses in perspective,
particularly for developing and justifying techniques
where more extensive imaging is used to improve the
accuracy of dose delivery.
This paper describes the experimental simulation of
radical prostate radiotherapy together with associated
CT scanning and portal imaging, with doses measured
using lithium fluoride TLD chips loaded in a male
anthropomorphic phantom. In anticipation of eventually
calculating second cancer induction probabilities, doses
were measured in organs (apart from the skin) to which
ICRP has ascribed tissue weighting factors [1]. Skin was
omitted because of the uncertainties arising from the
dose gradient at the surface due to build up effects at
megavoltage energies, the identification of the effective
point of measurement of a TLD chip under these
conditions and the large number of surface measure-
ments which would in any case be necessary in order to
form an unambiguous estimate of total skin dose.
Furthermore, the skin has the lowest fatal cancer
probability coefficient of all the main organs identified
by ICRP and is not as critical as other tissues when
considering second cancers following radiotherapy.
External beam radiotherapy of the prostate was chosen
for study because it is the most common cancer in UK
men and advances in treatment have shown an increase
in 10 year survival from 20% in 1971 to 50% in 2001 [16].
Brenner et al [17], in a comparison of second cancers in
patients treated for prostate carcinoma by radiotherapy
488
R M Harrison, M Wilkinson, A Shemilt et al
compared with surgery, have shown that there is a small
but significant increase in solid tumours from 5 years
following radiotherapy, principally showing as carcino-
mas of the bladder, rectum and lung, and sarcomas in or
near the treatment field. A further study has also shown
a significant increase in second rectal cancers [18].
Measurement of organ doses from CT scanning and
portal imaging allows doses to be calculated for any
number and combination of these imaging techniques,
thus facilitating the estimation of the organ dose
contribution from image-intensive IGRT.
Materials and methods
Radiotherapy
The position of the prostate and surrounding critical
organs such as the bladder and rectum were established
within a male RANDO phantom (The Phantom
Laboratory Incorporated, Salem, NY) and a treatment
plan for isocentric external beam prostate radiotherapy
developed according to protocols in place in this centre,
using a Helax-TMS treatment planning system
(Nucletron B.V., Veenendaal, The Netherlands) and CT
localization scans of the phantom obtained with a
Siemens Emotion Duo CT scanner (Siemens Medical
Solutions, Erlangen, Germany).
The simulated prostate treatment consists of two
phases. In phase 1, 64 Gy is delivered in 32 fractions to
the planning target volume (PTV), followed in phase 2 by
10 Gy in 5 fractions to a reduced target volume. The
prostate, bladder and rectum were identified on slabs 32,
33 and 34 of the phantom, following consideration of a
series of patient CT scans, and a three-field plan for both
phases developed as shown in Figure 1. Field sizes and
gantry angles for both treatment phases are given in
Table 1. Anteroposterior (AP) and lateral digitally
reconstructed radiographs were also generated
(Figure 2).
The phantom was loaded with Harshaw TLD-100
(LiF:Mg:Ti) chips (Thermo Electron Corporation, Solon,
OH), positioned within each organ of interest at the
centre of each phantom slab and secured with tissue
equivalent plugs. Each chip had a unique identifier and
calibration factor. Anatomy texts were used to check the
position of organs and tissues, and sampling points were
selected to give several measurements for each critical
organ. Table 2 shows the numbers of TLD chips
contributing to the dose measurement for each organ
or tissue.
Treatment was delivered using a Siemens PRIMUS H1
linear accelerator (Siemens Medical Solutions, Erlangen,
Germany) operating at 15 MV. For the simulation of this
treatment, the phantom was irradiated to 0.1 of the total
target dose (i.e. 7.4 Gy) and the subsequent measured
photon doses scaled to give the doses which would have
been received for a full clinical treatment.
Neutron dose contribution
At 15 MV, photoneutron production in the linear
accelerator head will contribute to the organ and tissue
The British Journal of Radiology, June 2006
Organ doses from prostate radiotherapy
(a)
Figure 1. Plan for prostate treatment simulation on slab 34 of a RANDO phantom (a) phase 1 treatment (64 Gy) and (b) phase 2
treatment to a smaller target volume (10 Gy).
doses. Measurement of individual organ doses due to
neutrons was not attempted, but instead calculated from
previously reported measurements of neutron fluence.
Expressions for direct, thermal and scattered neutrons
from medical linear accelerators, derived from Monte
Carlo calculations by McCall [19, 20] were used to
calculate the neutron fluence at the centre point of each
slice of the RANDO phantom. They are reproduced in
Equation (1) where the three terms give direct, scattered
and thermal neutron components, respectively. The total
neutron fluence per photon dose at the isocentre (W), is
given by:

105 mAs; slice width 5 mm; pitch 1.5; 33 slices). Ten
W~ aQ=4pd 2 zð5:4aQ=S Þzð1:26Q=S Þ ð1Þ
where a is the transmission factor for neutrons in head
shielding (taken as 0.85); Q is the neutron source strength
per unit photon dose at the isocentre for a Siemens
PRIMUS accelerator (taken as 0.2161012 neutrons Gy21)
[21]; d is the distance from the source; and S is the
treatment room surface area (180 m2).
Conversion of W to equivalent dose per Gy at the
isocentre used the fluence-to-dose conversion factors
given by NCRP [14], assuming a mean neutron energy of
0.5 MeV for an accelerator running at 15 MV [22, 23] and
taking values of radiation weighting factor wR of 20 and 5
for fast and thermal neutrons, respectively. Although the
ionization suffered by lithium fluoride TLD will include
contributions from (n,p) and (n,c) reactions, this has been
ignored in practice because the wR values ascribed to the
Table 1. Treatment parameters for the simulation of prostate radiotherapy
Total target dose (Gy) AP field size (cm6cm)
Phase 1 64 6.866.7
Phase 2 10 4.964.7
The British Journal of Radiology, June 2006
(b)
neutron absorbed dose are significantly higher than
unity.
CT scanning for localization
Using the same TLD loading scheme as for the
radiotherapy treatment, the phantom was scanned using
the standard departmental protocol for prostate localiza-
tion, using a Siemens Emotion Duo scanner (Siemens
Medical Solutions, Erlangen, Germany). The irradiation
consisted of an AP topogram (130 kV; 30 mAs; topogram
length 512 mm) followed by an axial scan (130 kV;
identical consecutive scans (topogram and axial) of the
phantom were carried out in order to ensure that the
doses at the TLD measurement positions were well
above the minimum detectable value. Measured doses
were subsequently scaled to give the doses received for a
single localization scan. (Doses from conventional
simulation were not measured, since a simulator is not
routinely used in this centre for prostate localization).
Portal imaging
Using the same TLD loading scheme as for the
radiotherapy treatment, the phantom was irradiated by
a 15 cm615 cm AP portal image exposure, as used in
conjunction with the Siemens portalvision electronic
portal imaging system (Siemens Medical Solutions,
Left lateral field size (cm6cm) Right lateral field size
(cm6cm)
6.766.6 6.766.6
5.564.6 5.564.7
489
 R M Harrison, M Wilkinson, A Shemilt et al

TLD calibration and dose measurement

Following irradiation, chips were left for a fixed time
of 48 h to allow the same fading time for all irradiations
and read out in a Harshaw 5500 TLD reader (Thermo
Electron Corporation, Solon, OH).
Individual chip calibration factors were established
based on irradiation of each chip to a known dose using
a Cs-137 calibration source. These factors provide a
correction for interchip variability within a batch. In
addition, batches of chips were irradiated within the
range of doses and energies used in the simulations in
order to generate a calibration curve. Doses were
measured with ionization chambers and electrometers
whose calibrations were traceable to national standards.

Figure 2. Digitally reconstructed AP radiograph of the Dose calculation

RANDO phantom with target and critical volumes outlined.
The prostate is shown in yellow with a red margin for the Raw data from each chip readout were modified by
clinical target volume (CTV). The bladder is shown in orange the individual chip calibration factor and the dose
and the rectum in dark red. The planning target volume calibration factor for the appropriate photon energy,
(PTV) is demarcated by yellow lines. giving the result as absorbed dose to water at the chip
position. Organ doses were then calculated as weighted
averages over all the TLD measurement points, using the
mass fraction of each organ within each slab [24]. For
Erlangen, Germany) mounted on the same linear accel- lung, bone surface and bone marrow, the mass fractions
erator as used for treatment, but running at 6 MV. given by Huda [25] were used. Doses to bone marrow
Typically, a clinical portal image would be acquired and bone surface were calculated from the average of
using 3 monitor units (MU). For simulation, 200 MU measurements to the positions of these organs within
were used in order to ensure a measurable dose to each slab and also, for comparison, from the average
distant organs and the subsequent doses scaled to give dose over the whole slab, in both cases using the same
the commensurate clinical exposure for a single portal mass fractions. The total doses to these organs were
image. A 10 cm610 cm lateral portal image exposure obtained by summing the dose contributions for each
was also simulated in the same way for comparison. slab. All organs, except for bone surfaces, were assumed
to be water-equivalent. Doses to bone were derived from
doses to water by application of mass stopping power
ratios (bone/water) of 0.91 and 0.93 for 6 MV and 15 MV,
respectively, and for CT energies a mass energy
Table 2. Thermoluminescent dosemeter (TLD) loadings for absorption coefficient ratio (bone/water) of 3.1 was
organs and tissues used. Chips which had received doses .2000 mGy were
not re-used.
Organ or tissue Number of RANDO slabs used
TLD chips for each organ or Calculated neutron doses for each organ were added
tissue to the photon contributions for the simulation of
treatment at 15 MV.
Gonads (testes) 4 35
Bone marrow and 101 5–6, 8–34
The current protocol in use in this centre defines the
bone surface maximum number of concomitant images which can be
Colon (and upper 60 24–31 acquired under a single authorization by the Practitioner,
large intestine) as defined under UK legislation [2]. It allows for the
Lung 59 10–20 following: (i) three localization episodes on the CT
Stomach 23 20–24 scanner; (ii) three pre-treatment verification episodes, at
Bladder 8 33–34 least one on the CT scanner and one on the treatment
Liver 36 19–25 unit; and (iii) eight verification episodes during the
Oesophagus 11 9–19
course of treatment. A localization or verification episode
Thyroid 12 9–11
Adrenals 2 21–22
on a CT scanner consists of a maximum of three
Brain 7 1, 3, 5 topograms, two CT scans up to 120 cm in length and
Upper large intestine 29 24–27 three single axial slices. A pre-treatment verification
Small intestine 20 24–30 episode on the treatment unit consists of a maximum of
Kidney 11 21–25 four portal images. A verification episode during
Pancreas 7 22–23 treatment delivery consists of a pre-treatment verifica-
Spleen 6 20–22 tion episode and a localization or verification episode on
Thymus 2 14 the CT scanner. Several combinations of CT scans and
Rectum 3 32–34
portal images are possible within this framework and
Total 401
organ doses for two possibilities were calculated: (i) 10

490 The British Journal of Radiology, June 2006

Organ doses from prostate radiotherapy
CT scans and 36 portal images, where portal imaging is
the primary method of verification; and (ii) 26 CT scans
and 4 portal images, where CT scanning is the primary
method of verification.
Results
Chip calibration
The uncertainty in a single TLD measurement was
found to be ¡4.3%, averaged over calibrations at all
photon energies. The minimum detectable dose was
1 mSv, taken to be the dose corresponding to three times
the mean background signal.
Neutron results
Figure 3 shows the neutron dose (in mSv Gy21 at the
isocentre) as a function of distance from the isocentre
normal to the central axis, i.e. along the long axis of the
phantom. This enabled a representative neutron dose to
be calculated for each slab. The decrease of dose with
distance is simply a consequence of an inverse square
term acting on the direct neutron component and
involving distance from the X-ray target.
Organ and tissue doses
Table 3 gives organ and tissue doses for the complete
radiotherapy treatment, single CT localization scan and
topogram, and single AP portal image. Organ doses
resulting from a lateral 10 cm610 cm portal exposure
were less than those arising from the 15 cm615 cm AP
exposure and the latter have been adopted as represen-
tative of prostate portal images. Doses are given for all
organs for which tissue weighting factors have been
identified [1], apart from the skin, and include remainder
organs. Apart from muscle, bone marrow and bone
surface, ranges of doses are also given for organs
Figure 3. Neutron dose (direct, scattered and thermal) as a
function of distance from the isocentre normal to the central
axis of the beam. These values have been calculated from
equation 1 using a50.85, Q50.2161012 neutrons Gy21 and
S 5 (1806104) cm2.
The British Journal of Radiology, June 2006
containing more than 5 TLD measurement points, or
where mean organ doses are .0.1 mSv. Although the
dose to the rectum is included within the estimate for the
colon, the rectal dose has been calculated separately,
because of its proximity to the prostate, and is included
in the footnote to Table 3. In fact, part of the rectum and
bladder were included in the target volume and received
measured doses in excess of 60 Gy and 73 Gy, respec-
tively. The neutron component of organ doses ranges
from 19–50 mSv for the whole treatment.
Since different approaches are taken when considering
high doses compared with the low doses usually
associated with radiation protection, some assumptions
need to be made when combining or comparing them.
Since most organ absorbed doses are ,500 mGy,
equivalent dose is used throughout to allow photon
and neutron components to be summed and effective
dose to be calculated for CT and portal exposures. For
consistency, it is assumed that wR51 for the high photon
doses to organs close to or within the target volume,
although it is accepted that the dose response function
for carcinogenesis may differ at low and high doses.
Table 4 gives the total organ and tissue doses for a
complete course of radiotherapy involving a total of 10
CT scans and 36 portal images. The data are also shown
graphically as total doses (Figure 4a) and as percentages
of the total doses from radiotherapy, CT scanning and
portal imaging for each organ (Figure 4b). Table 5 gives
organ doses for 26 CT scans and 4 portal images.
Figure 5 shows total doses and doses as percentages of
the total organ dose, respectively. A subset of these data
have previously been reported in preliminary form [26].
Approximate doses to organs in other positions may
be estimated from Figure 6, which shows a plot of mean
dose for each slab.
Within and around the target volume in slab 34, doses
to the chip positions were estimated from the treatment
planning system dose calculation algorithm and
summed for both phases of treatment. It was assumed
that the planned dose distribution was delivered exactly.
Figure 7 shows the difference between the delivered and
measured doses at 11 points as a fraction of the delivered
dose.
For reasons given above, effective dose is an inap-
propriate quantity for use in radiotherapy treatment.
However, the concept may be applied, with caution, to
diagnostic X-ray examinations such as CT scanning and,
in view of the relatively low doses involved, to portal
imaging. Applying the tissue weighting factors and
remainder organ rules stipulated by ICRP [1], the
measured organ doses for a CT localization scan
(including topogram), single AP portal image (3 MV
15 cm615 cm at 6 MV) and single lateral portal image
(3 MV 10 cm610 cm at 6 MV) gave effective doses of
4.2 mSv, 2.3 mSv and 0.8 mSv, respectively.
Discussion
Uncertainties in dose estimation
Organ doses have been measured under ideal condi-
tions. There was no movement of the phantom during
treatment, unlike the interfraction and intrafraction
491
 R M Harrison, M Wilkinson, A Shemilt et al

Table 3. Organ and tissue doses for complete radiotherapy treatment, single CT localization scan and single anteroposterior
(AP) portal image
Organ or tissue Equivalent dose (mSv)2

Radiotherapy treatment Radiotherapy Total radiotherapy Single CT localization Single 15 cm615 cm

(74 Gy at 15 MV) treatment (dose range scan and AP AP portal image
(74 Gy at in parenthesis) topogram (6 MV)
15 MV) (dose range in
parenthesis)
Photon contribution Neutron
contribution

Bladder 28750 48 28800 (4000–73200) 11.6 (10–12) 27.8 (24–61)

Colon1 2600 45 26501 3.11 2.81
Lung 62 40 102 (30–120) ,0.1 ,0.1
Stomach 95 43 138 (50–170) 0.2 (0.08–0.5) 0.11 (0.07–0.19)
Liver 78 42 120 (40–190) 0.1 (0.04–0.7) 0.08 (0.05–0.18)
Oesophagus 59 40 99 (30–90) ,0.1 ,0.1
Thyroid 45 38 83 (30–90) ,0.1 ,0.1
Bone marrow3 1350 42 13903 3.33 4.503
Bone surface3 719 27 7463 5.93 2.363
Adrenals 89 42 132 0.2 0.10
Brain 29 34 63 (28–32) ,0.1 0.20 (0.19–0.22)
Small intestine1 242 45 2871 2.21 0.971
Kidney 90 43 133 (60–160) 0.3 (0.2–0.7) 0.14 (0.11–0.28)
Muscle 100 18 118 1.5 0.32
Pancreas 103 43 146 (60–160) 0.2 (0.2–0.3) 0.12 (0.11–0.16)
Spleen 73 42 115 (50–140) 0.1 ,0.1
Thymus 31 20 51 (30–90) ,0.1 ,0.1
1
Dose to the colon is a weighted average over the lower and upper large intestine as recommended by ICRP [1]. However, the
rectum receives a large dose from prostate radiotherapy because of its proximity to the target volume. Thus the rectal dose
has been estimated separately as 24 000 mSv. The ranges of dose for LLI, ULI and SI were (120–1040 mSv), (70–480 mSv) and
(110–480 mSv), respectively, for radiotherapy, (1–15 mSv), (0.3–1.8 mSv) and (0.4–13 mSv), respectively, for CT and (0.25–
33 mSv), (0.15–0.43 mSv) and (0.17–19 mSv), respectively, for a single AP portal image.
2
For radiotherapy measurements, doses are quoted to 3 significant figures or 50 mSv for high doses. For the CT results, doses are
quoted to 0.1 mSv, since the measured doses were for 10 CT scans. For the portal image doses, results are quoted to 0.01 mSv,
since the delivered doses were 67 times the doses for a typical portal image.
3
Dose ranges for these tissues are not given, since they vary over the whole range of doses delivered, depending on the location
within the body and the proximity to the target volume.

Table 4. Organ and tissue doses for complete radiotherapy treatment, 10 CT localization scans and 36 portal images
Organ or tissue Total radiotherapy Dose from 10 CT scans % of total dose Dose from 36 % of total dose due Total dose1
dose1 (mSv) and topograms1 (mSv) due to CT AP portal to portal images (mSv)
images1 (mSv)

Bladder 28800 116 0.4 999 3.3 29900

Colon 2650 31.2 1.1 102 3.7 2780
Lung 102 0.3 0.3 1.1 1.1 103
Stomach 138 1.7 1.2 3.8 2.6 144
Liver 120 1.4 1.1 3.0 2.4 124
Oesophagus 99 0.2 0.2 3.1 3.0 102
Thyroid 83 ,0.1 0.05 3.3 3.8 86
Bone marrow 1390 33 2.1 91 6.0 1520
Bone surface 746 59 6.6 84.8 9.5 890
Adrenals 132 1.7 1.2 3.8 2.7 138
Brain 63 ,0.1 0.0 7.3 10.4 70.3
Small intestine 287 21.6 6.3 35 10.2 344
Kidney 133 3.1 2.2 5.0 3.5 141
Muscle 118 14.6 10.1 11.6 8.1 144
Pancreas 146 2.3 1.5 4.4 2.9 153
Spleen 115 1.4 1.2 3.1 2.6 120
Thymus 51 0.2 0.4 0.5 0.97 51.7
1
For radiotherapy measurements, doses are quoted to 3 significant figures or 50 mSv. For the CT results, doses are quoted to
0.1 mSv or three significant figures, since the measured doses were for 10 CT scans. For the portal image doses, results are
quoted to 0.1 mSv or three significant figures, since the delivered doses were 67 times the doses for a typical portal image.

492 The British Journal of Radiology, June 2006

Organ doses from prostate radiotherapy
Figure 4. Organ doses from radiotherapy (photons and
neutrons), 10 CT scans and 36 portal images for each organ
or tissue (a) total doses for each organ and (b) %
contribution of radiotherapy, CT scanning and portal
imaging. Mean bladder, colon and rectal doses were
29.9 Gy, 2.8 Gy and 25 Gy, respectively.
movement uncertainties and changes in shape which
might be seen in clinical treatments. In spite of making
401 individual TLD measurements, each critical organ
has been sampled fairly coarsely. Nevertheless, the
measurement accuracy (4.3% for a single TLD measure-
ment) is adequate, bearing in mind the subsequent use of
such data in risk estimation where considerable uncer-
tainties exist. In addition to the uncertainty per TLD
measurement, there are several other sources of uncer-
tainty in organ dose estimation. There will be genuine
dose variations within most organs because of organ
extent, proximity to the target volume and other dose
gradients associated with the edges of CT and portal
image fields. These are reflected in the ranges given in
parenthesis in Table 3. In addition, there are uncertain-
ties associated with the location of organs and the choice
of sampling points within them. These are difficult to
quantify. Finally, measurements on a single size and
shape of anthropomorphic phantom will not represent
the interpatient variation seen in practice.
The British Journal of Radiology, June 2006
Estimating doses to bone surfaces and red bone
marrow is particularly difficult because of the extended
and complex distribution of these tissues throughout the
body. Although dosemeters were placed close to selected
bone locations, the uncertainties in dose estimation for
marrow and surfaces are likely to be greater than for
more well-defined organs. As pointed out by Golikov
and Nikitin [24], the mass fractions for bone given by
Huda and Sandison [25] and used here, refer to the mass
distribution of the whole skeleton, which may differ
from the corresponding distribution of bone surface
cells. Doses to bone surfaces and marrow are also some
of the higher organ doses measured. This emphasises the
potential difficulty of using this technique as a reliable
basis for absolute risk assessment and suggests that the
technique may be more appropriate for the calculation of
relative risks, for example between two different treat-
ment modalities or between treatment and imaging
exposures, where dose measurement locations and
assumptions about organ position and extent are
invariant.
For regions in or near the target volume, TLD
measurement is likely to be considerably less accurate
than calculation using the well-developed algorithms
used in current treatment planning systems in combina-
tion with accurate ionization chamber measurements of
dose distributions in water. Hence, for organs and tissues
close to the target volume, use of the treatment planning
system is probably the best method for organ dose
calculation. However, planning systems are not designed
for calculation of doses to distant organs and do not, in
any case, include neutron dose estimates.
Calculated organ neutron doses are subject to several
assumptions inherent in the values given for the terms in
Equation (1). Furthermore, it has been assumed that the
estimated dose to a slab is appropriate for all organs
within the slab, whereas, in practice, this assumption will
overestimate the neutron dose to centrally placed organs.
Whilst negligible in comparison with doses to the target
and its immediate surroundings, the neutron contribu-
tion forms a substantial fraction of the total dose to
remote organs and tissues, although this would only
apply for treatments at energies in excess of approxi-
mately 10 MV.
Dose and risk estimates
The bladder and rectum, as expected, receive the
highest dose, since part of the bladder wall and rectum
will be contained within the target volume for this
treatment. Considerable variation in bladder and rectal
volumes, both within and between patients, have been
reported [27] and this will lead to corresponding
variations and uncertainties in bladder and rectal doses
– and risks – in practice. The observation of second
bladder and rectal cancers in patients who have
previously received radiotherapy for prostate cancer is
consistent with these high doses [17], since it implies that
parts of each organ may receive doses near the peak of
the dose–response curve.
In Figures 4 and 5, the numbers of concomitant
exposures have deliberately been set to correspond to
the upper limits of what is expected in this centre and it
493
 R M Harrison, M Wilkinson, A Shemilt et al

Table 5. Organ and tissue doses for complete radiotherapy treatment, 26 CT localization scans and 4 portal images
Organ or tissue Total radiotherapy Dose from 26 CT % of total Dose from 4 % of total dose Total dose1
dose1 (mSv) scans and dose due AP portal due to portal (mSv)
topograms1 (mSv) to CT images1 (mSv) images

Bladder 28800 302 1.0 111 0.36 29200

Colon 2650 81.1 3.0 11.4 0.41 2740
Lung 102 0.9 0.87 0.1 0.11 103
Stomach 138 4.4 3.0 0.4 0.29 143
Liver 120 3.6 2.9 0.3 0.27 124
Oesophagus 99 0.6 0.6 0.3 0.34 100
Thyroid 83 0.1 0.12 0.4 0.42 83
Bone marrow 1392 85.6 5.7 18.0 1.2 1500
Bone surface 746 153 17 9.4 1.0 908
Adrenals 132 4.4 3.2 0.4 0.3 137
Brain 63 ,0.1 0.01 0.8 1.2 64
Small intestine 287 56.1 16.1 3.9 1.1 347
Kidney 133 8.0 5.6 0.6 0.38 142
Muscle 118 38.2 24 1.3 0.58 158
Pancreas 146 6.0 3.9 0.5 0.32 153
Spleen 115 3.6 3.0 0.3 0.29 119
Thymus 51 0.4 0.78 0.1 0.11 51.5
1
For radiotherapy measurements, doses are quoted to 3 significant figures or 50 mSv. For the CT results, doses are quoted to
0.1 mSv or three significant figures, since the measured doses were for 10 CT scans. For the portal image doses, results are
quoted to 0.1 mSv or three significant figures, since the delivered doses were 67 times the doses for a typical portal image.

is seen that only for bone surfaces do the concomitant
exposures exceed 10% of the total organ dose for both
cases illustrated, with red bone marrow contributing
.10% for the case of treatments in which portal imaging
is the verification technique used.
To place concomitant doses in perspective, we may
consider the excess relative risk (ERR) associated with a
population receiving radiotherapy and associated con-
comitant exposures, compared with a hypothetical
population receiving radiotherapy only, assuming that
the probability of cancer induction per unit dose is
constant. The ERR is given approximately by:
ðDT,radio zDT,concom Þ{DT,radio DT,concom
ERRT & ~ ð2Þ
DT,radio DT,radio
where: DT,radio and DT,concom are the doses to organ T
from radiotherapy only and concomitant only exposures,
respectively.
Taking the two examples of concomitant exposures (10
CT scans + 36 portal images and 26 CT scans + 4 portal
images), in both cases, ERRs for most organs are ,0.1,
with bone surfaces, small intestine and muscle ,0.3. A
single CT scan or portal image will give an ERR ,0.01 for
all organs.
In comparison with work by Waddington and
McKenzie [28] the effective dose for a male AP pelvic
portal image (15 cm615 cm at 6 MV) is higher (0.78 mSv
MU21) compared with their 0.34 mSv MU21). However,
no account has been taken here of the fraction of bladder
wall volume which lies within the target volume and for
which the risk of radiation carcinogenesis may be
reduced because of cell kill. This would have the effect
of reducing the effective tissue weighting factor for the
bladder and hence the effective dose. Moreover, our
sampling of the relevant bladder tissue is coarse
compared with dose estimations using the dose plotting
facilities of a treatment planning system. Waddington
and McKenzie assumed that 25% of the bladder was
outside the treated volume and a similar assumption
would reduce our figure of effective dose to 0.42 mSv
MU21. They also assumed that any cancer cells induced
in the bladder tissue within the target would have a
negligible chance of survival. Whilst simple exponential
cell kill would support this assumption, there is
conflicting evidence from epidemiological studies, sum-
marized by Hall and Wuu [4]. We find an effective dose
of 0.27 mSv MU21 for the lateral portal field, compared
with 0.32 mSv MU21 [28] which possibly reflects the
smaller field size. In any case, estimates of effective dose
will vary with patient size and target volume, so that
these results should be considered to be representative
rather than definitive.
It has previously been argued [15] that for legislative
purposes under the UK IR(ME) Regulations, a single
justification for a pre-determined combination of radio-
therapy and concomitant exposures is practically more
appropriate than the justification of individual concomi-
tant exposures, provided that the contributions of the
latter are appreciated. These results support this sugges-
tion, since even for a relatively large number of
concomitant exposures, the contribution to the overall
organ dose, whilst not negligible, is nevertheless small.
However, it should be noted that for radiotherapy
treatments carrying a good prognosis, especially for
younger patients and children, the added risks of
concomitant exposures may need to be considered
carefully in the justification process. This is especially
true for treatments which involve imaging at each
fraction, either by CT or portal imaging, such as the
developing techniques associated with image-guided
radiotherapy.
The combination of radiotherapy and concomitant
doses to give an unambiguous estimate of risk of second
malignancy is not trivial and, because of the inappropri-
ateness of the current definition of effective dose, no
attempt has been made in this paper to do so.

494 The British Journal of Radiology, June 2006

Organ doses from prostate radiotherapy
Figure 5. Organ doses from radiotherapy (photons and
neutrons), 26 CT scans and 4 portal images for each organ or
tissue (a) total doses for each organ and (b) % contribution
of radiotherapy, CT scanning and portal imaging. Mean
bladder, colon and rectal doses were 29.2 Gy, 2.7 Gy and
24.3 Gy, respectively.
The British Journal of Radiology, June 2006
Nevertheless, the doses given here allow the total organ
and tissue doses from various combinations of radio-
therapy, localization and verification exposures to be
estimated for this particular treatment.
Conclusions
Doses to organs and tissues for which ICRP have
identified fatal cancer probabilities have been measured
using TLD for a radical radiotherapy treatment of the
prostate delivering a total of 74 Gy to the PTV.
Independently, doses to the same organs and tissues
have been measured from concomitant CT and portal
imaging for localization and verification. Combinations
of numbers of CT scans and portal images with the
radiotherapy treatment itself has allowed total organ
doses to be calculated.
Realistic numbers of concomitant exposures give a
small but significant contribution to the total dose to
most organs and tissues. Generally this is in the range 1–
10%, but can be as high a 20% for bone marrow and bone
surfaces. At 15 MV, a significant proportion of the doses
to organs distant from the target volume is due to
photoneutrons produced in the treatment head.
These data may be useful for giving a realistic
perspective on the contribution of concomitant doses to
the overall organ dose burden for this treatment, for
protection of the patient and compliance with the
requirements of UK Regulations [2]. Whilst doses from
similar abdominal radiotherapy treatments may be
inferred from these data, further work will be necessary
to establish organ and tissue doses for radiotherapy of
other parts of the body, both for adults and children.
Acknowledgments
We would like to thank members of staff of the
Regional Medical Physics Department and the Northern
Centre for Cancer Treatment who contributed to aspects
of this work.
Figure 6. Mean dose in each slice vs
distance from isocentre normal to
the central axis, as a percentage of
the target dose (74 Gy). Each slab of
the RANDO phantom is 2.5 cm thick.
Error bars represent¡1 SD of the
doses measured in each slab.
495

References
1. ICRP. 1990 Recommendations of the International
Commission on Radiological Protection. Pergammon
Press, 1991. Annals of the ICRP; 21 (1–3).
2. IR(ME)R. Ionising Radiation (Medical Exposure)
Regulations. HMSO, London, 2000. SI 2000 No. 1059.
3. NRPB. Risks of second cancers in therapeutically irradiated
populations. National Radiological Protection Board,
Chilton, Didcot, Oxon OX11 0RQ, UK, 2000. Documents
of the NRPB; vol. 11 no. 1.
4. Hall JD, Wuu C-S. Radiation-induced second cancers: the
impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys
2003;56:83–8.
5. Fung AYC, Grimm S-YL, Wong JR, Uematsu M. Computed
tomography localization of radiation treatment delivery
versus conventional localization with bony landmarks. J
App Clin Med Phys 2003;4:112–9.
6. Lewis DG, Swindell W, Morton EJ, Evans PM, Xiao ZR. A
megavoltage CT scanner for radiotherapy verification. Phys
Med Biol 1992;37:1985–99.
7. Jaffray DA, Siewerdsen JH, Wong JW, Martinez AA. Flat-
panel cone-beam computed tomography for image guided
radiation therapy. Int J Radiat Oncol Biol Phys 2002;53:
1337–49.
8. Mackie TR, Holmes T, Swerdloff S, et al. Tomotherapy: a
new concept for the delivery of dynamic conformal radio-
therapy. Med Phys 1993;20:1709–19.
9. Miralbell R, Lomax A, Cella L, Schneider U. Potential
reduction of the incidence of radiation-induced
second cancers by using proton beams in the treatment of
pediatric tumors. Int J Radiat Oncol Biol Phys 2002;54:
824–9.
10. Followill D, Geis P, Boyer A. Estimates of whole-body
dose equivalent produced by beam intensity modulated
conformal therapy. Int J Radiat Oncol Biol Phys 1997;38:
667–72.
11. Mutic S, Low DA. Whole-body dose from tomotherapy
delivery. Int J Radiat Oncol Biol Phys 1998;42:229–32.
12. Verellen D, Vanhavere F. Risk assessment of radiation-
induced malignancies based on whole-body equivalent
dose estimates for IMRT treatment in the head and neck
region. Radioth Oncol 1999;53:199–203.
13. Kry SF, Salehpour M, Followill DS, et al. The calculated risk
of fatal secondary malignancies from intensity-modulated
radiation therapy. Int J Radiat Oncol Biol Phys 2005;62:
1195–203.
14. NCRP. Limitation of exposure to ionizing radiation. NCRP
Report 116. Bethesda, MD 20814: National Council on
Radiation Protection and Measurements, 1993.
496
R M Harrison, M Wilkinson, A Shemilt et al
Figure 7. Delivered - measured
dose as a fraction of the delivered
dose for 11 TLD measurement points
within slab 34 of the RANDO phan-
tom. Error bars represent the uncer-
tainty in a single TLD measurement
(¡ 4.3%).
15. Harrison RM. Second cancers following radiotherapy: a
suggested common dosimetry framework for therapeutic
and concomitant exposures. Br J Radiol 2004;77:986–90.
16. UK CR. Cancer incidence, survival and mortality in the UK
and EU. In: Toms JR, editor. CancerStats Monograph 2004.
Cancer Research UK, 2004.
17. Brenner DJ, Curtis RE, Hall EJ, Ron E. Second malignancies
in prostate carcinoma patients after radiotherapy compared
with surgery. Cancer 2000;88:398–406.
18. Baxter NN, Tepper JE, Durham SB, Rothenberger DA,
Virnig BA. Increased risk of rectal cancer after prostate
radiation: a population-based study. Gastroenterology
2005;128:819–24.
19. McCall RC, McGinley PH, Huffman KE. Room scattered
neutrons. Med Phys 1999;26:206–7.
20. McCall RC, Jenkins TM, Shore RA. Transport of accelerator
produced neutrons in a concrete room. IEEE Trans Nuc Sci
1979;NS-26:1593–602.
21. Followill DS, Stovall MS, Kry SF, Ibbott GS. Neutron source
strength measurements for Varian, Siemens, Elekta and
General Electric linear accelerators. J App Clin Med Phys
2003;4:189–94.
22. Palta JR, Hogstrom KR, Tannanonta C. Neutron leakage
measurements from a medical linear accelerator. Med Phys
1984;11:498–501.
23. Ongara C, Leon JR, Perez J, Zanini A, Burn K. Monte Carlo
simulation and experimental evaluation of photoneutron
spectra produced in medical linear accelerators.
Proceedings of the 1999 Particle Accelerator Conference.
New York: IEEE, 1999.
24. Golikov VY, Nitikin VV. Estimation of the mean organ
doses and the effective dose equivalent from RANDO
phantom measurements. Health Physics 1989;56:111–5.
25. Huda W, Sandison GA. Estimation of mean organ doses in
diagnostic radiology from RANDO phantom measure-
ments. Health Physics 1984;47:463–7.
26. Harrison RM, Wilkinson M, Shemilt A, Rawlings DJ, Moore
M, Lecomber AR. Estimating second cancer risk following
radiotherapy: organ doses from prostate radiotherapy and
concomitant exposures. Biomedizinishe Technik
2005;50(Suppl. vol 1 Part 1):768–9.
27. Lebesque JV, Bruce AM, Kroes AP, Touw A, Shouman RT,
van Herk M. Variation in volumes, dose-volume histo-
grams, and estimated normal tissue complication probabi-
lities of rectum and bladder during conformal radiotherapy
of T3 prostate cancer. Int J Radiat Oncol Biol Phys
1995;33:1109–19.
28. Waddington SP, McKenzie AL. Assessment of effective
dose from concomitant exposures required in verification of
the target volume in radiotherapy. Br J Radiol 2004;77:
557–61.
The British Journal of Radiology, June 2006
The British Journal of Radiology, 79 (2006), 497–503

Intensity-modulated radiation therapy in the treatment of gastric

cancer: early clinical outcome and dosimetric comparison with
conventional techniques
1,2
M T MILANO, MD, PhD, 3M C GAROFALO, MD, 1S J CHMURA, MD, PhD, 1K FARREY, MS, 1C RASH, CMD,
4
R HEIMANN, MD, PhD and 1A B JANI, MD

1
Department of Radiation and Cellular Oncology, University of Chicago, Chicago, IL 60637,
2
Department of Radiation Oncology, University of Rochester, Rochester, NY 14642, 3Department of
Radiation Oncology, University of Maryland, Baltimore, MD 21201 and 4Division of Radiation
Oncology, University of Vermont, Burlington, VT 05401, USA

ABSTRACT. The purpose of this study was to assess the efficacy and toxicity of intensity-
modulated radiation therapy (IMRT) in the treatment of gastric cancer. Seven patients
with gastric cancer were treated with IMRT. Six patients (all Stage III) received post-
operative chemoradiotherapy with concurrent 5-fluorouracil and leucovorin. One
received planned pre-operative radiation, though did not proceed to surgery. All
patients were planned to receive 50.4 Gy in 1.8 Gy fractions. IMRT planning was
compared with opposed anterior-posterior: posterior-anterior (AP/PA) and 3-field
conventional three-dimensional plans. When compared with either AP/PA or 3-field
plans, IMRT significantly reduced the volume exceeding the threshold dose of the liver
and at least one kidney. Target coverage with IMRT was excellent, with 98¡1% of the
target receiving >100% of the dose. Compared with AP/PA and 3-field plans, IMRT
plans had a greater percentage of target receiving the prescribed dose, but also a
greater volume receiving .110% of the dose. IMRT was well tolerated; no patients Received 7 June 2005
developed acute gastrointestinal toxicity greater than grade 2. All seven experienced Revised 23 August 2005
grade 2 nausea, three had grade 2 diarrhoea and two had grade 2 oesophagitis. Accepted 7 October 2005
Weight loss ranged from 0–12% (mean 6.1% and median 5.8%). IMRT in the treatment
DOI: 10.1259/bjr/43441736
of gastric malignancies reduces the mean and above threshold doses to critical normal
tissues. In an initial cohort of seven patients, 50.4 Gy delivered by IMRT is well tolerated ’ 2006 The British Institute of
and safe. Radiology

Gastric carcinoma is an uncommon malignancy in
North America, yet it represents the eighth leading cause
of cancer death in the USA. In 2004, approximately 21 860
new cases of gastric cancer were estimated to occur in the
USA and approximately 11 550 deaths are expected as a
result [1]. Due to the lack of a cost-effective screening
tool, gastric cancer is often diagnosed at an advanced
stage. Surgery is the cornerstone of the treatment for
resectable advanced stage gastric cancer. Post-opera-
tively, the locoregional failure rate is approximately 40%
and approximately half of those failures represent the
only site of failure [2, 3]. In the setting of locally
advanced (T3–4 and/or node positive) non-metastatic
disease, adjuvant chemoradiotherapy has been estab-
lished as the standard of care in the USA based upon the
recently reported results of the Gastric Surgical Adjuvant
Trial Intergroup 0116 trial [4]. This trial demonstrated a
statistically significant benefit in relapse-free survival
(48% versus 31%, p,0.001) and overall survival
(50% versus 41%, p50.01) with the use of adjuvant
Address correspondence to: Ashesh B Jani, University of Chicago,
Department of Radiation and Cellular Oncology, MC 9006, Chicago,
IL 60637, USA.
chemoradiotherapy when compared with surgery alone.
Adjuvant therapy also reduced the percentage of failures
attributable to local failure (29% vs 19%). Since this
landmark trial was reported, the increased use of
radiation therapy in the adjuvant treatment of gastric
cancer prompted the publication of a consensus report
reviewing the details related to radiotherapy delivery
technique [5]. According to this consensus report,
‘‘parallel-opposed AP/PA fields are considered the most
practical arrangement for the overwhelming majority of
post-operative adjuvant radiotherapy cases’’.
The toxicity associated with adjuvant chemoradiation
using traditional techniques is significant. Treatment
volumes in the post-operative setting are necessarily large
to account for the patterns of failure established in
previous surgical studies [2, 3]. Typical target volumes
include the stomach bed (to include surgical clips), a
portion of the left hemi-diaphragm, and draining lympha-
tics at risk. The standard target dose of 45 Gy well exceeds
the tolerance of several surrounding critical normal tissues
(most notably the kidneys and liver). As a result,
conventional treatment volumes are often tailored out of
the fear of potential kidney and liver damage. By under-
dosing portions of the target, local control and survival

The British Journal of Radiology, June 2006 497


may be compromised. Through inverse planning, inten-
sity-modulated radiation therapy (IMRT) allows for more
conformal dose delivery and selective sparing of critical
structures such as the kidneys and liver, and may therefore
allow for more complete target coverage to full-dose.
Locoregional control may be improved through better
target coverage and treatment toxicity may be reduced
through the use of IMRT. We previously published our
experience of using IMRT in the treatment of pancreatic
and anal malignancies [6, 7]. Since 2001, our institution has
routinely treated gastric cancer patients with post-opera-
tive chemotherapy combined with IMRT. Here we report
the early clinical outcome of our single institution
experience of seven patients treated for gastric cancer
with IMRT, with emphasis on toxicity outcome, and the
results of a dosimetric comparison with traditional radio-
therapy delivery techniques.
Materials and methods
Between March 2001 and April 2004, seven consecu-
tive non-metastatic patients with adenocarcinoma of the
stomach were treated with IMRT. Six patients received
post-operative chemoradiotherapy (CRT) with concur-
rent 5-fluorouracil and leucovorin, using chemotherapy
doses and scheduling as described in the recently
reported Intergroup trial [4]. Two of these six patients,
treated in early 2001, were not offered additional
chemotherapy following radiation based on physician
preference. One 87-year-old patient with symptomatic
bleeding received planned pre-operative radiation (with-
out concurrent chemotherapy because of comorbidities),
since upfront surgery would have been delayed until a
pacemaker revision. However, she never proceeded to
surgery because of the comorbidities. All patients were
planned to receive 50.4 Gy in 1.8 Gy fractions, except for
the patient treated pre-operatively, who was prescribed
45 Gy to an initial planned target volume (PTV) with a
5.4 Gy boost to the gross tumour. For the purpose of
dosimetric comparisons (described below) this patient
was planned similarly to the others (i.e. 50.4 Gy with no
boost PTV). The higher dose of 50.4 Gy was chosen in an
attempt to improve locoregional control, and because the
IMRT dose–volume histograms (DVHs) suggested that it
could be safely delivered.
All six patients who underwent resection had a D2
resection, including subtotal (n54) or total
gastrectomy (n52), resection of perigastric and second
echelon lymph nodes as described by the Japanese
classification [8, 9].
Table 1. Typical input parameters for normal structures for IMRT planning
Structure Patient 1
% of tissue volume Maximum dose to
to receive .dose (Gy) structure (Gy)
Right kidney ,35% .18 35
Left kidney ,40% .20 48
Liver ,80% .25 45
Spinal cord ,65% .18 35
IMRT, intensity-modulated radiation therapy.
498
M T Milano, M C Garofalo, S J Chmura et al
Simulation and target contouring
Patients underwent CT based simulation in the supine
position (PQ5000 CT Simulator; Marconi Medical
Systems, Cleveland, OH) with 4 mm CT slices. A custom
immobilization device (Alpha Cradle; Smithers Medical
Product, Inc., North Canton, OH) was used to minimize
set-up variability. The pre-operative CT scan was image
correlated to the CT simulation scan using the AcQSim
VoxelQ software package. The PTV and normal struc-
tures (kidneys, liver and spinal cord) were manually
contoured onto the CT scan slices following the ICRU 50
recommendations [10]. The clinical target volume (CTV)
was contoured on axial CT scan slices. The CTV typically
included the original tumour volume, operative bed (as
defined by the operative note, pathologic findings,
surgical clips and discussion with the surgeon) and
the draining lymphatics at risk. The gross tumour
volume (GTV) was entered on the single patient treated
prior to planned surgery. The radiation dose was
prescribed to a PTV, which was generated by expanding
the CTV by 1 cm. The PTV design incorporated set-up
uncertainty and organ motion [11]. Normal structures
were also entered, including the kidneys, liver and spinal
cord.
IMRT planning
IMRT plans were generated using commercial inverse
planning software (CORVUS, versions 3.0-5.0; NOMOS
Corp., Sewickley, PA), which produces optimal inten-
sity-modulated profiles using a simulated annealing
algorithm. Dynamic multileaf collimators were used to
shape the fields. Eight to nine-field coplanar plans were
used. Typical angles were 40 ˚ increments starting from 0 ˚
to 320 ˚ (US convention). However, posterior fields that
overdosed one kidney were usually removed (hence
making an 8-field plan) or angled away from the
kidney.
The PTV and normal structure dose–volume con-
straints were iteratively adjusted to ensure optimal target
coverage while minimizing dose to the kidneys, liver and
spinal cord, thus optimizing the PTV and normal tissue
DVHs [12]. Typical input parameters were defined on an
individual basis for each patient, and thus varied from
patient to patient, depending on the geometry of the PTV
relative to the normal structures. Table 1 exemplifies the
input parameters for the normal structures in two
different patients. Generally, the input parameters for
the PTV were ,2% of the PTV receiving ,50.4 Gy with a
maximum of 54.0 Gy.
Patient 2
% of tissue volume to Maximum dose to
receive .dose (Gy) structure (Gy)
,5% . 8 50
,5% . 8 50
,45% .35 50
,30% .35 42
The British Journal of Radiology, June 2006
IMRT in the treatment of gastric cancer
The IMRT plans were optimized to minimize the
volume of PTV receiving ,95% of the prescribed dose
and the volume receiving . 110% of the prescribed dose.
Generally, the prescription dose was normalized to the
90–95% isodose line. The initial plans were considered
acceptable with ,2–4% of the PTV receiving ,100% of
the prescribed dose, ,5% of the PTV receiving .110%
and ,1% receiving .115%.
Because the normal tissue DVHs generated by
CORVUS ignores regions of overlap with the PTV, the
CORVUS IMRT plans were exported into PlanUNC [13]
for the purpose of dosimetric comparison with conven-
tional radiation plans.
Comparison of three-dimensional AP/PA plans to
IMRT plans
Three-field (3F) and opposed anterior-posterior:
posterior-anterior (AP/PA) (2F) 3DRT plans were gen-
erated for comparison with IMRT plans for the last 6 of 7
patients treated with IMRT (for one patient, the CT scan
with contoured structures was unable to be recovered
from the electronic archives). The 3F plan included a left
lateral and AP/PA fields. 3F and 2F plans were
generated using PlanUNC [13]. Segmented fields, vari-
able weighting of fields and wedges were used to
optimize the plan so as to improve dose homogeneity.
With the 3F plan, plans were chosen to minimize right
kidney dose without compromising PTV coverage. At a
minimum, 2F and 3F plans were acceptable only if .95%
of the volume received .98% of the prescribed dose.
Normalization was typically set at 99–100% of the
prescription dose. Custom blocks were used (1 cm
margin around the PTV in each beam’s eye view). All
fields were coplanar. 6 MV and 18 MV photons were
used with the 3F and 2F plans, while 6 MV photons were
used with the IMRT plans. DVHs were obtained for the
PTV, kidneys, liver and spinal cord.
Acute toxicity was scored using RTOG morbidity
scoring criteria [14]. Dosimetric endpoints for the target
and critical structures were compared using the two
tailed paired t-test.
Results
Seven patients with gastric cancer were treated with
IMRT. The patient characteristics are outlined in Table 2.
All six post-operatively treated patients had pathological
stage IIIa-IIIb disease. Tumours were located in the
antrum (n54) and lesser curvature (n53).
Dosimetric comparison between conventional and
IMRT plans
To demonstrate the differences in dose distribution,
Figure 1 shows isodose curves on an axial slice for one
patient for IMRT, 2F and 3F plans. Additionally, Figure 2
shows the DVH curves for the organs at risk for one
representative patient. Table 3 summarizes the mean
doses to the PTV, kidneys and liver for 2F, 3F and IMRT
plans. Compared with the 2F plan, IMRT significantly
The British Journal of Radiology, June 2006
Table 2. Patient characteristics
Number 7
Age (years) 54.5–86.6
average 62.8
T stage T2 3
T3 4
N stage N0 0
N1 3
N2 4
Stage IIIa 5
Stage IIIb 1
Not surgically staged 1
Total gastrectomy 2
Subtotal gastrectomy 4
Negative margins 4
Positive margins 2
reduced the mean dose to the left kidney. Compared
with the 3F plan, IMRT significantly reduced the mean
dose to the liver and right kidney, while the reduction in
the dose to the left kidney was not significant.
Table 4 summarizes the dosimetric endpoint analysis.
Table 4a summarizes the volume of the PTV receiving
greater than 50.4 Gy and 55.4 Gy. Table 4b summarizes
the volume of critical structures receiving greater than
the threshold dose [15]. Compared with 2F planning,
IMRT significantly reduced the volume of left kidney
receiving .20 Gy. Compared with 3F planning, IMRT
significantly reduced the volume of right kidney receiv-
ing .20 Gy, and the volume of liver receiving .30 Gy.
The mean PTV dose as well as the volume receiving
.55.4 Gy were greater for the IMRT plans, though
compared with 3F, neither parameter was significantly
different. The maximum doses for 2F, 3F and IMRT plans
were 109%¡3%, 108%¡3% and 115¡2%, respectively,
with IMRT being significantly greater than 2F of 3F.
Clinical/toxicity outcome
In this small cohort of patients, 3 of 7 are long term
survivors (.2 years after diagnosis) and remain without
evidence of disease. One died more than a year after
diagnosis, after developing a malignant pleural effusion
at 9 months. Two died from metastatic disease and rapid
deterioration at 7 months and 18 months following
diagnosis (both patients did not receive chemotherapy
immediately following radiation); it is not known if these
two patients had a component of local and/or regional
failure. The 87-year-old patient treated with planned pre-
operative radiation had a stroke preceded by a fall and
development of an interventricular haemorrhage
(5 months after completing radiotherapy) and died
8 months after diagnosis.
All patients completed their planned course of treat-
ment with no planned or unplanned treatment breaks
and no reduction or discontinuation of chemotherapy.
Chemoradiotherapy with IMRT was well tolerated with
no grade > 3 acute toxicity occurring during radio-
therapy. Acute gastrointestinal and haematological toxi-
city is summarized in Table 5. Acute weight loss (up to
1 month post-radiation) ranged from no loss (1 patient)
to a maximum of 12.3% weight loss, with a mean
499

(a)
(c)
Figure 1. Isodose curves on an axial slice for a representative patient for: (a) 2F (AP/PA) plan, (b) 3F plan, and (c) IMRT plan.
percentage of 6.1%¡4.7% and median of 5.8%. Skin
toxicity was grade 0–1 in all patients. Of the four patients
who received chemotherapy after radiation, two (50%)
had grade 3 haematological (WBC) toxicity, occurring
after the last dose of chemotherapy in both patients. No
abnormalities were detected on laboratory assessment of
kidney function (as compared with pre-treatment/base-
line values) for any of the seven patients, either during
treatment or on last available clinic follow-up visit; one
patient had elevated liver enzymes approximately
4 months after the completion of RT.
Although no late toxicity has been seen, the number of
long term survivors is too small in this series to permit
detailed analysis.
Discussion
IMRT in the treatment of gastric cancer has the potential
of lowering treatment related toxicity. To date, little has
been published on the use of IMRT with gastric cancer.
The University of Heidelberg has published two
papers comparing the dose distribution of IMRT versus
other planning modalities. In the first study, one patient
500
M T Milano, M C Garofalo, S J Chmura et al
(b)
was planned to 45 Gy with the following approaches: 8
field step and shoot IMRT (planned with CORVUS using
multiple couch angles), conventional 4-field box (4F), 4F
with an off-kidney boost after 16.2 Gy and a non-
coplanar 4F plan with a 90 ˚ couch kick and gantry angle,
to direct the beam off of the caudal kidneys [16]. IMRT
reduced the median dose to the kidneys (particularly the
left kidney) and liver; the doses that were exceeded by
30% and by 60% of the volumes (both liver and kidneys)
were also reduced with IMRT. In a follow-up paper with
15 patients, IMRT was compared against AP/PA, 4F and
serial tomotherapy [17]. IMRT reduced dose to the left
kidney (at the expense of greater spinal cord dosing) as
compared with 4F, with a slight reduction as compared
with AP/PA. Compared with 4F, IMRT more consis-
tently produced high quality plans, as evidenced by the
reduced standard deviation of the mean doses and doses
that were exceeded by 30% and by 60% of the volumes.
The dosing of normal tissues was not significantly
different between IMRT and serial tomotherapy, though
serial tomotherapy was superior with respect to con-
formality and homogeneity.
Princess Margaret Hospital planned 20 gastric cancer
patients to receive 45 Gy with conventional 5F coplanar
The British Journal of Radiology, June 2006
IMRT in the treatment of gastric cancer

(a) (b)

(c) (d)

Figure 2. Dose–volume histogram (DVH) curves for the organs at risk for a representative patient. (a) Liver, (b) spinal cord,
(c) left kidney, and (d) right kidney. [y-axis: %Volume; x-axis: Dose (Gy); Colour Scheme- Blue: opposed anterior-posterior:
posterior-anterior (AP/PA) plan; Red: 3F Plan; Green: IMRT plan].

Table 3. Mean dose to structures (percentage of prescribed dose)

Structure 2F 3F IMRT p-values*

2F vs 3F 2F vs IMRT 2F vs 3F

PTV 102.0¡0.7% 103.5¡1.6% 105.7¡1.3% NS 0.0014 0.044

Right kidney 24.8¡14.9% 26.7¡13.0% 18.9¡9.6% 0.021 NS NS
Left kidney 62.7¡30.2% 47.3¡24.2% 38.4¡19.1% NS 0.0049 NS
Liver 41.6¡9.5% 67.9¡17.8% 44.6¡6.4% 0.019 NS 0.0065
Values are given as mean¡standard deviation.
2F, AP/PA two-field three-dimensional plan; 3F, three-field conformal three dimensional plan; NS, not significant (p.0.05);
IMRT, intensity-modulated radiation therapy.
*Two-tailed paired t-test.

plans versus 7–9 field IMRT plans, generated by the sparing in 69% liver sparing in 71%, and spinal cord
CADPLAN Helios planning system [18]. Three reviewers sparing in 74%.
examined the plans, with IMRT being preferred in 89%, An Australian study compared AP/PA with a 3D
with subjectively better PTV coverage in 86%, kidney conformal technique using a mono-isocentric split field

The British Journal of Radiology, June 2006 501

 M T Milano, M C Garofalo, S J Chmura et al

Table 4. Volume of structures receiving greater than the threshold dose for 50.4 Gy treatment (a) Target
Volume above given dose (%) p-values*
Structure Dose (Gy)
2F 3F IMRT 3F vs IMRT 2F vs IMRT 2F vs 3F

PTV 55.4 1.3¡2.4 0.7¡1.6 4.4¡3.9 NS NS NS

PTV 50.4 84.1¡11.1 94.9¡4.0 98.1¡1.2 NS 0.016 0.022

(b) Organs at Risk

Right 20.0 20.1¡18.0 20.9¡16.6 11.6¡8.3 0.027 NS NS
kidney
Left kidney 20.0 67.6¡34.7 52.0¡31.9 52.1¡28.9 NS NS NS
Liver 30.0 33.7¡12.6 63.6¡28.5 18.9¡12.9 0.010 0.030 0.0062
Spinal cord 45.0 55.0¡22.3 4.1¡7.3 0 NS 0.00039 0.00033
Values are given as mean¡standard deviation.
2F, AP/PA two-field three-dimensional plan; 3F, three-field conformal three-dimensional plan; NS, not significant (p.0.05);
IMRT, intensity-modulated radiation therapy.
*Two-tailed paired t-test.

Table 5. Acute gastrointestinal (GI) and haematological
toxicity
Grade 0 1 2 .2
Upper GI 0 0 7 0
Lower GI 3 1 3 0
Oesophagitis 4 1 2 0
Haematological* 2 2 3 0
*Toxicity during radiation.
technique [19]. The 3D conformal technique improved
mean and threshold dose to the kidneys and spinal cord,
but not the liver. Memorial Sloan Kettering recently
investigated the use of IMRT in four patients with gastric
lymphoma, whose PTV had a high degree of overlap
with the kidneys. IMRT tended to improve kidney and
liver sparing [20].
The present study confirms the efficacy of IMRT in
reducing kidney and liver doses. IMRT, planned with
dynamic multileaf collimation, was compared with both
the standard 2F AP/PA approach as well as a 3-field
technique. As in the Heidelberg study, we used
CORVUS planning software. Two differences are that
our IMRT planning used dynamic multileaf collimation,
and that the IMRT dose matrices were exported into
PLUNC, allowing inclusion of normal structure volumes
that overlap with PTV. We also prescribed to a higher
dose, 50.4 Gy versus 45.0 Gy.
Our data demonstrate that IMRT offers better sparing
of the right kidney compared with conventional 3F
planning, with significantly lower mean dose and
volume above threshold dose [15]; and IMRT offers
better sparing of the left kidney as compared with 2F
planning, with lower mean dose and reduced volume
above threshold dose (the latter of which was not
significant). IMRT also affords liver sparing. As expected
from the use of the additional field in the 3F arrange-
ment, the liver dose increases dramatically with attempts
to improve target coverage as compared with 2F
planning. IMRT achieves similarly excellent target cover-
age as compared with 3F planning, while reducing the
mean liver dose and volume above threshold dose [15].
In part, these improvements in relation to kidney and
liver dose reduction with the use of IMRT may be due to
reducing entrance and exit dose to these organs from
beam angle selection.
To our knowledge, this study is the first to report
clinical outcome in gastric cancer patients treated with
IMRT. With respect to acute toxicity, patients fared
remarkably well (albeit in a very small series), particu-
larly since a higher than standard dose (50.4 Gy vs
45.0 Gy) was administered. There was no grade 3 or
greater toxicity during radiation and all patients com-
pleted their planned course of chemoradiotherapy. In the
Intergroup trial, 33% experienced grade 3 or greater
gastrointestinal toxicity and 64% completed the treat-
ment as planned. Grade 3 haematological toxicity was
seen in the week following post-radiation chemotherapy
in 2/4 patients, a percentage similar to that seen in the
Intergroup trial (54% grade 3 haematological toxicity).
In summary, IMRT offers improved sparing of normal
structures, allowing a dose of 50.4 Gy to be delivered to
the PTV. In this small series of patients, treatment was
well tolerated with acceptable toxicity, seemingly
improved as compared with the Intergroup trial. IMRT
in the treatment of gastric cancer warrants further study.
References
1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA
Cancer J Clin 2005;55:10–30.
2. Gunderson LL, Sosin H. Adenocarcinoma of the stomach:
areas of failure in a re-operation series (second or
symptomatic look) clinicopathologic correlation and impli-
cations for adjuvant therapy. Int J Radiat Oncol Biol Phys
1982;8:1–11.
3. Landry J, Tepper JE, Wood WC, et al. Patterns of failure
following curative resection of gastric carcinoma. Int J Radiat
Oncol Biol Phys 1990;19:1357–62.
4. Macdonald JS, Smalley SR, Benedetti J, et al.
Chemoradiotherapy after surgery compared with surgery
alone for adenocarcinoma of the stomach or gastroesopha-
geal junction. N Engl J Med 2001;345:725–30.
5. Smalley SR, Gunderson L, Tepper J, et al. Gastric surgical
adjuvant radiotherapy consensus report: rationale and
treatment implementation. Int J Radiat Oncol Biol Phys
2002;52:283–93.
6. Milano M, Jani A, Farrey K, et al. Intensity modulated
radiation therapy (IMRT) in the treatment of anal cancer:

502 The British Journal of Radiology, June 2006

IMRT in the treatment of gastric cancer
toxicity and clinical outcome. Int J Radiat Oncol Biol Phys
(In Press).
7. Milano MT, Chmura SJ, Garofalo MC, et al. Intensity-
modulated radiotherapy in treatment of pancreatic and bile
duct malignancies: toxicity and clinical outcome. Int J Radiat
Oncol Biol Phys 2004;59:445–53.
8. Ichikura T, Tomimatsu S, Uefuji K, et al. Evaluation of the
New American Joint Committee on Cancer/International
Union against cancer classification of lymph node metas-
tasis from gastric carcinoma in comparison with the
Japanese classification. Cancer 1999;86:553–8.
9. Kunisaki C, Shimada H, Nomura M, et al. Comparative
evaluation of gastric carcinoma staging: Japanese classifica-
tion versus new american joint committee on cancer/
international union against cancer classification. Ann Surg
Oncol 2004;11:203–6.
10. International Commission on Radiation Units and
Measurements (ICRU). Report Number 50. Prescribing,
recording and reporting photon beam therapy.
Washington, D.C.: ICRU; 1993.
11. Chen GT, Jiang SB, Kung J, et al. Abdominal organ motion
and deformation: implications for IMRT. Int J Radiat Oncol
Biol Phys 2001;51:210 (abstract).
12. Lawrence TS, Kessler ML, Ten Haken RK. Clinical inter-
pretation of dose-volume histograms: the basis for normal
tissue preservation and tumor dose escalation. Front Radiat
Ther Oncol 1996;29:57–66.
The British Journal of Radiology, June 2006
13. Sailer SL, Chaney EL, Rosenman JG, et al. Three dimen-
sional treatment planning at the University of North
Carolina at Chapel Hill. Semin Radiat Oncol 1992;2:267–73.
14. Trotti A, Byhardt R, Stetz J, et al. Common toxicity criteria:
version 2.0. an improved reference for grading the acute
effects of cancer treatment: impact on radiotherapy.
Int J Radiat Oncol Biol Phys 2000;47:13–47.
15. Emami B, Lyman J, Brown A, et al. Tolerance of normal
tissue to therapeutic irradiation. Int J Radiat Oncol Biol
Phys 1991;21:109–22.
16. Lohr F, Dobler B, Mai S, et al. Optimization of dose
distributions for adjuvant locoregional radiotherapy of
gastric cancer by IMRT. Strahlenther Onkol 2003;179:557–63.
17. Wieland P, Dobler B, Mai S, et al. IMRT for postoperative
treatment of gastric cancer: covering large target volumes in
the upper abdomen: a comparison of a step-and-shoot and
an arc therapy approach. Int J Radiat Oncol Biol Phys
2004;59:1236–44.
18. Ringash J, Perkins G, Lockwood G, et al. IMRT for adjuvant
radiation in gastric cancer: a referred plan? Int J Radiat
Oncol Biol Phys 2003;57:S381–2.
19. Leong T, Willis D, Joon DL, et al. 3D conformal radio-
therapy for gastric cancer--results of a comparative plan-
ning study. Radiother Oncol 2005;74:301–6.
20. Della Bianca C, Hunt M, Furhang E, et al. Radiation
treatment planning techniques for lymphoma of the
stomach. Int J Radiat Oncol Biol Phys 2005;62:745–51.
503
 The British Journal of Radiology, 79 (2006), 504–509

Changes in applicator positions and dose distribution between

high dose rate brachytherapy fractions in cervix carcinoma
patients receiving definitive radiotherapy
1
M GARIPAĞAOĞLU, MD, 1N TUNÇEL, PhD, 1M G DALMAZ, MD,
2
H GÜLKESEN, MD,
1
A TOY, BSc,
1
A Ü KIZILDAĞ, BSc and 1F G KÖSEOĞLU, MD

Akdeniz University School of Medicine, Departments of 1Radiation Oncology and 2Bioistatistics,

Antalya, 07070 Turkey

ABSTRACT. This study examines the change of applicator geometry and its effect on
rectal/rectum (R) and bladder (B) doses, and obtained radiobiological equivalent doses
(RED), between each high dose rate (HDR) brachytherapy (BT) fraction in cervical
carcinoma patients. BT using a tandem (T) and two ovoids (O) is included, and any
discrepancies in applicator positions among the fractions were calculated. Whether the
change of applicator position had an effect on the calculated R and B doses was
analysed. Furthermore, the relationship between the size of tumour, the magnitude of
displacement and the change in R and B doses was also investigated. Lastly, the
changes in R and B RED were noted. The average magnitude of displacement was
between 2.0 mm and 16.9 mm, showing time trend. There was no relationship
between tumour size and the magnitude of discrepancy of Left O, Right O, T, R, B, and
neither change in R and B doses (p.0.05). The mean differences of R and B doses were Received 28 January 2005
between 49–78 cGy, and 70–84 cGy, respectively. The magnitude of discrepancy and Revised 24 October 2005
changes in doses showed no correlation (p.0.05). There were no significant differences Accepted 5 December 2005
in REDs for bladder (p50.8) and rectum (p50.2). In conclusion, there were significant
DOI: 10.1259/bjr/33762931
differences in the applicator positions R and B and R and B doses among the fractions,
which confirm the necessity of treatment planning in each HDR BT fraction. However, ’ 2006 The British Institute of
the total calculated R and B REDs did not show a remarkable difference. Radiology

Radiotherapy consisting of external radiotherapy and
brachytherapy (BT) is the only curative treatment
method for locally advanced stage cervical carcinoma
[1–4]. The natural steep dose gradient of brachytherapy
dose distribution allows a higher dose to the target while
not exceeding the tolerance limit of normal rectal (R) and
bladder (B) tissues. The optimal integration of bra-
chytherapy and external radiotherapy is the main
defining factor of radiotherapy treatment characteristics,
such as the total point A dose, total paracentral dose and
total treatment time which are independent prognostic
factors in the treatment of cervical carcinoma as well as
tumour and patient related factors [1, 2, 5–7]. It was
claimed that in comparison with low dose rate (LDR) BT,
high dose rate (HDR) BT has a physical dose distribution
advantage. However, some authors claim that HDR BT
has a radiobiological disadvantage [8–15], although
performing HDR BT with multiple small fractions could
alter this radiobiological disadvantage [5, 11, 16, 17].
Critical organs, namely R and B, are very close to the
applicators and high dose region of BT. A small change
in the distance from a particular point to the active
sources may cause a great dose difference due to the BT
Address correspondence to: Dr Melahat Garipağaoğlu, Acibadem
Oncology and Neurological Science Hospital, Department of
Radiation Oncology, Inonu cad. Okur sok. No:20 Kozyatağı 34742
Istanbul, Turkey.
having a large dose gradient near the region close to the
active sources [14, 18]. Therefore, a precise dose calcula-
tion of normal tissue as well as the target is critical for
successful treatment in HDR BT. The aim of this study is
to detect the change of applicator geometry between
HDR BT fractions and its effect on the calculated R and B
doses, and the given radiobiological effective doses
(RED)s, using tandem (T) and ovoid (O) in cervical
carcinomas.
Methods
The orthogonal radiographs of patients receiving
intracavitary brachytherapy using T and two Os were
used for this retrospective study.
External radiotherapy was administered with a 1.8 Gy
fraction size, totalling 46–50.4 Gy to the whole pelvis.
BT was done five times using Ir192 Microselectron
HDR machine with 6 Gy fraction size, starting after the
completion of at least 40 Gy of external radiotherapy in
order to achieve enough tumour regression to perform
an optimal application, and BT fractions were used twice
a week to reduce the total treatment time.
The application was done after the patient had fasted
for 12 h. Laxatives were used for the elimination of rectal
matters, urinary Foley catheter was used to keep the
bladder empty during the whole procedure, and the

504 The British Journal of Radiology, June 2006

Position and dose variations among HDR brachytherapy fractions
balloon of the Foley catheter was filled with 7 cm3 of an
opaque solution.
Before the BT application, the same procedure of
sedation and analgesia was given in all fractions of each
patient. An extra analgesic was given when it was
needed after the applicator placement.
After the completion of the application, one tandem
and two ovoids were fixed to each other and a specially
designed table used until completion of treatment, to
ensure the immobilization of the patient during simula-
tion, treatment and transport.
In addition, all the applications were given by the
same experienced physician using the same kind of
packing technique.
Orthogonal X-rays were taken for each fraction. A
reference volume definition for R and B dose calculations
were carried out according to ICRU-38 recommenda-
tions. Treatment planning and optimization were
applied in each fraction. Special effort was made not to
exceed rectal and bladder tolerance levels.
In order to determine the changes of applicator
positions between the BT fractions, pelvic bone reference
points were used as fixed points, applicators (T, left O,
right O), B and R reference points were used as non-fixed
points (Figure 1). A Cartesian coordinate centre (for x, y
and z axes) was chosen using orthogonal radiographs
(anterior-posterior (AP) and lateral) (Figure 1). The
position of fixed and non-fixed points were measured
according to centre of Cartesian coordinate centre in x, y
and z axes. The distances between the fixed and non-
fixed points were measured on the x and y axes on AP
Figure 1. The fixed points, which are pelvic bone reference points; the non-fixed points, namely left ovoid (LO), right ovoid
(RO), tandem stopper (S), bladder (B) and rectal (Rg) reference points are shown on (a) anterior-posterior and (b) lateral
diagram.
The British Journal of Radiology, June 2006
radiographs and in z axis on lateral radiograms. The
differences in the measured distances were calculated
between the first fractions and subsequent fractions.
Whether the magnitude of displacement showed a time
trend was investigated. Furthermore, the magnitudes of
the displacements on the x, y and z axes were used to
calculate the resultant vector. The relationship between
the initial tumour size and the magnitude of resultant
vectors was also examined.
For this study, to evaluate whether the change of
applicator positions to the bony pelvis has an effect on
the calculated R and B doses, both active source positions
and calculated treatment time of first fraction were
repeated for subsequent fractions of each patient, and
inferential R and inferential B dose calculations were
made. Then the difference between the ICRU-38 recom-
mended R and B doses in the first fraction and the
subsequent inferential R and B doses was determined
[19]. The relationship between changes in subsequent
inferential R and B doses and initial tumour size
obtained from MR images was also evaluated.
It was assumed that geometric variations might lead to
a change in calculated doses. The relationship between
the magnitude of displacement of the applicators and the
inferential doses were examined.
In an effort to determine whether changes in R and B
doses have an effect on R and B REDs, R and B REDs of
each fraction were calculated using inferential R and B
doses, according to the ‘‘linear quadratic model,’’ then
added [20]. The chosen alpha/beta value was 3 for B and
R, as in previous studies [21]. The change in total REDs
505
 M Garipağaoğlu, N Tunçel, M G Dalmaz et al

of R and B calculated in concordance with the linear
quadratic model was investigated.
Statistical analysis
Statistical analysis was performed using SPSS 10.0
software, the normality of samples was analysed by a
Shapiro-Wilk test. Statistical significance of the displace-
ment and dose changes between fractions was tested
using the paired t-test or a Wilcoxon signed-rank test.
Spearman’s rho test was used for correlations; p,0.05 is
considered to be statistically significant.
Results
13 out of 75 total fractions were excluded from the
study because of low quality X-rays or the use of the
applications other than tandem and ovoids. 15 distances
of each fraction were measured on AP and lateral
graphs, with a total of 930 distances recorded.
The average magnitude of discrepancy in the x, y and z
axes and resultant vector are presented in Table 1. The
mean discrepancy of left O, right O, T, B and R reference
points between the fractions were between 10.0 mm and
19.4 mm. In the first fraction, the distance of the T from
the non-fixed point was 37¡3 mm (mean¡standard
deviation), and was 31¡3 mm in the fourth fraction
(p50.012). In other words, the stopper of T moved 6 mm
superior towards the end of the treatment. No other
statistically significant time trend or movement to a
certain direction of non-fixed points could be determined
(data not shown). There was no significant relationship
between the initial tumour size and the magnitude of
resultant vectors of left O (r50.239, p50.132), right O
(r50.24, p50.13), T (r50.036, p50.821), R (r50.035,
p50.834) and B (r50.309, p50.059).
The differences calculated between inferential R and B
doses and R and B doses in the first fraction are
presented in Table 2. The mean differences for B and R
were between 78–149 cGy and 70–84 cGy, respectively.
A correlation was seen between initial tumour size and
changes in inferential R doses (r50.414, p50.005), but no
correlation was seen in the B doses (r5–0.075, p50.621).
The magnitude of resultant vectors and changes in doses
for R (r50.455, p50.005), and B (r5–0.418, p50.007)
showed correlation (Figure 2).
The difference between inferential dose and admini-
stered dose for B was greater than 60 cGy in 15 (100%) of
the patients, and 6 (40%) of these patients had a
difference greater than 120 cGy. The difference for the
R was greater than 60 cGy in 10 (66%) patients; 6 (40%) of
them had a difference greater than 120 cGy.
The median of differences between total REDs and
total inferential REDs for the R for each patient was
11 Gy, whereas it was 4 Gy for the B as BED values. The
differences were not statistically significant.
Discussion
Significant geometric variations were seen in all three
applicators’, namely left O, right O and T positions,
between the HDR brachytherapy fractions. As noted
previously in either LDR or HDR brachytherapy, these
position differences were more than 1 cm in 60% of the
applications [22–29].

Table 1. Absolute discrepancy on x, y, z axes and displacement in resultant vector (mean, mm)
Delta x Delta y Delta z Resultant 95% CI SD
vector

Second fraction (n515)

Left ovoid 10.0 5.5 8.3 16.6 11.0–22.2 10.8
Right ovoid 11.4 5.5 7.3 17.2 11.2–23.2 11.8
Stopper 9.8 7.6 5.7 16.9 11.3–22.5 10.9
Bladder 8.9 2.0 4.9 12.5 8.1–16.9 8.1
Rectum 8.4 5.9 6.8 13.9 9.9–17.9 7.4
Third fraction (n514)
Left ovoid 11.3 6.2 7.9 16.9 10.9–22.9 11.1
Right ovoid 12.3 6.4 6.5 17.6 11.8–23.4 11.0
Stopper 10.1 6.5 6.4 16.7 11.7–21.7 9.2
Bladder 11.0 4.2 8.3 15.8 7.8–23.8 14.3
Rectum 8.3 6.8 5.9 15.4 11.0–19.8 8.2
Fourth fraction (n511)
Left ovoid 9.5 6.1 9.9 16.7 12.1–21.3 7.5
Right ovoid 12.7 6.3 10.9 19.4 14.0–24.8 8.8
Stopper 7.7 6.8 9.9 15.6 8.6–22.6 11.7
Bladder 10.9 3.7 5.8 15.1 5.3–24.9 15.4
Rectum 7.2 6.1 7.7 13.1 9.5–16.7 5.6
Fifth fraction (n55)
Left ovoid 14.0 4.2 5.7 16.8 5.6–28.0 12.5
Right ovoid 16.9 5.9 3.2 17.1 6.9–27.3 11.4
Stopper 13.7 8.2 4.1 17.6 7.6–27.6 11.1
Bladder 6.3 4.3 8.5 10.0 5.4–14.6 4.7
Rectum 9.2 8.9 8.1 17.3 11.3–22.9 6.1
CI, confidence interval; SD, standard deviation.

506 The British Journal of Radiology, June 2006

Position and dose variations among HDR brachytherapy fractions

Table 2. The difference between inferential rectal and started after the completion of 40 Gy external radio-
bladder doses, calculated using first fraction active sources therapy. The stopper of tandem moved approximately
positions and treatment times 6 mm superior towards the end of the treatment. We
speculate that this finding supports continuing tumour
Mean 95% CI SD
dose shrinkage in a period of brachytherapy fractions [24, 27,
29].
Second fraction (n515) In comparison with the magnitude of changes in B
Bladder 92 37–147 99
position, the changes in R position were higher. The B
Rectum 77 35–119 73
Third fraction (n514)
was kept empty and the ICRU 38 recommended
Bladder 149 79–219 126 B reference point is related to the balloon of the Foley
Rectum 84 34–135 88 catheter, therefore showing that the change in B position
Fourth fraction (n511) was small. The ICRU 38 recommended R reference point
Bladder 78 36–119 69 is defined according to the position of the applicators, so
Rectum 83 15–152 113 the changes in applicator positions also affected the
Fifth fraction (n55) changes in ICRU 38 recommended R reference point, as
Bladder 99 38–160 79 expected [28].
Rectum 70 24–116 55
The inferential R and B doses were helpful for
CI, confidence interval; SD, standard deviation. predicting R and B doses of subsequent fractions when
treatment planning and dose calculation were not
performed for remaining fractions. In addition, positional
Furthermore, in comparison with T and R applica- differences were seen, and R and B doses differed among
tions, the magnitude of differences was lower in the the fractions. (This difference was higher than 60 cGy,
present study, compatible with other reports regarding T which is 10% of prescribed dose in 10 out of 15 patients in
and O applications [22–26, 29]. B doses and 15 out of 15 patients in R doses, and higher
Some authors claim that insufficient fixation of than 120 cGy, which is 20% of prescribed dose in 10 out of
applicators both to each other and to the table is a main 15 patients in B doses and 15 out of 15 patients in R doses).
potential reason for geometric variation [26, 29]. The precise physical dose calculation is more important in
Although the variations were higher in non-fixed HDR than LDR brachytherapy because of its claimed
applications than in fixed and the changes in O were radiobiological disadvantage. Therefore, the prescribed
higher than in T in non-fixed applications, there was still dose should be adjusted considering normal tissue doses.
a significant variation in the present study. This magnitude of differences in normal tissue dose
Research also showed that another reason for dis- illustrates the necessity of treatment planning and dose
placement was anaesthesia given among the fractions calculation in each fraction of HDR brachytherapy, which
[27, 28]. In the present study, the same kind of is in agreement with other studies [24, 25, 27–29].
anaesthesia was used in all fractions of each patient. Furthermore, there was a relationship between the
However, there were significant geometric variations in initial tumour size and the change in R dose, as seen in
the applicator positions among the fractions. the literature [24, 28]. However, there was no association
In order to achieve enough tumour response to between the B dose and the initial tumour size.
perform an optimal application, brachytherapy was Presumably, the reason is that the ICRU 38 recommended

Figure 2. Simple scatter plot gra-

phics of relationship between resul-
tant vector and dose change in the
rectum (r520.418, p50.007).

The British Journal of Radiology, June 2006 507


R reference point is defined according to Os location,
therefore tumour shrinkage has an effect on R point. On
the contrary, ICRU 38 recommended B reference point is
dependent on Foley catheter balloon position without
using the applicators.
There were remarkable relationships between the
magnitudes of geometric displacement (as a resultant
factor) and dose variation for R and B. Both tumour
control and radiation injury increased with increasing
radiation doses. B and R are the main dose limiting
organs in the pelvis. For this reason, brachytherapy is
helpful for increasing central tumour doses while
external radiotherapy with midline shielding is helpful
for increasing parametrial doses without increasing R
and B doses. The hypothesis of using brachytherapy to
increase central tumour dose and of using external
radiotherapy with midline shielding to increase para-
metrial doses, without increasing R and B doses has
many shortcomings, such as the increase of rectal
complications, the difficulty of defining vagina location,
which kind of block technique (i.e. stepwedge or
rectangle) should be used and what percentage of
isodose is best to integrate with external radiotherapy
successfully [3, 5, 28, 30–35]. Moreover, according to the
results of the current study and previous studies, the
calculated doses differ among HDR fractions using T and
O [26–29]. For these reasons, the design of the midline
shield according to isodose distribution of the first BT
fraction could result in an inappropriate isodose dis-
tribution, due to changes in subsequent fraction’s
isodose distribution.
The total inferential R and B REDs were relatively
similar. However, one should keep in mind that the
acceptable R and B doses were calculated ICRU 38
recommended R and B reference points and based on the
points, rather than volume. As shown in other studies,
these doses do not represent the real R and B doses [36–
39]. On the other hand, volume based three-dimensional
planning for each fraction is not possible in most centres
using HDR.
The results of the current study suggest that proper
treatment planning, dose calculation and optimization
according to normal tissue doses for each fraction are
necessary in cervical carcinoma patients receiving HDR
brachytherapy using T and O.
References
1. Lanciano R. Optimizing radiation parameters for cervical
cancer. Semin Radiat Oncol 2000;10:36–43.
2. Nori D, Dasari N, Allbright RM. Gynecologic brachyther-
apy I: Proper incorporation of brachytherapy into the
current multimodality management of carcinoma of the
cervix. Semin Radiat Oncol 2002;12:40–52.
3. Logsdon MD, Eifel PJ. Figo IIIB squamous cell carcinoma of
the cervix: an analysis of prognostic factors emphasizing
the balance between external beam and intracavitary
radiation therapy. Int J Radiat Oncol Biol Phys
1999;43:763–75.
4. Green JA, Kirwan JM, Tierney JF, Symonds P, Fresco L,
Collingwood M, et al. Survival and recurrence after
concomitant chemotherapy and radiotherapy for cancer of
the uterine cervix: a systematic review and meta-analysis.
Lancet 2001;358:781–6.
5. Nag S, Erickson B, Thomadsen B, Orton C, Demanes JD,
Petereit D. The American Brachytherapy Society
508
M Garipağaoğlu, N Tunçel, M G Dalmaz et al
recommendations for high-dose-rate brachytherapy for
carcinoma of the cervix. Int J Radiat Oncol Biol Phys
2000;48:201–11.
6. Montana GS, Martz KL, Hanks GE. Patterns and sites of
failure in cervix cancer treated in the U.S.A. in 1978. Int J
Radiat Oncol Biol Phys 1991;20:87–93.
7. Katz A, Eifel PJ. Quantification of intracavitary brachyther-
apy parameters and correlation with outcome in patients
with carcinoma of the cervix. Int J Radiat Oncol Biol Phys
2000;48:1417–25.
8. Scalliet P, Gerbaulet A, Dubray B. HDR versus LDR
gynecological brachytherapy revisited. Radiother Oncol
1993;28:118–26.
9. Stitt JA, Fowler JF, Thomadsen BR, Buchler DA, Paliwal BP,
Kinsella TJ. High dose rate intracavitary brachytherapy for
carcinoma of the cervix: the Madison system: I. Clinical and
radiobiological considerations. Int J Radiat Oncol Biol Phys
1992;24:335–48.
10. Hsu WL, Wu CJ, Jen YM, Yen SH, Lin KT, Ger LP, et al.
Twice-per-day fractionated high versus continuous low
dose rate intracavitary therapy in the radical treatment of
cervical cancer: a nonrandomized comparison of treatment
results. Int J Radiat Oncol Biol Phys 1995;32:1425–31.
11. Orton CG. Fractionated high dose rate versus low dose rate
cervix cancer regimens. Br J Radiol 1991;64:1165–6.
12. Dale RG. The use of small fraction numbers in high dose-
rate gynaecological afterloading: some radiobiological con-
siderations. Br J Radiol 1990;63:290–4.
13. Eifel PJ. High-dose-rate brachytherapy for carcinoma of the
cervix: high tech or high risk? Int J Radiat Oncol Biol Phys
1992;24:383–6.
14. Thomadsen BR, Shahabi S, Stitt JA, Buchler DA, Fowler JF,
Paliwal BR, et al. High dose rate intracavitary brachyther-
apy for carcinoma of the cervix: the Madison system: II.
Procedural and physical considerations. Int J Radiat Oncol
Biol Phys 1992;24:349–57.
15. Sminia P, Schneider CJ, Fowler JF. The optimal fraction size
in high-dose-rate brachytherapy: dependency on tissue
repair kinetics and low-dose rate. Int J Radiat Oncol Biol
Phys 2002;52:844–9.
16. Brenner DJ, Hall EJ. Fractionated high dose rate versus low
dose rate regimens for intracavitary brachytherapy of the
cervix. I. General considerations based on radiobiology. Br J
Radiol 1991;64:133–41.
17. Orton CG, Seyedsadr M, Somnay A. Comparison of high
and low dose rate remote afterloading for cervix cancer and
the importance of fractionation. Int J Radiat Oncol Biol Phys
1991;21:1425–34.
18. King GC, Bloomer WD, Kalnicki S, Faul CM, Gooding WE,
Stockstill TF, et al. Point dose variations with time during
traditional brachytherapy for cervical carcinoma. Med
Dosim 2000;25:77–80.
19. ICRU International Commission of Radiation Units and
Measurements. Dose and Volume Specifications for report-
ing intracavitary therapy in gynaecology, ICRU report 38,
1985 Bethesda, MD.
20. Fowler JF. The linear-quadratic formula and progress in
fractionated radiotherapy. Br J Radiol 1989;62:679–94.
21. Sood B, Garg M, Avadhani J, Gorla G, Malhotra H, Guha C,
et al. Predictive value of linear-quadratic model in the
treatment of cervical cancer using high-dose-rate
brachytherapy. Int J Radiat Oncol Biol Phys 2002;54:
1377–87.
22. Grigsby PW, Georgiou A, Williamson JF, Perez CA. Grigsby
Anatomic variation of gynecologic brachytherapy prescrip-
tion points. Int J Radiat Oncol Biol Phys 1993;27:725–9.
23. Corn BW, Galvin JM, Soffen EM, Henze G, Schwaibold F.
Positional stability of sources during low-dose-rate bra-
chytherapy for cervical carcinoma. Int J Radiat Oncol Biol
Phys 1993;26:513–8.
The British Journal of Radiology, June 2006
Position and dose variations among HDR brachytherapy fractions
24. Hoskin PJ, Cook M, Bouscale D, Cansdale J. Changes in
applicator position with fractionated high dose rate
gynaecological brachytherapy. Radiother Oncol
1996;40:59–62.
25. Kim RY, Meyer JT, Plott WE, Spencer SA, Meredith RF,
Jennelle RL, et al. Major geometric variations between
multiple high-dose-rate applications of brachytherapy in
cancer of the cervix: frequency and types of variation.
Radiology 1995;195:419–22.
26. Kim RY, Meyer JT, Spencer SA, Meredith RF, Jennelle RL,
Salter MM. Major geometric variations between intracavi-
tary applications in carcinoma of the cervix: high dose rate
vs. low dose rate. Int J Radiat Oncol Biol Phys
1996;35:1035–8.
27. Bahena JH, Martinez A, Yan D, Mele E, Edmunson G,
Brown D, et al. Spatial reproducibility of the ring and
tandem high-dose rate cervix applicator. Int J Radiat Oncol
Biol Phys 1998;41:13–9.
28. Hellebust TP, Dale E, Skjonsberg A, Olsen DR. Inter fraction
variations in rectum and bladder volumes and dose
distributions during high dose rate brachytherapy treat-
ment of the uterine cervix investigated by repetitive CT-
examinations. Radiother Oncol 2001;60:273–80.
29. Datta NR, Kumar S, Das KJ, Pandey CM, Halder S,
Ayyagari S. Variations of intracavitary applicator geometry
during multiple HDR brachytherapy insertions in carci-
noma cervix and its influence on reporting as per ICRU
report 38. Radiother Oncol 2001;60:15–24.
30. Huang EY, Lin H, Hsu HC, Wang CJ, Chen HC, Sun LM, et
al. High external parametrial dose can increase the
probability of radiation proctitis in patients with uterine
cervix cancer. Gynecol Oncol 2000;79:406–10.
31. Ting JY, Wolfson AH, Wu X, Fiedler JA, Abdel-Wahab M,
Yang CC, et al. Bladder and rectal doses from external-
beam boosts after gynecologic brachytherapy. Radiology
1998;209:825–30.
The British Journal of Radiology, June 2006
32. Kim CR, Eaton BA, Stevens KR Jr. Localization of the apex
of the vagina: implications for radiation therapy planning.
Radiology 1999;212:155–8.
33. Roeske JC, Mundt AJ, Halpern H, Sweeney P, Sutton H,
Powers C, et al. Late rectal sequelae following definitive
radiation therapy for carcinoma of the uterine cervix: a
dosimetric analysis. Int J Radiat Oncol Biol Phys
1997;37:351–8.
34. Chao KS, Williamson JF, Grigsby PW, Perez CA.
Uterosacral space involvement in locally advanced carci-
noma of the uterine cervix. Int J Radiat Oncol Biol Phys
1998;40:397–403.
35. Wolfson AH, Abdel-Wahab M, Markoe AM, Raub WJ Jr,
Diaz D, Desmond JJ, et al. A quantitative assessment of
standard vs. customized midline shield construction for
invasive cervical carcinoma. Int J Radiat Oncol Biol Phys
1997;37:237–42.
36. Gebara WJ, Weeks KJ, Hahn CA, Montana GS, Anscher MS.
Computed axial tomography tandem and ovoids (CATTO)
dosimetry: three-dimensional assessment of bladder and
rectal doses. Radiat Oncol Investig 1998;6:268–75.
37. Fellner C, Potter R, Knocke TH, Wambersie A. Comparison
of radiography- and computed tomography-based treat-
ment planning in cervix cancer in brachytherapy with
specific attention to some quality assurance aspects.
Radiother Oncol 2001;58:53–62.
38. Serkies K, Badzio A, Jereczek-Fossa B, Tarnawska Z,
Nowak R, Szewczyk P, et al. Rectal doses in intracavitary
brachytherapy of gynecological malignancies: comparison
of two dosimetric methods. Radiother Oncol 2001;58:37–41.
39. Dale E, Hellebust TP, Skjonsberg A, Hogberg T, Olsen DR.
Modeling normal tissue complication probability from
repetitive computed tomography scans during fractionated
high-dose-rate brachytherapy and external beam radio-
therapy of the uterine cervix. Int J Radiat Oncol Biol Phys
2000;47:963–71.
509
 The British Journal of Radiology, 79 (2006), 510–517

Clinical and cellular ionizing radiation sensitivity in a patient with

xeroderma pigmentosum
1
C F ARLETT, PhD, 2P N PLOWMAN, MD, FRCP, FRCR, 2P B ROGERS, MB BS, MRCP, FRCR, 3C N PARRIS, PhD,
3
F ABBASZADEH, MSc, 4M H L GREEN, PhD, 5,6T J MCMILLAN, PhD, 6C BUSH, BSc, 7N FORAY, PhD and
1
A R LEHMANN, PhD

1
Genome Damage & Stability Centre, University of Sussex, Falmer, Brighton BN1 9RQ,
2
Radiotherapy/Clinical Oncology, St Bartholomew’s Hospital, 25 Bartholomew Close, West
Smithfield, London EC1A 7BE, 3Division of Biosciences, School of Health Sciences and Social Care,
Brunel University, Kingston Lane, Uxbridge, Middlesex UP8 3PH, 4School of Pharmacy and
Biomolecular Sciences, University of Brighton, Cockcroft Building, Lewes Road, Brighton BN2 4GJ,
5
Institute of Environmental and Natural Sciences, Faraday Building, Lancaster University, Lancaster
LA1 4YA, 6The Institute of Cancer Research, Royal Cancer Hospital, 15 Cotswold Road, Sutton,
Surrey SM2 5NG, UK, 7INSERM U647, ID17, European Synchrotron Research Facility, Rue Jules
Horowitz, BP220 – 38043 Grenoble, France

ABSTRACT. XP14BR is a cell line derived from a xeroderma pigmentosum (XP) patient
from complementation group C. The patient was unusual in presenting with an
angiosarcoma of the scalp, treated by surgical excision and radiotherapy. Following
38 Gy in 19 fractions with 6 MEV electrons, a severe desquamation and necrosis of the
underlying bone ensued, and death followed 4 years later. The cell line was
correspondingly hypersensitive to the lethal effects of gamma irradiation. We had
previously shown that this sensitivity could be discriminated from that seen in ataxia-
telangiectasia (A-T). The cellular response to ultraviolet radiation below 280 nm (UVC)
was characteristic of XP cells, indicating the second instance, in our experience, of dual
cellular UVC and ionizing radiation hypersensitivity in XP. We then set out to evaluate
any defects in repair of ionizing radiation damage and to verify any direct contribution
of the XPC gene. The cells were defective in repair of a fraction of double strand breaks,
with a pattern reminiscent of A-T. The cell line was immortalized with the vector
pSV3neo and the XPC cDNA transfected in to correct the defect. The progeny derived
from this transfection showed the presence of the XPC gene product, as measured by Received 28 October 2005
immunoblotting. A considerable restoration of normal UVC, but not ionizing radiation, Revised 23 December 2005
sensitivity was observed amongst the clones. This differential correction of cellular Accepted 9 January 2006
sensitivity is strong evidence for the presence of a defective radiosensitivity gene,
DOI: 10.1259/bjr/83726649
distinct from XPC, which is responsible for the clinical hypersensitivity to ionizing
radiation. It is important to resolve how widespread ionizing radiation sensitivity is ’ 2006 The British Institute of
amongst XP patients. Radiology

Xeroderma pigmentosum (XP) is a rare autosomal
recessive disease [1] characterized by clinical and cellular
sensitivity to ultraviolet (UV) radiation. The patients
show hypersensitivity to sunlight and extreme suscepti-
bility to sunlight-induced skin cancer [2, 3]. The majority
of cases are defective in one of seven genetically distinct
nucleotide excision repair genes (complementation
groups A–G [3]) which confer increased cellular suscepti-
bility to UV-induced killing and mutation [4–7]. In
addition, cells are also hypersensitive to agents which
generate bulky lesions in DNA [8]. Approximately 20%
of cases are competent in excision repair, but defective
in daughter strand repair [9] by virtue of a defect in a
Current address for Dr P B Rogers: Royal Berkshire Hospital,
London Road, Reading, Berkshire RG1 5AN, UK.
DNA polymerase (g) [10]. Cells from such patients are
minimally sensitive to the lethal effects of UV [11], but
are hypermutable [12, 13].
Since ionizing radiation mainly produces single and
double DNA strand breaks [14, 15] and types of base
damage repaired by the alternative pathway of base
excision repair [16, 17], hypersensitivity to ionizing
radiation would not be anticipated as a feature in the
XP syndrome. However, we have reported cellular
hypersensitivity to the lethal effects of gamma radiation
in fibroblasts from XP complementation group G, XP3BR
[18] derived from a patient who never came to radio-
therapy. There are two reports of patients from com-
plementation group C who have been exposed to
radiotherapy. DiGiovanna et al [19] recorded the
uneventful radiotherapy of such a patient with an

510 The British Journal of Radiology, June 2006

Ionizing radiation hypersensitivity in xeroderma pigmentosum
inoperable spinal cord astrocytoma, no assessment of
cellular sensitivity to either UVC or ionizing radiation
was provided (fibroblast cell line 5 XP23BE). In the
second example [20], the patient died 3 months after a
course of combined radiotherapy and chemotherapy
for a thalmic glioma. Again, no assessment of cellular
sensitivity was reported (fibroblast cell line 5 XP233VA).
We detail here radiosensitivity at both the clinical and
cellular level in an XP patient assigned to genetic
complementation group C. The homozygous mutation
in the XPC gene generates a stop codon at codon 718,
resulting in a truncated protein missing the C-terminal
233 of the 940 amino acids. To date, this patient is unique
in carrying this mutation (see XP mutation database:
URL http://xpmutations.org). The cell lines XP23BE and
XP233VA acted as reference controls. Earlier published
observations reported the clinical hypersensitivity in this
patient as one of a series of four sensitive individuals
encountered amongst 2000 paediatric radiotherapy
patients during a 20 year period at St Bartholomew’s
Hospital [21]. Discrimination from ataxia-telangiectasia
(A-T) was achieved on the basis of clinical criteria and by
demonstrating a normal level of radiation resistant DNA
synthesis.
We report here details of the pathology consequent to
radiotherapy, further characterization of the cell line as
XP, DNA repair studies following ionizing radiation and
evidence for the existence of a separate radiosensitivity
genetic defect.
Materials and methods
The patient
A clinical description of the patient, a female of
Pakistani origin, is provided in Salob et al [22] together
with details of the significant reduction to 9% in
nucleotide excision repair in fibroblasts (designated 86/
0029) and assignment to complementation group C. She
was considered unusual in presenting with both XP
and hypoplastic anaemia. At the age of 14 years, she
developed an angiosarcoma of the scalp overlying
the right parieto-occipital bone, a feature thought to be
unique in XP [23]. Previous skin (epithelial) tumours
had been treated with excision or with 5-fluorouracil
cream.
Table 1. Designations of the cell lines compared in this investigation
Cell line
XP14BR.1 (586/0029), XP14BR.2
XP14BRneo17
1BR.3, 142BR, 149BR, 250BR, 251BR, GF11, MRC5
1BR.3-G, 1BR.3neo, 142BRneo, MRC5V1, GM0637
XP23BE, XP233VI, XP15BR
AT1BR, AT2BR
T921A, T943A, T941B, T948B
XP, xeroderma pigmentosum; A-T, ataxia-telangiectasia.
The British Journal of Radiology, June 2006
Cells
Untransformed fibroblast cell lines from reference
normal control donors, 1BR.3, 142BR, 149BR, 250BR
and 251BR, were established in Brighton using standard
procedures [24]. The normal cell line used at Sutton was
GF11. Fibroblasts of the patient as used in the original
complementation assignment [22] designated 86/0029
(5XP14BR.1), were obtained from Dr P Botcherby (Guy’s
Hospital, London, UK). A second biopsy taken from an
uninvolved region of skin during radiotherapy at St
Bartholomew’s Hospital generated a second cell line,
XP14BR.2, in Brighton. Reference XP complementation
group C fibroblast cultures from the studies noted above
were obtained from Dr K Kraemer, NIH, Bethesda, MD
(XP23BE) [19] and Dr A Sarasin, Villejuif, France
(XP233VI) [20]. XP15BR is a reference XP cell line,
generated in Brighton, from complementation group A
(N Jaspers, Rotterdam, personal communication) cell line.
Two A-T cell lines AT1BR and AT2BR [25] were used for
scaling purposes. All these cells were maintained in
culture using standard materials and procedures [11, 24]
and were verified to be free of mycoplasma infection. All
biopsies were obtained under the approved ethical
procedures in place in the relevant institution at the time.
Fibroblast cell lines were transformed using the
immortalizing vectors pSV3gpt and pSV3neo in
Brighton in a programme using standard procedures
[26]. Both pre-crisis (not immortal) and post-crisis
(immortal) cultures of single cell origin were available
for investigation. Reference normal material was gene-
rated using this vector from the cell lines 1BR.3 (51BR.3-
G, from pSV3gpt and 1BR.3neo from pSV3neo), and
142BR (5142BRneo, from pSV3neo) in addition to
XP14BR (5XP14BRneo17, from pSV3neo). Immortalized
control cell lines MRC5V1 (from Dr P Debenham, former
MRC Radiobiology Unit, Harwell, UK) and GM0637
(from NIGMS, Camden, NJ, USA) were also employed.
The untransformed fibroblast cell line MRC5 was
obtained from C Babbs, Huntingdon Research.
Table 1 summarizes the designations of the strains
used.
Cell killing
Details of sources and dosimetry for UVC and a Cobalt
60 gamma ray facility, together with the design of cell
Properties
Fibroblast lines derived from separate biopsies of the subject of
this investigation
SV40-transformed derivative of XP14BR.2
Reference normal fibroblast lines
Reference SV40-transformed normal fibroblast lines
Reference XP fibroblast lines
Reference A-T fibroblast lines
Derivatives of XP14BRneo17, transfected for the XPC gene.
Clones T921A and T943A expressed significant XPC protein,
and T941B a low amount
511

killing experiments, have been described in detail else-
where [24, 27]. The response to the cross-linking agent
nitrogen mustard (Sigma Chemical Co., Poole, UK) was
verified using standard protocols [28].
DNA repair
We have used two methods for the evaluation of repair
following ionizing radiation. In the first, double strand
break (DSB) induction and repair were evaluated at the
Institute of Cancer Research, Sutton, using pulsed field
gel electrophoresis (PFGE) [29–31], with a repair time of
4 h. A separate series of experiments evaluating DSB
induction and repair was performed at Institut Gustave-
Roussy, Villejuif, France. The experimental procedures
are described in Foray et al [32]. A Caesium 137 radiation
source was used for the Villejuif experiments. Here, the
maximum repair time was 24 h. The control cell line was
149BR.
A second method, based upon a process of reduction
in the frequency of excess chromosome fragments [33],
was also undertaken at Villejuif. The premature chromo-
some condensation (PCC) technique permits the mea-
surement of chromosome breaks in cells in G0/G1. Cells
were irradiated in plateau phase of growth (6 Gy, c-ray
Figure 1. The irradiated scalp portal. The irradiated skin
remained depilated and became frankly ulcerated and
necrotic in the operation wound site.
512
C F Arlett, P N Plowman, P B Rogers et al
137Cs, 4 ˚C) and were then returned to 37 ˚C for repair.
Chromatin condensation was generated by fusion of
fibroblasts in plateau phase of growth with synchronized
mitotic cells by using PEG 6000, as described previously
[33], but with the notable exception that CHO cells were
replaced by HeLa cells. After hypotonic treatment, cells
were fixed in methanol:acetic acid and stained with
Giemsa. At least 30 fusions were analysed by means of
light microscopy. PCC data were expressed in chromo-
some fragments in excess (ECF) (total number PCC
fragments minus 46). The reference normal cell line was
149BR.
Transfection with the XPC gene
The SV40-transformed XP14BRneo17 cell line was
transfected with plasmid containing either the full-length
XPC cDNA, generously supplied by P van der Spek
(Rotterdam), or a clone truncated at the 5’ end of the
open reading frame, which nevertheless can complement
the UV sensitivity of an XPC cell line [34]. Each plasmid
was cotransfected with pSVgpt using calcium phosphate
precipitation and selection was applied for the co-
transfecting gpt marker using medium containing
mycophenolic acid, aminopterin and xanthine. When
colonies became visible to the naked eye, they were
picked individually using cloning rings and expanded
maintaining the selection for the gpt marker.
XPC protein measurements
Pellets of different cells were dissolved directly in
Laemmli buffer and 15 mg protein run on 7.5% SDS-
PAGE gels, followed by immunoblotting. The blots were
probed with a polyclonal antibody raised against a
peptide containing the C-terminal 18 amino acids of the
XPC protein [35] and developed with the ECL kit
(Amersham).
Results
Radiation pathology
The angiosarcoma of the scalp was excised and the
patient was referred for radiotherapy because of the
closeness of the marginal excision. 6 MEV electrons were
employed, which have a depth of penetration of
approximately 2 cm in soft tissue and less in bone. Due
to the characteristics of the electron beam, any further
depth would be fast diminishing.
After 38 Gy radiation applied in 19 fractions, the child
developed severe moist desquamation in the region in
the scalp (Figure 1). Ulcerated and underlying necrotic
bone at the depth of the ulcer was revealed and this
never healed. This reaction is very far outside the normal
expected range of response to this clinical radiotherapy
dose prescription.
2 years following radiotherapy, the patient suffered a
Grand Mal convulsion, and a CT scan of the brain
showed extensive cerebral oedema arising from an
oedematous focus directly beneath the radiation portal
The British Journal of Radiology, June 2006
Ionizing radiation hypersensitivity in xeroderma pigmentosum
Figure 2. CT brain scan demonstrating widespread hemi-
spheric oedema beneath the irradiated portal. Noteworthy is
the widespread extent of the oedema throughout the
hemisphere, not withstanding the restricted portal size, the
moderate prescription dose (38 Gy) and the low energy
electron portal (6 MEV) – that had to penetrate bone first. It
was this unrelenting cerebral oedema, emanating from radio
necrosis of brain under the radiation portal, that caused
death.
(Figure 2). However, the area of the oedema continued to
spread through most of the hemisphere over the next few
months. Steroids were employed and taken for a period
Figure 3. Ultraviolet radiation below 280 nm (UVC) clonal
survival of XP14BR fibroblasts. Key: XP14BR.1 (biopsy
1); XP14BR.2 (biopsy 2, mean ¡ standard error of 4
experiments). Normal fibroblast lines: 1BR.3 (mean
¡ standard error of 3 experiments), 142BR,
250BR, 251BR. Additional XP cell lines:
XP15BR, XP23BE. Single experiments except
where indicated.
The British Journal of Radiology, June 2006
of 22 months, but the child died from raised intercranial
pressure and extending left hemisphere oedema 4 years
after the radiotherapy. A diagnosis of ‘‘spreading
cerebral necrosis, consequent upon the radiotherapy,’’
was made.
Cell killing
The responses of XP14BR and a set of control and
reference fibroblast cell lines to the lethal effects of UVC
are illustrated in Figure 3. As anticipated XP14BR.2
fibroblasts and the earlier derived replicate culture,
XP14BR.1 (86/0029) proved to be hypersensitive.
Hypersensitvity was, as expected, also demonstrated
with the reference cell lines XP23BE and XP15BR used
for scaling purposes. With gamma irradiation, both
versions of XP14BR were indistinguishable in their
response and proved to be almost as sensitive as the
two reference A-T cell lines (AT1BR and AT2BR). Two
other cell lines derived from representatives of comple-
mentation group C (XP23BE and XP233VI), where no
clinical sensitivity had been reported, were normal in
their responses (Figure 4) as was XP15BR. With mito-
mycin C, no distinction in response between XP14BR.2
and the normal cell line 1BR.3 was observed (data not
shown). For nitrogen mustard, a comparison between the
immortalized XP14BRneo17 and the control lines
Figure 4. Ionizing radiation clonal survival of XP14BR
fibroblasts. Key as for Figure 3: Additional XP line;
XP233VA; Ataxia telangiectasia fibroblast lines;
AT1BR; AT2BR. Means ¡ standard error based
on numbers of experiments as follows: XP14BR.1, 1;
XP14BR.2, 6; 1BR.3, 58; 142BR, 8; 250BR, 4; 251BR, 13;
XP23BE, 2; XP233VA, 2; AT1BR, 5; AT2BR, 3.
513

Figure 5. Nitrogen mustard sensitivity of SV40 transformed
XP14BR fibroblasts. Key: XP14BRneo17; Normal trans-
formed fibroblast lines; MRC5V1; GM0637;
1BR.3-G. Means ¡ standard error of 2–4 experiments.
MRC5V1, GM0637 and 1BR.3-G (Figure 5) suggested no
increase in sensitivity in XP14BRneo17.
DNA Repair
DSB were measured using PFGE. Immediately after
irradiation there was a slight, but not significant,
reduction in the level of induced damage in XP14BR in
comparison with the control, GF11, over the dose range
of 0–40 Gy. There was no evidence for a difference
between the two cell lines after a 4 h period for repair
(data not shown).
When, however, a time course of repair was generated
(Figure 6) DSB repair in XP14BR.2 was more rapid at
earlier times than in control cells. A similar pattern has
been reported for A-T cells. At later times, however, the
residual levels of DSB were greater than in normal cells,
but less than in A-T cells. As a second measure of repair,
the time course of repair of chromosomal damage, as
reflected in the reduction of excess chromosome frag-
ments, revealed a small, but consistent, deficiency in
XP14BR.2 (Figure 7).
Transfection with the XPC gene
A series of single cell derived clones were established
from the transfection of XP14BRneo17 with XPC cDNA.
XPC protein was analysed by immunoblotting with an
antibody to the C-terminus of the XPC protein. The
XP14BR cell line has a mutation which truncates this
protein [36], so that it lacks the C-terminus against which
514
C F Arlett, P N Plowman, P B Rogers et al
Figure 6. DNA double strand break (DSB) repair rate after
30 Gy delivered at 4 ˚C. Repair data are expressed as the
fraction of radiation-induced DNA fragments (FAR) migrat-
ing in the gel that remained at the indicated repair times.
Each point represents the mean ¡ standard error of at least
three replicate experiments. Key: XP14BR.2;
149BR, Normal non-transformed fibroblast line.
the antibody was raised. As expected, no XPC protein
was detected in the XP14BRneo17 recipient cell line
(Figure 8, lane 3). However, the expected 125 kDa band
of XPC protein was restored to the level present in
MRC5V1 (lane 4) in clones T921A (lane 1) and T943A
(lane 2), but not in T941B or T948B (lanes 5 and 6). No
protein band was visible in the T941B or T948B
transfectants, using a low exposure of the X-ray film.
Overexposure of the film revealed a very low level of C-
terminal XPC protein in transfectant T941B, but not in
T948B (Figure 8, lower panel, lane 5).
Cell killing in transformed and transfected cells
In response to UVC, there was a significant enhance-
ment of survival of T921A and T943A (the clones
expressing substantial amounts of XPC protein) above
XP14BRneo17. However, a similar level of restoration of
UV resistance was also seen in clone T941B, which only
expressed low levels of XPC protein (Figure 8). The UVC
sensitivity recorded for these three clones was inter-
mediate between XP14BRneo17 and GM0637 and
MRC5V1 (Figure 9), the concurrent control normal cell
lines.
The response of XP14BR.2, XP14BRneo17, three nor-
mal cell lines, 1BR.3, 142BR and MRC5 and their
immortalized derivatives; 1BR.3neo, 142BRneo and
MRC5V1 to graded doses of gamma irradiation is
illustrated in Figure 10a. The enhancement of survival
The British Journal of Radiology, June 2006
Ionizing radiation hypersensitivity in xeroderma pigmentosum
Figure 7. Rate of repair of chromosome breaks after 6 Gy
delivered at 4 ˚C. Repair data are expressed as percentage of
premature chromosome condensation fragments in excess
(ECF) at the repair times indicated. Each point represents the
mean ¡ standard error of at least three replicate experi-
ments. Key: XP14BR.2; 149BR, Normal non-
transformed fibroblast line.
after ionizing radiation as a consequence of transformation
was first described by Arlett et al [37]. This is particularly
evident in XP14BR, with the consequence that the level of
discrimination between it and the reference normals is less
than in the corresponding untransformed cell lines.
However, unlike the case with UVC, after exposure to
gamma radiation no enhancement of survival above the
Figure 8. Immunoblotting for the C-terminal end of the XPC
protein. Lanes 1 and 2, T921A, T943A (transfectants); lane 3,
XP14BRneo17 (recipient); lane 4, MRC5V1 (control); lanes 5
and 6, T941B and T948B (transfectants). Upper panel, 20 s
exposure, lower panel, 5 min exposure.
The British Journal of Radiology, June 2006
Figure 9. Ultraviolet radiation below 280 nm (UVC) sensi-
tivity of XP14BRneo17 and its transfectants. Means ¡
standard error of the number of experiments indicated.
Key: XP14BRneo17 (non-transfected, 1 expt);
T921A (2 expts); T943A (4 expts); T941B
(2 expts) (all transfectants); MRC5V1 (1 expt);
GM0637 (2 expts).
parental cell line was observed. Indeed, all four clones
appeared slightly more sensitive (Figure 10b).
Discussion
Our patient was unusual amongst XPs in presenting
with hypoplastic anaemia, and was originally thought to
be a case of Fanconi’s anaemia [22]. Her angiosarcoma
was also considered to be unique [23] in association with
XP. The clinical radiosensitivity and consequent patho-
logy which followed treatment of this scapular lesion is
peculiar to this individual [21].
At the cellular level, the response to UVC is unexcep-
tional. However, the hypersensitivity to gamma radia-
tion has, in our hands, only been demonstrated in one
previous instance using fibroblasts (XP3BR) from an
individual from complementation group G [18]. The lack
of clinical sensitivity in the two other representatives of
complementation group C, XP23BE and XP233VI, which
act as reference controls, was reflected in their normal
responses in our cellular assays (Figure 4). The cellular
sensitivity of XP14BR to ionizing radiation may be
contrasted to the lack of sensitivity to cross-linking
agents and implies a cellular defect in repair specific to
gamma radiation.
The study of repair of DSB in XP14BR generates results
reminiscent of those obtained with A-T cells; an early,
more rapid repair than in normals culminating at 24 h
515

with less repair [32]. When combined with the outcome
of the study of repair of chromosome fragments, these
results are consistent with a defect in the repair of a small
fraction of DSB as has been found for A-T cells [32, 38].
We have already acknowledged that the cellular
radiosensitivity in XP14BR may be discriminated from
that manifest in A-T cells by the normal level of radiation
resistant DNA synthesis [21]. This alone is suggestive of
the possibility of the presence of another non-A-T
radiosensitivity gene, but first it was necessary to verify
that the XPC mutation [36] itself was not responsible for
the hypersensitivity.
The correction, by transfection with the XPC gene, of
the XP genetic defect was, as expected, successful in
restoring cellular resistance to UVC irradiation. In one
transfectant, T941B, substantial resistance to UVC irra-
diation was restored, even though only very low levels of
XPC protein were detected in the transfectant. This
suggests that XPC protein is not rate-limiting for
neuclotide excision repair and that this low level is
sufficient to deal with significant levels of UV damage.
The psv3neo immortalization of XP14BR itself pro-
duced a large enhancement of ionizing radiation
resistance. This effect of transformation, reported earlier
[37], is specific to ionizing radiation and not UV, and
may be generated by the abrogation of functional p53.
However, the failure to enhance ionizing radiation
resistance in clones, which, in addition to showing an
enhancement of their UV response are also positive for
the UVC gene product, supports the contention that
XP14BR is a double mutant combining defects in both
XPC and an unknown radiosensitivity gene. The resolu-
tion of this proposed genetic defect and the identification
of the gene(s) becomes a priority.
The other unusual clinical features exhibited by the
patient remain unexplained. There was no evidence for
more specific defects in immunity.
The present demonstration of both clinical and cellular
radiosensitivity in XP14BR raises the question of the
frequency of such radiosensitivity in this syndrome.
While XP14BR may be unique in being a double mutant,
we have reported a previous instance of cellular
sensitivity in XP3BR [18]. There was no relevant radio-
therapy in this case. In this report we provide data
516
C F Arlett, P N Plowman, P B Rogers et al
Figure 10. Ionizing radiation sensi-
tivity of XP14BR, XP14BRneo17, and
transfectants. Means ¡ standard
error of the number of experiments
indicated. A. Effect of SV40 trans-
formation on radiosensitivity.
Key: XP14BR (6 expts);
XP14BRneo17 (1 expt);
1BR.3 (58 expts);
1BR.3neo (4 expts);
142BR (8 expts);
142BRneo (3 expts);
M R C 5 ( 1 e x p t ) ;
MRC5V1 (2 expts). B.
Comparison of transfectants with
X P 1 4 B R n e o 1 7 . K e y :
XP14BRneo17 (1 expt);
T921A (2 expts); T943A
(3 expts); T941B (2 expts);
T948B (1 expt).
showing normal cellular responses for XP15BR, from
complementation group A and XP23BE and XP233VI
from complementation group C. The record of the
outcome of radiotherapy in the patient who generated
the cell line XP23BE is unambiguous in excluding any
clinical sensitivity [19]. The clinical information of the
outcome of combined chemotherapy and radiotherapy for
the second XPC patient is not so clear cut since the child
died 3 months after treatment [20]. In the absence of any
cellular radiosensitivity there remains the possibility of
sensitivity to components of the chemotherapy regimen.
As far as we are aware, the patient from whom cell line
XP15BR was generated has never come to radiotherapy. In
the light of the outcome of radiotherapy in our present
case and the observation of cellular radiosensitivity in
XP3BR, it becomes important to resolve the extent of
radiosensitivity in XP. This is reinforced by the possibility
that these patients, because of their high frequency of
tumours, are more likely to experience radiation.
Acknowledgments
We are grateful to Randy Legerski and Peter van der
Spek for the XPC cDNA plasmids, and to Peter van der
Spek for the anti-XPC antibody. NF was supported by an
INSERM Interface grant, Electricité de France, Ligue
Nationale Contre le Cancer and ARC fundations. CFA,
ARL and MHLG were supported in part by Euratom
grant B16-E1042-UK.
References
1. Bootsma D, Kraemer KH, Cleaver JE, Hoeijmakers JHJ.
Nucleotide excision repair syndromes: xeroderma pigmen-
tosum, Cockayne Syndrome, and trichothiodystrophy. In:
Volgelstein B, Kinzler KW, editors. The genetic basis of
human cancer. New York, NY: McGraw-Hill, 1998:245–74.
2. Kraemer KH, Lee MM, Scotto J. Xeroderma pigmentosum.
Cutaneous, ocular and neurologic abnormalities in 830
published cases. Arch Dermatol 1987;123:241–50.
3. Arlett CF, Lehmann AR. Xeroderma pigmentosum,
Cockayne syndrome and trichothiodystrophy: sun sensiti-
vity, DNA repair defects and skin cancer. In: Eeles R, Ponder
B, Easton DF, Horwich EA, editors. Genetic predisposition to
cancer, 2nd edn. London: Arnold, 2004:214–31.
The British Journal of Radiology, June 2006
Ionizing radiation hypersensitivity in xeroderma pigmentosum
4. Gartler SM. Inborn errors of metabolism at the cell culture
level. In: Fishbein M, editor. Second International
Conference on Congenital Malformations. New York, NY:
International Medical Congress, 1964:94.
5. Cleaver JE. Deficiency in repair replication of DNA in
xeroderma pigmentosum. Nature 1968;218:652–6.
6. Maher VM, McCormick JJ. Effect of DNA repair on the
cytotoxicity and mutagenicity of UV irradiation and of
chemical carcinogens in normal and xeroderma pigmento-
sum cells. In: Yuhas JM, Tennant RW, Regan JB, editors.
Biology of radiation carcinogenesis. New York, NY: Raven
Press, 1976:129–45.
7. Arlett CF, Harcourt SA. Variation in response to mutagens
amongst normal and repair-defective human cells. In:
Lawrence CW, editor. Induced mutagenesis. Molecular
mechanisms and their implications for environmental
protection. New York, NY: Plenum Press, 1983:249–66.
8. Maher VM, McCormick JJ, Grover P, Sims P. Effect of DNA
repair on the cytotoxicity and mutagenicity of polycyclic
hydrocarbon derivatives in normal and xeroderma pig-
mentosum human fibroblasts. Mutation Res 1977;43:117–38.
9. Lehmann AR, Kirk-Bell S, Arlett CF, Paterson MC, Lohman
PHM, de Weerd-Kastelein EA, et al. Xeroderma pigmento-
sum cells with normal levels of excision repair have a defect
in DNA synthesis after UV-irradiation. Proc Natl Acad Sci
USA 1975;72:219–23.
10. Masutani C, Kusumoto R, Yamada A, Dohmae N, Yokoi M,
Yuasa M, et al. The XPV (xeroderma pigmentosum variant)
gene encodes human DNA polymerase eta. Nature
1999;399:700–4.
11. Arlett CF, Harcourt SA, Broughton BC. The influence of
caffeine on cell survival in excision-proficient and excision-
deficient xeroderma pigmentosum and normal human cell
strains following ultraviolet light irradiation. Mutation Res
1975;33:341–6.
12. Maher VM, Ouellette LM, Curren RD, McCormick JJ.
Frequency of ultraviolet light-induced mutations is higher
in xeroderma pigmentosum variant cells than in normal
human cells. Nature 1976;261:593–5.
13. Myhr BC, Turnbull D, DiPaolo JA. Ultraviolet mutagenesis
of normal and xeroderma pigmentosum variant human
fibroblasts. Mutation Res 1979;62:341–53.
14. Ward JF. DNA damage and repair. In: Glass WA, Varma
MN, editors. Physical and chemical mechanisms in mole-
cular radiation biology. New York, NY: Plenum Press,
1991:403–21.
15. Goodhead DT. Initial events in the cellular effects of
ionising radiations: clustered damage in DNA. Int J
Radiat Biol 1994;65:7–17.
16. Tchou J, Kasai S, Shibutani M-H, Chung J, Laval J, Grollman
AP, Nishimura S. 8-Oxoguanine (8-hydroxyguanine) DNA
glycosylase and its substrate specificity. Proc Natl Acad Sci
USA 1991;88:4690–4.
17. Ward JF. DNA damage produced by ionizing radiation in
mammalian cells: identities, mechanisms of formation and
repairability. Progress in Nucleic Acid Research
1988;35:95–125.
18. Arlett CF, Harcourt SA, Lehmann AR, Stevens S, Ferguson-
Smith MA, Morley WN. Studies on a new case of
xeroderma pigmentosum (XP3BR) from complementation
group G with cellular sensitivity to ionizing radiation.
Carcinogenesis 1980;1:745–51.
19. DiGiovanna JJ, Patronas N, Katz D, Abangan D, Kraemer
KH. Xeroderma pigmentosum: spinal cord astrocytoma
with 9 year survival after radiation and isotretinoin
therapy. J Cutaneous Med Surg 1998;2:153–8.
20. Giglia G, Bouffet E, Jouvet A, Ohgaki H, Kleihues P, Sarasin
A. Molecular analysis of glioma and skin-tumour altera-
tions in a xeroderma-pigmentsum child. Int J Cancer
1999;81:345–50.
The British Journal of Radiology, June 2006
21. Rogers PB, Plowman PN, Harris SJ, Arlett CF. Four
radiation hypersensitivity cases and their implications for
clinical radiotherapy. Radiother Oncol 2000;57:143–54.
22. Salob SP, Webb DKH, Atherton DJ. A child with xeroderma
pigmentosum and bone marrow failure. Br J Dermatol
1992;126:372–4.
23. Leake J, Sheehan MP, Rampling D, Ramani P, Atherton DJ.
Angiosarcoma complicating xeroderma pigmentosum.
Histopathology 1992;21:179–81.
24. Arlett CF, Harcourt SA. Survey of radiosensitivity in a
variety of human cell strains. Cancer Res 1980;40:926–32.
25. Henderson L, Cole H, Arlett CF, James SE, Cole J, Lehmann
AR, et al. Diagnosis of ataxia-telangiectasia by T-lympho-
cyte cloning assay. Lancet 1985;ii:1242.
26. Mayne LV, Priestley A, James MR, Burke JF. Efficient
immortalization and morphological transformation of
human fibroblasts by transfection with SV40 DNA linked
to a dominant marker. Exptl Cell Res 1986;162:530–8.
27. Arlett CF, Harcourt SA, Cole J, Green MHL, Anstey AV. A
comparison of the response of unstimulated and stimulated
T-lymphocytes and fibroblasts from normal, xeroderma
pigmentosum and trichothiodystrophy donors to the lethal
action of UV-C. Mutation Res 1992;273:127–35.
28. Teo IA, Arlett CF. The response of a variety of human
fibroblast cell strains to the lethal effects of alkylating
agents. Carcinogenesis 1982;3:33–7.
29. Blöcher D, Einspenner M, Zajackowski J. CHEF electro-
phoresis, a sensitive technique for the determination of
DNA double-strand breaks. Int J Radiat Biol 1989;56:437–48.
30. Whitaker SJ, McMillan TJ. Oxygen effects for DNA double-
strand breaks determined by pulse field gel electrophoresis.
Int J Radiat Biol 1992;61:29–41.
31. Whitaker SJ, McMillan TJ. Pulsed-field gel electrophoresis
in the measurement of DNA double-strand break repair in
xrs-6 and CHO cell lines: DNA degradation under some
conditions interferes with the assessment of double-strand
break rejoining. Radiat Res 1992;130:389–92.
32. Foray N, Priestley A, Alsbeih G, Badie C, Capulas EP, Arlett
CF, et al. Hypersensitivity of ataxia-telangiectasia fibro-
blasts to ionizing radiation is associated with a repair
deficiency of DNA double-strand breaks. Int J Radiat Biol
1997;72:271–83.
33. Badie C, Iliakis N, Foray N, Alsbeih G, Cedervall B,
Chavaudra N, et al. Induction and rejoining of DNA
double strand breaks and interphase chromosome breaks
after exposure to x-rays in one normal and two hypersensi-
tive human fibroblast cell lines. Radiat Res 1995;144:26–35.
34. Legerski R, Peterson C. Expression cloning of a human
DNA repair gene involved in xeroderma pigmentosum
group C. Nature 1992;359:70–3.
35. van der Spek PJ, Eker A, Rademakers S, Visser C, Sugasawa
K, Masutani C, et al. XPC and human homologs of RAD23:
intracellular localization and relationship to other nucleo-
tide excision repair complexes. Nucleic Acids Res
1996;24:2551–9.
36. Chavanne F, Broughton BC, Pietra D, Nardo T, Browitt A,
Lehmann AR, et al. Mutations in the XPC gene in families
with xeroderma pigmentosum and consequences at the cell,
protein and transcript levels. Cancer Res 2000;60:1974–82.
37. Arlett CF, Green MHL, Priestley A, Harcourt SA, Mayne
LV. Comparative human cellular radiosensitivity: I. The
effect of SV40 immortalisation on the gamma-irradiation
survival of skin derived fibroblasts from normal indivi-
duals and from ataxia-telangiectasia patients and hetero-
zygotes. Int J Radiat Biol 1988;54:911–28.
38. Riballo E, Kuhne M, Rief N, Doherty AJ, Smith GCM, Recio
M-J, et al. A pathway of double strand break rejoining
dependent upon ATM, Artemis and proteins locating to c-
H2AX foci. Mol Cell 2004;16:715–24.
517
 The British Journal of Radiology, 79 (2006), 518–521

SHORT COMMUNICATION

Survival and initial chromatid breakage in normal and tumour

cells exposed in vitro to gamma rays and carbon ions at the HIRFL
1,2
Y JIANSHE, PhD, MSc, 1L WENJIAN, PhD, 1J XIAODONG, MSc,
1
J XIGANG, PhD,
1
G CHUANLING, BSc,
1
W WEI, PhD and 3G QINGXIANG, BSc

1
Institute of Modern Physics, Chinese Academy of Sciences, 2Life Science School of North West
Normal University and 3Life Science School of Lanzhou University, Lanzhou, China

ABSTRACT. Human hepatoma and normal liver cells were irradiated with 12C6+ ion
beams (linear energy transfer (LET)596 keV mm21) and c-rays at the Heavy Ion Research
Facility in Lanzhou (HIRFL). The numbers and types of chromatid breaks were detected
using the premature chromosome condensation technique. Irradiation with 12C6+ ions
Received 9 June 2005
produced a majority of isochromatid break types, while chromatid breaks were Revised 9 November 2005
dominant for irradiation with c-rays. Experimental results showed that the initial level Accepted 15 November
of chromatid breaks is clearly related to the absorbed dose from 12C6+ ions and c-rays. 2005
The 12C6+ ions are relatively more effective at inducing initial chromatid breaks when
DOI: 10.1259/bjr/99734289
compared with the c-rays. A relative biological effectiveness (RBE) of about 2.5 resulted
for the induction of initial chromatid breaks by 12C6+ ions relative to c-rays in both cell ’ 2006 The British Institute of
lines. Radiology

Heavy charged particle beams have been applied in
tumour radiotherapy because of their clear radiobio-
logical response in the destruction of malignant cells
[1]. The induction of chromatid breaks in the cell by
radiation is considered to be a major cause of damage,
which can lead to cell death. It has been reported by
Kawata et al [2] that low linear energy transfer (LET)
radiation induced mostly chromatid breaks, whereas the
most isochromatid breaks were generated by high LET
radiation. Previous studies [3] by the present authors
have shown that the chromatid break repaired more
easily than the isochromatid break. A major project
involving tumour radiotherapy using heavy ions is
currently being conducted at the Heavy Ion Research
Facility in Lanzhou (HIRFL). In the present communi-
cation, the use of the premature chromosome conden-
sation technique to measure the level of G2 initial
chromatid breaks of human hepatoma cells and normal
liver cells is reported.
It is a well known fact that the risk of liver cancer is
high in Asia, especially in China. For this reason, one
normal and one tumour human liver cell line were
selected for exposure to carbon ions generated by the
HIRFL. The main aim was to investigate and quantify the
relationship between the radiation dose and the level of
initial chromatid breakage. Such information is needed
for predicting the radiosensitivity of these two liver cell
lines, and for evaluating the relative biological effective-
ness (RBE) of carbon ions for inducing the chromatid
breaks. The experiments provided necessary and valu-
able data on liver cell radiosensitivity prior to the start of
clinical radiotherapy. It was found that the two cell lines
have moderate radiosensitivity when exposed to c-rays
[3].
Materials and methods
Cell culture and irradiation
The human hepatoma cell line SMMC-7721 and the
normal liver cell line L02 (obtained from the Chinese
Center for Type Culture Collection, CCTCC) with
moderate radiosensitivity, were grown in RPMI-1640
medium supplemented with 10% fetal calf serum at 37 ˚C
in 5% CO2; additionally, 0.25 U ml21 of insulin was
added to the L02 culture medium.
Exponentially growing cells were irradiated with
either c-rays obtained from a 60Co source (dose rate
0.2 Gy min21) or 12C6+ ion beams generated by the
HIRFL facility with a dose range from 0 to 8 Gy. The
initial energy of the 12C6+ ions was 80.55 MeV u21, which
was attenuated by the 13.58 mm Lucite (r51.2 g cm23)
to 20 MeV u21 before it reached the cells. The LET was
96 keV mm21 when the carbon ions interacted with the
cells located in the region of the Bragg peak and the
LET was calculated by the Trim Program 92. Dosimetry
was performed with an air ionizing chamber and the
uniformity of the carbon ion beams was found to be 85%,
as measured with CR39 track etch plastic.
Colony assay
After exposure to radiation, cells were washed three
times using PBS (pH57.0), then trypsinized and put into
5 ml culture medium, and the cell density counted by
using the light microscope. At each dose point 200, 400,
or 600 cells were plated into 35 mm diameter culture
dishes. 5 ml RPMI-1640 medium supplemented with

518 The British Journal of Radiology, June 2006

Short communication: Survival and initial chromatid breakage

10% fetal calf serum was added and cells were incubated
at 37 ˚C in 5% CO2 until colonies formed (7–14 days). Two
dishes were plated per dose point and the experiments
were repeated three times. All the data are the mean¡
standard deviation. Data were fitted using Origin 7.0
(Original Lab, America); the data from the cells exposed
to c-rays were fitted using the non-linear fit program,
while the data from the cells exposed to carbon ions were
fitted using the linear
2
fit program. The linear-quadratic
formula (S~e aD bD ) was used for the non-linear fit,
where a and b represent the probability of a single or
double photon interaction with the chromatids.

Chromosome preparation
Calyculin-A (BIOMOL America) was used as the PCC
inducer, which was dissolved in 100% ethanol as a
1 mmol l21 stock solution. In order to achieve the
chromatid breaks, 50 nmol l21 of Calyculin-A was added
to cell cultures 5 min before irradiation. Cells were then
incubated for a further 30 min at 37 ˚C in 5% CO2.
Chromosome spread was harvested by swelling the
cells in 75 mmol l21 KCl for 20 min at 37 ˚C and fixed
with Carnoy’s fixation. A final wash and fixation in
the same fixative agent was completed before dropping
the cells onto a glass slide and drying them in hot
humidity.
The chromosome was stained with 5% Giemsa for
20 min. According to the standard criteria [4], more than
40 G2 phase cells were scored for each dose level
investigated. Briefly, a chromatid discontinuity or mis-
alignment of the distal part to the lesion, or a non-stained
region longer than the chromatid width was considered
to be a chromatid break. Isochromatid breaks were
considered to be two breaks occurring at the same
position on the two sister chromatids, i.e. a penetrated
lesion through the two q arms or p arms of the
chromosome was regarded to be an isochromatid break.
The total chromatid breaks were calculated by summing
the numbers of chromatid and isochromatid breaks. 20
non-irradiated cells were scored and the mean of these
chromatid breaks was recorded. There were very few
spontaneous chromatid breaks. The number of chroma-
tid breaks observed in the irradiated cells was reduced
by the mean number of chromatid breaks observed in the
non-irradiated cells, and this number forms the experi-
mental data from which the results were evaluated. A
small number of chromatid exchanges were also noticed,
but these were not considered to be significant and did
not have a linear relationship with absorbed dose.
Results
Survival fraction of two cell lines exposed to
gamma rays and carbon ions
Survival curves for SMMC7721 and L02 cells exposed
to c-rays and carbon ions are shown in Figure 1. Several
trends can be seen: (1) the two cell lines have different
survival curves for c-rays and carbon ion irradiation;
(2) the cells irradiated with c-rays have a linear quadratic
survival curve, while it is almost linear for carbon ion
Figure 1. Survival curve of SMMC7721 and L02 cells exposed
to gamma rays and carbon ions. Survival fraction value is
the mean¡standard deviation. The survival curve of cells
exposed to gamma rays was linear-quadratic, and was
almost linear when cells were exposed to carbon ions. The
survival fraction of SMMC7721 cells and L02 cells are sig-
nificantly different when exposed to either gamma rays or
carbon ions.
irradiation; (3) the surviving fraction of L02 cells was
higher than that of the SMMC7721 cells for both
exposure scenarios. For the cells exposed to c-rays, the
fit parameters a and b were 0.03 and 0.06 for SMMC7721
cells and 0.04 and 0.05 for L02 cells, respectively.
Initial chromatid and isochromatid breaks after
irradiation of G2-PCC
Figure 2 shows the chromatid and isochromatid break
frequency of hepatoma and liver cell lines after irradia-
tion with c-rays and 12C6+ ion beams at different
absorbed dose. For both cell lines, the G2 chromatid
breaks and the isochromatid breaks increase linearly
with the absorbed dose. The number of chromatid breaks
and isochromatid break frequency in the SMMC7721 cell
line was greater than that of the L02 cell line for each
dose. The yields of two types of chromatid breaks
irradiated with 12C6+ ion beams were more than that of
c-rays.
For both cell lines, the number of chromatid breaks
induced by the carbon ions were about 2.5 times higher
relative to the c-ray exposure. Hence, the biological
effectiveness of carbon ions for inducing the chromatid
breaks is about 2.5 times higher than the c-rays.
The proportion of chromatid breaks in two cell lines
exposed to c-rays and 12C6+ ions
Figure 3 shows the proportion of two types of
chromatid breaks of two cell lines exposed to c-rays
and 12C6+ ion beams. It can be seen in Figure 3 that
for 12C6+ and c-ray irradiation of cells, the dominant

The British Journal of Radiology, June 2006 519


Figure 2. Correlation between absorbed dose and chroma-
tid breaks. In both (a) L02 and (b) SMMC7721 cell lines, more
isochromatid than chromatid breaks were seen when cells
were exposed to carbon ions, and fewer isochromatid than
chromatid breaks seen when cells were exposed to gamma
rays. Fewer breaks were induced in L02 cells than in
SMMC7721 cells after either carbon ion or gamma ray
exposure. All the data were the mean¡standard deviation.
breakage types were isochromatid and chromatid,
respectively.
Discussion
Heavy ions have the advantage of destroying the
tumour cells very effectively. Kawata et al [2] reported
that the cell survival fraction of human fibroblast
AG1552 was very much lower after exposure to heavy
520
Y Jianshe, L Wenjian, J Xiaodong et al
Figure 3. Proportion of two types of chromatid breaks. For
cells irradiated with gamma rays, the number of chromatid
breaks in both cell lines was much more than that of
isochromatid breaks; for cells exposed to carbon ions, the
isochromatid breaks were much more frequent than the
chromatid breaks. They were significantly different.
ions. In the present study, it was observed that the
surviving fraction of L02 and SMMC7721 cells irradiated
with carbon ions were lower than for cells exposed to c-
rays. The survival curve for two cell types exposed to
carbon ions was linear while it was linear quadratic for c-
ray exposure.
After irradiation, G2 chromatid breaks and isochro-
matid breaks increased linearly with the absorbed dose
for both cell lines, the increased ratio of L02 and
SMMC7721 cell lines were 2.5 and 3.5 for chromatid
breaks and 4.5 and 7 for isochromatid breaks, which was
consistent with data from Kawata et al [2]. The yields of
the two types of chromosome breaks irradiated with
12 6+
C ion beams were more than that of c-rays. The
relative biological effectiveness (RBE) of 12C6+ ion beams
in inducing the chromatid breaks was found to be about
2.5. These results are in agreement with several previous
studies [2, 5, 6, 7].
The main type of chromatid break was isochromatid
for cells exposed to 12C6+ ion beams, while the chromatid
breaks predominated in cells exposed to c-rays. Kawata
et al [2] reported that chromatid breaks dominated after
low LET irradiation, while isochromatid breaks domi-
nated for high LET exposures, such as heavy ions,
suggesting that most isochromatid breaks resulted from
two separate breaks on sister chromatids induced by
independent electron tracks. For low LET radiation
types, not enough energy can be deposited to penetrate
sister chromatids simultaneously, so most breaks were
found to be chromatid in nature. For cells exposed to
heavy ions, more isochromatid breaks were induced.
This explains why heavy ions have the higher RBE,
which is relevant for the destruction of tumour cells.
Conclusion
Our results suggest that heavy ions generated by the
HIRFL have a higher efficiency in killing tumour cells,
The British Journal of Radiology, June 2006
Short communication: Survival and initial chromatid breakage
and a high RBE for the induction of the chromatid
breaks. Also, the chromatid breaks are tightly correlated
with the cell surviving fraction and radiosensitivity.
Acknowledgment
This work was supported jointly by National Natural
Science Foundation of China and National Basic
Research Program of China through fellowship
No. 10335050 and 2003CCB00200. We express our thanks
to all the workers in HIRFL, for providing us with the
high quality heavy ion beams.
References
1. Kraft G. Tumor therapy with heavy charged particles.
Progress in Particle Nuclear Physics 2000;45:s473–544.
The British Journal of Radiology, June 2006
2. Kawata T, Durante M, Frusawa Y, et al. Dose-response of
initial G2-chromatid breaks induced in normal human
fibroblasts by heavy ions. Int J Radiat Biol 2001;77:165–74.
3. Yang JS, Li WJ, Zhou GM, et al. A comparative study on
radiosensitivity of various tumor cells and human normal
liver cells. World J Gastroenterol 2005;11:4098–101.
4. Savage JR. Classification and relationships of induced
chromosomal structural changes. J Med Genet 1976;13:
103–22.
5. Kawata T, Gotoh E, Durante M, et al. High-LET radiation-
induced aberrations in prematurely condensed G2 chromo-
some of human fibroblasts. Int J Radiat Biol 2000;76:929–37.
6. Kawata T, Durante M, Furusawa Y, et al. Rejoining of
isochromatid breaks induced by heavy ions in G2-phase
normal human fibroblasts. Radiat Res 2001;156:598–602.
7. Kawata T, Ito H, Uno T, et al. G2 chromatid damage and
repair kinetics in normal human fibroblast cells exposed to
low- or high-LET radiation. Cytogenet Genome Res 2004;104:
211–5.
521
 The British Journal of Radiology, 79 (2006), 522–525

SHORT COMMUNICATION

Usefulness of mini-tracheostomy and torque controlled insertion

of applicator in fractionated endobronchial brachytherapy
1
K KISHI, MD, PhD, 2T YOSHIMASU, MD, PhD, 1S SHIRAI, MD, 3Y MINAKATA, MD, PhD,
1
M KIMURA, MD,
1 1 1
PhD, T SONOMURA, MD, PhD, Y SHIOYAMA and M SATO, MD, PhD

Departments of 1Radiology, 2Thoracic and Cardiovascular Surgery and 3Third Department of

Internal Medicine, Wakayama Medical University (WMU), Kimiidera 811-1, Wakayama City,
641-0012, Japan

ABSTRACT. Endobronchial brachytherapy was developed as effective treatment of

endobronchial cancer and fractionated schedule is applied to decrease late toxicity.
However, repeated bronchofiberscopy is onerous to the patient and restricts the
treatment schedule itself. We applied mini-tracheostomy for a ready access route, and a
torque controlled technique for easy insertion of the endobronchial applicator. Eight Received 18 May 2005
patients with tracheobronchial cancer invasion were treated with endobronchial Revised 14 January 2005
brachytherapy of 18–30 Gy/3–5 fractions/1.5–2.5 weeks (median 24 Gy/4 fractions/ Accepted 24 January 2006
2 weeks) at reference points of 5 mm from the bronchial surface. The averaged
DOI: 10.1259/bjr/31613651
individual irradiation and single session times were 4 min and 24 min, respectively.
There were no procedure-related complications. These technical improvements may ’ 2006 The British Institute of
facilitate flexible fractionated dose prescriptions. Radiology

Fractionated endobronchial brachytherapy was devel-
oped for both effective palliation of endobronchial cancer
invasion [1–3], and eradication of localized superficial
endobronchial cancers [4, 5]. This was performed with an
endobronchial applicator inserted into the target lumen
by a transoral or transnasal bronchofiberscopic approach
[1–3, 5–8].
A shortcoming of this technique is the need to repeat
the transoral or transnasal endobronchial approaches.
Irrespective of possible fractionation patterns to decrease
late radiation toxicity [9], a commonly used fractionation
schedule has been once weekly [2, 4, 5, 7, 8, 10, 11].
To reduce the burden of this therapy we introduced a
ready access procedure through a mini-tracheostomy,
created without cartilage damage and maintained using
a recently devised fast and less invasive device com-
posed of a cannula-introducer assembly, which was
originally devised for emergency use [12] and has been
applied for various purposes [13, 14]. Usually the mini-
stoma spontaneously closes in a few days. Additionally,
the style of the applicator was curved to enable a torque
control manoeuvre to insert it smoothly under fluoro-
scopy. The clinical feasibility and usefulness of these
procedures were evaluated clinically.
Patients
Between January 1999 and June 2002, eight patients
with inoperable advanced cancer growing into the
endobronchial space, which developed 2.5–103.6 (med-
ian 46.7) months after disease onset, were subjected to
this brachytherapy treatment (Table 1) in our depart-
ment of radiation oncology, where 60–80 patients a year
were treated (to #7). The cancer ingrowths were the only
clinical manifestation at the time of the referral. The
participants consisted of six males and two females aged
51–70 years (median 55.9 years). Karnofsky performance
status of each patient was 80% or more. There were three
tracheal cancers, three lung cancers, and one mediastinal
cancer and metastatic cancer each. All had previously
received external beam radiation therapy at doses of
50.4–70 Gy (median 60 Gy). Therapeutic goals of the
brachytherapy included symptomatic relief from bron-
chial obstruction in four and prevention of further
trachobronchial obstruction in all (Table 1). Patients
were first evaluated by bronchofiberscopy to determine
indications and target locations [11].
Informed consent was obtained prior to treatment. A
choice of two treatment protocols had been offered to the
patients: once a week 5 Gy brachytherapy for 4 weeks each
with bronchofiberscopy according to many other reports
[2, 4, 5, 7, 8, 10, 11], or the present protocol of twice a week
less than 2.5 weeks after creation of a mini-stoma.
Methods
Mini-tracheostomy
Under local anaesthesia, according to the general
insertion procedure [15], the cricothyroid membrane
was penetrated with a needle-guidewire set of the Mini-
Trach II system (Mini-Trach II; Portex Ltd, Kent, UK),

522 The British Journal of Radiology, June 2006

Short communication: Mini-tracheostomy for endotracheobronchial radiotherapy

(a) (b)

(c)

Figure 1. (a) A cannula of Mini-Trach II was inserted to maintain a mini-stoma. (b) At fractionated endotracheobronchial
radiotherapy, the cannula was temporarily withdrawn and the endobronchial or endotracheal mini-applicator was gently
inserted through the mini-stoma (arrow) using a curved stylet (long arrow) under fluoroscopic monitoring. The soft elastic wings
were stretched during insertion (shadowed arch indicated with a broken arrow). (c) Measuring wire with dots at 1 cm intervals
was inserted in the bronchial applicator, the wings of which were opened to support the applicator position (Note the different
opened width of the soft wings). Reference points were determined according to the vertical distance from the bronchial
surface (one reference point is shown by an asterisk, the depth: vertical line).

and then a cannula-introducer assembly was introduced
into the tracheal lumen with Seldinger’s procedure.
Thus, a mini-stoma was created and then maintained
with the cannula (Figure 1a).
Applicator deployment
Patients were pre-medicated with inhalation of 0.2 g
lidocaine via a nebulizer and intramuscular injection of
0.5 mg of atropine sulphate and 25 mg of hydroxyzine.
An endobronchial mini-applicator (Clinical Supply,
Tokyo, Japan) composed of a stylet, an elastic hard
polyethylene inner tube and a soft silicone outer tube
with self-expandable wings stretched by the inner tube
(Figure 1b) [5, 7, 8], was introduced into the trachea via
the stoma and inserted to the target lumen under
fluoroscopic guidance, paying attention not to stimulate
the tracheobronchical wall. The inner tube was curved to
enable torque manipulation to reach the lesion more
easily (Figure 1c) [16]. Then the inner tube was with-
drawn to allow the wings to open softly and hold the
catheter position in the centre of the lumen. Routine
lidocaine spray topical anaesthesia, which is an irritant,
was not necessary during the procedure because of the
smooth insertion.
Before each session, the cannula was temporarily
withdrawn, and after each session the cannula was
re-inserted to maintain the stoma until the final
session, after which the stoma was left for spontaneous
obliteration.
Dose prescription
Treatment was scheduled twice weekly to a total dose
of 18–30 Gy (median 24 Gy) in 3–6 fractions (median 4
fractions) with a single fraction dose of 5–6 Gy (median
6 Gy) for 1.5–2.5 weeks. These prescriptions were
determined based on the necessity of treatment and
previously irradiated dose at the reference point of the
brachytherapy. In case no. 6, it was vital to suppress the
widespread intratracheal tumour relapse, even though
the total dose came to surpass the tolerance level of the

The British Journal of Radiology, June 2006 523


normal tissue. The reference point was at 5 mm depth
from the bronchial (tumour) surface (Figure 1c).
The mean volume with the prescribed dose was
11 cm3, ranging from 4.3 cm3 to 18.3 cm3. The biological
effective doses (BED) of the present schedules were the
same as those corresponding to the once weekly
schedules both at a/b ratio 3 and at 10.
Follow-up
Patients were followed up at our out-patient clinic.
Bronchofiberscopic follow-up study and X-ray CT study
were scheduled for the third month, and every sixth
month thereafter.
Results
Mini-tracheostomy was successfully performed and
maintained in all patients. In the fractionated endobron-
chial treatment, the mean time from insertion to the
targeted location of the applicator under fluoroscopic
guidance was 8 s (range 6–18 s). The mean dwelling time
of source in the irradiating position and the averaged
single session time of each fractionated brachytherapy
were 3.1 min (range 1 min 57 s to 8 min 6 s) and 24 min
(range 14–35 min), respectively, excluding the time for
treatment planning using the PLATO system. There were
no procedure-related complications. The stomas closed
in 3 days.
All patients obtained symptomatic improvement, and
four patients showed histologically negative broncho-
fiberscopic biopsy specimens.
During the mean follow-up period of 18.9 months, no
patient developed local relapse on the irradiated lesion.
Three patients died of metastatic disease and two patients
died of aggravated respiratory distress due to disease
progression. One patient died of pneumonia due to
tracheal fistula. Another died of haematoemesis at home.
One patient is alive at 31 months presently without disease.
Case presentations
A 53-year-old-man (patient no. 4) complaining of
severe dyspnoea due to obstruction of the trachea and
bilateral main bronchi was admitted to our hospital. He
underwent 60 Gy in 30 fractions of external beam
radiotherapy combined with intravenous administration
of cisplatinum and doxorubicin. To the persisting
tumour, an endobronchial brachytherapy of 24 Gy in
four fractions was scheduled. 1 month after the
brachytherapy, no tumour or viable cells were found in
the trachea. He was asymptomatic for 12 months.
A 52-year-old-woman (patient no. 7) complaining of
mild dyspnoea due to metastatic left bronchial tumour
obstruction was indicated for endobronchial brachy-
therapy. Since undergoing low anterior resection for
rectal cancer 6.7 years previously, she had experienced
gamma-knife therapy for brain metastases, conformal
radiotherapy for pelvic relapse and right mediastinum
mass, and stereotactic body irradiation for lung metas-
tases. She underwent endobronchial irradiation of 24 Gy
524
K Kishi, T Yoshimasu, S Shirai et al
in four fractions at the outpatient clinic. The dyspnoea
disappeared in 2 weeks. She survived 43.8 months after
the brachytherapy.
A 57-year-old-woman (patient no. 8) suffered from
dyspnoea and obstinate cough due to relapsed tracheal
cancer. 8 years previously she had undergone non-
curative surgical resection of the tumour followed by
70 Gy in 35 fractions of external beam radiotherapy.
During the recent 2 years, two metastatic nodules, both
in the left lung, were excised and one nodule in the right
lung was treated with stereotactic radiotherapy. A laser
treatment and stent placement were attempted to relieve
the present symptoms but were ineffective. The previous
dose in the relapsed lesion was 40 Gy. An endobronchial
irradiation of 24 Gy in four fractions was planned. She
continues to maintain a good performance status for over
30.4 months after the brachytherapy.
Discussion
The safety and less invasive nature of the mini-
tracheostomy procedure have been well established [12].
The insertion procedure was fast and simple, and the
reported mean procedural time was only 6.7 s from
incision to removal of the guidewire [17]. During the
endobronchial treatment, the mini-stoma was safely used
as a ready access route. Through the mini-stoma, the
applicator was inserted quickly and smoothly under
fluoroscopy to the target position using a torque control
technique. The patients were freed from the repeated
burden of transoral or transnasal endobronchial manipula-
tion under bronchofiberscopy. Thus, this fast fluoroscopic
procedure would save time, cost and effort of the patients.
Furthermore, the fast and convenient fluoroscopic
procedure allows flexible schedules of fractionation
unrestricted by bronchofiberscopy availability. Despite
biological knowledge that fractionation decreases the late
radiation toxicity [9], most of the reported schedules
have been restricted to once a week: for instance, 30 Gy
in two fractions [18], 30 Gy in 6 fractions [19], 22.5 Gy in
three or 10 Gy in two [11], 20 Gy in five [10], 10 Gy in
one or 18 Gy in three [8] and 14 Gy in 2 weeks as a boost.
To our knowledge, only one schedule has been described
in which 25 Gy for weekly five fractions was given [20].
Another merit of this procedure was that the appli-
cator insertion to deployment was carried out under
smooth fluoroscopic guidance, and anaesthesia such as
local lidocaine spray, which is irritating to the airway,
was not necessary.
It is difficult to discuss the precise effects, including
late toxicity, because of the limited number of observa-
tions in the present study and dosimetric inconsistency
between reports due to different depth of reference
points [1, 5]. However, efforts to control late radiation
toxicity should be continued.
Some patients survived for relatively long periods,
suggesting that an appropriate selection of patients
might result in a good outcome of the endotracheal
brachytherapy in some cases, but at present any such
conclusion is unwarranted due to the small number of
cases studied. The possible benefits of well-fractionated
endobronchial brachytherapy need to be determined in
future investigations.
The British Journal of Radiology, June 2006
Short communication: Mini-tracheostomy for endotracheobronchial radiotherapy
References
1. Stout R, et al. Clinical and quality of life outcomes in the
first United Kingdom randomized trial of endobronchial
brachytherapy (intraluminal radiotherapy) vs. external
beam radiotherapy in the palliative treatment of inoperable
non-small cell lung cancer. Radiother Oncol 2000;56:323–7.
2. Hennequin C, et al. [Endobronchial brachytherapy: techni-
que and indications]. Cancer Radiother 2003;7:33–41.
3. Ikeda H. [Developments in brachytherapy]. Gan To Kagaku
Ryoho 1995;22:1317–23.
4. Perol M, et al. Curative irradiation of limited endobronchial
carcinomas with high-dose rate brachytherapy. Results of a
pilot study. Chest 1997;111:1417–23.
5. Fuwa N, et al. The treatment results of 40 patients with
localized endobronchial cancer with external beam irradia-
tion and intraluminal irradiation using low dose rate (192)Ir
thin wires with a new catheter. Radiother Oncol
2000;56:189–95.
6. Marsh BR. Bronchoscopic brachytherapy. Laryngoscope
1989;99(7 Pt 2 Suppl. 47):1–13.
7. Nomoto Y, et al. [Endobronchial brachytherapy with high
dose rate 192Ir afterloading technique using a new
applicator]. Nippon Igaku Hoshasen Gakkai Zasshi
1996;56:42–7.
8. Nomoto Y, et al. High dose rate endobronchial brachyther-
apy using a new applicator. Radiother Oncol 1997;45:33–7.
9. Nag S, et al. Consensus guidelines for high dose rate remote
brachytherapy in cervical, endometrial, and endobronchial
tumors. Clinical Research Committee, American
Endocurietherapy Society. Int J Radiat Oncol Biol Phys
1993;27:1241–4.
The British Journal of Radiology, June 2006
10. Freitag L, et al. Sequential photodynamic therapy (PDT)
and high dose brachytherapy for endobronchial tumour
control in patients with limited bronchogenic carcinoma.
Thorax 2004;59:790–3.
11. Celebioglu B, et al. High dose rate endobronchial bra-
chytherapy effectively palliates symptoms due to inoper-
able lung cancer. Jpn J Clin Oncol 2002;32:443–8.
12. Cushing M. Minitrach in airway obstruction. Anaesthesia
1986;41:774.
13. Daykin AP. A new use of a ‘Minitrach’ introducer.
Anaesthesia 1993;48:538.
14. Bonde P, et al. Sputum retention after lung operation:
prospective, randomized trial shows superiority of prophy-
lactic minitracheostomy in high-risk patients. Ann Thorac
Surg 2002;74:196–202; discussion 202–3.
15. Corke C, Cranswick P. A Seldinger technique for mini-
tracheostomy insertion. Anaesth Intensive Care
1988;16:206–7.
16. Kishi K, et al. Treatment of malignant tracheobronchial
stenosis by Dacron mesh-covered Z-stents. Cardiovasc
Intervent Radiol 1994;17:33–5.
17. Slots P, et al. Retrograde intubation with a Mini-Trach II kit.
Acta Anaesthesiol Scand 2003;47:274–7.
18. Delclos ME, et al. Endobronchial brachytherapy with high-
dose-rate remote afterloading for recurrent endobronchial
lesions. Radiology 1996;201:279–82.
19. Lorchel F, et al. [High dose rate brachytherapy: a potentially
curative treatment for small invasive T1N0 endobronchial
carcinoma and carcinoma in situ]. Rev Mal Respir
2003;20:515–20.
20. Yokoyama A, et al. Endobronchial brachytherapy for centrally
located early-stage lung cancer. Haigan 1997;16:55–60.
525
 The British Journal of Radiology, 79 (2006), 526–528

CASE REPORT

MR findings of penile lymphoma

K-H CHIANG, MD, P-Y CHANG, MD, S-K LEE, MD, P-S YEN, MD, C-M LING, MD, C-C LIN, MD, C-C LEE, MD
and A S-B CHOU, MD

Department of Radiology, Buddhist Tzu Chi General Hospital, Hualien, Taiwan;

ABSTRACT. Penile lymphoma is extremely rare and secondary involvement of the penis
by lymphoma may be due to retrograde spread or to direct extension from Received 8 March 2005
neighbouring organ. The appearance of penile lymphoma varies and can be mistaken Revised 9 May 2005
for other soft tissue tumours. We report on a case with malignant lymphoma of the Accepted 23 May 2005
penis. MRI findings revealed soft-tissue mass of homogeneous isointensity around the
DOI: 10.1259/bjr/55555394
middle to distal part of penis on T1 weighted imaging and T2 weighted imaging. It was
well encapsulated, minimally enhanced and distinct from corpus cavernosum and ’ 2006 The British Institute of
corpus spongiosum. Radiology

Penile lymphoma is an extremely rare neoplasm.
Nodules, ulcers and diffuse penile swelling have been
reported as the presentation of penile lymphoma, and
the penile shaft is the most common site of involvement
[1–4]. Full physical examination and radiological image
investigations, including CT, MRI and PET (positron
emission tomography), should be undertaken to stage
the patient. Chemotherapy has the advantage of treating
both the primary lesion and any systemic disease.
Herein, we report on a case with malignant lymphoma
of the penis and its imaging findings.
Case report
the corpus cavernosum and corpus spongiosum
(Figure 1).
Cystoscopy with transrectal needle biopsy of prostate
and scrotal incisional biopsy revealed a picture of large
B-cell lymphoma, but there was no obvious abnormal
finding of prostate while reviewing the MR images. CT
of the chest, abdomen and pelvis showed only the penile
mass (Figure 2) and bone marrow biopsy was negative.
He underwent six courses of systemic chemotherapy
with a CHOP regimen and tolerated the whole course of
chemotherapy well. The penile mass disappeared gra-
dually after six cycles of chemotherapy. At regular
follow-up evaluations, the patient is free from recurrence
and dissemination 8 months after the diagnosis.

A 77-year-old man presented with a more than 2-week

history of a painless mass at penile base. He denied
sexual exposure, previous sexually transmitted diseases,
recent trauma and underlying medical disease. No
difficulty with urination had been experienced. There
was no fever, night sweats, fatigue or body weight loss in
the previous 6 months.
Physical examination revealed a firm nodule at penile
base. The remainder of his physical examination was
uneventful. Haematological and chemical panel values
were within normal limits except elevated prostate
specific antigen (PSA) and lactic acid dehydrogenase
(LDH) (PSA: 23.0 ng ml21, normal range: 0–4 ng ml21;
LDH: 751 IU l21, normal range: 15–400 IU l21).
Pelvic MRI without and with Gd-DTPA enhancement
was performed by a 1.5-Tesla MR scanner (Signa Excite;
GE Medical System, Milwaukee, WI). The study showed
soft-tissue mass of homogeneous isointensity around the
middle to basal part of penis on T1 weighted and T2
weighted images. It was well capsulated, minimally
enhanced by contrast medium, and distinct from both
Address correspondence to: Dr Andy Shau-Bin Chou, Department
of Radiology, Buddhist Tzu Chi General Hospital, 707, Section 3,
Chung Yang Road, Hualien, Taiwan.
Discussion
Penile lymphoma is extremely rare and secondary
involvement of the penis by lymphoma may be due to
retrograde spread (haematic or lymphatic) or to direct
extension from neighbouring organ [1–3, 5]. The most
commonly affected site was the shaft, followed by the
glans penis [6]. Diffuse large cell lymphoma was the
most common histological subtype [7].
The appearance of penile lymphoma varies: it may
appear as a mass, as plaques or ulcers in the skin of the
organ, or as diffuse penile swelling [8]. Full physical
examination and radiological image investigation,
including CT, MRI and PET, should be undertaken to
stage the patient [4]. Excision biopsy is essential to
achieve the correct diagnosis, and histological analysis
must include immunohistochemical tests to differentiate
lymphoma from undifferentiated sarcomas or carcino-
mas and to distinguish between B- and T-cell lympho-
mas [7].
As in our case, MRI provides better tissue-contrast
than CT. The margins between the mass and corpus
cavernosum and corpus spongiosum are clear. MRI
images of penile lymphoma had not been documented. It

526 The British Journal of Radiology, June 2006

Case report: Lymphoma of the penis

(a) (b)

(c) (d)

Figure 1. (a) Axial T1 weighted image (repetition time (TR)/echo time (TE) 5 467/8) showed soft-tissue mass of homogeneous
isointensity around the middle to basal part of penis. (b) Sagittal T2 weighted image (TR/TE 5 2350/87) showed the extension of
the lesion. The capsule of corpus spongiosum is intact. (c) Coronal T1 weighted image (TR/TE 5 2650/87) showed the relationship
between the mass and the corpus cavernosum and corpus spongiosum. (d) Coronal T1 weighted image after Gd-DTPA
demonstrated the mass without obvious enhancement.

The British Journal of Radiology, June 2006 527


Figure 2. Contrast-enhanced CT showed soft-tissue mass of
homogeneous isodense around the basal part of penis. There
is also no obvious enhancement of the mass.
should be homogeneous isointensity on T1 weighted and
T2 weighted images and minimally enhanced by contrast
medium.
Lymphoma of the penis seems to behave in a less
aggressive pattern [9]. Chemotherapy has the clear
advantage of obtaining good cosmetic and functional
results; it will also be effective in patients with occult
disseminated lymphomas [10]. Radical surgery should
be used only after the failure of other modalities [7].
528
K-H Chiang, P-Y Chang, S-K Lee et al
In conclusion, penile lymphoma is extremely rare and
can be mistaken for other soft tissue tumours. The
possibility of this diagnosis should be considered when
evaluating a nodule or mass in the penis. MRI may play a
role in the pre-treatment evaluation of penile lymphoma
for local tumour involvement.
References
1. Lo HC, Yu DS, Lee CT, Chen A, Chang SY, Sun GH.
Primary B cell lymphoma of the penis: successful treatment
with organ preservation. Arch Androl 2003;49:467–70.
2. Wang HT, Lo YS, Huang JK. Primary lymphoma of the
penis. J Chin Med Assoc 2003;66:379–381.
3. Pomara G, Cuttano MG, Tripodo C, Carlino F, Selli C.
Primary T-cell rich B-cell lymphoma of the penis: a first
case. BJU Int 2003;91:889.
4. Israel O, Keidar Z, Bar-Shalom R. Positron emission
tomography in the evaluation of lymphoma. Semin Nucl
Med 2004;34:166–79.
5. Nakayama F, Sheth S, Caskey CI, Hamper UM. Penile
metastasis from prostate cancer: diagnosis with sonogra-
phy. J Ultrasound Med 1997;16:751–3.
6. Gough J. Primary reticulum cell sarcoma of the penis. Br J
Urol 1970;42:336–9.
7. el-Sharkawi A, Murphy J. Primary penile lymphoma: the
case for combined modality therapy. Clin Oncol
1996;8:334–5.
8. Bunesch Villalba L, Bargallo Castello X, Vilana Puig R,
Burrel Samaranch M, Bru Saumell C. Lymphoma of the
penis: sonographic findings. J Ultrasound Med
2001;20:929–31.
9. Moreno Aviles J, Salinas Sanchez AS, Gomez Gomez G,
Server Falgas [Tumor of the penis: primary clinical
manifestation of a lymphoma]. Actas Urol Esp 1988;12:
488–90.
10. Arena F, di Stefano C, Peracchia G, Barbieri A, Cortellini P.
Primary lymphoma of the penis: diagnosis and treatment.
Eur Urol 2001;39:232–5.
The British Journal of Radiology, June 2006
The British Journal of Radiology, 79 (2006), 529–535

CASE REPORT

Deep tracheal laceration after balloon dilation for benign

tracheobronchial stenosis: case reports of two patients
1
Y H KIM, MD, 1D J SUNG, MD,
1
S B CHO, MD,
1
K B CHUNG, MD,
1
S H CHA, MD,
2
H S PARK, MD and
3
J W UM, MD

Departments of 1Radiology, 2Urology and 3Surgery, Korea University College of Medicine, Seoul,
Korea

ABSTRACT. We report two cases of deep tracheal laceration in female patients after
balloon dilation for benign tracheobronchial stenosis. Immediate post-procedure Received 17 February 2005
bronchoscopy and CT including 3D reconstructions showed deep lacerations in the Revised 6 May 2005
posterior tracheal wall. Clinically, the patients’ dyspnoea subsided and there has been Accepted 23 May 2005
no recurrence during follow-up after balloon dilation. On the follow-up
DOI: 10.1259/bjr/17839516
3D-reconstructed CT scans obtained 2 months and 8 months following balloon dilation,
respectively, the lacerations had healed completely and there was considerable ’ 2006 The British Institute of
improvement in lumen size. Radiology

Since Cohen et al [1] initially reported balloon dilation
of tracheobronchial stenosis in 1984, endoscopically or
fluoroscopically guided balloon dilation has become an
accepted treatment for benign tracheobronchial stenosis
[2–7]. The major anticipated advantages of balloon
dilation are lower morbidity and mortality than those
of corrective surgery or bougienage.
Although balloon dilation is safe and effective,
excessive balloon inflation may theoretically lacerate or
rupture the airway, thereby causing bleeding, pneu-
mothorax, pneumomediastinum or mediastinitis. As far
as we know, there is only one report [7] briefly
describing superficial or deep lacerations of the tracheo-
bronchial tree after balloon dilation for benign tracheo-
bronchial strictures. However, there was neither a
detailed description of the laceration nor information
regarding the patients’ detailed clinical outcomes.
We present two cases of deep tracheal laceration
confirmed on post-procedure bronchoscopy and CT
scans following balloon dilation for benign tracheobron-
chial stenosis, as well as the detailed clinical data.
Case reports
Case 1
A 31-year-old female complained of aggravated cough,
dyspnoea, and a 3 kg weight loss over a 7 month period.
She had a past history of pulmonary tuberculosis 10
years ago. Chest radiography was normal. However,
Address correspondence to: Ji Hoon Shin, MD, Department of
Radiology, Asan Medical Center, University of Ulsan College of
Medicine, 388-1, Pungnap-2dong, Songpa-gu, Seoul 138-736, Korea.
This investigation has disclosed a potential conflict; none of the
other authors have identified a conflict of interest.
bronchoscopy with biopsy and chest CT including
three-dimensional (3D) reconstructions (Figure 1) showed
a 6 cm long tracheal fibrotic stenosis and a 1.5 cm long
right main bronchial fibrotic stenosis, which was covered
with whitish-yellow elevated lesions. The diameters of the
narrowed segments of the trachea and right main
bronchus were 6 mm and 2 mm, respectively, while those
of the normal segment of the trachea and right main
bronchus were 16 mm and 11 mm, respectively. A
pulmonary function test (PFT) showed a forced expiratory
volume in one second (FEV1) of 1.3 l (45%, predicted) and a
forced vital capacity (FVC) of 2.0 l (54%, predicted).
The detailed technique of balloon dilation is the same
as described in the previous report [7]. We initially
dilated the right main bronchial stenosis and subse-
quently the tracheal stenosis. A 6 mm diameter balloon
catheter was used first for the severe right main
bronchial stenosis in order to provide passage of the
larger balloon catheter. Then the balloon (Boston
Scientific/Medi-tech, Watertown, MA; 10 mm in dia-
meter and 4 cm long for the bronchial stenosis, 18 mm in
diameter and 10 cm long for the tracheal stenosis) was
slowly inflated manually using a diluted water-soluble
contrast medium until the waist formation in the inflated
balloon catheter at the stricture segment disappeared.
There was not much resistance during balloon inflation.
Immediately after balloon dilation, the patient com-
plained of mild chest pain and expectorated some blood-
tinged sputum. On bronchoscopy (not shown) obtained
immediately after balloon dilation, a tracheal laceration
was observed at the posterior tracheal wall. Its length
and depth were estimated at approximately 5 cm and
8 mm, respectively, on bronchoscopy. For further eva-
luation of the extent of the laceration and possible
associated complications, CT scanning was performed
and revealed a deep, longitudinal laceration on the

The British Journal of Radiology, June 2006 529


(a)
Figure 1. Case 1. (a) Anteroposterior and (b) lateral views of the three-dimensional (3D) reconstruction CT, obtained 4 days
before balloon dilation, show a 6 cm long tracheal stenosis (arrowheads in (a) and (b)) from the mid-trachea to the carina and a
1.5 cm long right main bronchial stenosis (arrows in (a)).
posterior tracheal wall with the pneumomediastinum
(Figure 2). As a result of the tracheal laceration and
separation of the tracheal wall at the tear point, the
diameter of the trachea became widened. The patient’s
vital signs were stable, she was without fever or chills
and there was no change in her haemoglobin level. Her
chest pain and blood-tinged sputum disappeared within
24 h. Therefore, she was prescribed only oral antibiotics
to prevent possible infection and was discharged from
the hospital 5 days after the procedure.
On follow-up CT scans (Figure 3) obtained 8 months
following balloon dilation, the deep laceration had
completely healed and the widened tracheal lumen
was maintained. Tests, also obtained 8 months following
balloon dilation, showed an FEV1 of 1.8 l (63%,
predicted) and an FVC of 3.3 l (89%, predicted); the
FEV1 and FVC increased as much as 18% and 35%,
respectively, compared with those of the predilation PFT.
The patient has maintained her symptomatic improve-
ment without recurrence for 10 months.
Case 2
A 27-year-old woman had complained of respiratory
difficulty for 3 months. She had a past history of tracheal
intubation due to decreased mental ability by reason of
530
J H Kim, J H Shin, T S Shim et al
(b)
diabetic ketoacidosis for 6 months. Bronchoscopy and CT
(Figure 4) obtained 3 weeks before balloon dilation
revealed two focal fibrotic stenoses in the mid and lower
levels of the trachea. The diameters of the narrowed
segments in the mid and lower levels of the trachea were
5 mm and 12 mm, respectively, while the diameter of the
normal segment in the trachea was 16 mm. PFT’s
performed 1 week before balloon dilation were as
follows: the FEV1 was 1.1 l (39%, predicted) and the
FVC was 2.3 l (65%, predicted).
Dilation was performed without much resistance,
using an 8 cm long and 16 mm diameter balloon for
the two focal tracheal stenoses. Immediately after balloon
dilation, the patient complained of mild chest pain. On
bronchoscopy and CT (Figure 5) obtained immediately
after balloon dilation, a deep longitudinal laceration
(4.5 cm in length, 1 cm in depth) was detected at the
mid-level of the posterior tracheal wall. Her vital signs
were stable without fever, and her chest pain disap-
peared within 24 h. She did not receive any further
therapy except for preventative oral antibiotics.
On follow-up CT scans (Figure 6) obtained 2 months
following balloon dilation, the deep laceration had
nearly disappeared and the tracheal lumen was
widened. Follow-up PFT, obtained 2 months after
balloon dilation, showed an FEV1 of 1.9 l (69%,
predicted) and an FVC of 2.3 l (65%, predicted); the
The British Journal of Radiology, June 2006
Case report: Deep tracheal laceration after balloon dilatation

(a) (b)

Figure 2. Case 1. Immediately after balloon dilation, the

lumens of the stenoses appear to be widened on (a) the
anteroposterior view of the 3D reconstruction CT. However,
a long and deep laceration (arrowheads in (b) and (c)) on the
posterior tracheal wall with pneumomediastinum (arrows in
(c)) is well visualized on (b) the lateral view of the 3D
(c) reconstruction CT and (c) axial CT scan.

The British Journal of Radiology, June 2006 531


(a)
Figure 3. Case 1. (a) Anteroposterior and (b) lateral views of the follow-up 3D reconstruction CT scans show marked
improvement of the right main bronchial (arrows in (a)) and tracheal stenosis (arrowheads in (a)) without further visualization
of the deep laceration on the posterior tracheal wall (arrowheads in (b)).
FEV1 had increased as much as 30% compared with that
of the pre-dilation PFT. The patient’s symptoms were
much improved and had not recurred for 5 months at the
time of writing after balloon dilation.
Discussion
Although expandable metallic stent deployment has
been an effective and minimally invasive procedure for
benign tracheobronchial stenosis and may overcome the
problem of short-term recurrence after balloon dilation,
stent placement has notable drawbacks, including
migration, recurrence of stenosis from tissue hyperpla-
sia, stent fracture and difficulty of stent removal [4, 7–
10]. Therefore, balloon dilation is usually performed as
the initial treatment for benign tracheobronchial stenosis
because it is a safe and simple procedure [2–7]. However,
several complications, such as bronchospasm or lung
atelectasis after balloon dilation, have been reported by
several authors [7, 11–13]. Furthermore, overdilation of
the stenosis may cause tracheobronchial rupture similar
to the rupture or bleeding reported for balloon dilation of
the upper gastrointestinal tract [14–16]. Although the
predominant fibrotic process can be successfully dilated
and has a potentially successful clinical outcome [3, 6],
there may be potential complications such as laceration
532
J H Kim, J H Shin, T S Shim et al
(b)
or even free perforation of the tracheobronchial tree after
sudden balloon dilation of fibrotic stenosis because of the
hardness or stiffness of the fibrotic stenosis.
To the best of our knowledge, Lee et al [7] initially
reported tracheal or bronchial laceration after balloon
dilation, i.e. two cases of deep mucosal laceration and 15
cases of superficial mucosal laceration of the 59 patients
treated with balloon dilation for benign tracheobronchial
stenosis. In their report, most patients (45 of 59, 76%) had
chronic fibrotic stenosis due to tuberculosis, and balloons
for bronchial and tracheal stenosis were 6–12 mm and 14–
20 mm in diameter, respectively. Their patients experi-
enced tracheobronchial laceration (25%, 17/59) with
relatively high frequency after balloon dilation, although
the procedures were performed with caution. However, in
Lee’s report, the clinical outcome and follow-up data of the
deep tracheal or bronchial laceration were not documen-
ted in detail. Interestingly, in our two cases the long and
deep tracheal laceration disappeared completely during
the follow-up period after balloon dilation. Lee et al [7]
also briefly stated that two deep lacerations left no
subsequent clinical sequelae. We assume that, if a deep
laceration heals soon without much growth of granulation
tissue, a good clinical outcome can be achieved after the
deep tracheobronchial laceration because the lumen of the
stenotic segment will be widened enough. Balloon dilation
for congenital tracheal stenosis is another illustration of
The British Journal of Radiology, June 2006
Case report: Deep tracheal laceration after balloon dilatation

Figure 4. Case 2. Anteroposterior view of the three-dimen- Figure 5. Case 2. Immediately after balloon dilation, a deep
sional (3D) reconstruction CT obtained 3 weeks before laceration (arrowheads) on the posterior tracheal wall is
balloon dilation, shows two focal stenoses (arrows) at the clearly visualized on the axial CT scan.
mid and lower levels of the trachea.

The British Journal of Radiology, June 2006 533


Figure 6. Case 2. (a) Anteroposterior view of 3D reconstruction and (b) axial scan of the follow-up CT show improvement of the
two focal tracheal stenoses (arrows in (a)) and reveal a completely healed deep laceration on the posterior tracheal wall.
further proof to support our assumption that rupture of
the complete cartilaginous rings by balloon dilation
represents a prerequisite step for increasing the luminal
diameter itself [17, 18].
As in our cases, even the deep and large tracheal
laceration can be treated conservatively if patients show
stable conditions and a minimal and asymptomatic
pneumomediastinum. However, if patients present with
acute respiratory distress secondary to a tracheal lacera-
tion, surgical or interventional treatment is indispensable
[19]. Bronchoscopy can detect early tracheal laceration
and determine the location. However, bronchoscopy
may not offer sufficient information about anatomical
location and morphology of deep tracheal laceration due
to limited sight of view. In contrast, CT including 3D
reconstructions can not only delineate the precise extent
and morphology of the deep laceration, but also detect
pneumomediastinum or mediastinal bleeding, which are
important factors in deciding the treatment plan.
References
1. Cohen MD, Weber TR, Rao CC. Balloon dilatation of
tracheal and bronchial stenosis. AJR Am J Roentgenol
1984;142:477–8.
534
J H Kim, J H Shin, T S Shim et al
2. Carlin BW, Harell JH, Moser KM. The treatment of
endobronchial stenosis using balloon catheter dilatation.
Chest 1988;93:1148–51.
3. Ferretti G, Jouvan FB, Thony F, Pison C, Coulomb M.
Benign noninflammatory bronchial stenosis: treatment with
balloon dilation. Radiology 1995;196:831–4.
4. Lee KW, Im JG, Han JK, Kim TK, Park JH, Yeon KM.
Tuberculous stenosis of the left main bronchus: results of
treatment with balloons and metallic stents. J Vasc Interv
Radiol 1999;10:352–8.
5. Carre P, Rousseau H, Lombart L, et al. Balloon dilatation
and self-expanding metal Wallstent insertion for manage-
ment of bronchostenosis following lung transplantation.
Chest 1994;105:343–8.
6. Sheski FD, Mathur PN. Long-term results of fiberoptic
bronchoscopic balloon dilation in the management of
benign tracheobronchial stenosis. Chest 1998;114:796–800.
7. Lee KH, Ko GY, Song HY, Shim TS, Kim WS. Benign
tracheobronchial stenoses: long-term clinical experience
with balloon dilation. J Vasc Interv Radiol 2002;13:909–14.
8. Kim JH, Shin JH, Shim TS, et al. Results of temporary
placement of covered retrievable expandable nitionl stents
for tuberculous bronchial strictures. J Vasc Interv Radiol
2004;15:1003–8.
9. Kim JH, Shin JH, Shim TS, et al. Efficacy and safety of a
retrieval hook for removal of retrievable expandable
tracheobronchial stents. J Vasc Interv Radiol 2004;15:
697–705.
The British Journal of Radiology, June 2006
Case report: Deep tracheal laceration after balloon dilatation
10. Petersen BD, Uchida BT, Barton RE, Keller FS, Rosch J.
Gianturco-Rosch Z stents in tracheobronchial stenoses. J
Vasc Interv Radiol 1995;6:925–31.
11. Brown SB, Hedlund GL, Glasier CM, Williams KD,
Greenwood LH, Gilliland JD. Tracheobronchial stenosis in
infants: successful balloon dilation therapy. Pediatr Radiol
1987;164:475–8.
12. Hebra A, Powell DD, Smith CD, Othersen HB Jr. Balloon
tracheoplasty in children: results of a 15-year experience. J
Pediatr Surg 1991;26:957–61.
13. Elkerbout SC, van Lingen RA, Gerriten J, et al. Endoscopic
balloon dilatation of acquired airway stenosis in newborn
infants: a promising treatment. Arch Dis Child
1993;68:37–40.
14. Kang SG, Song HY, Lim MK, Yoon HK, Goo DE, Sung KB.
Esophageal rupture during balloon dilation of strictures of
benign or malignant causes: prevalence and clinical
importance. Radiology 1998;209:741–6.
The British Journal of Radiology, June 2006
15. Solt J, Bajo J, Szabó, Horváth ÖP. Long-term results of
balloon catheter dilation for benign gastric outlet stenosis.
Endoscopy 2003;35:490–5.
16. Kim JH, Shin JH, Di ZH, et al. Benign duodenal strictures:
treatment by means of floroscopically guided balloon
dilation. J Vasc Interv Radiol 2005;16:543–8.
17. Brown SB, Hedlund GL, Glasier CM, Williams KD,
Greenwood LH, Gilliland JD. Tracheobronchial stenosis in
infants: successful balloon dilation therapy. Radiology
1987;164:475–8.
18. Kim HJ, Shin JH, Hong SJ, Park SJ, Lee SY, Song HY.
Treatment of congenital tracheal stenosis with balloon-
assisted posterior tracheal splitting and temporary place-
ment of a covered retrievable metallic stent. J Vasc Interv
Radiol 2005;16:287–91.
19. Madden BP, Datta S, Charokopos N. Experience with
ultraflex expandable metallic stents in the management of
endobronchial pathology. Ann Thorac Surg 2002;73:938–44.
535
’ 2006 The British Institute of Radiology The British Journal of Radiology, 79 (2006), 536

Correspondence
(The Editors do not hold themselves responsible for opinions expressed by correspondents)

Any fool can understand IMRT (Received 13 January 2006 and accepted 27 January 2006)
DOI: 10.1259/bjr/77476500

The Editor—Sir,
In his paper, Intensity-modulated radiation therapy
References
(IMRT): a clinical reality for cancer treatment, ‘‘any fool 1. Webb S. Intensity-modulated radiation therapy (IMRT): a
can understand this’’ (The 2004 Silvanus Thompson clinical reality for cancer treatment, ‘‘any fool can under-
Memorial Lecture) [1], Prof. Webb directs your readers to stand this’’. Br J Radiol 2005;78:S64–S73.
the famous book by Silvanus Thompson, Calculus made 2. Thompson S. Calculus made easy. Gardner M, editor.
Palgrave Macmillan, 1999.
easy, from which his ‘‘any fool…’’ quotation is taken. I,
too, would strongly recommend this book to anyone
interested in mathematics, but would caution that the
book of that title to be ordered via Amazon, as suggested Author’s reply
by Prof. Webb, is not the original book as displayed in
his Figure 2, but is a re-write of the book by Martin
Gardner [2]. Martin Gardner is an outstanding mathe- The Editor—Sir,
matician and popularizer of mathematics, but I have to Thank you for this useful clarification. As may be seen
say I do not believe he has left very much of the original from my Figure 2 [1], I had a version of the original when
or entirely succeeded in keeping its spirit in his new I was writing. It was probably not too wise to mention
Amazon anyway and I concur with Peter that obtaining
version. I would recommend looking for a second-hand
sight of the original is sound advice.
copy of the original. It was this book which the teenage
Yours etc.,
Richard Feynman, having discovered that ‘‘calculus is a
S WEBB
big thing,’’ found and taught himself from. As Feynman
later wrote: ‘‘I have since realised that that particular Institute of Cancer Research
calculus book had its especially screwy methods…and it Royal Marsden NHS Trust
invented proofs that weren’t proofs…and there were Downs Road
errors in proofs…but of course the proofs were not Sutton
important’’. Yes indeed – but Thompson’s was one of the Surrey
earliest attempts to present in an accessible form a SM2 5PT
subject once thought too difficult to be attempted outside UK
a university mathematics course.
Yours etc., (Received and accepted 30 January 2006)
P DAWSON DOI:
DOI: 10.1259/bjr/77476500
10.1259/bjr/69383091

Directorate of Imaging
UCL Hospitals NHS Trust References
235 Euston Road 1. Webb S. Intensity-modulated radiation therapy (IMRT): a
London clinical reality for cancer treatment, ‘‘any fool can under-
NW1 2BU stand this’’. Br J Radiol 2005;78:S64–S73.

536 The British Journal of Radiology, June 2006

The British Journal of Radiology, 79 (2006), 537–539
CASE OF THE MONTH
Neck pain: an unusual presentation of a common disease
A C PANKHANIA, MBChB, MRCS, T PATANKAR, MBChB, DMRD, DMRE,
Department of Neuroradiology and Neuropathology, Hope Hospital, Salford, Manchester M6 8HD,
UK
A 68-year-old man presented with 1 month history of
neck pain, progressively worsening sensory dysfunction
in the right hand, weakness of both hands and difficulty
walking. On clinical examination he was found to have
reduced power in both upper and lower limbs, with up-
going plantar reflexes.
MRI was performed using a Philips Gyroscan 1.5 T
machine using sagittal T1 weighted turbo spin echo (TSE;
repetition time (TR) 400/echo time (TE) 10, matrix 512,
field of view (FOV) 2556255, slice thickness 3.0 mm/
0.3 mm, number of signal averages (NSA) 4), T2
weighted TSE (TR 3500/TE 120, matrix 512, FOV
2556255, slice thickness 3.0 mm/0.3 mm, NSA 4) and
axial T2 weighted turbo field echo (TFE, TR 7.8/TE 3.9,
flip angle 45 ˚, matrix 512, FOV 2256225, slice thickness
3.5 mm/21.8 mm, NSA 3) sequences. Post-contrast
sagittal (as above) and axial T1 weighted TFE (TR 9.4/
Address correspondence to: Dr Tufail Patankar, 30 Windy Hill
Drive, Bolton BL3 4TH, UK.
The British Journal of Radiology, June 2006
DNBE, FRCR and D DU PLESSIS, MRCPath
Received 15 November
2004
Revised 10 June 2005
Accepted 11 July 2005
DOI: 10.1259/bjr/28763793
’ 2006 The British Institute of
Radiology
TE 4.6, flip angle 25 ˚, matrix 512, FOV 24562.4, slice
thickness 4.0 mm/22.0 mm, NSA 4) sequences were
also performed from C3 to D1 level. MRI demonstrated a
destructive lesion involving the right facet joint of C4/5
associated with a medially placed extradural mass of
intermediate signal on T1 weighted images, intermediate
heterogeneous signal on T2 weighted images which
showed peripheral contrast enhancement post-
gadolinium (Figure 1). Enhancement was also present
in the joint and surrounding soft tissues. The soft tissue
mass was compressing and displacing the spinal cord
and intrinsic high signal was present in the cord on T2
weighted sequences. Similar but less severe changes
were also present in the right C2/C3 facet joint.
A CT scan performed to look for bony changes
revealed subtle eggshell calcification noted around the
extradural mass and well-defined erosive changes
involving the facets of C2/C3, C3/C4 and C4/C5
(Figure 2).
What is the differential diagnosis?
537
 A C Pankhania, T Patankar and D Du Plessis

Figure 1. MRI showing a destructive lesion involving the right facet joint of C4/5 associated with a medially placed extradural
mass compressing the spinal cord and of intermediate signal on (a) T1 weighted images, (b) heterogeneous intermediate signal
on T2 weighted images which showed (c) peripheral contrast enhancement post-gadolinium (arrow demonstrates well-defined
facet joint erosion).

Figure 3. Low power image showing tophaceous deposits

Figure 2. CT scan shows well defined erosive change
involving the right C4/C5 facet joint.
[black arrows] against a background of fibrosis and focal
chronic inflammation (haematoxylin and eosin stain, original
magnification 650). Inset: Birefringent needle shaped urate
crystals [white arrow] demonstrated by polarised light
(original magnification 6630).

538 The British Journal of Radiology, June 2006

Case of the month: Neck pain
Imaging findings suggested a diagnosis of spinal gout.
The patient had no history or evidence of gout and
denied weight loss or trauma. Subsequently, the serum
urate level was found to be 0.49 mmol l21 (normal
values 0.24–0.50 mmol l21). Inflammatory markers were
normal except for a minimally raised C-reactive protein,
which was 22 mg l21 (normal ,10 mg l21).
Posterior surgical decompression and debulking of the
extradural mass was undertaken. Histology of the mass
showed areas of chronic inflammation and necrosis.
Some of the areas of the necrosis showed birefringent
needle shaped crystalline structures consistent with
urate crystals (Figure 3). The diagnosis was therefore
made of gout related arthropathy with tophus formation.
A good post-operative recovery was made and the
patient received medical therapy for gout.
Discussion
Gout is a common metabolic disorder characterized by
episodes of recurrent arthritis and the presence of
monosodium urate in the affected tissues. The disease
tends to affect distal joints but involvement of the axial
skeleton, though rare, has been reported [1, 2] with less
than 40 cases reported in the world literature [3]. The
distribution between cervical, thoracic and lumbar spine
is debatable [1, 2].
The imaging features of spinal gout can be non-specific
and can mimic infectious, inflammatory, degenerative or
neoplastic disease. The MR appearances are defined to a
great extent by the tophus but are variable. The tophi are
low to intermediate signal on T1 weighted images, but
The British Journal of Radiology, June 2006
may be homogeneously low or high in signal on T2
weighted images [2, 4]. The most common pattern is
homogeneous intermediate signal on T1 and heteroge-
neous intermediate to low signal on T2 weighted imaging
[2], which are similar to the appearances seen on MR in
our patient. The variability of signal characteristics on
MR is thought to be due to variable levels of calcium
deposition within the tophus [2, 4]. The tophus may
show homogeneous enhancement or heterogeneous
peripheral enhancement following gadolinium [2] as in
this case.
Our case is unusual in its unilateral involvement of the
facets and that the patient had no radiological or clinical
evidence of gout [2, 5]. Infection and neoplastic process
were excluded on imaging because of multilevel invol-
vement centred on the facet joints. An inflammatory
condition such as rheumatoid disease was considered
unlikely as there was no involvement of other joints.
References
1. Duprez TP, et al. Gout in the cervical spine: MR pattern
mimicking disk vertebral infection. AJNR Am J Neuroradiol
1996;17:151–3.
2. Hsu C-Y, et al. Tophaceous gout of the spine: MR imaging
features. Clin Radiol 2002;57:919–25.
3. Barrett K, Miller ML, Wilson JT. Tophaceous gout of the
spine mimicking epidural infection: report and review of the
literature. Neurosurgery 2001;48:1170–3.
4. Yu JS, et al. MR imaging of tophaceous gout. AJR Am J
Roentgenol 1997;168:523–7.
5. Kaye PV, Dreyer MD. Spinal gout: an unusual clinical and
cytological presentation. Cytopathology 1999;10:411–4.
539
’ 2006 The British Institute of Radiology
Book reviews
Computer imaging: digital image analysis and processing.
By S E Umbaugh. pp. 696, 2005 (CRC Press, Taylor &
Francis Group, Boca Raton, FL), £38.99
ISBN 0-8493-2919-1
This is a comprehensive text-book dealing with all
aspects of the acquisition and processing of visual
information by computer. The treatment given is
oriented towards application of the techniques and
examples are taken from a wide range of fields, which
includes medical imaging but deals with other areas
such as the entertainment industry and space explora-
tion. The book is intended primarily for use by university
students and staff in teaching image processing courses
or research, but will be useful for professionals working
in the commercial sector, government research or the
health service. The book is designed for use by scientists
and engineers. It is well written and a strong computing
background is not required in order to gain a conceptual
understanding of the subject from the book. However,
far more can be gained if the reader has computing and
programming skills. Tutorial exercises are included with
each chapter to provide hands-on experience and enable
the reader to gain insight into the use of various image
analysis and processing algorithms. The book is illu-
strated throughout with images to demonstrate the
effects of the processing techniques described. Since the
book is not specifically aimed at the medical sector, it is
not a book to dip into to find the techniques applicable to
a particular aspect of medical imaging, but rather to
obtain a grounding in image analysis and processing
techniques.
The book is divided into four sections. The introduc-
tion to computer imaging presents a global picture to
enable the reader to gain an understanding of the overall
process. It contains the basic concepts required to
understand computer imaging, including optics, imag-
ing analysis and presentation. The second and third
sections on digital image analysis and image processing
make up over three-quarters of the text. Section 2
describes the tools, concepts and models required for
analysing digital images, including segmentation, trans-
forms and feature analysis. Section 3 on image proces-
sing starts with visual perception and discusses the
application of processing images for human consump-
tion, including topics such as enhancement, restoration
and compression. Section 4 contains information on
computer vision and image processing tools software
developed at the author’s department. A windows
version of the software is on a CD accompanying the
book. This has a menu-driven user interface to facilitate
applications and is designed to allow the reader to apply
the algorithms. A knowledge of C and C++ program-
ming will allow the user to develop the algorithms
further for their own applications.
The book provides a valuable text for those wishing to
study and develop skills in image analysis and processing.
It differs from other texts by including an insight into the
application of the techniques in many different fields.
C J MARTIN
540
The British Journal of Radiology, 79 (2006), 540–541
Coronary radiology. By M Oudkerk. pp. x+254, 2004
(Springer-Verlag, Berlin, Heidelberg, New York),
£107.50
ISBN 3-540-43640-5
This book provides an overview of the various imaging
modalities used to assess the coronary vessels. Five main
areas are covered: anatomy, invasive imaging, non-
invasive imaging, calcification and multidimensional
computed coronary visualization. The invasive imaging
section includes intracoronary ultrasound. Non-invasive
imaging principally outlines coronary CT, but also
addresses MR coronary angiography and electron beam
CT. The calcification chapter makes up a significant
proportion of the book and outlines coronary calcifica-
tion imaging, its epidemiology and the clinical implica-
tions. In depth coverage of imaging rendering techniques
is found in the final chapter. This particular section
would also be useful to imaging in other areas of the
body.
The large sections are divided into clearly set short
chapters which lend this book well to reading in multiple
short time periods. Furthermore, the text flows well and
it is an easy read. Correlation between modalities is
frequently displayed particularly with regard to conven-
tional angiography positioning and CT angiography
positioning. This is evident in the demonstration of
anatomy with direct correlations between diagrammatic
angiographic positioning with conventional angiograms
and CT angiographic images displayed together. The
images are clear, up to date and well set out. The use of
diagrams to assist explanation is frequent and works
well. References are displayed throughout the text with
lists found at the end of each chapter. Overall this book
allows the reader to gain an understanding rather than
simply use for reference.
B HOLLOWAY
Molecular basis of breast cancer: prevention and treatment.
By J Russo and I H Russo. pp. xiv+448, 2004 (Springer-
Verlag, Berlin, Heidelberg, New York), £154.00
ISBN 3-540-00391-6
This magnum opus covers the entire molecular basis of
breast cancer in ten self-contained chapters.
Epidemiological considerations in breast cancer are
covered in chapter one. This reviews geographical
influences, ionizing radiation, electromagnetic field,
reproductive aspects and environmental factors, includ-
ing smoking and alcohol. This is an interesting introduc-
tion to the subject.
The development of the breast is described in the
second chapter in great detail. A knowledge of breast
development is critical in understanding breast cancer
risks, such as its inverse relationship with early parity.
The level and depth of information is excellent. The
changes in breast structure from intrauterine life to old
age are described in detail.
Endocrine control of breast development is covered in
chapter three. The response of breast tissue to oestrogen
The British Journal of Radiology, June 2006
BJR
The British Journal
of Radiology

July 2006
Volume 79
Issue 943
July 2006, Volume 79, Issue 943

● Hypoxia in biology and medicine: the legacy of L H Gray

● Controversies in non-accidental head injury in infants

● Advanced imaging applied to radiotherapy planning in head and

neck cancer: a clinical review

● Radiofrequency ablation in pig lungs: in vivo comparison of

internally cooled, perfusion and multitined expandable
electrodes

● PET/CT detects abdominal wall and port site metastases of

colorectal carcinoma

● Quantitative colorectal cancer perfusion measurement by

multidetector-row CT: does greater tumour coverage improve
measurement reproducibility?

● Prophylactic implantation of inferior vena cava filter during

interventional radiological treatment for deep venous
thrombosis of the lower extremity

● A survey of MRI quality assurance programmes

● Influence of menopausal status and use of hormone replacement

therapy on radiation dose from mammography in routine breast
screening

● Kodak EDR2 film for patient skin dose assessment in cardiac

catheterization procedures

● Randomized phase II study of GM-CSF to reduce mucositis caused by accelerated

radiotherapy of laryngeal cancer

● Magnetic resonance urography: a pictorial overview

● Focal pancreatic lesion: can a neoplasm be confidently excluded?

The British Journal of Radiology, 79 (2006), 545–549
COMMENTARY
Hypoxia in biology and medicine: the legacy of L H Gray
1
P P DENDY, PhD and 2P WARDMAN, DSc
1
1A Coppice Avenue, Great Shelford, Cambridge CB2 5AQ and 2Gray Cancer Institute, PO Box 100,
Mount Vernon Hospital, Northwood, Middlesex HA6 2JR, UK
On 10 November 2005, the centenary of the birth of
Louis Harold Gray FRS (1905–1965), a meeting was held
at Blenheim Palace near Oxford to mark the anniversary
and to bring together scientists who had worked with
Gray and others specializing in the field of tumour
hypoxia – which is arguably his main legacy. This
Commentary cannot summarize all the presentations,
but aims to provide, in this context, a brief overview of
the current understanding of hypoxia in relation to
radiotherapy and other wider biological and medical
implications. We illustrate how progress in imaging
hypoxia and understanding changes in gene expression
linked to hypoxia provide important avenues towards
molecular targeting in radiation oncology, and should
eventually lead to individualization of patient treatment
to an extent that Gray could only have dreamed of.
The scientific career of L H Gray
Barry Michael opened the meeting by outlining Gray’s
scientific career. ‘‘Hal’’, as he was known to colleagues,
studied physics under Rutherford at the Cavendish
Laboratory, Cambridge. His mentors at Cambridge
included J J Thompson, Chadwick, Cockcroft, C T R
Wilson, Aston and Kapitza; a veritable ‘‘who’s who’’ of
physicists of the early 1930s. At an early stage, Gray
decided he wished to use his knowledge in a practical
way and turned to biology and medicine. He was elected
President of the British Institute of Radiology for 1949–
1950 and in 1952 he delivered the 32nd Silvanus
Thomson Memorial Lecture, entitled The initiation and
development of cellular damage by ionising radiations [1]. The
link between tumour blood supply, oxygen delivery and
radiosensitivity had been discussed by Mottram in this
Journal as early as 1936 [2], but in 1953 and 1955 Gray
and his colleagues published two seminal papers [3, 4],
recognizing that it might be possible to manipulate the
oxygen status to improve radiotherapy. Thus the first
paper [3] stated: ‘‘Consideration is given to the supply of
oxygen to tissues as a factor in radiotherapy, and it is
The British Journal of Radiology, July 2006
Received and accepted 15
March 2006.
DOI: 10.1259/bjr/13634453
’ 2006 The British Institute of
Radiology
concluded, on the basis of existing knowledge, that in
certain circumstances the effectiveness of X-ray treat-
ment might be increased if the patient were breathing
oxygen at the time of irradiation.’’ In the later paper [4],
Tomlinson and Gray showed that the histological pattern
of necrosis in sections of some human lung tumours was
consistent with that expected if the supply of oxygen was
the limiting factor determining the onset of necrosis. For
the remainder of his life, Gray led an enthusiastic team
establishing radiobiology as a new, rigorous, scientific
discipline. They worked in many areas, but the challenge
of hypoxic cells always played a central role and, to this
day, remains an increasingly complex problem.
Hypoxia and radiotherapy: the position in the
late 1990s
The Fifth Edition of Eric Hall’s monograph [5]
summarizes the state of knowledge on hypoxic cells
and hypoxic cell radiosensitizers in the late 1990s. Some
key points are:
‘‘Oxygen ‘‘fixes’’ (i.e. makes permanent) the damage
produced by free radicals. In the absence of oxygen,
damage produced by indirect action may be repaired.
Chronic hypoxia results from the limited diffusion range
of oxygen through respiring tissue. Acute hypoxia is a
result of the temporary closing of tumour blood vessels
and is therefore transient. … There is good evidence that
human tumours contain hypoxic cells … evidence
includes histologic appearance, oxygen probe measure-
ments, the binding of radioactive nitroimidazoles …
There is clinical evidence that hypoxia may play an
important role in malignant progression.’’
Hall also summarizes progress on radiosensitizing
hypoxic cells, and work towards hypoxia-selective
cytotoxins such as tirapazamine. He describes how
‘‘Adams and his colleagues listed properties that would
be essential for a clinically useful hypoxic cell sensitizer.
Firstly, it had to selectively sensitize hypoxic cells at a
concentration that would result in acceptable toxicity in
545

normal tissues … be chemically stable and not subject to
rapid metabolic breakdown … highly soluble in water or
lipids … capable of diffusing a considerable distance
through a non-vascularized cell mass to reach hypoxic
cells … effective at … doses of a few grays.’’
Several drugs of the nitroimidazole family had been
tested as radiosensitizers. Misonidazole had a higher
electron affinity and was more effective than metroni-
dazole, being very effective in cells in culture and with
animal tumours but with poor results in clinical trials.
Related compounds evaluated included etanidazole and
nimorazole, which had less toxicity because of shorter
biological half-lives or reduced passage across the blood–
brain barrier. Nimorazole was shown to be of benefit in
head-and-neck cancer in Danish trials, consistent with an
earlier meta-analysis by the Danish group of all
randomized trials of hypoxia modification indicating
an odds ratio of 1.3 [6]. A conclusion from these studies
has been that the clinical trials of hypoxic cell radio-
sensitizers were compromised by a lack of knowledge of
the hypoxic status of an individual patient’s tumour.
Now, however, newer ‘‘bioreductive’’ drugs selectively
toxic to hypoxic cells without radiation have been
identified. Currently, tirapazamine is the lead compound
in clinical trial. Interestingly, these bioreductive drugs
also have a dependency on electron affinity for their
activity.
Changes in gene expression
In recent years, hypoxia-induced changes in gene
expression have been extensively demonstrated; two
speakers (Bradly Wouters and Christopher Pugh) dis-
cussed recent developments. Hypoxia inducible factor
(HIF)-1 is a heterodimeric transcription factor made up
of a and b subunits and it was first recognized as the
DNA binding factor that mediates the hypoxia-induced
expression of the erythropoietin gene. HIF-1 may
promote either directly or indirectly the expression of
as many as 60 target genes. Hydroxylation of HIF-1a to
an inactive form has an absolute requirement for
molecular oxygen and prolyl hydroxylase enzymes:
HIF-1a escapes inactivation in hypoxia and hence can
bind HIF-1b, form the functional HIF-1 complex and
drive gene expression. Hypoxia-inducible genes are
known to be involved in regulation of biological
processes associated with malignancy [7]. HIF-1 can
regulate expression of many enzymes in the glycolytic
pathway, as well as processes involved in genetic
instability, tissue invasion and metastases. HIF-1 also
has a clear role in the regulation of genes involved in
angiogenesis, both in normal development and in
tumours [8]. In all the above roles HIF-1 is likely to
promote tumour growth but there are a few situations
where HIF-1 (and other isoforms of HIF) can have a
negative regulatory effect on tumour response.
The two speakers also discussed different genetically-
related issues. Bradly Wouters pointed out that other
mechanisms are required to explain the biological
response to acute, rather than chronic, hypoxia. He
presented new evidence that a potentially important
point for regulating gene expression that is able to
respond rapidly to changes in the microenvironment is
546
P P Dendy and P Wardman
the process of RNA translation (protein synthesis).
Global mRNA translation is severely, but reversibly,
inhibited during hypoxic conditions, but Wouters and
colleagues have shown [9] that, in HeLa cells and
prostate carcinoma cells in vitro, this averaging process
obscures wide variations in behaviour at the level of
individual genes. Indeed a significant number of mRNA
species are not dependent on the translation factors that
are inhibited during hypoxia and in this efficiently
translated fraction of mRNA, 120 genes were more than
4-fold up-regulated by hypoxia.
Pugh pointed out that although HIF is frequently up-
regulated in cancer, genetic studies have not always
supported a simple model in which up-regulation of HIF
promotes a specific biological process associated with
malignancy, e.g. angiogenesis, and hence tumour growth
[8]. He postulated that genetic mutation may affect the
function of an extensive physiological pathway. Thus
clonal selection of a particular property affects a package
of properties which, individually, could contribute
positively, negatively, or not at all to the overall
advantage driving selection of the clone. For example,
cellular proliferation, HIF activation and angiogenesis
might be co-selected because they were linked by
pathways that operate physiologically to preserve oxy-
gen homeostasis. Pugh concluded that understanding
the HIF system has many implications for cancer biology
and interference with this system may have therapeutic
uses. However, we will need to know a lot more about
the consequences of intervention at any particular point
in the development of malignancy to be certain of a net
positive benefit.
Labelling and mapping hypoxic cells
There have been important developments in markers
of hypoxia, especially markers detectable by immuno-
histochemistry. These markers require no additional
intervention beyond an initial pre-treatment biopsy,
which is used to generate formalin-fixed or frozen
sections, and may be very suitable for widespread
clinical use. These markers are more applicable than
oxygen electrodes and provide a high resolution assay of
the distribution of hypoxia at the microregional level.
Albert van der Kogel discussed imaging the dynamics of
tumour hypoxia using immunohistochemical markers
[10]. Pimonidazole was injected before biopsy into patients
with head and neck cancer, and this can be regarded as an
exogenous marker of hypoxia. Following sectioning,
carbonic anhydrase CA9 (a HIF-dependent gene that has
been investigated as a putative endogenous hypoxia
marker) and blood vessels could also be visualized by
immunohistochemistry. Beautiful triple-staining images of
pimonidazole, CA9 and vessels were shown. Kaanders et
al [10] confirmed the widespread variation in hypoxia in
human tumours of the head and neck – in 43 squamous
cell carcinomas, tumour area staining positive for pimo-
nidazole ranged from 0.3% to 17.2%. Follow up of patients
after treatment showed a significant correlation between
vascular density and locoregional control, and a significant
negative correlation between pimonidazole binding and
both locoregional control (15 months) and disease-free
survival (2 years).
The British Journal of Radiology, July 2006
Commentary: Hypoxia – the legacy of L H Gray
In the laboratory, more sophisticated experiments are
possible because of the use of experimental models. For
example, van Laarhaven et al [11] injected two nitroimi-
dazole markers of hypoxia, CCI-103F and pimonidazole,
before and after treatment with nicotinamide and
carbogen (95% O2, 5% CO2) singly and in combination
into two different murine carcinomas. Bromo-
deoxyuridine and Hoechst 33342 were also used as
proliferation and perfusion markers, respectively. The
two tumours showed extensive differences in vascular
architecture, distribution patterns of hypoxia and bromo-
deoxyuridine labelling. All treatment combinations
caused a decrease in the hypoxic fraction, but the
responses of the two tumours were quantitatively very
different.
As an alternative to image analysis of histological
sections, flow cytometry was used by Bennewith and
Durand [12] to obtain important information about the
transient nature of hypoxia. Pimonidazole was injected
hourly into a human tumour xenograft for 8 h. This
causes a time-integrated asymptotic rise in the number of
cells showing the hypoxic marker. 1 h before sacrifice, a
second hypoxic marker (CCI-103F) was injected.
Examination by flow cytometry of the single and dual
fluorescence peaks from a large number of cells showed
that substantial numbers of cells that had been hypoxic
were no longer hypoxic immediately before sacrifice.
This dynamic behaviour is likely to have a big influence
on tumour management and particularly the application
of hypoxia-directed therapies.
Methods of monitoring hypoxia using non-invasive
imaging techniques are being explored. Nitroimidazoles
labelled with I-123 may be suitable for conventional
gamma camera imaging or with F-18 for PET imaging. A
recent paper by Nöth et al [13] described the use of 15C5-
loaded alginate capsules as fluorine-19 oxygen sensors in
MRI for in vivo determination of tumour oxygenation
during growth and in response to carbogen breathing.
Molecular targeting in radiation oncology
Addressing this subject, Michael Baumann pointed out
that radiotherapy is very efficient at reducing a tumour
to a small number of surviving clonogenic cells. There is
therefore scope for developing novel therapeutic agents
which, although perhaps not curative in themselves, may
be highly effective if used in combination with an
appropriate therapeutic regimen. Hypoxia is known to
have a role in: (i) selecting for cells that have lost
sensitivity to the tumour suppressor gene p53;
(ii) regulating genes involved in drug resistance; (iii) a
tendency to select for a more malignant phenotype;
(iv) increasing the mutation rate; (v) increasing expres-
sion of genes associated with angiogenesis; and
(vi) tumour invasion. There is therefore considerable
potential for developing new approaches to therapy
based on targeting hypoxia. Two examples discussed by
Baumann are summarized below.
A number of groups have shown a link between
expression of epidermal growth factor receptor (EGFR)
and tumour growth. For example, Ang et al [14] showed
that in a large series of patients with head and neck
squamous cell carcinomas, EGFR expression was a
The British Journal of Radiology, July 2006
strong independent prognostic indicator of overall and
disease-free survival. It was also a robust predictor of
locoregional relapse, but not of distant metastases. The
authors recommended that EGFR immunohisto-
chemistry should be considered for selecting patients
for more aggressive combined therapies or enrolment in
trials targeting EGFR signalling pathways. Baumann and
Krause [15] had reviewed the evidence that inhibition of
EGFR can increase radiosensitivity of clonogenic cells
and tumour cell proliferation, and Krause et al [16] have
shown that in nude mice the anti-EGFR monoclonal
antibody can significantly reduce the tumour control
dose with decreased repopulation and reoxygenation.
The second example targets tumour angiogenesis.
Adjuvant inhibition of the vascular endothelial growth
factor receptor (VEGFR) with a tyrosine kinase inhibitor
after fractionated irradiation prolongs tumour growth
[17]. Zips et al [18] have confirmed this finding for a
human squamous cell carcinoma grown in nude mice
and exposed to the VEGFR inhibitor for 75 days after
irradiation, but showed that the 50% tumour control
dose was no different. The authors conclude that
recurrences depend on VFGF-driven angiogenesis but
surviving tumour cells retain their clonogenic potential
during this treatment.
Other strategies for targeting therapies based on
hypoxia, not discussed in detail at the meeting, include
hypoxia-activated prodrugs, hypoxia-selected gene ther-
apy, and the use of genetically-engineered anaerobic
bacteria [19]. Molecular-based approaches targeted to
hypoxia-mediated processes add a further dimension to
the complexity of the problem since overall and disease-
free survival, locoregional control, and control of
metastases rarely give concordant results in animal
experiments and clinical trials.
Clinical trials
Van der Kogel also outlined the background to the
present Phase 3 clinical trials of ARCON (accelerated
radiotherapy with carbogen and nicotinamide) in head
and neck and bladder cancers [20]. This concept, which
was pioneered at the Gray Laboratory, is designed to
attack tumour cells that may have more than one
resistance mechanism (i.e. acute and chronic hypoxia)
and therefore may be responsive to combined modality
treatments. Furthermore, tumour repopulation is a major
cause of poor local tumour control, which generally
deteriorates as treatment time increases (at fixed dose).
Thus the first arm of the ARCON attack is to use
accelerated fractionated radiotherapy with several frac-
tions per day. Inhalation of carbogen is designed to
decrease diffusion-limited hypoxia and nicotinamide is
used to decrease perfusion-limited hypoxia, although it
may also have other effects.
ARCON is a good example of translational research.
Pre-clinical studies, mainly on animals, have shown that
each of the three components can be effective, both in
isolation and in combination. For example, tumour
control rate for a mouse mammary carcinoma was the
same with the combination treatment with almost 50%
less radiation dose. Phase 1 and 2 clinical trials showed
that the ARCON regimen was feasible and tolerable in
547

patients and produced promising results in terms of
tumour control. For the Phase 3 trial that is nearing
completion, selection of patients has been mainly based
on clinical and histopathological tumour characteristics.
It is already clear that, whatever the outcome, improved
selection of patients based on measures of hypoxia will
be desirable, if not essential, to achieve good therapeutic
outcomes, as may now be appreciated from previous
hypoxic cell radiosensitizer trials.
Widening horizons
Two speakers took a broader view of some aspects of
radiobiology. Using the title Science spanning the genera-
tions, Ian Stratford reviewed a number of other areas in
which early work by Gray was taken forward by
colleagues and later generations with varying degrees
of success. One phenomenon that attracted the attention
of Gray and co-workers was the effect of nitric oxide on
the radiosensitivity of tumour cells [21, 22]. Nitric oxide
appeared to be able to take the place of oxygen in
enhancing sensitivity to X-rays in bacteria, plant mer-
istem and mammalian tumour cells. Furthermore, cell
survival curves showed that sensitization by nitric oxide
was exactly the same as that achieved with oxygen when
added to anoxic conditions. It is now known that nitric
oxide is a biochemical signalling molecule, generated in
tissues by nitric oxide synthase (NOS). NOS levels are
elevated in a wide range of tumour types (see, for
example, Ambs et al [23]). Stratford showed that
cytokine-mediated induction of NOS radiosensitized
tumour cells at intermediate oxygen tensions in vitro,
hence suggesting that varying levels of NOS in human
tumours could be a significant factor in radiotherapy.
Another topic studied by Gray was the dissociation of
oxyhaemoglobin [24]; Adams, Stratford and colleagues
subsequently explored enhancement of hypoxia by
compounds such as BW12C which influenced this
dissociation. They showed that these compounds could
increase tumour hypoxia and thereby be useful when
combined with hypoxia-selective bioreductive drugs
[25]. An alternative approach has been taken in the
development of efaproxiral (RSR13), which reduces
haemoglobin–oxygen binding to decrease hypoxia [26];
this agent is now in clinical trial in combination with
radiotherapy.
Stratford also discussed the extensive research on the
development of bioreductive drugs worldwide, some
pioneered at the Gray Laboratory [19, 25, 27], and the
potential for utilizing the presence of hypoxia in tumours
to drive selective gene therapy [28]. Finally, he returned
to the impact of modulating HIF function on tumour
response to radiotherapy, presenting data to support the
hypothesis that in HIF-1 deficient tumours, hypoxic cells
have an extremely limited lifetime and therefore may not
contribute to radiation sensitivity [29].
Gillies McKenna outlined the research programme of
the newly-established Department of Radiation
Oncology and Biology at the University of Oxford,
which will incorporate the Gray Laboratory when it is
moved from its current location at Mount Vernon
Hospital. The vision of the Department is to continue
the tradition of the ‘‘Gray Lab’’; be research focused and
548
P P Dendy and P Wardman
clinically-relevant; learn lessons from the past; be flexible
to encompass all relevant areas in the future; comple-
ment existing groups in Oxford and elsewhere in the UK;
and be a centre for training all health care personnel in
radiation oncology and biology. Multidisciplinary colla-
boration will be a key feature of the Department, and two
examples were given.
First, in the 21st century it will be necessary to move
on from empirical therapy, tailored by experience, to
targeted therapy. Maximizing the information from
tumour imaging will involve diagnostic radiologists,
medical physicists, radiographers, and others. Pushing
back the boundaries of physical treatment involves
radiotherapists, radiation physicists and industry. To
these groups must be added a broad spectrum of
biologists who will open up new frontiers. McKenna
and his colleagues are seeking to define biological
processes that regulate the radiation responsiveness of
solid tumours, and take agents that target these
processes into the clinic in order to improve therapeutic
outcome.
Second, there will be strong emphasis on transla-
tional research. Examples of areas where laboratory
workers and clinicians must work in close collaboration
include signal transduction inhibitors with radiation,
systematic targeting of radiation with Auger electron-
emitting radionuclides, molecular targets for modifying
the radiation response, and the molecular basis of
metastases.
Conclusions
Hal Gray was a true polymath: introducing their
biography of him, Loutit and Scott discussed Gray’s
‘‘escape from this confinement’’ of the ‘‘cult of specia-
lization for learning more and more about less and less’’
[30]. One of us (PW), writing in 1982 in a booklet to
commemorate the silver jubilee of the opening of the
Gray Laboratory, asserted ‘‘… [Gray] must have been the
first – and quite possibly the last – scientist to have a
thorough appreciation of current activity in all four
sectors of radiation research – physics, chemistry,
biology and medicine.’’ While no one subject can sum
up his work as a scientist, overcoming tumour hypoxia is
a challenge that requires such a broad approach. Had he
not died so young he might well have received a Nobel
Prize. Some of his contributions are now well under-
stood, for example radiation dosimetry, where Gray’s
name is recorded for posterity as the SI unit of absorbed
dose. Other areas, especially hypoxia in tumours,
represent a far more complex problem than Gray could
possibly have imagined, with heterogeneity of behaviour
in almost every aspect investigated.
In future, human tumours must be better characterized
on an individual basis. Performing clinical trials on
unselected patients who have a mixture of hypoxic and
better-oxygenated tumour cells runs the clear risk of
rejecting a treatment that could be of significant benefit
to a sub-set of patients. Loutit and Scott described Gray
as ‘‘the Fellow [of The Royal Society] who fathered
radiobiology’’ [30], but it is arguable that drawing
attention to the importance of tumour hypoxia is the
most important legacy of L H Gray.
The British Journal of Radiology, July 2006
Commentary: Hypoxia – the legacy of L H Gray
Acknowledgments
We thank Prof. I J Stratford for helpful comments on a
draft manuscript. PW is supported by Cancer Research
UK.
References
1. Gray LH. The initiation and development of cellular
damage by ionising radiations. Br J Radiol 1953;26:609–18.
2. Mottram JC. A factor of importance in the radiosensitivity
of tumours. Br J Radiol 1936;9:606–14.
3. Gray LH, Conger AD, Ebert M, Hornsey S, Scott OCA. The
concentration of oxygen dissolved in tissues at the time of
irradiation as a factor in radiotherapy. Br J Radiol
1953;26:638–48.
4. Tomlinson RH, Gray LH. The histological structure of some
human lung cancers and the possible implications for
radiotherapy. Br J Cancer 1955;9:539–49.
5. Hall EJ. Radiobiology for the radiologist, 5th edn.
Philadelphia, PA: Lippincott, Williams and Wilkins,
2000:108–9, 458.
6. Overgaard J, Hansen HS, Overgaard M, Bastholt L,
Berthelsen A, Specht L, et al. A randomized double-blind
phase III study of nimorazole as a hypoxic radiosenstizer of
primary radiotherapy in supraglottic larynx and pharynx
carcinoma. Results of the Danish Head and Neck Cancer
Study (DAHANCA) Protocol 5-85. Radiother Oncol
1998;46:135–46.
7. Harris AL. Hypoxia – a key regulatory factor in tumour
growth. Nat Rev Cancer 2002;2:38–47.
8. Pugh CF, Ratcliffe PJ. Regulation of angiogenesis by
hypoxia – role of the HIF system. Nat Med 2003;9:677–84.
9. Koritzinsky M, Seigneuric R, Magognin MG, van der
Beuchen T, Lambin P, Wouters BG. The hypoxic proteome
is influenced by gene specific change in m-RNA transcrip-
tion. Radiother Oncol 2005;76:177–86.
10. Kaanders JH, Wijffels KI, Marres HA, Ljungkvist AS, Pop
LA, van den Hoogen FJ, et al. Pimonidazole binding and
tumour vascularity predict for treatment outcome in head
and neck cancer. Cancer Res 2002;62:7066–74.
11. van Laarhoven HWM, Bussink J, Lok J, Punt CJA,
Heerschap A, van Der Kogel AJ. Effects of nicotinamide
and carbogen in different murine colon carcinomas:
immunohistochemical analysis of vascular architecture
and microenvironmental parameters. Int J Radiat Oncol
Biol Phys 2004;60:310–21.
12. Bennewith KL, Durand RE. Quantifying transient hypoxia
in human tumour xenografts by flow cytometry. Cancer Res
2004;64:6183–9.
13. Nöth U, Rodrigues LM, Robinson SP, Jork A, Zimmermann
U, Newell B, Griffiths JR. In vivo determination of tumour
oxygenation during growth and in response to carbogen
breathing using 15C5 loaded alginate capsules as fluorine-
19 magnetic resonance imaging oxygen sensors. Int J Radiat
Oncol Biol Phys 2004;60:909–19.
14. Ang KK, Berkey BA, Tu X, Zhang HZ, Katz R, Hammond
EH, et al. Impact of epidermal growth factor receptor
expression on survival and pattern of relapse in patients
The British Journal of Radiology, July 2006
with advanced head and neck carcinoma. Cancer Res
2002;62:7350–6.
15. Baumann M, Krause M. Targeting the epidermal growth
factor in radiotherapy: radiobiological mechanisms, pre-
clinical and clinical results. Radiother Oncol 2004;72:
257–66.
16. Krause M, Ostermann G, Petersen C, Yaromina A, Hessel F,
Harstrick A, et al. Decreased repopulation as well as
increased reoxygenation contribute to the improvement in
local control after targeting of the EGFR by C225 during
fractionated irradiation. Radiother Oncol 2005;76:162–7.
17. Williams KJ, Telfer BA, Brave S, Kendrew J, Whittaker L,
Stratford IJ, et al. ZD6474, a potent inhibitor of vascular
endothelial growth factor signaling, combined with radio-
therapy: schedule-dependent enhancement of antitumor
activity. Clin Cancer Res 2004;10:8587–93.
18. Zips D, Hessel F, Krause M, Schiefer Y, Hoinkis C, Thames
HD, et al. Impact of adjuvant inhibition of vascular
endothelial growth factor receptor tyrosine kinase on
tumour growth delay and local tumour control after
fractionated irradiation of human squamous cell carcinoma
in nude mice. Int J Radiat Oncol Biol Phys 2005;61:908–14.
19. Brown JM, Wilson WR. Exploiting tumour hypoxia in
cancer treatment. Nat Rev Cancer 2004;4:437–47.
20. Kaanders JH, Bussink J, van der Kogel AJ. ARCON A novel
biology-based approach in radiotherapy. Lancet Oncol
2002;3:728–37.
21. Gray LH, Green FO, Howes CA. Effect of nitric oxide on the
radiosensitivity of tumour cells. Nature 1958;182:952–3.
22. Dewey DL. The effect of oxygen and nitric oxide on the
radiosensitivity of human cells in tissue culture. Nature
1960;186:780–2.
23. Ambs S, Bennett WP, Merriam WG, Ogunfusika MO, Oser
SM, Khan MA, et al. Vascular endothelial growth factor and
nitric oxide synthetase expression in human lung cancer
and the relation to p53. Br J Cancer 1998;78:223–39.
24. Gray LH, Steadman JM. Determination of the oxyhaemo-
globin dissociation curves for mouse and rat blood. J
Physiol 1964;175:161–71.
25. Adams GE, Stratford IJ. Bioreductive drugs for cancer
therapy – the search for tumour specificity. Int J Radiat
Oncol Biol Phys 1994;29:231–8.
26. Choy H, Nabid A, Stea B, Scott C, Roa W, Kleinberg L, et al.
Phase II multicenter study of induction chemotherapy
followed by concurrent efaproxiral (RSR13) and thoracic
radiotherapy for patients with locally advanced non-small-
cell lung cancer. J Clin Oncol 2005;23:5918–28.
27. Foster JL, Conroy PJ, Searle AJ, Willson RL. Metronidazole
(Flagyl): characterization as a cytotoxic drug specific for
hypoxic tumour cells. Br J Cancer 1976;33:485–90.
28. Dachs GU, Patterson AV, Firth JD, Ratcliffe PJ, Townsend
KM, Stratford IJ, et al. Targeting gene expression to hypoxic
tumor cells. Nat Med 1997;3:515–20.
29. Williams KJ, Telfer BA, Xenaki D, Sheridan MR, Desbaillets
I, Peters HJ, et al. Enhanced response to radiotherapy in
tumours deficient in the function of hypoxia-inducible
factor-1. Radiother Oncol 2005;75:89–98.
30. Loutit JF, Scott OCA. Louis Harold Gray 1905–1965. Elected
F.R.S. 1961. Biogr Mem Fellows R Soc 1966;12:195–217.
549
 The British Journal of Radiology, 79 (2006), 550–553

COMMENTARY

Controversies in non-accidental head injury in infants

N STOODLEY, FRCS, FRCR

Consultant Neuroradiologist, Department of Neuroradiology, Frenchay Hospital, Bristol BS16 1LE,

UK

ABSTRACT. Non-accidental head injury in infants is not uncommon and is associated

with significant morbidity and mortality. It is therefore important to identify it at the Received 23 January 2004
earliest opportunity so that appropriate intervention can be made which protects the Revised 2 February 2006
child from further harm. The whole topic is controversial and the aim of this paper, in Accepted 27 March 2006
question and answer format, is to review some of the more controversial areas to give
DOI: 10.1259/bjr/23921951
an overview of the neuroimaging features of this condition. The author has drawn on
his clinical and medicolegal experience of these cases, and the review is based upon ’ 2006 The British Institute of
questions commonly encountered in Court. Radiology

Child abuse is a controversial subject and aspects
relating to non-accidental head injury (NAHI) are no
exception, not least because the limited evidence base
hampers professionals working in the field.
The main role of the Expert Witness is to interpret and
explain to the Court material which is within the Expert’s
area of expertise, but which is outwith the experience of
the Judge or the jury. The Expert Witness is in the
privileged position of being allowed to give opinion
evidence to the Court, but this brings a responsibility to
ensure that the opinion given is both reasonable and
capable of withstanding logical analysis. An Expert
Witness has an overriding duty to the Court that takes
precedence over any obligation to those Instructing the
Expert. The Expert should be independent, impartial and
confine their opinions to their particular area of
expertise. All sides in Court, will rightly question this
evidence and the Expert must be able to justify their
opinion. However, just as lawyers’ opinions on the
interpretation of points of law sometimes differ, doctors
reading the same scientific papers may come to different
conclusions; hence controversy!
What are the ‘‘typical’’ neuroimaging features
of NAHI?
There are no absolutely pathognomonic features that
could only be due to NAHI in all cases. Why should
there be when we are only looking at the effects of a
traumatic episode on specific tissues in a biological
system? The brain has limited ways in which it can
respond to various insults, so it is not surprising that
some cases of accidental trauma show imaging appear-
ances very similar to those seen in NAHI.
A fairly common finding in cases of NAHI is of
shallow subdural haematomas (SDH) at various, sepa-
rate sites over the cerebral hemispheres and often in the
posterior fossa. The subdural blood may be of different
ages, although this is by no means always the case. The
most common site for SDH following both accidental
trauma and NAHI is over the cerebral convexities. SDH
following all but severe accidental trauma, however,
tends to be unifocal whereas SDH in NAHI tends to be
multifocal. SDH at sites other than over the convexities,
especially if in the posterior interhemispheric fissure or
posterior fossa, is more likely to occur following NAHI
than following accidental trauma [1]. Subdural blood can
be seen following an impact injury from any cause, but it
is usually related to the site of the impact and is often
associated with a fracture. Subdural blood seen away
from the point of impact is unusual in accidental trauma,
unless severe.
Focal parenchymal lesions such as contusions, haema-
tomas and shearing injuries may be seen following both
accidental and non-accidental head trauma; when seen
following accidental trauma there is usually a very clear
history of a major traumatic event. Shearing injuries
(diffuse axonal injury or the larger gliding contusions)
were thought to be very common in NAHI. However,
recent neuropathological and neuroimaging evidence
suggests that axonal injury is uncommon even in infants
who die following NAHI and that hypoxic–ischaemic
changes are more common in these infants [2, 3].
Often there is evidence of reduced grey-white differ-
entiation focally or more generalized on head scans
following NAHI and, given the neuropathology, it
would seem likely that these scan changes are related
to hypoxic–ischaemic changes in the brain.
There is no absolute scientific evidence that
shaking causes the injuries in ‘‘shaken baby
syndrome’’ is there, doctor?
Preferable terms are, abusive head trauma or NAHI
which describe what has happened, but do not make any
assumptions about mechanism. However, shaking may
be an important mechanism in the majority of cases.
There must be something very different about the

550 The British Journal of Radiology, July 2006

Commentary: Non-accidental infant head injury
mechanism of injury in NAHI as opposed to accidental
head trauma as the clinical presentation is often
different, the neuroimaging appearances are different
and the outcome, both in terms of mortality and
morbidity in the short and the long term, is different
[4]. The main differentiating factor could well be the
different mechanism of injury.
There is no absolute scientific proof that shaking
causes these injuries because it is not possible to perform
the relevant scientific experiment, of shaking infants of
different ages and sizes with different degrees of force
and then performing sequential scans; and then studying
them clinically and correlating this with scan appear-
ances. In the absence of such data it is necessary to rely
on various other sources of evidence including animal
experiments, from the neuroimaging appearances of
infants following witnessed accidental trauma and
imaging evidence from cases of NAHI in which there
has been a reliable confession.
Primate experiments have been performed in which
the animals were subjected to pure translational and/or
rotational forces without impact. The head injuries seen
in the animals were very similar to those seen in NAHI
[5]. Scans of infants and children who have sustained
head injuries following accidental head trauma may
show the same pattern of SDHs at different sites (as in
NAHI), but usually only where the accidents involved
severe forces such as following road traffic accidents or
major falls. Most domestic accidents do not involve this
degree of force and, given the number of domestic
incidents that occur daily involving children bumping
their heads, the fact that our departments are not
inundated with scan requests for infants suggests that
the majority of these falls are neurologically benign, as
has been shown in the literature [6].
There is debate as to whether shaking alone is
sufficient to cause all of the features of NAHI or whether
some impact (even against a soft surface) is always
necessary. Some of the animal work cited above suggests
that impact is not required. There are biomechanical
models that show that the amount of force generated
during a pure shake is much less than following an
impact, but there are also reports in the literature of fatal
cases of NAHI where there has been no evidence of
impact injury even on post-mortem. The focal brain
injuries sometimes seen in NAHI may reflect associated
impact or the result of very severe shaking.
Some perpetrators admit to shaking and many a loving
parent must have been pretty close to it, having been up
for hours at night with an inconsolable infant that they
have done everything to try to settle. It is possible that
only the minority of cases are due to wilful cruelty and
some authors have advocated changing the way in which
we classify abuse partly because of factors such as these
[7].
The subdurals are very thin and not causing
significant mass effect, so what is the cause of
the presenting symptoms?
These infants present with varied and non-specific
signs and symptoms which may be the same or similar to
those seen in infants who are unwell for a variety of
The British Journal of Radiology, July 2006
reasons. These range from being ‘‘off-colour’’ and
grizzly, off feeds to vomiting, with various degrees of
reduced levels of consciousness, through fits to frank
coma and death. The severity of the symptoms and signs
does not relate to the size or number of SDH, but does
have some relation to the degree of associated brain
hypoxic–ischaemic injury and/or swelling. Those infants
presenting in coma and fitting are more likely to have
generalized brain changes on their initial CT scans when
compared with infants presenting with lesser symptoms
[8]. There have also been a few publications showing that
the pattern of changes on diffusion weighted imaging in
NAHI is that of hypoxic–ischaemic change rather than
diffuse axonal (shearing) injury [9].
This neuroimaging evidence is supported by the
neuropathological literature. Geddes et al published
two papers after studying a cohort of fatal cases of
NAHI. One paper [2] was a review of 53 cases, 37 infants
ranging in age from 20 days to 9 months and 16 children
ages ranging from 13 months to 8 years. The authors
demonstrated that most of the brain damage seen in
these fatal cases was due to hypoxic vascular damage
rather than traumatic axonal damage. Traumatic diffuse
axonal injury was only seen in three cases.
A second paper by this group [3] was based upon 39
infants aged 9 months or under where the authors
compared the neuropathology with that seen in 14
control patients. Traumatic diffuse axonal injury was
only found in two of the NAHI patients, both of who had
suffered severe head injury with multiple fractures.
If hypoxic–ischaemic injury is the major cause of death
in fatal cases, it is logical to assume that it is an important
factor in non-fatal cases, and that the degree of
associated hypoxic–ischaemic change largely determines
presenting symptoms in the short term as well as
outcome in terms of mortality and morbidity. If this is
correct then it is likely that infants who sustain a severe
injury will lose consciousness at the time of the injury
and some may not recover. At the other end of the
spectrum, the hypothesis would suggest that there
should be a group of infants who sustain an injury
which may cause a minor change in behaviour insuffi-
cient to cause their carers to take them to a doctor and
where the brain (and therefore the infant) recovers from
the injury spontaneously.
In between these extremes lies a group of infants in
whom there is a probably quite a marked change in
behaviour following the injury. In less susceptible
infants, the brain may be able to recover spontaneously
from the injury whereas in more susceptible infants some
threshold may be reached beyond which the brain
cannot recover from the insult by itself. This latter group
of infants would go on to deteriorate clinically and
present with an encephalopathic illness. In this way,
hypoxic–ischaemic change following NAHI may be in
some way analogous to birth asphyxia where infants
may develop hypoxic–ischaemic encephalopathy some
hours after the insult.
It is the different degree and extent of hypoxic–
ischaemic injury that determines whether and how these
infants will present to medical attention and that the
SDHs are just markers of the mechanism of injury.
Sometimes the SDHs appear so insignificant on the scan
551

of an extremely sick infant that they may be overlooked
and their true significance not realised.
What else could have caused the SDH?
All other possible causes of subdural effusions and
haematomas have to be considered in all of these cases.
These other causes can usually be diagnosed or excluded
with a combination of proper history taking and relevant
investigations. Severe previous accidental or birth
trauma can lead to the development of SDH. SDH is
probably quite common following all modes of delivery
but, in otherwise normal infants, they do not persist
beyond the first few weeks of life [10]. Infections such as
meningitis can lead to infected subdural collections
(empyemas) and these could be associated with some
haemorrhage, although clinical experience would sug-
gest that bleeding into empyemas is not very common.
Infants with coagulopathies may bleed spontaneously
or after minimal trauma, but this again is surprisingly
uncommon. The pattern of intracranial haemorrhage in
infants with clotting disorders also tends to be different
as intraparenchymal haemorrhage is much more com-
mon than SDH in these infants [11]. Congenital abnorm-
alities including vascular malformations or metabolic
conditions such as glutaric aciduria may predispose to
SDH.
When did the bleeding occur?
Estimating the age of subdural blood on CT and MR is
not precise because various factors influence the appear-
ance of blood on both modalities. On CT, acute blood
should be of high attenuation, but this assumes that the
blood has clotted, that the patient is not severely anaemic
and that there has been no significant dilution of the
subdural blood by CSF following a traumatic tear of
the arachnoid. In severe shaking injuries, CSF dilution of
the subdural blood may make acute SDH appear of low
attenuation and be misinterpreted as chronic. The time
course of attenuation changes is also variable depending
on such factors as the volume of blood and the
haemoglobin level of the patient at the time of the bleed.
Different factors apply to the dating of blood on MR
scans, but when there is a combination of CT and MR
scans it is often possible to give a range of probable ages
for the subdural collections present. This does not
necessarily help with the timing of the injury, however,
as the SDH is unlikely to be responsible for the
presenting symptoms. Usually, the history gives a better
assessment of the timing of injury than the radiology
because an infant that has suffered a NAHI of sufficient
severity to lead to admission to hospital is in my
experience extremely unlikely to have behaved comple-
tely normally after that injury was inflicted.
What about re-bleeding into chronic SDH?
In the elderly population, minimal trauma can trigger
re-bleeding into a chronic SDH. The basic pathology
must be similar in infants, but the incidence of chronic
552
N Stoodley
subdural haematoma in infants is extremely low. In the
elderly population, the brain is undergoing involutional
change, so any blood within the subdural space may
persist allowing the haematoma to become chronic. In a
normal infant, the situation is completely different as the
brain is growing rapidly and this growth may act to
minimize the potential for any material to persist in the
subdural space. If the brain does not grow properly due
to a congenital or acquired insult of sufficient severity,
SDH may also become chronic in infants. The most
common previous insult is NAHI. The presence of a
chronic subdural haematoma in an infant who has not
had an insult such as severe birth injury, a major
accident, meningitis or a known underlying brain
metabolic abnormality should raise the possibility of
previous NAHI [12].
What degree of force is required to produce
these injuries?
This is unknown. These injuries are not seen following
‘‘normal’’ domestic trauma or rough play, otherwise
there would be many cases every day throughout the
UK. Therefore, it is likely the degree of force is such that
an independent witness would realise that it was likely
to cause harm. By the same token, it is possible to
severely injure an infant without intending to cause them
harm. The active intent in most of these cases is to stop
them crying.
As we see similar appearances to NAHI following
severe accidental trauma, it is sometimes suggested that
the degree of force involved in NAHI must be similar to
that occurring during road accidents. This is incorrect, as
the mechanism of the two injuries is different: high
velocity impact with acceleration/deceleration in acci-
dents; lower velocity rotation and apnoea in NAHI.
The radiological investigation of NAHI
This has been comprehensively covered in recent
publications [13, 14] and the rationale for the imaging
approach discussed [15], emphasising the importance of
using both CT and MR in these cases. CT should be the
initial investigation because it is good at demonstrating
acute blood, it is more widely available than MR and it is
much simpler to perform a CT scan on a sick infant. MR
is better at showing older collections of blood, blood in
sites not well seen on CT (such as the middle cranial
fossa), low volume haematomas and, of course, is
vastly superior to CT in the demonstration of parenchy-
mal brain injuries. Both modalities are therefore
required for a full neuroradiological assessment of these
infants.
Many of these cases are obvious, others are not and we
won’t find the evidence unless we look. Think of SDH in
the same way as metaphyseal and rib fractures: we look
for these fractures assiduously with skeletal surveys but,
although they are markers of mechanisms of injury, they
are of little long-term consequence. Shouldn’t we be
looking for the SDH that is also marker of a mechanism
of head injury that causes most of the morbidity and
mortality in these cases?
The British Journal of Radiology, July 2006
Commentary: Non-accidental infant head injury
Conclusions
To miss abuse risks sending the child back into an
abusive environment; to suggest abuse where there has
been none can tear a family apart. These difficult
decisions can only be taken if we have sufficient
information on which to base them. The radiologist
may be the first clinician to suggest the possibility of
child abuse and may therefore become involved in the
legal proceedings that inevitably, and rightly, follow
many of these cases.
In the summer of 2005, four joined appeals against
convictions in cases of alleged non-accidental head injury
were heard in the Court of Appeal. The appellants had
been convicted of murder (1); manslaughter (2) and
grievous bodily harm (1). The Court heard from 21
experts and the Judgment [16] contains much invaluable
guidance on the approach to take in possible NAHI
cases. The Judgment emphasises that each of these cases
is fact-specific and each should be determined on their
individual facts, and also that not all cases where the so-
called triad (encephalopathy, subdural and retinal
haemorrhage) is present will be due to NAHI. Having
heard the evidence in these four cases, one murder
conviction was reduced to manslaughter, two convic-
tions were quashed (manslaughter and grievous bodily
harm) and an appeal against a conviction for man-
slaughter was dismissed.
References
1. Ewing-Cobbs L, Prasad M, Kramer L, Louis PT,
Baumgartner J, Fletcher JM, et al. Acute neuroradiologic
findings in young children with inflicted or non inflicted
traumatic brain injury. Child’s Nervous System 2000;16:
25–34.
2. Geddes JF, Hackshaw AK, Vowles GH, Nickols CD,
Whitwell HL. Neuropathology of inflicted head injury in
children. 1. Patterns of brain damage. Brain 2001;124:
1290–8.
The British Journal of Radiology, July 2006
3. Geddes JF, Hackshaw AK, Vowles GH, Nickols CD,
Whitwell HL. Neuropathology of inflicted head injury in
children. II: Microscopic brain injury in infants. Brain
2001;124:1299–306.
4. Duhaime AC, Christian CW, Moss E, Seidl TS. Long term
outcome in infants with the shaking impact syndrome.
Pediatr Neurosurg 1996;24:292–8.
5. Ommaya AK, Gennarelli TA. Cerebral concussion and
traumatic unconsciousness. Brain 1974;97:633–54.
6. Warrington SA, Wright CM. Accidents and resulting
injuries in premobile infants: data from the ALSPAC study.
Arch Dis Childhood 2001;85:104–7.
7. Southall DP, Samuels MP, Golden MH. Classification of
child abuse by motive and degree rather than type of injury.
Arch Dis Childhood 2003;88:101–4.
8. Kemp AM, Stoodley N, Cobley C, Coles L, Kemp KW.
Apnoea and brain swelling in non-accidental injury. Arch
Dis Childhood 2003;88:472–6.
9. Biousse V, Suh DY, Newman NJ, Davis PC, Mapstone TB,
Lambert SR. Diffusion weighted magnetic resonance
imaging in shaken baby syndrome. Am J Ophthalmol
2002;133:249–55.
10. Whitby EH, Griffiths PD, Rutter S, Smith MF, Sprigg A,
Ohadike P, et al. Frequency and natural history of subdural
haemorrhages in babies and relation to obstetric factors.
Lancet 2004;362:846–51.
11. Vorstman EBA, Anslow P, Keeling DM, Haythornthwaite
G, Bilolikar H, McShane T. Brain haemorrhage in five
infants with coagulopathy. Arch Dis Childhood
2003;88:1119–21.
12. Feldman KW, Bethel R, Shugerman RP, Grossman DC,
Grady MS, Ellenbogen RG. The cause of infant and toddler
subdural hemorrhage: a prospective study. Pediatrics
2001;108:636–46.
13. Jaspan T, Griffiths PD, McConachie NM, Punt JAG.
Neuroimaging for non-accidental head injury in childhood:
A proposed protocol. Clin Radiol 2003;58:44–53.
14. Kemp AM. Investigating subdural haemorrhage in infants.
Arch Dis Childhood 2002;86:98–102.
15. Stoodley N. Neuroimaging in non-accidental head injury: if,
when, why and how. Clin Radiol 2005;60:22–30.
16. R v Harris and Ors [2005] EWCA Crim 1980.
553
 The British Journal of Radiology, 79 (2006), 554–561

REVIEW ARTICLE

Advanced imaging applied to radiotherapy planning in head and

neck cancer: a clinical review
1
K NEWBOLD, MRCP, FRCR, 2M PARTRIDGE, PhD, CPhys, 2G COOK, MD, FRCP, FRCR, 1S A SOHAIB, BSc, MRCP,
2 1 1
FRCR, E CHARLES-EDWARDS, MSc, CSci, P RHYS-EVANS, FRCS, K HARRINGTON, MRCP, FRCR and
1
C NUTTING, MD, MRCP, FRCR

1
The Royal Marsden NHS Trust, Fulham Road, London SW3 6JJ and 2The Royal Marsden NHS Trust,
Downs Road, Sutton, Surrey SM2 5PT, UK

ABSTRACT. Head and neck squamous cell carcinoma represents an ideal model to
investigate the application of recent advances in medical imaging to radiotherapy
planning. Tumours usually remain localized, and are potentially curable with local
radiation. The steep radiation dose–response relationships support the strategies of
radiation dose escalation to increase local control. Two-dimensional simulator-based
planning and CT planning have significant drawbacks in terms of accurate target Received 3 February 2005
volume definition. MRI has enhanced soft tissue delineation, but has to be fused with Revised 23 March 2006
CT to allow dose calculation. Functional imaging using dynamic contrast enhanced CT Accepted 6 April 2006
or MRI sequences may allow improved knowledge of tumour function. Positron
DOI: 10.1259/bjr/48822193
emission tomography (PET) may allow further physiological information to be
determined. This review summarizes the current techniques in clinical development in ’ 2006 The British Institute of
this area. Radiology

Head and neck cancer (HNC) is the seventh most
common cancer in the UK. Whilst early stage disease has
a high cure rate, stage III and IV HNC still have poor
rates of local control and survival [1]. Surgery and/or
radiotherapy aim to achieve locoregional control, and
where it enables organ preservation with maintenance of
function, radiotherapy is the modality of choice. Failure
to achieve locoregional control may be due to geogra-
phical miss of primary tumour or nodes due to under-
staging disease, intrinsic radioresistance of the tumour,
or factors such as hypoxia and proliferation known to
reduce radiosensitivity of a tumour. Three-dimensional
conformal radiotherapy (3D-CRT) and intensity-modu-
lated radiotherapy (IMRT) aim to achieve better loco-
regional control and improve survival by radiation dose
escalation, but such techniques demand more accurate
localization of tumour and normal tissues with non-
invasive imaging techniques.
Anatomical imaging for head and neck
radiotherapy planning
Current standard radiotherapy imaging
In the UK, the most common methods of target
localization for radiotherapy planning include the use
of the simulator and CT scanning. In the simulator,
standard radiotherapy field borders are placed on
identifiable bony landmarks, and simple outlines are
taken at levels though the treatment volume. For most
patients, this approach gives the radiation oncologist a
‘‘one size fits all’’ treatment. There is limited opportunity
for individualization of treatment volumes, although
simple customization is possible with reference to bony
landmarks within individual cases. It results in treatment
fields with large margins around tumours to account for
uncertainty of target position and conservative doses due
to uncertainties in dose to organs at risk (OARs).
CT planning provides a 3D representation of the target
volume and OARs allowing improved target definition
and also accurate OAR delineation. The use of iodinated
contrast agents increases sensitivity, and CT with
contrast remains the best modality for defining nodal
disease, cartilage invasion, or bone destruction [2, 3].
Reconstruction of high resolution coronal and sagittal
sections may aid RT planning. CT images are most
accurate when tumour interfaces are with air cavity, fat,
or bone such as in the paranasal sinuses or neck, but are
less accurate when the tumour interfaces with normal
soft tissue such as in the tongue, or when there may be
mucosal spread for example in the larynx and hypo-
pharynx [4]. CT planning generates electron density data
for photon dose calculation, and calculation of dose–
volume histograms. CT planning allows accurate con-
formal shaping of the radiotherapy portals around the
target and shielding of the OARs (Figure 1). CT planning
removes many of the uncertainties of position of tumour
and OAR. However, CT planning per se does not
necessarily allow radiation dose escalation, especially if
the adjacent OARs are neurological. The spinal cord,
brain stem and optic nerves are assumed to have a serial

554 The British Journal of Radiology, July 2006

Review article: Advanced imaging applied to radiotherapy planning in HNC
organization of functional subunits, where inactivation
of a single subunit causes loss of function of the whole
organ. In this setting, the risk of late toxicity is
determined by the maximum dose received by the
OAR, and therefore partial reductions of the volume of
tissue irradiated, such as those offered by 3D-CRT, do
not reduce the risk of late toxicity. In such cases the more
advanced radiation delivery techniques such as IMRT,
which can produce dose distributions with concave
isodose surfaces, may be required [5, 6].
Use of MRI for treatment planning in head and
neck cancer
The main advantage of MRI over CT is significantly
greater soft tissue contrast, permitting better definition of
disease extent and OAR [7]. T1 weighted images give
good anatomical detail whilst T2 weighted images can
differentiate between normal and pathological tissues.
Image contrast can be enhanced by intravenous gadoli-
nium. Artefacts arising from dental amalgam can be
made significantly less conspicuous on MRI than CT.
MRI is the modality of choice for imaging tumours of the
base of tongue and lesions arising at the skull base [8, 9],
also visualization of important OAR such as the orbit,
optic nerves and chiasm and central nervous system.
Emami et al showed that the co-registration of two
anatomical modalities, MRI and CT improved the target
definition of cancer of the nasopharynx [10].
The use of MRI in radiotherapy treatment planning is
limited by the presence of geometrical distortions, includ-
ing inhomogeneities within the magnetic field and
non-linearities in the spatial-encoding gradients. These
distortions increase as the distance from the isocentre
increases and have been reported to be 5 mm at 12 cm
from isocentre [11]. These therefore are of clinical
significance for head and neck planning, and need to be
either removed by a post-processing technique [12, 13], or
by CT/MR fusion and image correction. Electron density
information necessary for treatment dosimetry cannot be
obtained from MRI, and the images have to be fused with
CT data if this information is to be used in treatment
planning calculations. Therefore fusion of distortion-
corrected MRI and CT images could provide both the
improved target definition and dosimetric accuracy
required for treatment planning.
The British Journal of Radiology, July 2006
Figure 1. Three-dimensional con-
formal radiotherapy planning. CT
plan of a left sided oropharyngeal
tumour. Blue colourwash: planning
target volume (PTV), isodose levels
indicated by coloured linear bound-
aries. ANT: anterior beam, LAO: left
anterior oblique, LPO: left posterior
oblique.
Intrinsic susceptibility-weighted or blood oxygenation
level dependent (BOLD) MRI exploits the differences in
magnetic susceptibility of oxyhaemoglobin and deoxy-
haemoglobin. Changes in blood oxygenation can there-
fore be characterized by looking at differences between
T2* weighted images during a change in oxygenation,
and may help to identify hypoxia in tumours. No
intravenous contrast agents are required; however,
BOLD effects are very short lived, and require rapid
sequences. Some of the changes in signal can be small
and are difficult to reproduce.
Functional imaging
Functional imaging is defined as characterizing tumours
radiologically in terms of their biochemistry or physiology.
Such imaging modalities include positron emission tomo-
graphy (PET), single photon emission computed tomo-
graphy (SPECT), magnetic resonance spectroscopy (MRS),
and dynamic contrast-enhanced MRI and CT. Functional
imaging may improve staging of disease by detecting
occult carcinoma, or give clearer delineation of areas of
previously known tumour (Figure 2). Furthermore, it may
provide information on tumour parameters such as blood
flow, vascular permeability, proliferation rate and oxyge-
nation. The introduction of functional imaging to radio-
therapy planning adds a new concept, termed the
biological target volume (BTV) [14].
PET
(18)F-Fluoro-2-deoxy-D-glucose (18FDG) is the most
commonly used imaging tracer in the diagnosis and
staging of HNC. Table 1 summarizes studies showing
the sensitivity and specificity in diagnosis of HNC which
is superior to CT and MR in assessing lymph nodes,
distant metastases and second primaries in a single
study [15–23], although the number of false positives
makes the specificity suboptimal.
PET-CT
PET alone has been of limited value in radiotherapy
planning because of its limited spatial resolution
555
 K Newbold, M Partridge, G Cook et al

(a) (b) (c)

18
Figure 2. FDG PET/CT case study. This patient presented with a right sided level II node containing metastatic squamous cell
carcinoma. Conventional imaging with a diagnostic contrast enhanced CT scan showed the right neck node but the primary
tumour was occult. (a) The non-contrast enhanced CT component of the PET/CT scan illustrates this (arrow points at node). (b)
18
FDG PET/CT demonstrated the right neck node and but also identified an area of 18FDG uptake in the right tongue base which
on biopsy was proven to be the primary site. (c) The planning target volume (PTV) defined in blue shows the target volume
planned with conventional data only whereas the PTV defined in green shows the expansion of this when the PET/CT identifies
the tongue base tumour and therefore includes the oropharynx and contralateral neck. The PTV increased by 467 cm3 with the
addition of data from the PET/CT.

(4–5 mm for 18FDG-PET) and a lack of anatomical
landmarks. Software image co-registration of non-con-
temporaneous PET images and CT or MRI is possible,
but may generate significant matching errors which
make it unsatisfactory for radiotherapy planning.
Integrated PET/CT scanners produced hardware fused
images (Figure 3) which reduces these errors and
increases accuracy compared with PET alone in head
and neck cancer (96% vs 90%, p 50.03, [15]). The CT
component of a standard PET/CT scan is not usually of
diagnostic resolution in order to minimize the radiation
dose received by the patient. Optimal CT scanning, for
example with multislice scanners, are sometimes advised
to supplement the PET/CT examination. PET/CT is
likely to have a maximum impact in the head and neck
region because accurate co-registration enables differ-
entiation between pathological and physiological areas
of 18FDG uptake. Common causes of false positives are
Waldeyers ring, salivary glands, brown fat and fast
twitch muscles, which have the potential to cloud the
diagnostic picture.
Using PET, an area of tracer uptake may be seen which
is distinct from, or overlap the conventionally defined
gross target volume (GTV). In lung cancer planning, the
addition of PET led to 26–100% of patients having a
change in radiotherapy management when compared
with CT plans alone [24]. The changes were not
consistent. 15–64% of patients showed an increase in
the planning target volume (PTV) and 21–36% had a
decrease. Ciernik et al [25] took 39 patients with mixed
primaries (12 with HNC) and compared GTV and PTV
when localizing using CT alone compared with 18FDG-
PET/CT. They observed that the GTV in HNC changed
in 32% with the PET data (either an increase or decrease)

18
Table 1. FDG-PET in staging head and neck cancer (HNC)
Author Year Number of T stage N stage Management
patients change (%)
Sensitivity (%) Specificity (%) Sensitivity (%) Specificity (%)

Schmid [16] 2003 48 8

Dizendorf [17] 2003 202 27
all sites
Kresnick [18] 2001 24 21
Hanasano [19] 1999 146 50 83 86 73
Kau [20] 1999 70 87 94
Nowak [21] 1999 71 87 67 80 92
Wong [22] 1997 54 67 100

556 The British Journal of Radiology, July 2006

Review article: Advanced imaging applied to radiotherapy planning in HNC
and the mean PTV change was 20%. Interestingly, the
interobserver variability was reduced when the PET data
were included compared with CT alone. Nishioka [26]
looked at image fusion between 18FDG-PET and MRI/CT
for radiotherapy planning in 12 patients with orophar-
yngeal carcinoma and nine patients with nasopharyngeal
carcinoma. They concluded that the fusion was useful in
GTV and CTV determination, and enabled sparing of
normal tissues. Scarfone et al observed an average
increase in GTV of 15% when defined by PET/CT versus
CT alone [27]. Table 2 summarizes radiotherapy plann-
ing studies with PET imaging.
Other PET tracers for functional imaging
DNA precursors, such as 11C or 124I labelled thymidine
or deoxyuridine, are incorporated into DNA during
repair or S phase and can be imaged to non-invasively
identify regions of cell proliferation. 11C-labelled methio-
nine or choline are substrates for protein synthesis and
have been tested in prostate cancer [28]. Imaging of
tumour hypoxia with tracers is now possible. An
example is the group of imidazole containing agents
(e.g. 18F-misonidazole) which are bioreductive molecules
that accept an electron to form a free radical that is
incorporated into the cell constituents under hypoxic
conditions. 62Cu-diacetylbis (N-4-methyl-thiosemicarba-
zone, Cu-ATSM) is a non-imidazole bioreducible radio-
pharmaceutical which has been used clinically to image
hypoxia and to define a potential target for therapy [29].
SPECT
SPECT produces a three-dimensional tomographic
image of the distribution of an injected radioisotope,
and therefore is of potential interest for radiotherapy
18
Table 2. FDG-PET and radiotherapy planning
Author Year n Tumour type
Scarfone [27] 2004 6 HNC
Ciernik [25] 2003 39 Various
Nishioka [26] 2002 21 HNC
n, number of patients; HNC, head and neck cancer; NSCLC, non small cell lung cancer; GTV, gross tumour volume; CTV, clinical
target volume; PTV, planning target volume.
The British Journal of Radiology, July 2006
18
Figure 3. Hardware-fused FDG-
PET/CT showing primary tumour in
the left oropharynx. The position of
the 18FDG uptake can be seen in
relation to the CT derived anatomy
in three dimensions.
planning. However, the detection of single photons,
rather than two coincident rays, reduces the spatial
resolution compared with PET. In view of this, SPECT
has remained in the research arena in the head and neck
region.
Magnetic resonance spectroscopy (MRS)
MRS provides a non-invasive method with which to
identify and quantify the presence of specific chemicals
within a tissue, e.g. tissue metabolites or the presence of a
specific drug. It could feasibly be used to identify
chemicals associated with hypoxic tissue. Current
limitations of MRS include sensitivity, limited spatial
localization and organ motion, but advances in data
acquisition techniques have the potential to make
spectroscopy an increasingly important clinical tool [30,
31].
Dynamic contrast-enhanced imaging of
tumour vascularity, blood flow and
permeability
Rapid scanning sequences during the administration
of contrast agents for CT and MRI have allowed dynamic
scans that can impart information regarding the biology
of the tumour and its microenvironment, such as blood
flow, vascular permeability, hypoxia and pH. Two
techniques are currently in clinical research.
Dynamic contrast enhanced (DCE) MRI
DCE-MRI involves the acquisition of multiple sequen-
tial MRI scans of an area of interest following injection of
a contrast agent. This enables study of pharmacokinetics
Fusion Results
Software fusion Modified GTV by a mean 15%
18
FDG-PET and CT increase
Hardware fusion GTV Change-56%, Reduced inter-
18
FDG-PET/CT observer variability
Software fusion Useful in GTV and CTV, and normal
18
FDG-PET with MRI/CT tissue sparing
557

of para-magnetic contrast agents and provides informa-
tion on tumour vascularity, blood volume and vessel
permeability. MRI sequences can be designed to be
sensitive to the initial, largely intravascular phase of
contrast delivery [32, 33]. Cooper et al [34] examined the
relationship between DCE-MRI parameters and
Eppendorf pO2 histographic measurements in 30
patients with cervical carcinoma and found a correlation
between maximum enhancement over baseline and rate
of enhancement, using T1 weighted sequences, with both
median pO2 and proportion of pO2 with values less than
5 mmHg. They concluded that DCE-MRI could be used
to measure hypoxia in human tumours in vivo. Hoskin et
al [35] examined tumour perfusion in patients with
advanced HNC using DCE-MRI (T1 weighted sequences)
and found a correlation between local tumour control
and maximum tumour enhancement following accele-
rated radiotherapy.
Perfusion CT
Similarly, rapid acquisition of images by spiral or
multislice CT, as contrast is given, can estimate tissue
perfusion based on the contrast density changes over
time [36]. Histological assessment of tumour neovascu-
larization such as microvessel density correlates with
contrast enhancement parameters in lung and renal
cancer [37, 38]. Hermans et al [39] used this method in
patients with HNC who had undergone radical radio-
therapy or chemotherapy and concluded that tumour
perfusion was an independent predictor of local control,
with decreased perfusion levels associated with a higher
local failure rate. Possibly these tumours had a reduced
blood supply rendering them relatively hypoxic and
therefore radioresistant.
Application of functional imaging to head and
neck cancer treatment planning
Hypoxia targeting
Low oxygen levels are associated with reduced
apoptotic potential, increased angiogenesis and
increased frequency of mutations [40], and are associated
with poor local control and survival. Hypoxic radio-
resistance may be overcome to some extent by increasing
tumour oxygenation, hypoxic cell sensitizers, or increas-
ing the radiation dose [41, 42]. Some studies both in vitro
and in vivo suggest that a radiation dose 2.5–3 times
current dose levels are required to overcome the effects
of hypoxia [43]. However, others suggest that more
modest dose increases of 1.2–1.5 times the primary dose
may result in equivalent tumour control [44]. These
doses, in the region of 90–100 Gy, are potentially
achievable to areas within tumours with concomitant
boost techniques deliverable with IMRT [45]. The spatial
and temporal stability of the hypoxic volume during
radiotherapy is critical for such approaches and is the
subject of ongoing studies. Chronic, or diffusion limited
hypoxia is defined by reduced pO2 over hours to days,
thought to be due to the distance from a capillary, the
oxygen content, the rate of blood flow and the oxygen
558
K Newbold, M Partridge, G Cook et al
metabolism of that capillary. Acute, or perfusion limited
hypoxia is defined when the variation is over minutes
and is thought to be due to intermittent reductions in
capillary flow. Dose escalation strategies based on
targeting areas of chronic hypoxia are only likely to
succeed if this is the dominant cause of treatment failure.
PET imaging with 18F-Miso, or CuATSM have been used
in this context to quantify hypoxia in head and neck
cancer [46, 47]. In a study by Taylor et al [48], BOLD MRI,
in conjunction with carbogen-breathing in patients with
head and neck carcinoma, suggested improved tissue
oxygenation and blood flow; however, difficulties with
measurement and reproducibility of BOLD signals have
made it impractical for radiotherapy planning.
Validating hypoxia imaging using histological
markers
Direct measurement of tissue oxygen tension is
possible using the Eppendorf polarographic electrode.
This technique is not ideal for validating hypoxia
because it is invasive, requires accessible tumours and
is highly user dependent. 50–150 readings are required
per sample, and spatial heterogeneity remains a problem
[49]. Sequential readings in assessment of temporal
changes are unreliable due to tissue damage following
initial formation of tracks [50]. Despite this, DCE-MRI
parameters have been correlated to Eppendorf measure-
ments in carcinoma of the cervix [34].
Exogenous and endogenous markers are more promis-
ing. Exogenous markers are chemicals that accumulate
or are bioreducible in hypoxic conditions, e.g. nitroimi-
dazoles [51]. These retained bioreductive products can be
detected by immunohistochemistry, e.g. pimonidazole
[52]. Pimonidazole staining has been correlated with
outcome in HNC [53]. Endogenous markers are gene
products that are up-regulated in the presence of
hypoxia. Aebersold et al reported 94% of a cohort of 98
oropharyngeal squamous cell carcinomas over-
expressed the transcription factor, hypoxia-inducible
factor 1-a (HIF 1-a) [54]. Carbonic anhydrase-9 (CA9)
has been shown to have the greatest magnitude of
expression in response to hypoxia among a range of 12
genes [55]. CA9 levels rise from 4 h to 24 h of levels of
pO2 at 20 mmHg and less. Expression has been
concordant with pimonidazole in head and neck cancer
[53], and with polarographic (electrode) measurements
[56] in cervical cancer.
The examination of tumour specimens for hypoxia
with the above markers can be used to validate imaging
methods performed in the pre-operative period, as long
as careful attention is paid to the orientation of the
specimen at the time of surgery and histological sections
are cut in the same plane as the test images.
What is the likely impact of functional imaging on
radiotherapy planning?
Functional imaging adds to anatomical imaging for
radiotherapy planning in a number of ways. First,
functional imaging may alter disease stage, which may
have a major impact on disease management. This is
The British Journal of Radiology, July 2006
Review article: Advanced imaging applied to radiotherapy planning in HNC
more likely to be an issue in diseases such as lung cancer
with a high metastatic potential, than with head and
neck cancer. Second, functional imaging may improve
our localization of the target volume by detection of
unexpected tumour extension, or presence of occult
locoregional metastases in lymph nodes. This will lead to
changes in GTV and CTV definition which may be
clinically significant. Finally, functional imaging may
determine radioresistant sub-regions within the conven-
tional GTV due to hypoxia or accelerated proliferation.
Such biological target volumes may be suitable for
radiation dose escalation delivered by techniques such
as simultaneous IMRT boosts (Figure 4). These allow
increase in total dose and dose-per-fraction, which
radiobiologically may be of particular benefit.
Conclusion
Several new imaging techniques, both anatomical and
functional are currently being evaluated for treatment
planning for head and neck cancer (Figure 5). As well as
improved conventional target volume definition, new
biological target volumes may be generated by these
The British Journal of Radiology, July 2006
Figure 4. Intensity-modulated
radiotherapy (IMRT) plan, illustrat-
ing dose boost to a biological target
volume (BTV).
Figure 5. Interaction of anatomical
and functional imaging modalities
and their compatibility with radio-
therapy planning (RTP).
imaging methods. Careful validation of these imaging
methods against histological parameters is urgently
required before they can be integrated into clinical
treatment planning. Functional imaging will most likely
be used in conjunction, rather than as an alternative to,
conventional imaging techniques.
References
1. Parker SL, Tong T, Bolden S, Wingo PA. Cancer statistics,
1997. CA Cancer J Clin 1997;47:5–27.
2. Yousem DM, Som PM, Hackney DB, Schwaibold F, Hendrix
RA. Central nodal necrosis and extracapsular neoplastic
spread in cervical lymph nodes: MR imaging versus CT.
Radiology 1992;182:753–9.
3. Curtin HD, Ishwaran H, Mancuso AA, Dalley RW, Caudry
DJ, McNeil BJ. Comparison of CT and MR imaging in
staging of neck metastases. Radiology 1998;207:123–30.
4. Hermans R, Feron M, Bellon E, Dupont P, Van den Bogaert
W, Baert AL. Laryngeal tumor volume measurements
determined with CT: a study on intra- and interobserver
variability. Int J Radiat Oncol Biol Phys 1998;40:553–7.
5. Guerrero Urbano MT, Nutting CM. Clinical use of
intensity-modulated radiotherapy: part I. Br J Radiol
2004;77:88–96.
559

6. Guerrero Urbano MT, Nutting CM. Clinical use of
intensity-modulated radiotherapy: part II. Br J Radiol
2004;77:177–82.
7. Khoo VS, Dearnaley DP, Finnigan DJ, Padhani A, Tanner
SF, Leach MO. Magnetic resonance imaging (MRI): con-
siderations and applications in radiotherapy treatment
planning. Radiother Oncol 1997;42:1–15.
8. Leslie A, Fyfe E, Guest P, Goddard P, Kabala JE. Staging of
squamous cell carcinoma of the oral cavity and oropharynx:
a comparison of MRI and CT in T- and N-staging. J Comput
Assist Tomogr 1999;23:43–9.
9. Chung NN, Ting LL, Hsu WC, Lui LT, Wang PM. Impact of
magnetic resonance imaging versus CT on nasopharyngeal
carcinoma: primary tumor target delineation for radio-
therapy. Head Neck 2004;26:241–6.
10. Emami B, Sethi A, Petruzzelli GJ. Influence of MRI on target
volume delineation and IMRT planning in nasopharyngeal
carcinoma. Int J Radiat Oncol Biol Phys 2003;57:481–8.
11. Aoyama H, Shirato H, Nishioka T, Hashimoto S, Tsuchiya
K, Kagei K, et al. Magnetic resonance imaging system for
three-dimensional conformal radiotherapy and its impact
on gross tumor volume delineation of central nervous
system tumors. Int J Radiat Oncol Biol Phys 2001;50:821–7.
12. Tanner SF, Finnigan DJ, Khoo VS, Mayles P, Dearnaley DP,
Leach MO. Radiotherapy planning of the pelvis using
distortion corrected MR images: The removal of system
distortion. Phys Med Biol 2000;45:2117–32.
13. Doran SJ, Charles-Edwards L, Reinsberg S, Leach MO. A
complete distortion correction for MR images: I Gradient
warp correction. Phys Med Biol (In press).
14. Ling CC, Humm J, Larson S, Amols H, Fuks Z, Leibel S, et
al. Towards multidimensional radiotherapy (MD-CRT):
biological imaging and biological conformality. Int J
Radiat Oncol Biol Phys 2000;47:551–60.
15. Schoder H, Yeung HW, Gonen M, Kraus D, Larson SM.
Head and neck cancer: clinical usefulness and accuracy of
PET/CT image fusion. Radiology 2004;231:65–72.
16. Schmid DT, Stoeckli SJ, Bandhauer F, Huguenin P, Schmid
S, von Schulthess GK, et al. Impact of positron emission
tomography on the initial staging and therapy in loco-
regional advanced squamous cell carcinoma of the head
and neck. Laryngoscope 2003;113:888–91.
17. Dizendorf EV, Baumert BG, von Schulthess GK, Lutolf UM,
Steinert HC. Impact of whole-body 18F-FDG PET on
staging and managing patients for radiation therapy. J
Nucl Med 2003;44:24–9.
18. Kresnik E, Mikosch P, Gallowitsch HJ, Kogler D, Wiesser S,
Heinisch M, et al. Evaluation of head and neck cancer with
18F-FDG PET: a comparison with conventional methods.
Eur J Nucl Med 2001;28:816–21.
19. Hanasono MM, Kunda LD, Segall GM, Ku GH, Terris DJ.
Uses and limitations of FDG positron emission tomography
in patients with head and neck cancer. Laryngoscope
1999;109:880–5.
20. Kau RJ, Alexiou C, Laubenbacher C, Werner M, Schwaiger
M, Arnold W. Lymph node detection of head and neck
squamous cell carcinomas by positron emission tomogra-
phy with fluorodeoxyglucose F 18 in a routine clinical
setting. Arch Otolaryngol Head Neck Surg 1999;125:1322–8.
21. Nowak B, Di Martino E, Janicke S, Cremerius U, Adam G,
Zimny M, et al. Diagnostic evaluation of malignant head
and neck cancer by F-18-FDG PET compared to CT/MRI.
Nuklearmedizin 1999;38:312–8.
22. Wong WL, Saunders M. The impact of FDG PET on the
management of occult primary head and neck tumours.
Clin Oncol (R Coll Radiol) 2003;15:461–6.
23. Fogarty GB, Peters LJ, Stewart J, Scott C, Rischin D, Hicks
RJ. The usefulness of fluorine 18-labelled deoxyglucose
positron emission tomography in the investigation of
560
K Newbold, M Partridge, G Cook et al
patients with cervical lymphadenopathy from an unknown
primary tumor. Head Neck 2003;25:138–45.
24. Paulino AC, Thorstad WL, Fox T. Role of fusion in
radiotherapy treatment planning. Semin Nucl Med
2003;33:238–43.
25. Ciernik IF, Dizendorf E, Baumert BG, Reiner B, Burger C,
Davis JB, et al. Radiation treatment planning with an
integrated positron emission and computer tomography
(PET/CT): a feasibility study. Int J Radiat Oncol Biol Phys
2003;57:853–63.
26. Nishioka T, Shiga T, Shirato H, Tsukamoto E, Tsuchiya K,
Kato T, et al. Image fusion between 18FDG-PET and MRI/
CT for radiotherapy planning of oropharyngeal and
nasopharyngeal carcinomas. Int J Radiat Oncol Biol Phys
2002;53:1051–7.
27. Scarfone C, Lavely WC, Cmelak AJ, Delbeke D, Martin WH,
Billheimer D, et al. Prospective feasibility trial of radio-
therapy target definition for head and neck cancer using 3-
dimensional PET and CT imaging. J Nucl Med
2004;45:543–52.
28. Schoder H, Larson SM. Positron emission tomography for
prostate, bladder, and renal cancer. Semin Nucl Med
2004;34:274–92.
29. Chao KS, Bosch WR, Mutic S, Lewis JS, Dehdashti F,
Mintun MA, et al. A novel approach to overcome hypoxic
tumor resistance: Cu-ATSM-guided intensity-modulated
radiation therapy. Int J Radiat Oncol Biol Phys
2001;49:1171–82.
30. Pirzkall A, McKnight TR, Graves EE, Carol MP, Sneed PK,
Wara WW, et al. MR-spectroscopy guided target delinea-
tion for high-grade gliomas. Int J Radiat Oncol Biol Phys
2001;50:915–28.
31. Pouliot J, Kim Y, Lessard E, Hsu IC, Vigneron DB,
Kurhanewicz J. Inverse planning for HDR prostate brachy-
therapy used to boost dominant intraprostatic lesions
defined by magnetic resonance spectroscopy imaging. Int
J Radiat Oncol Biol Phys 2004;59:1196–207.
32. Padhani AR, Husband JE. Dynamic contrast-enhanced MRI
studies in oncology with an emphasis on quantification,
validation and human studies. Clin Radiol 2001;56:
607–20.
33. Padhani AR. MRI for assessing antivascular cancer treat-
ments. Br J Radiol 2003;76 Spec. No. 1:S60–80.
34. Cooper RA, Carrington BM, Loncaster JA, Todd SM,
Davidson SE, Logue JP, et al. Tumour oxygenation levels
correlate with dynamic contrast-enhanced magnetic reso-
nance imaging parameters in carcinoma of the cervix.
Radiother Oncol 2000;57:53–9.
35. Hoskin PJ, Saunders MI, Goodchild K, Powell ME, Taylor
NJ, Baddeley H. Dynamic contrast enhanced magnetic
resonance scanning as a predictor of response to accelerated
radiotherapy for advanced head and neck cancer. Br J
Radiol 1999;72:1093–8.
36. Axel L. Cerebral blood flow determination by rapid-
sequence computed tomography: theoretical analysis.
Radiology 1980;137:679–86.
37. Tateishi U, Nishihara H, Watanabe S, Morikawa T, Abe K,
Miyasaka K. Tumor angiogenesis and dynamic CT in lung
adenocarcinoma: radiologic-pathologic correlation. J
Comput Assist Tomogr 2001;25:23–7.
38. Jinzaki M, Tanimoto A, Mukai M, Ikeda E, Kobayashi S,
Yuasa Y, et al. Double-phase helical CT of small renal
parenchymal neoplasms: correlation with pathologic find-
ings and tumor angiogenesis. J Comput Assist Tomogr
2000;24:835–42.
39. Hermans R, Meijerink M, Van den Bogaert W, Rijnders A,
Weltens C, Lambin P. Tumor perfusion rate determined
noninvasively by dynamic computed tomography predicts
outcome in head-and-neck cancer after radiotherapy. Int J
Radiat Oncol Biol Phys 2003;57:1351–6.
The British Journal of Radiology, July 2006
Review article: Advanced imaging applied to radiotherapy planning in HNC
40. Dachs GU, Chaplin DJ. Microenvironmental control of gene
expression: implications for tumor angiogenesis, progres-
sion, and metastasis. Semin Radiat Oncol 1998;8:208–16.
41. Horsman MR, Overgaard J. Overcoming tumour radiation
resistance resulting from acute hypoxia. Eur J Cancer
1992;28A:2084–5.
42. Overgaard J, Hansen HS, Overgaard M, Bastholt L,
Berthelsen A, Specht L, et al. A randomized double-blind
phase III study of nimorazole as a hypoxic radiosensitizer
of primary radiotherapy in supraglottic larynx and pharynx
carcinoma. Results of the Danish Head and Neck Cancer
Study (DAHANCA) Protocol 5-85. Radiother Oncol
1998;46:135–46.
43. Steel. G. Basic clinical radiobiology, 3rd edn. London:
Hodder Arnold, 2002.
44. Popple RA, Ove R, Shen S. Tumor control probability for
selective boosting of hypoxic subvolumes, including the
effect of reoxygenation. Int J Radiat Oncol Biol Phys
2002;54:921–7.
45. Nutting CM, Corbishley CM, Sanchez-Nieto B, Cosgrove VP,
Webb S, Dearnaley DP. Potential improvements in the
therapeutic ratio of prostate cancer irradiation: dose escala-
tion of pathologically identified tumour nodules using
intensity modulated radiotherapy. Br J Radiol 2002;75:
151–61.
46. Rasey JS, Koh WJ, Evans ML, Peterson LM, Lewellen TK,
Graham MM, et al. Quantifying regional hypoxia in human
tumors with positron emission tomography of [18F]
fluoromisonidazole: a pretherapy study of 37 patients. Int
J Radiat Oncol Biol Phys 1996;36:417–28.
47. Fujibayashi Y, Taniuchi H, Yonekura Y, Ohtani H, Konishi
J, Yokoyama A. Copper-62-ATSM: a new hypoxia imaging
agent with high membrane permeability and low redox
potential. J Nucl Med 1997;38:1155–60.
48. Taylor NJ, Baddeley H, Goodchild KA, Powell ME,
Thoumine M, Culver LA, et al. BOLD MRI of human
The British Journal of Radiology, July 2006
tumor oxygenation during carbogen breathing. J Magn
Reson Imaging 2001;14:156–63.
49. Becker A, Hansgen G, Bloching M, Weigel C,
Lautenschlager C, Dunst J. Oxygenation of squamous cell
carcinoma of the head and neck: comparison of primary
tumors, neck node metastases, and normal tissue. Int J
Radiat Oncol Biol Phys 1998;42:35–41.
50. Rudat V, Vanselow B, Wollensack P, Bettscheider C,
Osman-Ahmet S, Eble MJ, et al. Repeatability and prog-
nostic impact of the pretreatment pO(2) histography in
patients with advanced head and neck cancer. Radiother
Oncol 2000;57:31–7.
51. Whitmore GF, Varghese AJ, Gulyas S. Reaction of 2-
nitroimidazole metabolites with guanine and possible
biological consequences. IARC Sci Publ 1986:185–96.
52. Azuma C, Raleigh JA, Thrall DE. Longevity of pimonida-
zole adducts in spontaneous canine tumors as an estimate
of hypoxic cell lifetime. Radiat Res 1997;148:35–42.
53. Kaanders JH, Wijffels KI, Marres HA, Ljungkvist AS, Pop
LA, van den Hoogen FJ, et al. Pimonidazole binding and
tumor vascularity predict for treatment outcome in head
and neck cancer. Cancer Res 2002;62:7066–74.
54. Aebersold DM, Burri P, Beer KT, Laissue J, Djonov V,
Greiner RH, et al. Expression of hypoxia-inducible factor-
1alpha: a novel predictive and prognostic parameter in the
radiotherapy of oropharyngeal cancer. Cancer Res
2001;61:2911–6.
55. Lal A, Peters H, St Croix B, Haroon ZA, Dewhirst MW,
Strausberg RL, et al. Transcriptional response to hypoxia in
human tumors. J Natl Cancer Inst 2001;93:1337–43.
56. Loncaster JA, Harris AL, Davidson SE, Logue JP, Hunter
RD, Wycoff CC, et al. Carbonic anhydrase (CA IX)
expression, a potential new intrinsic marker of hypoxia:
correlations with tumor oxygen measurements and prog-
nosis in locally advanced carcinoma of the cervix. Cancer
Res 2001;61:6394–9.
561
 The British Journal of Radiology, 79 (2006), 562–571

Radiofrequency ablation in pig lungs: in vivo comparison of

internally cooled, perfusion and multitined expandable
electrodes
1,2
J M LEE, MD, 1,2J K HAN, MD,
1,2
J M CHANG, MD,
1,2
S Y CHUNG, MD,
1,2
S H KIM, MD,
1,2
J Y LEE, MD
and 1,2B I CHOI, MD

1
Department of Radiology, and Institute of Radiation Medicine, Seoul National University College
of Medicine, 28 Yongon-dong, Chongno-gu, Seoul 110-744 and 2Clinical Research Institute, Seoul
National University Hospital, Seoul, Korea

ABSTRACT. The purpose of this study was to compare the amounts of in vivo
coagulation obtained by radiofrequency (RF) ablation in porcine lung, using three
types of electrodes. 15 in vivo ablation procedures were performed in the lungs of five
pigs using three kinds of currently available RF devices under CT guidance. After
placing an electrode in the lung, three ablation zones were created at each of three
different regimens: Group A: RF ablation with an internally cooled electrode; Group B:
RF ablation with a perfusion electrode, with instillation of 0.9% NaCl solution at a rate
of 1.5 ml min21; Group C: RF ablation with a multitined expandable electrode.
According to the manufacturer’s recommendations, RF application times were 12 min
in group A and 20 min in group B. In group C, RF energy was delivered for 7 min after a
mean temperature of 110 ˚C was reached at 5 cm deployment. 36 min after the
procedures, contrast-enhanced CT scans were obtained to evaluate the volume of zone
of coagulation, and lungs were harvested for gross measurements. After macroscopic
and histopathological analyses of 5 mm-thick lung sections, diameters, volumes and
variation coefficients of regions of central coagulation were assessed. During RF
ablation, the perfusion electrode allowed a larger energy delivery than the
internally cooled or the multitined expandable electrodes, i.e. 33.6¡4.7 kJ in group A,
40.0¡8.2 kJ in group B and 23.5¡6.1 kJ in group C (p,0.05). On gross observation, the
cut surface of the gross specimen containing RF-induced coagulation showed that the
ablated tissue appeared to be a central, firm, dark-brown area surrounded by an
irregular outer margin (approximately 3–10 mm thick) of bright red tissue. In vivo studies
Received 12 September
showed that RF ablation using the perfusion electrode achieved larger coagulation 2005
volume than RF ablation using the other electrodes (p,0.05): 7.2¡4.1 cm3 in group A; Revised 24 November 2005
16.9¡5.5 cm3 in group B; 7.5¡3.3 cm3 in group C. The corresponding variation Accepted 19 January 2006
coefficients were 0.55, 0.31, and 0.45, respectively. Our study shows that RF ablation
DOI: 10.1259/bjr/51844219
using a perfusion electrode achieves a larger coagulation volume with an irregular
margin than RF ablation using internally cooled or multitined expandable electrodes in ’ 2006 The British Institute of
the porcine lung. Radiology

Lung cancer is one of the major leading causes of
cancer-related mortality, with more than 1.2 million new
cases of lung cancer being diagnosed worldwide each
year [1–3]. In addition, the lung is also the second most
frequent site of metastatic disease. Until recently,
therapeutic options for primary lung cancers, depending
on tumour grading and staging and the presence of
comorbidities, included a combination of surgical resec-
tion, chemotherapy, and/or XRT. Indeed, surgical resec-
tion is the treatment of choice for early-stage non-small
cell lung cancer (NSCLC), but unfortunately, patients
with NSCLC are frequently poor surgical candidates due
to coexistent medical diseases such as chronic obstruc-
tive bronchopneumopathy or cardiac disease [4]. Several
This study was supported by grant No. 21-2005-021-0 from the
Seoul National University Hospital Research Fund.
studies have documented survival benefits in patients
with pulmonary metastases with favourable histologies
who received complete resection as compared with
unresectable individuals [5, 6]. In patients with lung
cancer who are not surgical candidates, the treatment
options are primarily XRT with or without chemo-
therapy. However, for stage I/II NSCLC, the effective-
ness for XRT relative to surgery remains uncertain [7],
and a meta-analysis of trials comparing primary treat-
ment with or without chemotherapy showed that
chemotherapy provided only a modest benefit [8].
Therefore, less invasive therapies that can accomplish
tumour destruction without the use of general anaes-
thesia may complement, improve, or even replace
existing therapies.
Radiofrequency (RF) ablation has received much
recent attention as a minimally invasive strategy for the

562 The British Journal of Radiology, July 2006

In vivo radiofrequency ablation in pig lungs
treatment of various neoplasms of the liver, kidney and
bone [9–12]. Also, a number of recent experimental
and clinical studies have demonstrated the feasibility
and safety of RF ablation for the treatment of inoperable
lung malignancies [13–18]. Some preliminary studies
have shown that RF ablation enables the successful
treatment of relatively small lung malignancies with a
high rate of complete response and acceptable morbidity
[15–18]. However, others have shown limitations in
achieving complete necrosis in large tumours measuring
3 cm or more in diameter [19–21]. Lee et al [20] treated 32
lesions by RF ablation and achieved complete necrosis in
100% of tumours smaller than 3 cm in diameter, and in
23% of larger tumours.
To successfully ablate malignant tumours, it is
essential to ensure the coagulation of the entire targeted
volume with as few complications as possible. Indeed,
the acquisition of a large volume of ablation in a safe
manner is of paramount importance if RF is to be
accepted as routine form of intervention. Previous
clinical studies on lung RF ablation have examined
internally cooled needle electrodes [18, 20] and multi-
tined expandable electrodes [19, 21]. Currently in Korea,
Valleylab, BerchtoldTM, and RITAH RF systems have
been used (personal communication with Drs KY Jin, YK
Kim, and GS Jung) for the treatment of primary and
secondary lung cancers. Recently, Lee et al [22, 23]
demonstrated improved RF ablation efficacy using saline
infusion to induce a large volume of coagulation both ex
vivo and in vivo rabbit lungs. However, no study has
compared the in vivo efficacies of different types of
electrodes in the lung using large animals. The purpose
of our study was to compare in vivo coagulation obtained
with currently available RF ablation devices, namely an
internally cooled needle electrode, a multitined expand-
able electrode and a perfusion electrode.
Materials and methods
Animals and preparations
The experimental protocol was approved by the
Animal Use and Care Administrative Advisory
Committee of our institution. All experiments were
performed according to a protocol approved by the local
institutional committee on animals, in accordance with
the general guidelines issued by the National Institute of
Health for the care of laboratory animals. Five female
farm pigs were used in this study (weight range 30–
40 kg). The animals were fasted overnight, but had free
access to water before the experiments. Each of the five
pigs was anaesthetised using an intramuscular injection
of 50 mg kg21 of ketamine hydrochloride (Ketamine;
Yuhan, Seoul, Korea) and 5 mg kg21 of Xylazine
(Rumpun, Bayer Korea), and prepared for RF ablation.
Booster injections of up to half of the initial dose were
administered as needed. Ringer’s lactate solution was
continuously infused during the experiment
(500 ml h21). Endotracheal intubation was performed
and anaesthesia was maintained with inhaled enfluorane
(Gerolan; Choongwae Pharma Corporation, Seoul,
Korea). Mechanical ventilation was used throughout
the procedure.
The British Journal of Radiology, July 2006
Cardiac and respiratory parameters were monitored
throughout the procedures. Each animal’s lateral hind-
quarters were shaved bilaterally, and two 8 cm 6 12 cm
wire-mesh grounding-pads coated with conductive gel
were placed on each hind limb. For the RF ablation
procedures, animals were placed in the supine position
on the CT scanner. Animals were euthanized approxi-
mately 1–2 h after the final RF ablation procedure with a
pentobarbital overdose of 60 mg kg21. The lungs were
then removed.
Study design
To minimize potential variations in the RF ablation
procedures, all procedures were performed by consensus
between two radiologists who had extensive routine
experience with all three systems. Electrodes were
placed under the same experimental conditions with
CT guidance (Somatom plus 4 scanner or Sensation 16;
Siemens Medical Solutions, Forchheim, Germany). CT
enabled an electrode tip to be positioned at least 2 cm
from the pleura, from a previous ablation site in the same
lung, and from large vessels (. 3 mm in diameter). RF
ablation was performed using one of the three RF
systems: Group A: a 200 W generator (CC3: Valleylab
TM
) and an internally cooled electrode; Group B: a 60 W
generator (HiTT 106 BerchtoldH, Tuttlingen, Germany)
and a perfusion electrode; Group C: a 150 W generator
(1500 TM model; RITAH medical Systems; Mountain
View, CA) and a multitined expandable electrode.
The RF systems, lung lobes, position and the order in
which each ablation procedure was performed were
randomly assigned. A new electrode was used for each
ablation session. All settings were performed according
to manufacturer’s recommendations. The automated
control mechanisms functioned by measuring the total
impedance between the electrodes for the BerchtoldH,
and ValleylabTM systems and on the temperature of the
electrode tip for the RITAH system. Applied current,
power output and impedance were continuously mon-
itored using a generator system during RF ablation and
were recorded. The technical aspects of the RF ablation,
including impedance and wattage changes, and the
dimensions of the RF-coagulated area for each system
were compared.
RF devices, ablation protocols and procedures
Group A (ValleylabTM system)
A 480 kHz generator (CC3; ValleylabTM) capable of a
maximum power of 200 W was used with a 1.6 mm-
diameter internally cooled electrode (single cool-tip
needle), which has a 3 cm-long active distal region
(Figure 1a). Electrode cooling was ensured by the
peristaltic perfusion of chilled saline using a peristaltic
pump (PE-PM; ValleylabTM), which allowed the elec-
trode to maintain a tip temperature of below 25 ˚C during
RF delivery. This RF system requires the applications of
four neural pads. The circuitry incorporated into the
generator allowed continuous monitoring of the impe-
dance between the active electrode and the grounding
563
 J M Lee, J K Han, J M Chang et al

Figure 1. Different electrodes used for monopolar radiofrequency ablation. (a) ValleylabTM internally cooled electrode with a
30 mm long exposed tip (Cool-tip). (b) BerchtoldH perfusion electrode with a 15 mm long exposed tip and side holes. (c) RITAH
multitined expandable electrode (Starburst XL) with nine curved tines.

pads. RF current was passed for 12 min at a maximum
generator setting for the impedance control method. This
method allows the maximum power to be delivered until
impedance rises to 10 V above the baseline value. At this
point, the current is switched off automatically to avoid a
further local increase in temperature, which would result
in tissue charring. 15 s later, the current is automatically
switched on again, thus being referred to as the pulsed
RF technique, which increases RF ablation area [24].
Group B (BerchtoldH system)
This perfusion RF system is supported by a 375 kHz
generator (HiTT 106; BerchtoldH) capable of 60 W at
maximum through a 1.7 mm-diameter monopolar elec-
trode with an active tip length of 1.5 cm. The electrode
used for this saline-enhanced technique was double
walled at its distal part, and the inner wall had small
perfusion holes (Figure 1b). A 0.9% saline solution was
used as the perfusion liquid. This flowed through the
hollow shaft of the electrode and permeated through the
perfusion holes into the space between the inner and
outer walls of the needle tip. Continuous interstitial
perfusion of saline was started at 30 s at 90 ml h21 prior
to RF application, and was maintained during RF
ablation using a digitally controlled syringe pump
(Pilot C; Frensenius Vial, Brezins, France). RF current
was applied for 20 min at 40 W, in accordance with the
manufacturer’s recommendations. The associated con-
trol mechanism stabilized RF power within moderate
impedance changes of between 100 V and 350 V. When
the impedance exceeded 900 V, an additional saline
bolus was administered [25].
Group C (RITAH system)
Model 1500TM (RITAH Medical systems) utilizes a
150 W generator operating at 460 kHz. In this RF system,
the expandable electrode (Starburst XL; Rita Medical
Systems) consists of an insulated outer needle with a
diameter of 2.2 mm that houses nine deployable curved
tines, which have a maximum diameter of 5 cm when fully
expanded (Figure 1c). The tips of five of these tines contain
a thermocouple and allow temperature monitoring during

564 The British Journal of Radiology, July 2006

In vivo radiofrequency ablation in pig lungs
the ablation procedure. Electrodes were progressively
extended deeper into the lung parenchyma with tempera-
ture monitoring, and power was controlled according to
average temperature. Tines were first deployed at 2 cm
with a pre-selected target temperature of 80 ˚C, then
advanced to 3 cm with a target temperature of 105˚C
and, finally, extended to 4–5 cm with a target temperature
of 110˚C. Target temperatures were maintained for 7 min
and then post-ablation temperatures were monitored [26].
Imaging follow-up
A multirow detector CT (Sensation 16, Siemens
Medical Solutions) was used to monitor ablations at
30–60 min after RF ablation. Axial CT scans were
obtained using a 0.75 mm detector collimation, a
reconstruction increment of 3 mm and a 1.0 pitch, and
included both lungs, before and after injecting 70 ml of
contrast medium (Ultravist 370H; Schering Korea, Seoul,
Korea). Contrast medium was injected at a rate of
2 ml s21 through an ear vein; post-contrast CT scans
were obtained at 60 s after contrast administration and
CT images were reconstructed at an interval of 3 mm in
the axial plane and at an interval of 1 mm to obtain a
high-quality data set for multiplanar reconstruction
images. The thin section data set was forwarded to a
PC containing dedicated 3D software (Rapidia,
INFINITT, Seoul, Korea), and data was reconstructed
into 3 mm-thick coronal and sagittal slices.
Assessment of coagulation zone (imaging and
pathological studies)
Pigs were euthanized after obtaining CT images. Once
harvested, lungs were serially sectioned at 5 mm inter-
vals along the axial plane. The histopathological study
included staining for mitochondrial enzyme activity,
which was performed by incubating thin representative
tissue sections for 30 min in 2% 2,3,5,-triphenyl tetra-
zolium chloride (TTC; Sigma, St Louis, MO), at 20–25 ˚C.
This test is a vitalline stain for mitochondrial enzyme
activity [27] and can be used to determine irreversible
cellular injury during the early stages of RF-induced
necrosis [28]. As the unstained area of an RF-induced
coagulation has been shown to correspond to the zone of
necrosis [28], two observers measured axial diameters
along the axis of the electrodes (Dax) and transverse
diameters (Dtr) of ablation areas in the axial plane before
reaching consensus. The number of slices containing an
RF ablated region determined the vertical diameter (Dv).
In addition, the slices were photographed using a digital
camera (Canon EOS 300D; Canon Inc., Tokyo, Japan),
and images were saved to image management software
(PhotoShop; Adobe, San Jose, CA). Area analysis was
performed on a computer equipped with NIH Image J
software (National Institutes of Health; http://rsb.info.
nih.gov/ij/) [29]. The area of coagulation, on each slice
was calculated using this computer program, and
volumes were calculated by multiplying areas by slice
thickness and summed to obtain total lesion volumes.
The shapes of RF-induced coagulations were charac-
terized using the ratio between the long axis diameter
The British Journal of Radiology, July 2006
and the short-axis diameter: Dax/Dtr. Thus, a ratio near
1 indicates a near spherical shape. In addition, volume
variations in each group were determined using coeffi-
cients of variation, calculated as follows: standard devia-
tion of the ablation volume/mean value of the ablation
volume. The closer this ratio is to 0, the more reproducible
the coagulation [26]. The RF-induced ablated regions of
representative cases in each group were fixed in 10%
formalin for routine histological processing, and processed
by paraffin sectioning and haematoxylin-eosin staining for
light microscopic studies.
On contrast-enhanced CT images, the diameters of
hyperattenuated ablation areas in treated lungs were
measured on axial images and on sagittal reconstructed
images using a dedicated 3D program (Rapidia:
INFINITT, Seoul, Korea). The Dax and Dtr values of
hyperattenuated coagulation were measured in axial slices
showing maximum ablation zone dimension, and Dv was
measured in sagittal reconstruction images showing
maximum ablation zone dimension. Using the Image J
software (http://rsb.info.nih.gov/ij/) [29], the area of
coagulation, on each CT image was calculated, and
volumes were calculated by multiplying areas by slice
thickness, summed to obtain total lesion volumes.
Statistical analysis
The dimensions of thermal ablation areas and the
technical parameters such as accumulated RF energy of
the three groups were averaged for each group and
compared using one-way analysis of variance (ANOVA)
test. In comparisons between groups, the Bonferroni
multiple comparison test was applied. The volume of the
hyperattenuated area seen at CT was correlated with the
volume of central white area of the RF-induced coagula-
tion at pathological specimen, and then the degree of
correlation between them analysed using Pearson’s
correlation coefficient. Values are expressed as
means¡SD. For all statistical analyses, a p-value of
,0.05 was considered significant. Statistics were per-
formed using the Instat program (GraphPad Software,
Inc., San Diego, CA).
Results
Evaluation of RF methods
All pigs tolerated the RF ablation procedures well. In
terms of the duration of energy application, the mean
duration of RF energy delivery was 12 min in Group A,
20 min in Group B and 23¡1.7 min in Group C (p,0.05).
Impedance values decreased linearly during the procedure
in group B. However, in Group C, impedance values
gradually increased over the first 10 min and fluctuated
between 150 V and 200 V over the final 10 min. In Group A,
the impedance intermittently increased and activated the
pulsed RF algorithm (Figure 2). In Group B, we were able
to deliver 40 W and keep the impedance below 150 V. As a
consequence, RF ablation using the perfusion electrode
(Group B) allowed larger energy delivery than RF ablation
using the internally cooled (Group A) or multitined
expandable (Group C) electrodes, i.e. 33.6¡4.7 kJ in
565
 J M Lee, J K Han, J M Chang et al

(a)

(b)

Figure 2. Graphic depiction of the electrical parameters during radiofrequency ablation. (a) Tissue impedance (lower row),
radiofrequency (RF) current (middle row), and power changes (upper row) during radiofrequency ablation using an internally
cooled electrode. (b) Power and impedance changes (lower row), and tissue temperature changes (upper row) during
radiofrequency ablation using the multitined expandable electrode.

566 The British Journal of Radiology, July 2006

In vivo radiofrequency ablation in pig lungs

Group A, 40.0¡8.2 kJ in Group B and 23.5¡6.1 kJ in
Group C (p,0.05).
CT and gross examination of the RF-induced
coagulation area
On gross observation, the cut surface of gross speci-
mens containing RF-induced coagulation showed that
ablated tissue had a two layered structure presenting as a
central, firm, dark-brown area surrounded by an
outer, irregular margin (approximately 3–10 mm thick)
of bright red tissue (Figure 3). When the specimens were
fixed with 2% 2,3,5-TTC to determine if mitochondrial
function was irreversibly damaged after RF ablation,
the central dark-brown area did not show any staining
with TTC (no enzyme activity), but the outer bright red
zone showed inhomogeneous staining with TTC in
red colour (admixture of stained and not-stained areas).
The viable lung tissue outside of this area also showed
homogeneous staining in red with TTC (positive enzyme
activity). The lung tissue on CT, RF-ablated regions
appeared as irregularly shaped hyperattenuated areas,
which agreed with gross examination findings. In
some cases, RF-induced hyperattenuated areas consisted
of heterogeneous central zones containing air and
surrounded by a broad hyperattenuated region
(Figure 3).
Evaluation of volumes of coagulation using gross
specimens and CT images
The mean Dax’s of the RF induced central white zone
(unstained area with TTC) in axial planes measured in
gross specimens of the three groups, were 1.7¡0.4 cm
in Group A, and 2.7¡0.5 cm in Group B, and
1.9¡0.5 cm in Group C (p50.018) (Table 1). The corre-
sponding mean Dtrs were 2.6¡0.6 cm, 3.5¡0.8 cm and
2.9¡1.0 cm (p.0.05), and the mean Dvs of ablated
regions were 3.1¡0.5 cm, 3.6¡0.5 cm and 2.9¡1.0 cm
(p.0.05) (Figure 3). Furthermore, the volumes of coagu-
lation areas in the three groups were 7.2¡4.1 cm3
in Group A, 16.9¡5.5 cm3 in Group B and 7.5¡3.3 cm3
in Group C. The differences were significant between
Groups A and B (p50.002), and between Groups B and C
(p50.018). The coefficients of variability for internally
cooled, perfusion and multitined expandable electrodes
were 0.55, 0.31, and 0.45, respectively (p.0.05).
Dax/Dt ratios of coagulated areas produced with each
RF system were 0.7¡0.1 in Group A, 0.78¡0.2 in Group
B and 0.74¡0.4 in Group C. Although there was no
significant difference in Dax/Dt ratios between the
groups (p.0.05), the perfusion electrode created more
irregularly bordered coagulations than the internally
cooled or multitined expandable electrodes. On CT,
the volumes of RF-induced hyperattenuated areas in
the three groups were 9.2¡4.5 cm3 in Group A,
20.4¡6.3 cm3 in Group B and 10.2¡4.0 cm3 in Group
C, respectively. The volume of RF-induced hyperattenu-
ated areas in Group B was larger than those of Groups A
or C (p,0.05). The volumes of RF-induced hyperattenu-
ated areas on CT were larger than the volumes of RF-
induced coagulation (central unstained area of the gross
specimen with TTC staining) measured on the gross
specimens, but there was a significant correlation
between them (p,0.05).
Histopathological results
Histological specimens showed central tissue loss at
the electrode insertion tract and in the inner ablation
zone; lung structures seemed to remain, but parenchy-
mal cells showed eosinophilic cytoplasm with pyknotic
nuclei and alveolar exudates, which correspond to the
known early changes of coagulation necrosis [30]. The
outer ablation zone showed haemorrhagic congestion,
which was accompanied by neutrophil infiltration, and
pulmonary alveoli were filled with exudates. In this
zone, there was an admixture of abnormal cells with
pyknotic nuclei and eosinophilic cytoplasm and normal
looking parenchymal cells. Beyond this area, lung tissue
showed normal histological findings, except acute
inflammatory cell infiltrate.
Discussion
The efficacy and safety of percutaneous RF ablation
have been firmly established for the treatment of hepatic
malignancies [9, 10]. However, this is not true in the lung
[18, 19]. The most promising application for lung RF
ablation is in the treatment and local control of primary

Table 1. Measured values of RF-induced coagulation necrosis in gross specimens in the three groups
Coagulation Group A Group B Group C P-value
(12 min RFA) (20 min RFA) (23 min RFA)

Dax (cm) 1.7¡0.4 2.7¡0.5 1.9¡0.5 p , 0.05+

Dtr (cm) 2.6¡0.6 3.5¡0.8 2.9¡1.0 p . 0.05
Dv (cm) 3.1¡0.5 3.6¡0.5 2.9¡1.0 p . 0.05
Volume (cm3) 7.2¡4.1 16.9¡0.5 7.5¡3.3 p , 0.05+
Ratio of Dax/Dtr 0.7¡0.1 0.78¡0.2 0.74¡0.4 p . 0.05
Coefficient of variation 0.55 0.31 0.45 p . 0.05
Mean applied energy (kJ) 33.6¡4.7 40.0¡8.2 23.5¡6.1 p , 0.01*
Dax, the axial diameter of the central coagulation along the axis of the electrodes; Dtr, transverse diameter of the central
coagulation in the axial plane; Dv, vertical diameter of the central coagulation. Group A, RF ablation using an internally
cooled electrode; Group B, RF ablation using a perfusion electrode; Group C, RF ablation using a multitined expandable
electrode. +Differences between groups A and B, and groups B and C were significant (p , 0.05). *Differences between groups
A and C, and groups B and C were significant (p,0.05).

The British Journal of Radiology, July 2006 567

 J M Lee, J K Han, J M Chang et al

(a) (b)

(c) (d)

(e)

568 The British Journal of Radiology, July 2006

In vivo radiofrequency ablation in pig lungs

Figure 3. Contrast-enhanced CT scans and photographs of lung treated by radiofrequency (RF) ablation using an internally
cooled electrode and a multitined expandable electrode in a pig model. (a) Contrast-enhanced axial CT scan obtained
immediately after RF ablation showing an RF-induced hyperattenuated region (arrows) in both lungs. Note that the
hyperattenuated region in the right lung treated with an internally cooled electrode is similar to that in left lung treated with a
multitined expandable electrode. Photographs of gross specimen containing RF-induced coagulation areas created with (b) an
internally cooled electrode and (c) a multitined expandable electrode. Ablated tissues appeared as a central, firm, dark-brown
area (asterisk) surrounded by an outer, irregular margin (arrows) of bright red tissue. (d) Contrast-enhanced coronal
reconstruction image obtained immediately after RF ablation using a perfusion electrode showing RF-induced hyperattenuated
regions in the left lung (arrows). Note that the hyperattenuated region in the left lung treated with a perfusion electrode is
larger than those produced using an internally cooled electrode (arrowheads) in the right lung. (e) Photographs of a gross
specimen containing an RF-induced coagulation area created with a perfusion electrode. Note that the short- and long-axis
diameters of the coagulation area were larger in Group B using a perfusion electrode than in the other Groups (a and c).

NSCLC. Published studies [16–21] indicate that,
although RF ablation has been offered to a heterogeneous
cohort of lung cancer patients, initial results are
encouraging. According to previous studies, significant
differences are apparent in terms of the complete tumour
necrosis rates of tumours smaller and larger than 3 cm in
diameter [20, 21]. This limitation of lung RF ablation is
primarily due to the fact that, when using a single
electrode device in monopolar mode, it is limited by the
precipitous drop in current density that occurs with
distance from the energy source, which makes the
periphery of the RF lesion particularly prone to vascular
cooling [10, 30]. To circumvent these problems, strategies
to increase the dimension of RF-induced ablation zones
are needed, and several researchers have suggested that
saline-enhanced RF ablation using a perfusion electrode
can expand RF-induced coagulation [30–32].
In our study, RF ablation according to manufacturer’s
recommendations leads to larger volumes of coagulation
using the perfusion electrode than using multitined or
internally cooled electrodes. These results confirm those
of previous studies [22, 23], i.e. that RF ablation using
perfusion electrodes produces significantly larger coa-
gulation areas than internally cooled or multitined
electrodes, when RF ablation is performed according to
manufacturer’s recommendations. In our study, the high
efficacy of the perfusion electrode in creating a large
volume coagulation could be attributed to the low
impedance kept during the ablation procedure, and high
energy delivery during the RF ablation. This phenom-
enon could be explained by the presence of highly ionic
saline around the electrode with continuous saline
infusion during the RF procedure, which improves
tissue conductivity [30, 31]. In addition, the larger
volumes of coagulation obtained with this device are
probably the result of the effects of heated fluid and its
higher thermal conductivity [22].
The shape of a coagulated area is at least as important
as the coagulation volume because the ablation of a
tumour requires that the entire tumour, and a safety
margin of grossly normal tissue, are encompassed by the
ablation. Thus, the creation of large but complex
asymmetric coagulation shapes does not reflect the
effectiveness of a device. In clinical practice, coagulation
shape is determined by the configuration of the RF
electrode, the location of the tumour, tumour consis-
tency, perfusion-mediated cooling effects and, in the
case of open-perfusion devices, the saline distribution
[26]. One potential disadvantage of RF ablation with
saline infusion is the possibility of irregular zones of
coagulation, due to uneven distribution of saline and a
higher complication rate [32–35]. According to a pre-
vious study of RF ablation using a perfusion electrode for
treating human pulmonary tumour by Kim et al [34],
hypertonic saline-enhanced RF ablation was powerful
and efficient in local ablation, but it was difficult to
predict the exact extent of ablation. In our study,
although there was no significant difference between
the Dax/Dtr ratio of RF-induced coagulation areas for
the three kinds of electrodes, the perfusion electrode
created rather irregular bordered coagulations compared
with the internally cooled or multitined expandable
electrodes.
Given that prospective surgical data demonstrate 3
and 2.4 fold increases in local-regional recurrence rates
for local wedge resection and segmental resection,
respectively, compared with lobectomy, RF ablation
alone for the treatment of primary lung cancer may not
be validated [36]. However, for tumours under 2 cm in
diameter (stage IA NSCLC), a recent study that
compared limited resection (segmentectomy) with
lymph node assessment versus lobectomy showed
equivalent 5-year survival and local recurrence rates
[37]. Recent studies on lung RF ablation [18–20] have led
to the opinion that, for small tumours (,3 cm), RF
ablation might provide a viable alternative to surgical
resection for local disease control, especially in the non-
surgical patient cohort. RF ablation can also be used in
conjunction with other treatment modalities. In addition
to the previously mentioned combination with XRT,
there are ongoing studies on combined RF ablation and
brachytherapy in patients with either metastatic lung
malignancies or a history of prior treatment that
precludes additional external beam radiotherapy [17].
The rationale involves the enhancement of local control
by magnifying the cytoreductive and radiation effect by
destroying the central hypoxic area of the target tumour.
As demonstrated in this study, the larger volume of
coagulation created by RF ablation using a perfusion
electrode may increase the clinical utility of RF ablation
therapy by allowing the successful treatment of larger
lung tumours, or by reducing the number of sessions
needed to treat a given tumour.
In our study, the measured volume of RF-induced
hyperattenuated region on CT was greater than that of
the RF-induced coagulation at gross specimen.
Differences between the measured volumes of RF-
induced hyperattenuation on CT and on gross specimens
were attributed to the fact that the hyperattenuated area
on CT corresponded to a region including both central

The British Journal of Radiology, July 2006 569


and peripheral discoloured zones of ablation area on
gross specimen; but on the gross specimens with TTC
staining, the only central whitish area showed no
enzyme activity (no staining).
Several limitations of this study must be addressed.
First, because of the small sample size, interpretations
are limited. Second, differences between the results
obtained for the three RF devices are valid in healthy
lungs, but not for lung tumours. Therefore, the extent to
which our findings reflect the clinical situation is limited.
Third, the duration of RF energy application was
different between the groups. It would be fair to compare
the same duration of energy application, because the
volume of coagulation using internally cooled electrodes
could be increased if RF energy is applied for more than
12 min. However, in our study, all RF ablation proce-
dures were performed according to manufacturer’s
recommendations as in previous studies of porcine liver
[25, 26]. Fourth, although there are modifications of
expandable electrodes and internally cooled electrodes
(cluster) available, we did not include those electrodes in
the present study. Given that the developmental speed of
RF technology is rapid, this study represents a snapshot
in time because further refinements and improvements
of current techniques will undoubtedly increase the
effectiveness and further expand the role of RF ablation.
In addition, we evaluated the geometry of RF-induced
coagulation by the two-dimensional measurements.
Three-dimensional virtual modelling of RF-induced
coagulation in the gross specimens would be beneficial
for the evaluation of coagulation geometry. However,
given the explanted lung collapse, its value may not be as
good as in the liver. Finally, large volumes of coagulation
may not always be beneficial or desirable. In certain
circumstances, coagulation extending beyond the
tumour boundaries could be detrimental if surrounding
structures are damaged or if insufficient tissue is
preserved to permit normal organ function.
In summary, RF ablation according the manufacturer’s
recommendations leads to larger volumes of coagulation
using perfusion electrodes. Based on our study results, we
believe that large lung tumours could be treated by RF
ablation using a perfusion electrode, with longer energy
application, more effectively than with other electrode
types.
References
1. Parkin DM. Global cancer statistics in the year 2000. Lancet
Oncol 2001;2:533–43.
2. Smith RA, Glynn TJ. Epidemiology of lung cancer. Radiol
Clin North Am 2000;38:453–70.
3. Landis SH, Murray T, Bolden S, Wingo PA. Cancer
statistics, 1998, CA. Cancer J Clin 1998;48:6–29.
4. Rajdev L, Keller SM. Neoadjuvant and adjuvant therapy of
non-small cell lung cancer. Surg Oncol 2002;11:243–53.
5. Kobayashi K, Kawamura M, Ishihara T. Surgical treatment
for both pulmonary and hepatic metastases from colorectal
cancer. J Thorac Cardiovasc Surg 1999;118:1090–6.
6. Kandolier D, Kormer E, Tuchler H, End A, Muller MR,
Wolner E, et al. Long-term results after repeated surgical
removal of pulmonary metastases. Ann Thorac Surg
1998;65:909–12.
7. Rowell NP, Williams CJ. Radical radiotherapy for stage I/II
non-small cell lung cancer in patients not sufficiently fit for
570
J M Lee, J K Han, J M Chang et al
or declining surgery (medically inoperable): a systematic
review. Thorax 2001;56:628–38.
8. Non-small Cell Lung Cancer Collaborative Group.
Chemotherapy for non-small cell lung cancer. Cochrane
Database of Systematic Reviews. (2) 2000;CD002139.
9. Rossi S, Garbagnati F, Rosa L, Azzaretti A, Belloni G,
Quaretti P. Radiofrequency thermal ablation for treat-
ment of hepatocellular carcinoma. Int J Clin Oncol
2002;225–35.
10. McGahan JP, Dodd GD. Radiofrequency ablation of the
liver: current status. AJR Am J Roentgenol 2001;176:3–16.
11. Gervais DA, McGovern FJ, Arellano RS, McDougal WC,
Mueller PR. Renal cell carcinoma: clinical experience and
technical success with radiofrequency ablation of 42
tumors. Radiology 2003;226:417–24.
12. Woertler K, Vestring T, Boettner F, Winkelmann W,
Heindel W, Lindner N. Osteoid osteoma: CT-guided
percutaneous radiofrequency ablation and follow-up in 47
patients. J Vasc Interv Radiol 2001;12:717–22.
13. Ahrar K, Price RE, Wallace MJ, Madoff DC, Gupta S,
Morello FA Jr, et al. Percutaneous radiofrequency ablation
of lung tumors in a large animal model. J Vasc Interv Radiol
2003;14:1037–43.
14. Lee JM, Jin GY, Li CA, Chung GH, Lee SY, Han YM, et al.
Percutaneous radiofrequency thermal ablation of lung VX2
tumors in a rabbit model using a cooled tip-electrode:
feasibility, safety, and effectiveness. Invest Radiol
2003;38:129–39.
15. Oyama Y, Nakamura K, Matsuoka T, Toyoshima M,
Yamamoto A, Okuma T, et al. Radiofrequency ablated
lesion in the normal porcine lung: long-term follow-up with
MRI and pathology. Cardiovasc Intervent Radiol
2005;28:346–53.
16. Dupuy DE, Zagoria RJ, Akerley W, Mayo-Smith WW,
Kavanagh PV, Safran H. Percutaneous radiofrequency
ablation of malignancies in the lung. AJR Am J
Roentgenol 2000;174:57–9.
17. Nishida T, Inoue K, Kawata Y, Izumi N, Nishiyama N,
Kinoshita H, et al. Percutaneous radiofrequency ablation of
lung neoplasms: a minimally invasive strategy for inoper-
able patients. J Am Coll Surg 2002;195:426–30.
18. Herrera LJ, Fernando HC, Perry Y, Gooding WE,
Buenaventura PO, Christie NA, et al. Radiofrequency
ablation of pulmonary malignant tumors in nonsurgical
candidates. J Thorac Cardiovasc Surg 2003;125:929–37.
19. Dupuy DE, Mayo-Smith WW, Abbott GF, DiPetrillo T.
Clinical applications of radiofrequency tumor ablation in
the thorax. Radiographics 2002:S259–69.
20. Lee JM, Jin GY, Goldberg SN, Lee YC, Chung GH, Han YM,
et al. Percutaneous radiofrequency ablation for inoperable
non-small cell lung cancer and metastases: preliminary
report. Radiology 2004;230:125–34.
21. Akeboshi M, Yamakado K, Nakatsuka A, Hataji O, Taguchi
O, Takao M, et al. Percutaneous radiofrequency ablation of
lung neoplasms: initial therapeutic response. J Vasc Interv
Radiol 2004;15:463–70.
22. Lee JM, Kim SW, Li CA, Youk JH, Kim YK, Jin Z, et al.
Saline-enhanced radiofrequency thermal ablation of the
lung: a feasibility study in rabbits. Korean J Radiol
2002;3:245–53.
23. Lee JM, Youk JH, Kim YK, Han YM, Chung GH, Lee SY, et
al. Radio-frequency thermal ablation with hypertonic saline
solution injection of the lung: ex vivo and in vivo feasibility
studies. Eur Radiol 2003;13:2540–7.
24. Goldberg SN, Stein MC, Gazelle GS, Sheiman RG, Kruskal
JB, Clouse ME. Percutaneous radiofrequency tissue abla-
tion: optimization of pulsed-radiofrequency technique to
increase coagulation necrosis. J Vasc Interv Radiol
1999;10:907–16.
The British Journal of Radiology, July 2006
In vivo radiofrequency ablation in pig lungs
25. Denys AL, De Baere T, Kuoch V, Dupas B, Chevallier P,
Madoff DC, et al. Radio-frequency tissue ablation of the
liver: in vivo and ex vivo experiments with four different
systems. Eur Radiol 2003;13:2346–52.
26. Pereira PL, Trubenbach J, Schenk M, Subke J, Kroeber S,
Schaefer I, et al. Radiofrequency ablation: in vivo compar-
ison of four commercially available devices in pig livers.
Radiology 2004;232:482–90.
27. Goldlust EJ, Placzynski RP, He YY, Hsu CY, Coldberg MP.
Automated measurement of infarct size with scanned
images of triphenyltetrazolium chloride-stained rat brains.
Stroke 1996;27:1657–62.
28. Goldberg SN, Gazelle GS, Compton CC, Mueller PR,
Tanabe KK. Treatment of intrahepatic malignancy with
radiofrequency ablation: radiologic-pathologic correlation.
Cancer 2000;88:2452–63.
29. ImageJ. http://rsb.info.nih.gov/ij/download.html/ [Accessed
27 March 2006].
30. Goldberg SN. Radiofrequency tumor ablation: principles
and techniques. Eur J Ultrasound 2001;13:129–47.
31. Schmidt D, Trubenbach J, Brieger J, Koenig C, Putzhammer
H, Duda SH, et al. Automated saline enhanced radio-
frequency thermal ablation: initial results in ex vivo bovine
livers. AJR Am J Roentgenol 2003;180:163–5.
The British Journal of Radiology, July 2006
32. Lee JM, Han JK, Choi SH, Kim SH, Lee JY, Shin KS, et al.
Comparison of renal ablation with monopolar radiofre-
quency and hypertonic-saline-augmented bipolar radio-
frequency: in vitro and in vivo experimental studies. AJR
Am J Roentgenol 2005;184:897–905.
33. Boehm T, Malich A, Reichenbach JR, Fleck M, Kaiser WA.
Percutaneous radiofrequency (RF) thermal ablation of
rabbit tumors embedded in fat: a model for RF ablation of
breast tumors. Invest Radiol 2001;36:480–6.
34. Kim TS, Lim HK, Kim H. Excessive hyperthermic necrosis
of a pulmonary lobe after hypertonic saline-enhanced
monopolar radiofrequency ablation. Cardiovasc Intervent
Radiol 2005;29:160–3.
35. Gillams AR, Lees WR. CT mapping of the distribution of
saline during radiofrequency ablation with perfusion
electrodes. Cardiovasc Intervent Radiol 2005;28:476–80.
36. Kodama K, Doi O, Higashiyama M, Yokouchi H.
Intentional limited resection for selected patients with T1
N0 M0 non-small-cell lung cancer: a single-institution
study. J Thorac Cardiovasc Surg 1997;14:347–53.
37. Okada M, Yoshikawa K, Hatta T, Tsubota N. Is segmen-
tectomy with lymph node assessment an alternative to
lobectomy for non-small cell lung cancer of 2 cm or
smaller? Ann Thorac Surg 2002;71:956–61.
571
 The British Journal of Radiology, 79 (2006), 572–577

PET/CT detects abdominal wall and port site metastases of

colorectal carcinoma
1 2 3
E GOSHEN, MD, T DAVIDSON, MD, D ADERKA, MD and 1S T ZWAS, MD

Departments of 1Nuclear Medicine, 2Radiology and 3Oncology, The Chaim Sheba Medical Center,
Tel-Hashomer and Sackler School of Medicine, Tel Aviv University, Israel

ABSTRACT. Abdominal wall metastases from colorectal cancer (CRC) may be resected
with curative results. Such lesions, often indicators of additional intra-abdominal
lesions, may appear in surgical scars, stomas and port site metastases after laparoscope-
assisted surgery (LAS). Post-operative changes, primarily surgical scars, alter local
physical findings making early detection of small lesions challenging. The purpose of
this study was to retrospectively evaluate the contribution of PET/CT to the diagnosis of
recurrent colorectal cancer in the post-operative abdominal wall. 120 patients were
referred for PET/CT with suspected recurrent CRC based on clinical, radiological or
laboratory findings. All underwent whole body PET/CT imaging. 12 of these 120 (10%),
were found to have abdominal wall lesions. A total of 16 abdominal wall lesions were
detected, located to surgical scars, stomas, drain and laparoscope ports. Additional
Received 10 August 2005
findings on PET/CT in this group included liver metastases, intra-abdominal lesions and Revised 1 November 2005
retroperitoneal lymph node involvement. In general, the patients in this small group Accepted 24 November
were young with high grade tumours presenting in advanced stages. In conclusion, 2005
PET/CT appears to be a sensitive tool for the diagnosis of abdominal wall recurrence of
DOI: 10.1259/bjr/25287790
CRC. The accuracy of localization afforded by the fused functional and anatomic
images makes PET/CT a likely tool for diagnosing abdominal wall lesions, including port ’ 2006 The British Institute of
site metastases of other aetiologies. Radiology

Disease recurrence in the abdominal wall from a primary
colorectal cancer is a poorly studied and little understood
phenomenon. Aggressive resection of disease restricted to
the abdominal wall and associated adherent viscera can
result in local disease control with little morbidity and no
mortality [1], yet abdominal wall metastases are often
indicators of recurrent intra-abdominal cancer. Certainly,
the timely diagnoses of abdominal wall lesions and
possible concomitant intra-abdominal disease are crucial
to optimize patient management.
Positron emission tomography (PET) with 18F-FDG
allows functional imaging of malignant tissue. The well-
documented mechanism of FDG uptake is based on
increased glucose metabolism and increased expression
of glucose transporters in malignant cells as compared
with normal tissue. Thus, the metabolic rate of suspi-
cious lesions as determined from PET can be used to
define the presence and extent of active disease. While
the sensitivity of dedicated PET has consistently been
reported to be in the range of 85–100% for the detection
of recurrent colorectal lesions the lack of anatomical
detail affects image interpretation and remains a major
limitation of PET. The combined PET/CT technique now
helps overcome this drawback by providing fused
images of functional PET and anatomic CT studies. The
almost synchronous image acquisition and exact co-
registration of anatomical and metabolic data improves
the anatomic localization of PET abnormalities and
reduces the number of equivocal PET interpretations
[2–4]. PET/CT imaging has been found to increases the
accuracy and certainty of locating lesions in colorectal
cancer with a consequent improvement in staging and
restaging accuracy from 78% to 89% [5].
This report describes the use of PET/CT in the
detection of abdominal wall lesions, including port site
metastases from colorectal carcinoma.
Patients and methods
This was a retrospective evaluation of positive PET
scans. Patients referred for PET/CT with a diagnosis of
colorectal cancer were listed in a prospective database
beginning 01/2004. Review of this database (10/2004)
indicated 120 patients, 12 of whom (3 female, 9 male,
aged 42–73 years, average 57 years) were found to have
recurrent disease in the abdominal wall. The study
population consisted of these 12 patients, who under-
went 13 PET/CT scans. Primary disease was located in
caecum (n53), transverse colon (n52), sigma (n54) and
rectum (n53). All presented initially with advanced
disease (all T3 or T4 according to TNM classification),
and 10 of the 12 had moderate to poor tumour
differentiation (1 was well differentiated, 1 unknown).
Only 5 of the 12 had documented initial lymph node
involvement. Three patients had intestinal perforations,
yet there was no evidence of peritoneal spread at
surgery. Three had liver metastases at presentation. All
were reported to have clean surgical margins. Of the 12
patients, 8 had originally undergone open laparotomy,
and 4 had laparoscope-assisted surgery (LAS). One
patient (pt #4) had both LAS and open surgery, and

572 The British Journal of Radiology, July 2006

PET/CT detects abdominal wall and port site metastases of CRC

Table 1. Demographic and clinical data

Patient sex, Primary site Clinical Surgery Tumour Stage Lymph Other Sites at CEA Treatment
age presentation differentiation nodes presentation

1(F, 42) sigma abdominal open* mod-poor T3N2M1 6 of 15 liver De Gramont

abscess
2(F, 67) cecum vomiting LAP mod-poor T3N1M0 3 of 39 31 5FU-LCV
3(M, 71) cecum obstruction open well-mod T3N2M0 4 of 8 5FU-LCV
4(M, 73) cecum obstruction open mod T4N2M0 Folfiri
LAP
5(M, 56) rectum perforation open* mod T3N0M0 0 of 5 no
6(M, 45) transverse open well diff. T3N1M0 3 of 9 7.9 Folfiri
7(M, 46) sigma into bladder open* mod T4N0M1 0 of 26 liver XRT+UFT
8(M, 73) sigma anaemia LAP mod T3N0M0 0 of 6 no
9(F, 49) rectal pain open NA NA NA appendix
10(M, 57) transverse pain LAP mod T3N1M0 1of 4 5FU-LCV
11(F, 59) sigma obstruction open mod T3N0M0 0 of 46 peritonitis refused
12(M, 46) rectum pain open mod T3N0M1 0 of 4 liver 29 Folfiri
*Patients with perforations at surgery.

also underwent two PET scans. The clinical data of these
patients are summarized in Table 1.
All 12 patients presented at follow-up with clinical,
radiological or laboratory findings suggestive of
recurrent disease. Four patients were referred for PET/
CT due to palpable masses, one had intestinal obstruc-
tion, one had CT findings suggestive of local recurrence
and the remaining six were referred for PET/CT due to
marker elevation. Patient #4 was referred both times due
to palpable abdominal lesions.
The interval between initial surgical procedures and
PET/CT for suspected recurrence in this group ranged
between 4 months and 37 months, the average being
15 months. The indications for scan and intervals from
surgery are presented in Table 2, together with PET/CT
results.
PET/CT imaging protocol
All patients were imaged with a Gemini PET/CT
Imaging System (Philips Co., USA) which provides
three-dimensional acquisition, processing and display of
CT, PET and PET/CT images, with 6.0 mm PET spatial
resolution and a dual slice MX800 EXP CT scanner. Whole
body PET/CT imaging was performed in fasting patients
(4–6 h) following intravenous administration of 370 MBq
(10 mCi) of 18F-FDG. Neither oral nor intravenous
contrasts were administered. After a 60 min uptake
period, during which patients were instructed to rest
silently, images were acquired. First, a CT surview,
(30 mA, 120 kVp, FOV 500 mm, length of scan 1.0–1.5 m
with a speed of 100 mm s21 and a spatial resolution of
1 mm) was performed from the orbital level to the level of
the proximal thigh. This was followed by a dual slice CT
(50 mAs per slice, 120 kVp, with a slice thickness of
6.5 mm, length of scan according to result of surview, with
a bed speed of 20 mm s21, rotation time of 0.75 s and a
pitch of 1.5, and FOV of 600 mm). Finally, the acquisition
of PET emission images was performed (2–3 min per bed
position of 8.4 cm). The total acquisition time, accumulat-
ing between 100 and 150 million useful events, varied
between 25 min and 35 min per patient. The CT data were
used for attenuation correction of PET emission images.

Table 2. Scan indications, recurrence time and PET/CT findings

18
Pt# Indication for Post-op interval Lesion # Lesion Size (cm) F-FDG uptake Abdominal wall Other recurrence sites
PET/CT (months) intensity location

1 CEA 10 1 1.3 + Scar None

2 CEA 9 2 1.4 +++ LAP Retroperitoneal LN
3 Ca-19.9 21 3 2.061.0 + Scar (hernia) Intra-abdominal spread
4 Palpable 28 4 1.663.0 +++ Scar Local recurrence
5 1.061.2 +++ Scar *
Palpable 4 6 2.063.0 +++ LAP Local recurrence
5 CT 10 7 1.0 +++ Stoma none
6 Obstruction 29 8 1.761.0 +++ Scar Local recurrence
37 9 2.063.0 +++ Stoma *
7 CEA 6 10 1.3 + Scar Liver
8 Palpable 27 11 567 +++ LAP none
9 Palpable 12 12 362 +++ Scar Liver
10 18 13 1.4 +++ LAP Intra-abdonimal spread
11 CEA 10 14 1.0 ++ Drain Liver
15 1.0 ++ Drain *
12 CEA 7 16 1.061.0 ++ Drain (hepatic) Local Recurrence
*Same site of recurrence as listed above in previous scan of same patient.

The British Journal of Radiology, July 2006 573


Non-attenuated data was reconstructed after scan acquisi-
tion was completed. Reconstruction of attenuation cor-
rected data was executed concurrently. The lower limit of
resolution of the PET system used is 0.45 cm.
Image interpretation
Both attenuation corrected and non-attenuation cor-
rected scans were coregistered with the CT for interpreta-
tion using Syntegra (Version 2.0j, Philips) software. All
studies were visually interpreted independently by
individual physicians from the two specialties (Nuclear
Medicine and Diagnostic Radiology). Studies were inter-
preted for PET alone and for combined PET/CT data. The
low-dose non-enhanced CT data was used only in
conjunction with PET data. Comparisons between the CT
data and PET regarding contributions in the specific
clinical context reported here were not made.
Evaluation of accuracy and statistical analysis
Results from combined PET/CT data were validated
by comparison with concurrent contrast-enhanced CT,
histopathological findings and at least 3 months of
clinical follow-up. The contrast-enhanced CT scans used
for comparison were not performed in our department,
and were provided by the patients. PET/CT findings
were interpreted on both lesion and patient levels, and
determined to be true-positive (presence of cancer), or
false-positive (increased FDG uptake unrelated to can-
cer). As the patients were selected on the basis of a
positive PET result, there can be no meaningful assess-
ment of true or false negatives.
Results
A total of 16 abdominal wall lesions were detected in
13 scans performed in 12 patients. Three patients had 2
Figure 1. (a) Non enhanced CT demonstrates small irregularity in right anterior abdominal wall. (b) Attenuation-corrected PET
image demonstrates focus of increased uptake of 18F-FDG in corresponding site in abdominal wall. The additional focus of intra-
abdominal uptake on the right is due to physiological uptake in the ureter. (c) Fused PET/CT image yields functional and
anatomic information.
574
E Goshen, T Davidson, D Aderka and S T Zwas
lesions each (pts 6, 11 and 4 who had 3 sites in 2 scans).
The lesions ranged in size from 1.0–567 cm, with most
(9/16) measured less than 1.5 cm. Intensity of 18F-FDG
uptake was variable, yet the majority of lesions (10/16)
had intense 18F-FDG avidity. The least uptake was
observed in the smallest lesions.
All lesions were noted from PET findings and were
subsequently localized anatomically by CT. Seven find-
ings were localized to mid-abdominal surgical scars (1/7
in the site of a hernia), four localized to laparoscope
insertion sites, two were in stomas and three in sites of
surgical drains (1/3 in a hepatic drain site). The non-
enhanced CT findings generally depicted heterogeneous
nodular lesions, described as soft tissue masses with ill
defined borders located near muscle. In some cases, fat
infiltration was observed adjacent to the soft tissue mass,
consistent with post-operative changes. No increased
uptake of 18F-FDG was observed in any such region of fat
infiltration. The abdominal lesions were divided nearly
equally between open-surgery scars and drain/LAP
sites. The lesions depicted essentially the same functional
and anatomical characteristics (demonstrated in
Figure 1). There were no suspicious findings on CT
which were undetected by PET (no negative PET scans),
and no CT findings which were below the resolution of
the PET. All abdominal wall lesions were previously
unreported.
Concomitant disease was present in 9 of 12 patients.
The PET/CT findings demonstrated additional sites of
active disease in 11 of 13 scans as follows: retroperitoneal
lymphadenopathy (n51), intra-abdominal spread (n52),
liver metastasis (n53) and local recurrence (n54),
including repeated findings twice in pt #4. Findings
from PET/CT are summarized in Table 2.
All PET/CT findings in this group were interpreted as
true-positive, i.e. consistent with presence of cancer. Two
abdominal wall lesions (pts #5 and #12) were highly
suspicious clinically, while all other occult sites of
abdominal wall and local recurrences were validated
histologically.
The British Journal of Radiology, July 2006
PET/CT detects abdominal wall and port site metastases of CRC
The patient profile from the data of this small group
included young age (8/12 below 60 years of age, of
whom 5 were in their 50s), advanced disease at
presentation (8/12 patients T3; 3/12 patients T4, and 1
unknown), with tumour differentiation being mostly
moderate to poor.
Discussion
Changes in tumour metabolism precede morphologic
changes, and the functional information obtained from
PET is therefore essentially independent of tumour
location and size. The additional anatomic information
provided by the CT contributed to the localization and
characterization of lesions, thereby increasing the accu-
racy of the PET diagnoses. Overall, PET scanning has
been reported to be contributory in the initial staging as
well as in the evaluation of recurrent colorectal carcino-
mas. Recent works have reported the successful use of
FDG-PET for the assessment of tumour aggressiveness,
staging of disease, evaluation of treatment response, and
the detection of recurrent disease [6]. While abdominal
wall lesions and port site metastases are well docu-
mented entities, to the best of our knowledge the use of
PET or PET/CT specifically addressing their diagnostic
contribution has not been previously described.
Abdominal wall recurrences have been reported to
occur in approximately 60% of patients with advanced
colorectal cancer [7]. The importance of early and
accurate diagnosis of such recurrence is not only in the
localization of disease, allowing resection with curative
intent, but also in the fact that abdominal wall metastases
are often indicators of recurrent intra-abdominal cancer
[7]. In the small study group evaluated here in fact, all
but three patients (9/12 patients, in 11/13 scans) had
additional sites of active disease. In three patients the
additional sites were previously diagnosed by CT (in two
patients referred for evaluation of palpable masses, and
one with obstruction on previous CT) whereas pre-
viously unsuspected sites of disease were disclosed in
another six patients. This finding is particularly striking
as it was observed in the asymptomatic patients referred
for evaluation due to elevated markers. In this subgroup,
the lack of disease documentation on their previous CT
was, in fact, the reason for referral to PET. When
dividing our population into subgroups according to
indication for the scan, this was also the largest
subgroup.
Abdominal wall lesions, long recognized in surgical
scars and stomas, have received renewed attention due
to the recognition of port site metastases in patients
following laparoscopic colorectal resections [1].
Although beneficial to the patient in the immediate
post-operative period, the initial descriptions of port site
lesions had put the adequacy of laparoscopic-assisted
colectomy for tumour under question [8]. The incidence
of port site metastases, however, is now recognized to be
close to the incidence of wound metastases after open
surgery [9]. Clinical evidence indicates that wound
recurrence rates are 0.60% and 0.85% for open and
laparoscopic colon cancer operations, respectively [10].
The group investigated here was derived from a larger
group of 120 patients with suspected recurrent CRC, yet
The British Journal of Radiology, July 2006
no evaluation of the entire population was performed,
particularly regarding initial surgical procedure.
Interestingly, however, the limited results in this small
group indicate the abdominal lesions to be divided
nearly equally between open-surgery scars and trocar/
LAP sites. Excluding the two lesions localized to stomas,
of the remaining lesions, half (7/16) were in incision sites
from open surgery, and half (7/16) in trocar and
laparoscope insertion sites. The lesions were essentially
the same functionally and anatomically, and differed, in
fact, only by location.
Increased uptake in stoma sites is frequently observed
and well recognized. Physiological intestinal uptake is
generally mild, but may be more increased if com-
pounded by inflammation, complicating scan interpreta-
tion. Anatomic localization of the uptake by CT, and
particularly the definition of soft tissue changes (nodular
lesions, or infiltration) surrounding the stoma, allows
clearer diagnoses with greater confidence when tumour
is suspected. The retrospective evaluation of all post-
operative CT images helps to discern physiological
findings from pathology, including inflammatory
changes (subcutaneous fat infiltration, or fluid collec-
tions suggestive of abscess).
Two tumours in stoma were demonstrated, and later
pathologically proven, in the small group reported here.
In both cases (pts 5 and 6) the uptake of 18F-FDG in the
stoma was intense, yet could have been confused with
inflammation. In these cases the additional anatomic
information provided by the CT served to differentiate
the findings from physiological uptake and inflamma-
tory changes. The intensely increased 18F-FDG uptake,
which was localized to ill-defined soft tissue masses in
the abdominal wall surrounding the stoma, allowed for
the diagnosis of viable tumour.
This is not a report on all CRC patients referred for
PET/CT, but rather a retrospective evaluation of 12
patients with abdominal wall lesions. Only two patients
presented with stoma lesions. No remarkable findings
were noted in stoma sites in the remainder of the group
(data not presented). As this is a report on the 12 patients
with positive scans, there is also no information regard-
ing the incidence of inflammatory uptake in stoma in the
overall CRC population evaluated.
This report does not attempt to compare the contribu-
tion of PET/CT with that of CT in the context of
abdominal wall lesions. Certainly, if abdominal wall
lesions are large enough to be detected on conventional
CT, and considering the high prevalence of concomitant
disease in the presence of abdominal wall metastases,
PET/CT may even be non-contributory and not cost-
effective. Certainly, if the lesions, particularly intra-
abdominal findings, are discovered (and duly reported)
from CT, the PET may be unnecessary as curative
options are limited. Our population was comprised of
patients who presented at follow-up with clinical,
radiological or laboratory findings suggestive of recur-
rent disease. Four patients were referred for PET/CT for
evaluation of palpable masses, one had intestinal
obstruction, one had equivocal CT findings suggestive
of recurrence. The remaining six were referred for PET/
CT due to marker elevation in the presence of ‘‘normal’’
CT.
575

Interestingly, while all abdominal wall lesions were at
least 1 cm in size, and could have been identified on the
concurrent contrast-enhanced CT, this was not the case.
Unfortunately, missed diagnoses occurred. The PET
consistently elucidated equivocal findings, and yielded
‘‘hot-spots’’ which easily directed attention to under-
lying lesions. The point which cannot be overstressed is
that PET ‘‘shines the light’’ on small, sometimes over-
looked, findings.
Port sites and open wounds are at equal risk of tumour
implantation, with wound implantation affected mostly
by the stage of the tumour and the operative techniques
[10]. While our group is of limited size, and a complete
demographic evaluation of the entire population (120
patients with suspected recurrent CRC) was not per-
formed, a patient profile can be outlined. In general, the
patients who developed abdominal wall recurrences
were mostly young, and presented initially with
advanced disease of moderate–poor differentiation. The
general population (120 patients) from which this
subgroup was derived includes patients in all age groups
referred for evaluation, regardless of the stage of CRC at
presentation. As such, this subgroup, with its relative
homogeneity, is not representative of the entire popula-
tion. While none were reported to have intra-abdominal
dissemination at surgery, local spread of microscopic
disease could not be excluded, particularly in two
patients (pts #1 and 11, who presented initially with
abscess, and peritonitis, respectively). As expected, the
abdominal wall lesions were indistinguishable clinically
whether they resulted in patients after open surgery or
developed in port sites after laparoscope-assisted proce-
dures.
The average time interval between resection of
primary tumour and PET/CT diagnosis of recurrence
was 15 months in this group (ranging 4–37 months). This
is in keeping with anecdotal reports of abdominal wall
recurrences. A case of a port site recurrence of colonic
adenocarcinoma with diffuse peritoneal carcinomatosis
was reported 1 month after laparoscopic-assisted right
hemicolectomy, suggesting that intraperitoneal dissemi-
nation and tumour implantation on surgical wounds
may have been the principal mechanism of recurrence
after laparoscopic surgery [9]. In another case reported, a
drain-site tumour recurrence was diagnosed 2 years after
right colon resection for adenocarcinoma of the ascend-
ing colon [11].
Itano et al [12] reported on a patient who developed
two subcutaneous metastases at two trocar sites.
Multiple lesions in the abdominal wall, as observed in
three of our patients (pt # 6, #11 and #4) are, therefore,
also in keeping with the literature. While the patient
group is too small to allow statistical evaluation, the fact
that three out of 12 patients had multiple abdominal wall
lesions is an interesting observation which may warrant
further investigation and should, perhaps, be taken into
consideration during follow-up of patients undergoing
such procedures.
Of special interest is the lesion detected in the insertion
site of a hepatic drain (pt #12). While a similar finding
has been previously described, the occurrence is rare. A
previous case has been reported in the literature of a
colon carcinoma metastasizing to a port site following
laparoscopic cholecystectomy has been reported. That
576
E Goshen, T Davidson, D Aderka and S T Zwas
case too was of an advanced tumour at the time of
diagnosis, in an immunosuppressed patient, who deve-
loped clinical recurrence within 3 months of her
procedure [13].
Previous reports of abdominal wall metastases after
laparoscopic procedures for colorectal cancer have
suggested the finding to be anecdotal [12–14]. In our
group, however, 12 of 120 (10%) of those initially
referred for investigation of suspected recurrent CRC
had PET/CT findings consistent with abdominal wall
lesions. Excluding the two lesions localized to stomas,
of the remaining lesions half (7/16) were in incision
sites from open surgery and half (7/16) in trocar and
laparoscope insertion sites. This appears to be in keeping
with large-scale studies that have shown the actual rate
of port site metastasis to be similar to that observed in
open surgery [12]. Furthermore, PET/CT disclosed
concomitant findings in the vast majority (75%) of our,
albeit small, group.
PET/CT is frequently indicated for the restaging of
disease in CRC patients prior to metastasectomy. As
limited disease is considered operable with curative
intent, the scan is performed with hopes of demonstrat-
ing only limited disease, thereby allowing the planned
surgery to proceed. While the presence of additional
findings may contraindicate surgery, abdominal wall
and port site metastases per se should not serve as
contraindications, but, rather, should be considered for
resection as well. The combined findings should be taken
into consideration when planning treatment and the
presence of a resectable abdominal wall lesion should
not exclude surgery as an option.
Conclusion
Although preventive measures have been proposed
[9], it seems likely that abdominal wall lesions, including
port site metastases, will continue to present diagnostic
challenges. A well-documented complication after
laparoscopic resection of intra-abdominal malignancies,
port site metastases have been reported in various other
clinical settings including breast cancer [15], gallbladder
and bile duct cancer [16, 17], renal cell carcinoma [18],
pancreatic cancer [19], upper gastrointestinal tract
cancers [20] and gynaecological tumours including
ovarian, cervical and endometrial cancer [21–23].
The metabolic images provided by PET scanning
appear to have a high positive predictive value regarding
the detection of abdominal wall metastases from colo-
rectal cancer. The implementation of PET/CT will likely
allow detection, characterization and anatomic localiza-
tion of such occult lesions, as well as the diagnosis of
additional concomitant lesions, in other clinical settings
as well.
References
1. Koea JB, Lanouette N, Paty PB, Guillem JG, Cohen AM.
Abdominal wall recurrence after colorectal resection for
cancer. Dis Colon Rectum 2000;43:628–32.
2. Beyer T, Townsend DW, Brun T, Kinahan PE, Charron M,
Roddy R, et al. A combined PET/CT scanner for clinical
oncology. J Nucl Med 2000;41:1369–79.
The British Journal of Radiology, July 2006
PET/CT detects abdominal wall and port site metastases of CRC
3. Schoder H, Erdi YE, Larson SM, Yeung HW. PET/CT: a
new imaging technology in nuclear medicine. Eur J Nucl
Med Mol Imaging 2003;30:1419–37.
4. Martinelli M, Townsend D, Meltzer C, Villemagne VV.
Survey of results of whole body imaging using the PET/CT
at the University of Pittsburgh Medical Center PET Facility.
Clin Positron Imaging 2000;3:161.
5. Cohade C, Osman M, Leal J, Wahl RL. Direct comparison of
(18)F-FDG PET and PET/CT in patients with colorectal
carcinoma. J Nucl Med 2003;44:1797–803.
6. Hustinx R. PET imaging in assessing gastrointestinal tumor.
Radiol Clin North Am 2004;42:1123–39.
7. Silecchia G, Perrotta N, Giraudo G, Salval M, Parini U,
Feliciotti F, et al; For the Italian Registry of Laparoscopic
Colorectal Surgery. Abdominal wall recurrences after
colorectal resection for cancer: results of the Italian registry
of laparoscopic colorectal surgery. Dis Colon Rectum
2002;45:1172–7.
8. Jacquet P, Averbach AM, Jacquet N. Abdominal wall
metastasis and peritoneal carcinomatosis after laparo-
scopic-assisted colectomy for colon cancer. Eur J Surg
Oncol 1995;21:568–70.
9. Curet MJ. Port site metastases. Am J Surg 2004;187:705–12.
10. Allardyce RA. Is the port site really at risk? Biology,
mechanisms and prevention: a critical view. Aust N Z J
Surg 1999;69:479–85.
11. Torzilli G, Cremascoli G, Cattaneo S, Stefanini P, Olivari N.
Drain-site tumour recurrence after laparotomy resection for
colorectal cancer. Eur J Surg Oncol 1999;25:546–7.
12. Itano O, Watanabe T, Jinno H, Suzuki F, Baba H, Otaka H. Port
site metastasis of sigmoid colon cancer after a laparoscopic
sigmoidectomy: report of a case. Surg Today 2003;33:379–82.
13. Neuhaus S, Hewett P, Disney A. An unusual case of port
site seeding. Surg Endosc 2001;15:896.
14. Lauter DM, Froines EJ. Initial experience with 150 cases of
laparoscopic assisted colectomy. Am J Surg 2001;181:398–403.
The British Journal of Radiology, July 2006
15. Mylonas I, Janni W, Friese K, Gerber B. Unexpected
metastatic lobular carcinoma of the breast with intraab-
dominal spread and subsequent port-site metastasis after
diagnostic laparoscopy for exclusion of ovarian cancer.
Gynecol Oncol 2004;95:405–8.
16. Nakagawa S, Tada T, Furukawa H, Abe M, Hatakeyama K.
Late-type recurrence at the port site of unexpected
gallbladder carcinoma after a laparoscopic cholecystect-
omy: report of a case. Surg Today 2000;30:853–5.
17. Sakata N, Suzuki M, Shibuya K, Takeda K, Matsuno S.
Unexpected bile duct carcinoma presenting with port-site
metastasis after laparoscopic cholecystectomy for cholecys-
tolithiasis. Hepatobiliary Pancreat Surg 2002;9:511–4.
18. Iwamura M, Tsumura H, Matsuda D, Kurosaka S, Yoshida
K, Baba S. Port site recurrence of renal cell carcinoma
following retroperitoneoscopic radical nephrectomy with
manual extraction without using entrapment sac or wound
protector. J Urol 2004;171:1234–5.
19. Jorgensen JO, McCall JL, Morris DL. Port site seeding after
laparoscopic ultrasonographic staging of pancreatic carci-
noma. Surgery 1995;117:118–9.
20. Shoup M, Brennan MF, Karpeh MS, Gillern SM, McMahon
RL, Conlon KC. Port site metastasis after diagnostic
laparoscopy for upper gastrointestinal tract malignancies:
an uncommon entity. Ann Surg Oncol 2002;9:632–6.
21. Viala J, Morice P, Pautier P, Castaigne D, Vanel D. CT
findings in two cases of port-site metastasis after laparo-
scopy for ovarian cancer. Eur J Gynaecol Oncol 2002;23:
293–4.
22. Gregor H, Sam CE, Reinthaller A, Joura EA. Port site
metastases after laparoscopic lymph node staging of
cervical carcinoma. J Am Assoc Gynecol Laparosc 2001;8:
591–3.
23. Muntz HG, Goff BA, Madsen BL, Yon JL. Port-site
recurrence after laparoscopic surgery for endometrial
carcinoma. Obstet Gynecol 1999;93(5 Pt 2):807–9.
577
 The British Journal of Radiology, 79 (2006), 578–583

Quantitative colorectal cancer perfusion measurement by

multidetector-row CT: does greater tumour coverage improve
measurement reproducibility?
1,2
V GOH, MA, MRCP, FRCR, 1S HALLIGAN, MD, FRCP, FRCR,
1
L GARTNER, MRCP,
1
P BASSETT, MSc and
1
C I BARTRAM, FRCP, FRCS, FRCR

1
Intestinal Imaging Centre, St Mark’s Hospital, Harrow, and 2Paul Strickland Scanner Centre, Mount
Vernon Hospital, Northwood, UK

ABSTRACT. The purpose of this study was to determine if greater z-axis tumour
coverage improves the reproducibility of quantitative colorectal cancer perfusion
measurements using CT. A 65 s perfusion study was acquired following intravenous
contrast administration in 10 patients with proven colorectal cancer using a four-
detector row scanner. This was repeated within 48 h using identical technical
parameters to allow reproducibility assessment. Quantitative tumour blood volume,
blood flow, mean transit time and permeability measurements were determined using
commercially available software (Perfusion 3.0; GE Healthcare, Waukesha, WI) for data
obtained from a 5 mm z-axis tumour coverage, and from a 20 mm z-axis tumour
coverage. Measurement reproducibility was assessed using Bland-Altman statistics, for
a 5 mm z-axis tumour coverage, and 20 mm z-axis tumour coverage, respectively. The
mean difference (95% limits of agreement) for blood volume, blood flow, mean transit
time and permeability were 0.04 (22.50 to +2.43) ml/100 g tissue; +8.80 (250.5 to +68.0)
Received 24 August 2005
ml/100 g tissue/min; –0.99 (28.19 to +6.20) seconds; and +1.20 (25.42 to +7.83) ml/100 g Revised 7 November 2005
tissue/min, respectively, for a 5 mm coverage, and 20.04 (22.61 to +2.53) ml/100 g Accepted 24 November
tissue; +7.40 (250.3 to +65.0) ml/100 g tissue/min; 22.46 (212.61 to +7.69) seconds; 2005
and 20.23 (28.31 to +7.85) ml/100 g tissue/min, respectively, for a 20 mm coverage,
DOI: 10.1259/bjr/18842556
indicating similar levels of agreement. In conclusion, increasing z-axis coverage
does not improve reproducibility of quantitative colorectal cancer perfusion ’ 2006 The British Institute of
measurements. Radiology

Perfusion imaging techniques have been advocated
increasingly for assessment of tumour response to
therapy because antiangiogenic and vascular targeting
drugs may not cause tumour shrinkage, despite being
therapeutically effective. As a result, conventional
response assessment based on changes in tumour size,
such as RECIST or WHO criteria [1, 2], may under-
estimate the effect of these agents [3]. For example,
Hurwitz et al reported a 5 month improvement in overall
survival in patients with advanced colorectal cancer who
had been treated with bevacizumab (Avastin; Genentech,
CA), an antivascular endothelial growth factor agent:
this was accompanied by a response rate of only 10%
using conventional size-based criteria [4].
Tumours are spatially and temporally heterogeneous;
at any time point they may recruit 20–85% of the
vasculature available to them [5]. Thus functional
assessment derived from a single tumour level (or time)
This research was supported by a grant from the Royal College of
Radiologists, London, UK.
Address correspondence to: Professor Steve Halligan, Intestinal
Imaging Centre, St Mark’s Hospital, Watford Road, Harrow,
Middlesex HA1 3UJ, UK. Current address: Specialist Radiology,
Level 2 Podium, University College Hospital, 235 Euston Road,
London NW1 2BU, UK.
may not reflect perfusion in the tumour as a whole.
Greater anatomical coverage is now possible with
multidetector row CT scanners and this is likely to
increase in the future as ever more detector rows become
possible. To date, there has been no attempt to determine
if increased tumour coverage results in decreased
measurement variability for colorectal cancer. We
obtained colorectal cancer perfusion measurements from
a single 5 mm axial slice, and from 4 contiguous 5 mm
slices, producing a z-axis coverage of 20 mm, to
determine if reproducibility is enhanced when a greater
volume of data is analysed.
Materials and methods
Subjects
This prospective study received local ethics committee
approval, and written informed consent was obtained
from all patients. 10 patients (mean age 67 years, range
39–84 years; 6 male, 4 female) with histologically proven
colorectal adenocarcinoma, and attending for a CT
staging examination, were recruited prospectively. Of
the 10 tumours examined, 5 were located in the rectum

578 The British Journal of Radiology, July 2006

Colorectal perfusion measurement by multidetector-row CT
and 5 in the sigmoid colon. Of these, 2 were T2, 5 were
T3, and 3 were T4 tumours (mean size 6.2 cm, range 5–
11.5 cm).
CT scanning
Following a 4 h fast, 1000 ml of water-soluble contrast,
2–4% meglumine and sodium diatrizoate (Gastrografin;
Bracco, Milan, Italy), was ingested orally 30 min prior to
CT scanning in order to opacify the small bowel, as per
normal practice in our institution. With the patient lying
supine on the scanner table, an 18 G venous cannula was
sited in the antecubital fossa. 20 mg of the spasmolytic
hyoscine N butylbromide (Buscopan; Boehringer
Ingelheim, Ingelheim am Rheim, Germany) was admi-
nistered intravenously to reduce bowel peristalsis.
Patient movement was minimized by placing a restrain-
ing band around the abdomen.
All patients were scanned using a four-detector row
CT scanner (Lightspeed Plus; GE Healthcare
Technologies, Waukesha, WI). A non-contrast abdom-
inal-pelvic study was performed initially in order to
confirm the location of the known colorectal tumour,
using the following parameters: slice thickness/interval
10 mm/5 mm, mode HS/speed 30 mm s21 (pitch 1.5),
120 kV, 180 mA, 0.6 s rotation speed, scan field of view
(SFOV) 50 cm, matrix 512 mm6512 mm. The images
were then inspected on the CT console by the super-
vising radiologist, the mid-tumour level identified, the
scan location noted and thus used to plan the subsequent
dynamic study.
The dynamic studies were performed using contrast
and image acquisition protocols in accordance with
manufacturer’s recommendations. A pump injector
(Percupump Touchscreen; EZ-EM, Westbury, NY) was
used to inject 100 ml of iopamidol 340 (Niopam 340;
Bracco) intravenously at a rate of 5 ml s21. Four
contiguous slices collimated to 5 mm each were obtained
at 1 s intervals through the mid-point of the tumour
using a ‘‘cine mode’’ (120 kV, 60 mA, SFOV 50 cm,
matrix 512 mm6512 mm). Data acquisition commenced
5 s following the start of intravenous injection, to obtain
baseline non-contrasted images, and lasted for a total
duration of 65 s.
All patients returned within 48 h of the initial study
for a second dynamic study specifically to assess
measurement reproducibility. Scans were acquired in
an identical fashion to the initial study. In particular, the
second non-contrast planning scan was compared with
that used for the initial study so that the tumour level
examined for each study could be matched. Intravenous
spasmolytic and contrast were administered exactly as
previously and data was acquired using technical
parameters identical to the initial dynamic study.
Image analysis
Image analysis was performed by a single radiologist
experienced in CT perfusion analysis. All 20 dynamic
studies (10 patients; 2 studies each) were analysed on a
stand-alone workstation (Advantage 4.1; GE Healthcare
Technologies) using commercial software based on
The British Journal of Radiology, July 2006
deconvolution analysis (Perfusion 3.0; GE Healthcare
Technologies). The initial 65 s dynamic study for each
patient was loaded into the software (Body tumour,
Perfusion 3.0) and a single 5 mm axial slice that best
visualized the tumour was chosen from the four axial
slices available. A processing threshold of 0–120
Hounsfield Units was selected so that the subsequent
analysis appropriately included soft tissue, both unen-
hanced and enhanced.
The arterial input was determined by manually
selecting a circular region of interest (ROI) from the
control panel and placing this, using a mouse, within
either the iliac or femoral arteries, whichever was best
visualized in the imaging plane. Arterial attenuation
change was determined over the 65 s acquisition by the
software. A time-attenuation curve was generated auto-
matically, and from this the timing of the end of the first
pass of contrast could be estimated by visual inspection
of the shape of the time-attenuation curve. Subsequent
selection of this time point separating the two phases of
contrast enhancement, as necessitated by the software
program, then permitted generation of the four perfusion
parametric maps (blood volume, blood flow, mean
transit time and permeability) for all of the tissues
within the imaging plane, within the processing thresh-
old selected. This arterial ROI was saved using the
software so that the exact same sized ROI could be
automatically placed in the same location in subsequent
analysis in an attempt to minimize measurement
variability due to this confounder.
A ROI was then drawn freehand around the peri-
pheral margin of the tumour using an electronic cursor
and mouse. Care was taken to exclude perirectal or
pericolonic fat and also intraluminal gas, a process that
was facilitated by viewing a cine-loop of the perfusion
acquisition in order to gauge the degree and margins of
patient movement during acquisition. A time-attenua-
tion curve for the selected tumour tissue and the four
perfusion parameters within the tumour ROI were then
generated. Mean values for the four tumour perfusion
parameters (blood volume, blood flow, mean transit time
and permeability) from this single 5 mm axial slice were
recorded for each individual patient.
Image analysis was repeated for the remaining three
contiguous axial slices in the same manner, recalling the
arterial ROI to minimize variability due to this. The
tumour ROI was drawn freehand due to morphological
differences in each slice. Mean values for all four
perfusion parameters were recorded for each of the
three contiguous slices. Then, in order to obtain overall
mean values for all four perfusion parameters for an
equivalent z-axis tumour coverage of 20 mm, values
were averaged from all four 5 mm axial slices, and
recorded for each individual patient.
Image analysis was performed by the same investi-
gator for the second set of dynamic scans obtained from
each of the 10 patients. Analysis was performed exactly
as previously, for a single 5 mm axial slice, most similar
to the initial analysis, and for all four slices. Mean values
for the four tumour perfusion parameters (blood volume,
blood flow, mean transit time and permeability) for a z-
axis tumour coverage of 5 mm, and for a z-axis tumour
coverage of 20 mm were recorded for each individual
patient.
579

Statistical analysis
The mean (standard deviation (SD)) of tumour blood
volume, blood flow, mean transit time and permeability
measurements from a 5 mm and 20 mm z-axis tumour
coverage were determined. Measurement reproducibility
was assessed using the Bland Altman test statistic [6, 7]:
the mean difference, standard deviation (SD) of the
differences, and 95% limits of agreement were calculated
for each of the four perfusion parameters (blood volume,
blood flow, mean transit time and permeability) for a
5 mm and 20 mm z-axis tumour coverage.
Results
The mean value (standard deviation, SD) for each
perfusion parameter and the mean difference, SD of the
differences, and 95% limits of agreement are summar-
ized in Table 1, with corresponding scatter and Bland-
Altman agreement plots for blood volume and flow
shown in Figures 1 and 2. Similar limits of agreement
were obtained between measurements from a 5 mm and
20 mm z-axis coverage, indicating no improvement in
measurement reproducibility.
Discussion
CT perfusion has been advocated to monitor the effects
of antivascular and antiangiogenic drugs because this
quantitative technique allows in vivo assessment of
tumour vascularity [8], and is more robust than
techniques such as dynamic contrast enhanced MRI [9].
In vivo functional assessment is advantageous because
histological methods that measure angiogenesis, such as
microvessel density counting, may not accurately reflect
functional tumour vascularity [5]. However, whether
assessment of tumour perfusion at a single tumour level,
rather than of the whole tumour, is representative and
adequate is questionable, particularly in the context of
therapeutic assessment.
Multidetector row CT permits a number of contiguous
axial images to be acquired at a given tumour level, and
z-axis coverage is contingent on the number of detector
rows. For example, the maximum tumour coverage
Table 1. 95% limits of agreement analysis showing the mean difference, standard deviation (SD) of differences and 95% limits
of agreement for repeated measurements of tumour blood volume, blood flow, mean transit time and permeability for a 5 mm
and 20 mm z-axis tumour coverage. The overall mean value (SD) for each perfusion parameter is also shown
Tumour perfusion z-axis coverage Mean (SD)
measurement (mm)
Blood volume 5 6.1 (1.3)
(ml/100 g tissue)
20 5.9 (1.3)
Blood flow (ml/100 g 5 91.1 (34.4)
tissue/min)
20 81.1 (28.9)
Mean transit time 5 7.3 (3.3)
(seconds)
20 8.5 (3.9)
Permeability (ml/100 g 5 14.1 (3.6)
tissue/min)
20 13.6 (2.9)
580
V Goh, S Halligan, L Gartner et al
achievable for a perfusion study on the four-detector
row scanner is 20 mm in the z-axis dimension, consisting
of 4 contiguous 5 mm axial slices in cine mode.
However, the rapid pace of technological advancement
has meant that a 40 mm acquisition in the z-axis
dimension, consisting of 8 contiguous 5 mm axial slices,
will soon be possible for a perfusion study with the latest
64 detector row scanners. It seems logical to assume that
by increasing tumour coverage, a more representative
assessment of global tumour perfusion would be
obtained, and that this would be less prone to measure-
ment error. However, this has not been proven to date
with colorectal cancer.
Reproducibility assessment encompasses quantifica-
tion of all the intrinsic and extrinsic factors that
contribute to measurement variability. These include
tumour heterogeneity, CT technique, software variabi-
lity, and observer variability [10]. While it is not possible
to separate these factors, it is possible to minimize effects
of certain factors to permit evaluation the effect of the
volume of data surveyed on reproducibility. For exam-
ple, it is possible to minimize the effects of observer
variability by using a single experienced observer to
obtain all measurements [11], and to minimize effects
from the technique and software evaluation by keeping
technical factors constant between studies. Thus, we
compared the reproducibility of measurements from a
5 mm z-axis tumour coverage with that from 4 contig-
uous 5 mm slices, the results from which had been
averaged to provide information equivalent to from a
20 mm z-axis tumour coverage to identify if any
improvement in reproducibility was achieved.
Overall, reproducibility was acceptable for all four
perfusion measurements obtained from both a 5 mm and
20 mm tumour coverage, and comparable with pre-
viously cited reproducibility levels using CT within the
intracranial and extracranial circulation in animal and
human studies [12–15]. For example, a variability of 13%
and 7%, assessed using analysis of variance, has been
quoted for repeated cerebral blood flow and blood
volume measurements, respectively, in rabbits [13],
while a variability of 14%, 20% and 18% has been quoted
for repeated blood flow, blood volume and permeability
measurements in the rabbit VX2 tumour [15]. Analysing
our data in the same manner using analysis of variance
Mean difference SD 95% limits of agreement
–0.04 1.26 –2.50 to +2.42
–0.04 1.31 –2.61 to +2.53
+8.8 30.2 –50.5 to +68.0
+7.4 29.4 –50.3 to +65.0
–0.99 3.67 –8.19 to +6.20
–2.46 5.18 –12.61 to +7.69
+1.20 3.38 –5.42 to +7.83
–0.23 4.12 –8.31 to +7.85
The British Journal of Radiology, July 2006
Colorectal perfusion measurement by multidetector-row CT

Figure 1. Blood volume: (a) scatter plot showing the distribution of values with a line of perfect agreement shown; and
(b,c) Bland-Altman agreement plots with the mean difference and 95% limits of agreement for a 5 mm (D) and 20 mm (O) z-axis
tumour coverage.

would have resulted in a variability of 23%, 14% and 17%
for colorectal blood flow, blood volume and perme-
ability, respectively. Reproducibilty of permeability
measurements is also comparable with that reported
for dynamic contrast enhanced MRI for a variety of
extracranial tumours, where a mean difference of 0.03,
95% confidence interval of 20.04 to +0.06, and coefficient
of variation of 29% was reported for log transformed
values of ktrans [16].
With reference to published data, the differences
between repeated blood flow measurements that we
observed may be sufficiently small to be overwhelmed
by the change in perfusion induced therapeutically by
antivascular targeting and antiangiogenic agents being
evaluated currently. For example, dynamic contrast
enhanced (DCE)-MRI data from a Phase I study of
combretastatin, a vascular targeting agent, have shown a
significant group mean measurement change of 37%
after drug administration [17]. Likewise, perfusion CT
data from a Phase I study of bevacizumab have shown
mean blood flow changes of the order of 40% [18], while
data from a study of the effects of chemoradiation have
shown a mean blood flow changes of 62% [19]. However,
for treatments that produce a smaller effect on tumour
vascularity, some caution must be applied to the in the
interpretation of results, as measurement change may
remain within the range of measurement variability. On
an individual patient basis, there may be a role for

The British Journal of Radiology, July 2006 581

 V Goh, S Halligan, L Gartner et al

Figure 2. Blood flow: (a) scatter plot showing the distribution of values with a line of perfect agreement shown; and
(b,c) Bland-Altman agreement plots with the mean difference and 95% limits of agreement for a 5 mm (D) and 20 mm (O) z-axis
tumour coverage.

assessing individual intrapatient measurement variabi-
lity, so that any response may be interpreted in the light
of this. An alternative would be to use surrogates that
show better reproducibility, for example semi-
quantitative measurements such as peak enhancement,
or standardized perfusion value. Further evaluation of
these strategies is required.
Reproducibility did not improve with increasing
tumour z-axis coverage. There may be several explana-
tions why data acquisition from a larger tumour volume
did not improve reproducibility. Most obviously, it may
be the case that colorectal spatial heterogeneity is
insufficient to influence data acquisition once the tumour
volume has achieved 5 mm in its z-axis dimension. This
is in contrast to preliminary data from lung tumours that
suggests improved reproducibility with increasing z-axis
acquisition [20]. However, the lung tumours studied
were large heterogeneous tumours with a necrotic
centre, unlike the colorectal tumours within this study,
and thus spatial heterogeneity may have been a major
factor contributing to measurement variability in these
tumours.
It is also possible that the variation in tumour
morphology over time encountered with colorectal
cancer may be partly responsible for the lack of
improvement observed. For example, bowel tumours

582 The British Journal of Radiology, July 2006

Colorectal perfusion measurement by multidetector-row CT
might change shape and position as a result of
peristalsis. Assessment of reproducibility from a single
5 mm-tumour level is straightforward, as it is usually
possible to ensure comparable tumour levels on both sets
of scans when choosing from the 4 slices available.
However, we found that ensuring the same tumour level
was assessed on all 4 slices was more difficult, because
there was no leeway in post-acquisition slice selection.
Tumour shape and position were unlikely to be identical
on both sets of scans, despite care in scan acquisition,
including administration of an antiperistaltic agent.
It is possible that the quantitative software analysis
method we used may have contributed to the lack of
improvement found with increased tumour coverage.
While the ROI used to define the arterial input may be
saved using the software for future use, to minimize
variability from ROI placement in subsequent analyses,
due to variation in tumour shape between studies, a ROI
has to be drawn freehand around the tumour on the
parametric perfusion maps, which are generated auto-
matically by the software to obtain the quantitative
parameters. Each colour pixel on these maps represents
an individual quantitative value of the parameter
measured (blood flow, blood volume, mean transit time
or permeability) and a mean parameter value is
generated for the ROI drawn. It is inevitable that some
variability will be introduced when drawing a ROI
(though this can be reduced by using the same
unblinded observer to create each ROI). Finally, our
patient numbers were small and inevitably reduced the
power of our experiment, but this was a pragmatic
consideration based on a natural reluctance to unneces-
sarily expose patients to ionizing radiation on two
separate occasions.
In summary, we found that increasing z-axis tumour
coverage did not improve reproducibility of measure-
ments of colorectal tumour perfusion. Measurement
reproducibility remains clinically acceptable with single
level measurements.
References
1. Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan
RS, Rubenstein L, et al. New guidelines to evaluate the
response to treatment in solid tumors. J Natl Cancer Inst
2000;92:205–16.
2. World Health Organisation. WHO hand book for reporting
results of cancer treatment. WHO offset publication
1979;no 48. Geneva, Switzerland.
3. Tozer GM. Measuring tumour vascular response to anti-
vascular and antiangiogenic drugs. Br J Radiol 2003;76:
S23–S35.
4. Hurwitz H, Fehrenbacher L, Novotny W, Cartwright T,
Hainsworth J, Heim W, et al. Bevacizumab plus irinotecan,
fluorouracil, and leucovorin for metastatic colorectal cancer.
N Engl J Med 2004;350:2335–42.
5. Endrich B, Vaupel P. The role of the microcirculation in the
treatment of malignant tumors: facts and fiction. In: Molls
The British Journal of Radiology, July 2006
M, Vaupel P, editors. Blood perfusion and microenviron-
ment of human tumors: implications for clinical radio-
oncology. Berlin, Germany: Springer-Verlag, 2001;19–39.
6. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986;1:307–10.
7. Bland JM, Altman DG. Applying the right statistics:
analyses of measurement studies. Ultrasound Obstet
Gynecol 2003;22:85–93.
8. Miles KA, Griffiths MR. Perfusion CT: a worthwhile
enhancement? Br J Radiol 2003;76:220–31.
9. Goh V, Halligan S, Balmer JA, Bartram CI. Colorectal cancer
perfusion measurement with multi-detector computed
tomography: measurement reproducibility and clinical
implications. In: Radiological Society of North America
90th Scientific Assembly and Annual Meeting Program;
2004. Chicago, USA: RSNA, 2004.
10. Bland M. An introduction to medical statistics, 3rd edn.
Oxford, UK: Oxford University Press, 2003.
11. Goh V, Halligan S, Hugill JA, Bassett P, Bartram CI.
Quantitative assessment of colorectal cancer perfusion
using MDCT: inter and intra-observer agreement. AJR
Am J Roentgenol 2005;185:225–31.
12. Nabavi DG, Cenic A, Dool J, Smith RM, Espinosa F, Craen
R, et al. Quantitative assessment of cerebral hemodynamics
using CT: stability, accuracy, and precision studies in dogs.
J Comput Assist Tomogr 1999;23:506–15.
13. Cenic A, Nabavi DG, Craen RA, Gelb AW, Lee TY. A CT
method to measure hemodynamics in brain tumors:
validation and application of cerebral flow maps. AJNR
Am J Neuroradiol 2000;21:462–70.
14. Gillard JH, Antoun NM, Burnet NG, Pickard JD.
Reproducibility of quantitative CT perfusion imaging. Br J
Radiol 2001;74:552–5.
15. Purdie TG, Henderson E, Lee TY. Functional CT imaging of
angiogenesis in rabbit VX2 soft tissue tumour. Phys Med
Biol 2001;46:3161–75.
16. Galbraith SM, Lodge MA, Taylor NJ, Rustin GJ, Bentzen S,
Stirling JJ, et al. Reproducibility of dynamic contrast
enhanced MRI in human muscle and tumours: comparison
of quantitative and semi-quantitative analysis. NMR
Biomed 2002;15:132–42.
17. Galbraith SM, Maxwell RJ, Lodge MA, Tozer GM, Wilson J,
Taylor NJ, et al. Combretastatin A4 Phosphate has tumor
antivascular activity in rat and man as demonstrated by
dynamic magnetic resonance imaging. J Clin Oncol
2003;21:2831–42.
18. Willett CG, Boucher Y, di Tomaso E, Duda DG, Munn LL,
Tong RT, et al. Direct evidence that the VEGF-specific
antibody bevacizumab has antivascular effects in human
rectal cancer. Nature Med 2004;10:145–7.
19. Sahani DV, Kalva SP, Hamberg LM, Hahn PF, Willett CG,
Saini S, et al. Assessing tumor perfusion and treatment
response in rectal cancer with multisection CT: initial
observations. Radiology 2005;234:785–92.
20. Ng QS, Goh V, Klotz E, Fichte H, Carnell D, Saunders M, et
al. Single versus four-level tumour perfusion measurement
using helical dynamic contrast enhanced MDCT: effect on
measurement agreement and potential for whole tumour
assessment. In: Proceedings for European Congress of
Radiology 2005; Vienna, Austria. Eur Radiol 2005;
15(Suppl. 1):414.
583
 The British Journal of Radiology, 79 (2006), 584–591

Prophylactic implantation of inferior vena cava filter during

interventional radiological treatment for deep venous
thrombosis of the lower extremity
T YAMAGAMI, MD, PhD, T KATO, MD, T HIROTA, MD, R YOSHIMATSU, MD, T MATSUMOTO, MD and
T NISHIMURA, MD, PhD

Department of Radiology, Graduate School of Medical Science, Kyoto Prefectural University of

Medicine, 465 Kajii-chyo, Kawaramachi-Hirokoji, Kamigyo, Kyoto, 602-8566, Japan

ABSTRACT. The purpose of this study was to evaluate the filtering effect of the Gunther
tulip retrievable vena cava filter (GTF) during treatment of deep venous thrombosis
(DVT) in the lower extremity using various interventional radiological procedures.
Subjects of the study were all 17 consecutive patients (8 women, 9 men; age range 18–
87 years; mean age 55.9 years) with symptomatic lower limb DVT referred for
interventional radiological treatment between February 2001 and September 2004. In
all of these patients, the GTF was implanted during interventional radiological
treatment. Trapped thrombus in the filter was evaluated with venocavography
performed repeatedly during the treatment for DVT. Implantation of a total of 29 GTFs
was successfully performed in the 17 patients. In 10 (58.8%), more than 2 filters were
subsequently implanted to prolong implantation time. Also in 10 patients, the DVT
resolved after therapy and retrieval of the final GTF was successful with one exception.
Worsening of or new formation of pulmonary embolism was avoided in all patients. In Received 30 August 2005
8 (47.1%) of the 17 patients, a trapped thrombus in the GTF was observed during Revised 13 November 2005
treatment for DVT. In six patients the trapped thrombus was large, filling more than Accepted 10 January 2006
half the height of the filter. In conclusion, we found that the GTF is effective in filtering
DOI: 10.1259/bjr/54853087
the relieved thrombus from DVT in the lower extremity and in protecting against
movement of the thrombus to the pulmonary artery during therapies with ’ 2006 The British Institute of
interventional radiological procedures. Radiology

The initial symptoms of deep venous thrombosis (DVT)
in the lower extremities may be quite severe [1, 2] and,
additionally, there is substantial risk of pulmonary
embolism [1, 3]. Moreover, in DVTs that remain, delayed
complications are possible, including the spectrum of
debilitating symptoms referred to as post-thrombotic
syndrome [4]. Therefore, treatment of DVT is necessary.
Traditionally, therapy for DVT has consisted of systemic
intravenous administration of heparin [5]. Other options
have been systemic intravenous administration of fibrino-
lytic agents such as urokinase or streptokinase [5–7].
Recombinant tissue plasminogen activator is gaining
attention as an effective fibrinolytic agent [8, 9]. Systemic
thrombolysis is more effective than heparinization, but less
effective than catheter-directed thrombolysis, which is an
interventional radiological treatment that has recently
become widely accepted as useful in treating DVT [4, 9–
11]. Most recently, other interventional radiological pro-
cedures such as manual aspiration of thrombus [12],
mechanical thrombectomy [4, 13, 14], percutaneous
transluminal angioplasty [13, 15] and self-expandable
metallic stent placement [4, 13, 15, 16] have sometimes
been combined with catheter-directed thrombolysis to
improve the quality of treatment for DVT.
Theoretically, a thrombus released from the DVT
during catheter-directed thrombolysis could move to
the pulmonary artery, which may cause pulmonary
embolism. Thus, some physicians implant inferior vena
cava filters prophylactically to prevent such an occur-
rence [17–20]. However, the prophylactic use of an
inferior vena cava filter is considered by some to be
unnecessary [1, 10, 11], with the result that this
procedure is currently very controversial. Nevertheless,
in our daily clinical work, we have experienced many
situations where, if a filter had not been implanted
during catheter-directed thrombolysis, severe pulmon-
ary embolism might have occurred. Many interventional
radiological procedures for DVT other than catheter-
directed thrombolysis have been developed [12–16], but
there is little information on the prophylactic use of
filters in these procedures [21].
When such prophylactic use of filters is required,
ideally, a permanent inferior vena cava filter would not
be placed, considering the long life expectancy of such
patients [22, 23]. Hence, temporary vena cava filters have
been widely used [17, 24]. However, paralleling the
increase in the use of temporary vena cava filters have
been reports describing complications related to their
insertion, mainly associated with their structure, where
part of the device projects from the insertion site [24–26].
Because of such problems, it would be advantageous to
use a retrievable vena cava filter, that is, a permanent
filter that can be retrieved when necessary. The Gunther
tulip retrievable vena cava filter (GTF) (Cook,

584 The British Journal of Radiology, July 2006

Inferior vena cava filter for DVT
Bjaeverskov, Denmark) is among the commercially
available retrievable vena cava filters.
The present study evaluated whether an inferior vena
cava filter could have a role in filtering a thrombus
released from the lower limb during treatment of lower
extremity DVT using various interventional radiological
procedures. Also examined were the feasibility and
safety of a GTF used for that purpose.
Subjects and methods
Patients
The cohort of this study was 17 consecutive patients (8
women, 9 men; age range 18–87 years; mean age
55.9 years) who were treated for symptomatic DVT of
the lower extremity, utilizing various interventional
radiological procedures between February 2001 and
September 2004, at our institution. In all of these patients,
traditional intravenously administered systemic antic-
oagulation and/or thrombolysis had been performed
first as a treatment for DVT. However, despite such
treatment, DVT developed. A GTF was implanted to
prevent the complication of pulmonary embolism during
the entire period of interventional radiological treatment
of DVT (Figure 1). Interval from initial symptoms to start
of therapy for DVT was within 1 week in 11 patients,
between 1 week and 4 weeks in 5 patients and after
6 weeks in 1 patient.
In all cases, the existence of lower extremity DVT was
diagnosed by ultrasound. When extension of the
thrombus from the lower extremity to the inferior vena
cava was suspected, enhanced abdominal CT was
performed. Intravenous ascending venography with
contrast injection into the vein of the foot was also
performed as an interventional radiological diagnostic
procedure.
Before starting therapies for DVT, enhanced chest CT
was performed in all 17 patients to evaluate the existence
of pulmonary embolism. Pulmonary embolism was
found in two patients, who received intravenously
administered anticoagulation and thrombolytic therapies
for pulmonary embolism, in addition to therapies for
DVT. Enhanced chest CT was also obtained after
therapies for DVT in all 17 patients.
Procedures
All interventional radiological procedures performed
as therapy for DVT and GTF placement and retrieval
were performed by one of three experienced interven-
tional radiologists in our institution after written
informed consent was obtained from each patient. The
consent included use of records, images, data, etc. for
research purposes. Our institution does not require
institutional review board approval for this type of
report. Principles of the Declaration of Helsinki were
followed.
The following are interventional radiological therapies
for DVT that were performed in our institution and that
were reviewed retrospectively in all 17 patients who
received such therapy.
The British Journal of Radiology, July 2006
In the angiography suite, catheter-directed thrombo-
lysis was performed after a 0.035-inch guide-wire was
advanced via a 5-French catheter over the thrombosed
segment and this catheter was exchanged with a
commercially available multiside-hole catheter (Cragg-
McNamara Valved Infusion Catheter; Micro
Therapeutics, Irvine, CA). Venous access was through
the femoral, jugular, or popliteal vein. After positioning
the tip of the multiside-hole catheter in the thrombosed
segment, a mixture of 120 000–240 000 IU urokinase and
1000–2000 IU heparin diluted by 20–40 ml of physiolo-
gical saline was manually administered over a period of
approximately 10 min. In the majority of patients (n516),
because the thrombus remained, 120 000–240 000 IU
urokinase per day was further administered persistently
via the catheter which remained positioned in the
thrombosed segment over a prolonged period, even
after the patient returned to the ward from the
angiography suite. Heparin was administered systemi-
cally during the lytic procedure and after lysis to keep
the partial thromboplastin time at 75–100 s.
When DVT was shown to be completely or almost
completely resolved on venography via the catheter
inserted for catheter-directed thrombolysis, lysis was
stopped. But if much of the thrombus remained as
shown on this venography, manual aspiration of the
venous thrombus and/or mechanical thrombectomy
were often combined concurrently with or after cathe-
ter-directed thrombolysis. For manual aspiration, after a
10-French or 11-French with 11 cm long sheath intro-
ducer was inserted from the femoral vein at the site of
the thrombosis, an 8-French with 25 cm long sheath was
advanced through the wider sheath into the segment
occupied by the thrombus. The 8-French sheath was
fitted with a syringe by which the clot was manually
aspirated. After the aspiration, the 8-French sheath was
withdrawn from the 10-French or 11-French sheath. If
clots of thrombus were observed, the 8-French sheath
was reinserted and the thrombus was aspirated. These
steps were repeated until the venous thrombus was no
longer observed by aspiration. Alternatively, in the more
recent cases, after a 6-French sheath introducer was
inserted from the femoral vein at the site of DVT, the
venous thrombus was aspirated utilizing a 6-French
catheter (Thrombuster; Kaneka, Tokyo, Japan). This
catheter is commercially available and was developed
exclusively for aspiration of a venous thrombus.
Mechanical thrombectomy was performed using a
commercially available thrombectomy device
(Hydrolyser; Cordis or Oasis; Boston Scientific) with
access from the femoral vein.
When an underlying stenosis or obstruction due to a
non-fresh thrombus remained, percutaneous transluminal
angioplasty for the segment that was narrowed due to
thrombus was occasionally performed. If necessary, a self-
expandable metallic stent was placed to obtain sufficient
relief of the lesion, particularly at the level of the iliac vein.
With the femoral or jugular venous approach, percuta-
neous transluminal angioplasty was performed with a
commercially available angioplasty balloon catheter after
advancing a 0.035-inch guidewire over the narrowed
lesion. The angioplasty balloon catheter was positioned
in the narrow segment, after which percutaneous trans-
luminal angioplasty was performed for the entire narrow
585
 T Yamagami, T Kato, T Hirota et al

(a) (b)

(c) (d)

586 The British Journal of Radiology, July 2006

Inferior vena cava filter for DVT

Figure 1. A 19-year-old man (patient no. 6). (a) Venography shows venous thrombus distributing from the left common iliac
vein to the left femoral vein (arrows). Note that Gunther tulip retrievable vena cava filter (GTF) was implanted at the inferior
vena cava with an approach from the right jugular vein. (b) Venography obtained immediately after catheter-directed
thrombolysis still shows that much of the thrombus remained. Note that multiple side holes of the catheter for catheter-directed
thrombolysis were positioned in the segment of the thrombus (arrows). (c) Venocavography obtained from the right femoral
vein shows that thrombus filling greater than the height of filter was captured in the filter (arrows). This image was obtained
just after manual aspiration, mechanical thrombectomy and percutaneous transluminal angioplasty were performed. These
interventional radiological procedures were performed after catheter-directed thrombolysis, as shown in Figure 1a,b, and
prolonged thrombolysis performed subsequently in the ward through the catheter with its tip positioned at the thrombosed
segment. (d) Roentgenogram shows a temporary filter (arrow) (Neuhaus Protect, Toray) that was inserted at the cephalad level
of GTF before GTF with thrombus was retrieved to be exchanged with a new GTF. (Continued)

During the entire period of therapy by the above-

(e)
Figure 1. (Cont.) (e) Venography through the right femoral
vein, obtained after GTF was exchanged for a new one
(arrow), shows no thrombus remaining in the inferior vena
cava. The new GTF was implanted to continue treatment for
the DVT, further avoiding the complication of pulmonary
embolism. 7 days after this, metallic stent placement was
performed, then venous thrombus resolved and rapid blood
flow in the previously obstructed segment due to thrombus
in the left common and external iliac and femoral vein was
obtained.
described interventional radiological procedures, the GTF
remained in place. When the period of implantation
greatly exceeded 10–14 days, the period cited in many
reports within which a filter could be withdrawn safely
[27–30], the first implanted GTF was retrieved and another
inserted near at a different site from the former GTF.
When the venous thrombus in the lower extremities
resolved after therapy, as shown on intravenous veno-
graphy of the lower extremity, the GTF was retrieved
and reinsertion was not required. However, when it
remained in spite of various therapies, the last GTF
inserted during therapy was left in the vena cava without
retrieval to play a role thereafter as a permanent inferior
vena cava filter.
After a venacavogram was performed to plan the
position of filter placement, the GTF was introduced
through the right internal jugular vein and was
positioned at the cephalad side of the top of the venous
thrombus. In cases when insertion from the right internal
jugular vein was difficult, the right subclavian vein was
used to introduce the GTF. In principle, the GTF was
positioned at the infrarenal inferior vena cava. The GTF
was placed through the sheath introducer according to
the manufacturer’s instructions. Retrieval was per-
formed using the GTF retrieval set supplied by the
manufacturer (Cook). Details of techniques to place and
retrieve the GTF are described elsewhere [29, 31].
Venocavography was performed repeatedly via the
inserted catheter positioned in the thrombosed segment
to perform catheter-directed thrombolysis continuously
over a prolonged period in the ward. After retrieving the
catheter for thrombolysis, venocavography was obtained
through an 18-gauge intravenous catheter (Surflo;
Terumo, Tokyo, Japan) or a 4-French catheter, which
was inserted from the femoral vein in which the
thrombus did not exist. In cases when a GTF was
exchanged or retrieved, just prior to retrieval, venocavo-
graphy was performed through the sheath. Trapped
thrombus in the filter was evaluated from all these
venocavograms obtained by various approaches. Also,
immediately after retrieving the GTF, venography was
performed to confirm that no thrombus remained where
the retrieved GTF had been implanted.

segment. Either the Ultra-thin diamond balloon catheter

(Boston Scientific, Watertown, MA) or the Power flex plus
balloon catheter (Cordis, Miami, FL) was used. For self-
expandable metallic stent placement, an Easywall stent
(Boston Scientific) or spiral Gianturco Z stent (Cook) was
implanted with the femoral venous approach. Placement
over the area of joint movement (i.e. hip joint) was avoided.
Investigated parameters
We retrospectively investigated the following: selected
interventional radiological treatments for DVT, state of
DVT after such therapies, number of GTFs implanted
and retrieved, rate of success in preventing through GTF

The British Journal of Radiology, July 2006 587

 T Yamagami, T Kato, T Hirota et al

placement the worsening or development of pulmonary
embolism and degree of trapped thrombus in the filter.
Results
As shown in Tables 1 and 2, implantation in the planned
position of a total of 29 GTFs was successful in 17 patients
with DVT in the lower extremity for which various
therapies utilizing interventional radiological procedures
were performed. Details of each GTF implanted are shown
in Table 3. In four patients, the GTF was placed at the
suprarenal inferior vena cava for the following reasons. In
two of those four patients, a venous thrombus existed in
the inferior vena cava just below the level of entry of the
renal vein, leaving insufficient space for implantation in
the infrarenal inferior vena cava. The third patient had a
double inferior vena cava with the thrombus in both. In the
fourth patient, placement of an additional GTF at the
cephalad level of the normally placed GTF was necessary.
Retrieval of the GTF was successful in all 22 attempts,
with one exception (Table 3). The implantation period per
filter before retrieval was 10.4¡5.0 days (mean¡SD; range
0–18 days; median 13 days). The period of treatment with
GTFs in place was 19.1¡8.2 days (mean¡SD; range 5–37
days; median 20 days). In 10 (58.8%) of the 17 patients,
more than 2 filters were subsequently implanted to
prolong implantation time. No complications were
encountered relating to filter insertion and retrieval. In
both patients having a pulmonary embolism at the
beginning of interventional radiological treatments for
DVT as shown in Table 1, the pulmonary embolism had
resolved at the end of therapy for DVT. Pulmonary
embolism did not occur in any of the 15 patients without
pre-existing pulmonary embolism.
On venocavography performed repeatedly during the
entire period of therapies for DVT with interventional
radiological techniques, in 8 (47.1%) of the 17 patients, a
trapped clot of thrombus was observed. Trapped throm-
bus filled , 1/4 the height of the filter in 1, from >1/4 to ,
1/2 in 1, and >1/2 but within the height of the filter in 4. In
2, the trapped thrombus occupied greater than the height
of the filter. Among those eight patients, in only one
patient was catheter-directed thrombolysis the only inter-
ventional radiological procedure performed. In the
remaining seven patients, other interventional radiological
procedures were added. In other words, a trapped
thrombus in the filter was confirmed in one (25.0%) of
four patients who received only catheter-directed throm-
bolysis. However, a trapped thrombus was observed in
seven (53.8%) of 13 patients in whom other interventional
radiological treatments were combined with catheter-
directed thrombolysis. The difference was not significant
(p50.34, Fisher’s exact probability test).
All filters containing a trapped thrombus as revealed on
venocavogram were retrieved without any complications,
including pulmonary embolism. Attempts were made to
decrease all trapped thrombi in the filter by catheter-
directed thrombolysis, whereby the catheter tip that had
been in the DVT lesion was repositioned to the filter. In
two cases, both of which had a trapped thrombus that
occupied greater than the height of the filter, manual
aspiration of the thrombus in the filter was additionally
performed using the sheath introducer inserted from the

Table 1. Characteristics of patients with deep venous thrombosis of the lower extremity
Patient no./sex/ Site of DVT Location of Interventional radiological procedures Existence of PE Final status of DVT
age (years) venous thrombus performed as therapy for DVT

1/m/87 bilateral IVC, CI, EI, F CDT, p-CDT, A, PTA, EMS absent resolved
decreased but still
2/f/62 left CI, EI, F, P CDT, p-CDT A, MT, PTA, EMS absent present
3/f/18 right F CDT, p-CDT, A, PTA absent resolved
decreased but still
4/f/83 right IVC, CI, EI, F CDT, A absent present
decreased but still
5/f/52 bilateral F, P CDT, p-CDT present present
6/m/19 left CI, EI, F CDT, p-CDT, A, MT, PTA, EMS absent resolved
7/f/41 bilateral EI, F CDT, p-CDT absent resolved
8/m/54 right F CDT, p-CDT absent resolved
decreased but still
9/m/76 right IVCa, CI, EI, F CDT, p-CDT absent present
10/m/21 left CI CDT, p-CDT, A absent resolved
decreased but still
11/f/80 bilateral IVC, CI, EI, F CDT, p-CDT, A, PTA, EMS absent present
12/m/63 left CI, EI, F CDT, p-CDT, A present resolved
decreased but still
13/f/75 left IVC, CI, EI CDT, p-CDT, A absent present
14/m/27 left CI, EI, F, P CDT, p-CDT, A, PTA absent resolved
decreased but still
15/m/77 right CI, EI CDT, p-CDT, A, PTA absent present
16/f/61 left EI, F CDT, p-CDT, A, PTA absent resolved
17/m/55 left EI, F CDT, p-CDT, A absent resolved
IVC, inferior vena cava; CI, common iliac vein; EI, external iliac vein; F, femoral vein; P, popliteal vein; CDT, catheter-directed
thrombolysis; p-CDT, prolonged thrombolysis performed from the catheter directly inserted in the thrombosed segment;
A, aspiration of venous thrombus; MT, mechanical thrombectomy; PTA, percutaneous transluminal angioplasty; EMS,
placement of self-expandable metallic stent; DVT, deep venous thrombus; PE, pulmonary embolism.
a
Case with double inferior vena cava.

588 The British Journal of Radiology, July 2006

Inferior vena cava filter for DVT

Table 2. Usage of Gunther retrievable vena cava filter

Patient no. No. filters used Final use of GTF as Duration of therapy under Trapped thrombus
a permanent protection from PE with GTF in GTF as revealed
vena cava filter prior to exclusive use of
conventional therapy (days)

1 2 No 8 Grade 1
2 2 Yes 24 Grade 3
3 2 No 23 Grade 3
4 1a Yes 7 Grade 0
5 1 Yes 5 Grade 0
6 2 Yesc 21 Grade 4
7 2 No 28 Grade 0
8 3b No 21 Grade 4
9 1 Yes 15 Grade 0
10 3 No 37 Grade 0
11 1 Yes 19 Grade 0
12 1 No 18 Grade 2
13 2 Yes 14 Grade 3
14 2 No 28 Grade 0
15 1 Yes 23 Grade 0
16 2 No 20 Grade 3
17 1 No 14 Grade 0
a
Cases requiring correction of filter placement.
b
One filter being placed during single interventional radiological procedure.
c
Case of failure to retrieve filter.
PE, pulmonary embolism; GTF, Gunther tulip retrievable vena cava filter.
Grade 0: No trapped thrombus is seen.
Grade 1: Trapped thrombus filled >1/4 the height of the filter.
Grade 2: Trapped thrombus filled from >1/4 to >1/2 the height of the filter.
Grade 3: Trapped thrombus filled >1/2 but within the height of the filter.
Grade 4: Trapped thrombus filled greater than the height of the filter.

Table 3. Characteristics of each Gunther retrievable vena cava filter

GTF implanted Number

Approach for implantation of GTF

right jugular vein 28
right subclavian vein 1
Level of IVC where GTF was placed
infrarenal 25
suprarenal 4
Final status of each filter
Left in IVC to use as a permanent filter thereafter 7
Failed attempt at retrieval after resolution of DVTa 1
Successfully retrieved 21
exchanged with another GTF to prolong the implantation period 11
temporarily placed during an IR procedure and retrieved immediately after that procedure 1
retrieved and further implantation was not performed as the DVT had resolved 9
GTF, Gunther tulip retrievable vena cava filter; IVC, inferior vena cava; DVT, deep venous thrombus; IR, interventional radiology.
a
Hook part of the filter had become attached to the wall of the inferior vena cava resulting in failure of retrieval.

femoral vein into the inside of the filter. As a result, two
GTFs were retrieved after the thrombus resolved following
attempts to decrease the size of the captured thrombus by
means such as catheter-directed thrombolysis. Regarding
the remaining six cases, the size of the thrombus in the
filter either did not change or had increased, so the filter
was replaced with a new one. In exchanging the filter, a
temporary filter (Antheor; Boston Scientific, Watertown,
MA, n54, or Neuhaus Protect, Toray Medical, Tokyo,
Japan, n51) or another GTF (n51) was temporarily placed
at the cephalad level of the GTF, trapping the thrombus
exclusively during the GTF retrieval procedure. Thus, the
thrombus dropped from the first GTF was captured
during GTF retrieval. In the case of another GTF used
temporarily, displacement of that freshly placed GTF
avoided by taking meticulous care in using the retrieval kit
during the procedure to remove the originally implanted
GTF. Venocavography after retrieval revealed no throm-
bus remaining where the first GTF had been placed in any
of the eight patients. Figure 1 shows one such case.
Discussion
There has been much controversy regarding the
prophylactic use of inferior vena cava filters to prevent
the occurrence of pulmonary embolism during catheter-
directed thrombolysis of lower extremity DVT [1, 9–11,

The British Journal of Radiology, July 2006 589


17–21]. Some researchers have advocated that such filter
placement is necessary during catheter-directed throm-
bolysis only in patients who have large, mobile, free-
floating thrombi within the inferior vena [1]. On the
other hand, there has been little discussion on the
necessity of filter placement during therapies for DVT
with interventional radiological procedures other than
catheter-directed thrombolysis [21].
Pulmonary embolism occurred in only two (0.9%) of
214 patients who received catheter-directed thrombolysis
for DVT in the lower extremity without inferior vena
cava filter placement, according to Bjarnason et al [1],
and in only 6 (1.3%) of 473 patients, according to
Mewissen et al [11]. However, one of the six cases was
a fatality. Some investigators [1, 9–11] insist that
prophylactic use of an inferior vena cava filter during
catheter-directed thrombolysis is unnecessary because of
such a low rate of pulmonary embolism.
In a summary of their experience with implantation of
temporary vena cava filters in 132 patients with DVT
who were receiving thrombolytic therapy, Thery et al
[32] observed a clot lysed by thrombolytic therapy in the
filter in 41 (31%) of the 132 patients. None of the 132
patients experienced pulmonary embolism. This sug-
gests that the filter might prevent pulmonary embolism
during thrombolytic treatment in at least 41 of these
patients, if not in all 132 patients.
Some researchers [17–21, 30, 33, 34] propose that the
prophylactic use of an inferior vena cava filter is
necessary in view of the high rate of trapped thrombus
in these filters during therapies for DVT, as revealed by
Thery et al [32], and attach great importance to the
potential mortality in such cases, although rare [11].
Based on these opinions, we perform therapy for DVT
under the prophylactic protection against pulmonary
embolism by using the GTF.
Ease and safety of insertion of GTF is well known [21,
27–31, 33, 35, 36], with some of these reports citing its use
as a non-permanent filter [21, 27, 29, 30, 33–36]. Millward
et al [33] reported that attempts at retrieval were
successful in 98% of 53 GTFs. According to some
previous reports [27–30], the maximal period of implan-
tation before successful retrieval was recommended as
no longer than 10–14 days. Although we had roughly
followed this recommended period to implant the GTF
in the patients presented here, successful retrieval of a
GTF after implantation longer than 14 days has been
reported, as described in some recently published
studies. Millward et al [33] noted that the maximal
period of implantation before successful retrieval was
25 days, and very recently Terhaar et al [37] reported a
period of 126 days. Also, we would like to mention that
other retrievable filters are now available that can be left
in place for much longer [38]. Few complications related
to GTF placement have been reported [33, 39]. Millward
et al [33] documented filter occlusion in two (5%) of 90
patients with an implanted GTF; no serious complica-
tions were documented. However, another case report
[39] described right atrial migration.
Our study also showed good results in insertion and
retrieval of GTFs. Insertion of all 29 GTFs and retrieval of
95.5% (21) of 22 GTFs were successfully accomplished.
The maximal period of implantation before success-
ful retrieval was 18 days. No complications were
590
T Yamagami, T Kato, T Hirota et al
encountered in filter placement and retrieval. In seven
patients in whom the lower extremity DVT remained in
spite of various treatments and in who the possibility of
pulmonary embolism remained, the GTF was used as a
permanent filter by simply leaving it in the inferior vena
cava. On the other hand, the GTF was retrieved in 9 of
the 10 patients as it was no longer needed after
successful resolution of the deep venous thrombus.
In the present study, in which a retrievable inferior
vena cava filter was placed in all 17 patients with DVT
who underwent various interventional radiological
treatments such as catheter-directed thrombolysis, no
worsening or development of pulmonary embolism in
any of the 17 patients was seen. However, in eight
(47.1%) patients a trapped thrombus in the filter was
confirmed during interventional radiological treatments.
This was despite adequate anticoagulation therapy. In
these eight patients, especially in the six in whom the
thrombus captured by the GTF filled greater than half
the height of the filter, it is possible that if the filters had
not been implanted, the thrombus might have moved to
the pulmonary artery causing pulmonary embolism.
Furthermore, we would like to note that the rate of
occurrence of trapped thrombus in the filter was higher
in cases that were treated not only with catheter-directed
thrombolysis, but additionally with other interventional
radiological procedures than in those solely treated by
catheter-directed thrombolysis (53.8% versus 25.0%). The
difference, however, was not statistically significant.
In addition, our results showed that the DVT did not
extend to the inferior vena cava in six (75.0%) of our
eight cases with a captured thrombus in the filter. This
might suggest that limiting the indication for a filter to
large, mobile, free-floating thrombi within the inferior
vena cava is insufficient to avoid a pulmonary embolism
in such situations.
We want to reiterate that, from the standpoint of
preventing pulmonary embolism during therapies for
DVT in the lower extremity with various interventional
radiological procedures, the inferior vena cava filter
played an important role in filtering a thrombus released
from the lower limb. However, the use of permanent or
temporary filters had some drawbacks, as previous
studies have suggested [22–26]. Hence, when interven-
tional radiological procedures are performed as therapy
for DVT, prophylactic placement of a retrievable filter,
such as GTF, should be employed, especially when
another interventional radiological procedure is added to
the catheter-directed thrombolysis, which is the proce-
dure most commonly performed.
References
1. Bjarnason H, Kruse JR, Asinger DA, Nazarian GK, Dietz CA
Jr, Caldwell MD, et al. Iliofemoral deep venous thrombosis:
safety and efficacy outcome during 5 years of catheter-
directed thrombolytic therapy. J Vasc Interv Radiol
1997;8:405–18.
2. Strandness DE Jr, Langlois Y, Cramer M, Randlett A, Thiele
BL. Long-term sequelae of acute venous thrombosis. JAMA
1983;250:1289–92.
3. Hirsch J, Hoak J. Management of deep vein thrombosis and
pulmonary embolism. Circulation 1996;93:2212–45.
4. Sharafuddin MJ, Sun S, Hoballah JJ, Youness FM, Sharp WJ,
Roh BS. Endovascular management of venous thrombotic
The British Journal of Radiology, July 2006
Inferior vena cava filter for DVT
and occlusive diseases of the lower extremities. J Vasc
Interv Radiol 2003;14:405–23.
5. Elliot MS, Immelman EJ, Jeffery P, Benatar SR, Funston MR,
Smith JA, et al. A comparative randomized trial of heparin
versus streptokinase in the treatment of acute proximal
venous thrombosis: an interim report of a prospective trial.
Br J Surg 1979;66:838–43.
6. Arnesen H, Hoiseth A, Ly B. Streptokinase or heparin in the
treatment of deep vein thrombosis. Acta Med Scand
1982;211:65–8.
7. Robertson BR, Nilsson JM, Nylander G. Value of strepto-
kinase and heparin in treatment of acute deep venous
thrombosis. Acta Chir Scand 1968;134:203–8.
8. Sugimoto K, Hofmann LV, Razavi MK, Kee ST, Sze DY,
Dake MD, et al. The safety, efficacy, and pharmacoeco-
nomics of low-dose alteplase compared with urokinase for
catheter-directed thrombolysis of arterial and venous
occlusions. J Vasc Surg 2003;37:512–7.
9. Grossman C, McPherson S. Safety and efficacy of catheter-
directed thrombolysis for iliofemoral venous thrombosis.
AJR Am J Roentgenol 1998;172:667–72.
10. Semba CP, Dake MD. Iliofemoral deep venous thrombosis:
aggressive therapy with catheter-directed thrombolysis.
Radiology 1994;191:487–94.
11. Mewissen MW, Seabrook GR, Meissner MH, Cynamon J,
Labropoulos N, Haughton SH. Catheter-directed thrombo-
lysis for lower extremity deep venous thrombosis: report of
a national multicenter registry. Radiology 1999;211:39–49.
12. Roy S, Laerum F. Transcatheter aspiration: the key to
successful percutaneous treatment of deep venous throm-
bosis? Acad Radiol 1999;6:730–5.
13. Bravo SM, Reinhart RD, Meyerovitz MF. Percutaneous
venous interventions. Vascular Med 1998;3:61–6.
14. Delomez M, Beregi JP, Willoteaux S, Bauchart JJ, d’Othee
BJ, Asseman P, et al. Mechanical thrombectomy in patients
with deep venous thrombosis. Cardiovasc Interv Radiol
2001;24:42–8.
15. AbuRahma AF, Perkins SE, Wulu JT, Ng HK. Iliofemoral
deep vein thrombosis: conventional therapy versus lysis
and percutaneous transluminal angioplasty and stenting.
Ann Surg 2001;233:752–60.
16. Nazarian GK, Bjarnason H, Dietz CA Jr, Bernadas CA,
Hunter DW. Iliofemoral venous stenoses: effectiveness of
treatment with metallic endovascular stents. Radiology
1996;200:193–9.
17. Neuerburg J, Günther RW. Developments in inferior vena
cava filters: a European viewpoint. Semin Intervent Radiol
1994;11:349–57.
18. Millward S. Temporary and retrievable inferior vena cava
filters: current status. J Vasc Interv Radiol 1998;9:381–7.
19. Emanuelli G, Segramora V, Frigerio C. Selected strategies in
venous thromboembolism: local thrombolytic treatment
and caval filters. Haematologica 1995;80:84–6.
20. Palombo D, Porta C, Brustia P, Peinetti F, Udini M, Antico
A, et al. Loco-regional thrombolysis in deep venous
thrombosis. Phlebologie 1993;46:293–302.
21. Yamagami T, Kato T, Iida S, Tanaka O, Nishimura T.
Retrievable vena cava filter placement during treatment for
deep venous thrombosis. Br J Radiol 2003;76:712–8.
22. Ferris E, McCowan TC, Carver DK, McFarland DR.
Percutaneous inferior vena cava filters: follow-up of seven
designs in 320 patients. Radiology 1993;188:851–6.
The British Journal of Radiology, July 2006
23. Decousus H, Leizorovics A, Parent F, et al. A clinical trial of
vena caval filters in the prevention of pulmonary embolism
in patients with proximal deep-vein thrombosis. N Engl J
Med 1998;338:409–15.
24. Lorch H, Welger D, Wagner V, Hillner B, Strecker EP,
Herrmann H, et al. Current practice of temporary vena cava
filter insertion: a multicenter registry. J Vasc Interv Radiol
2000;11:83–8.
25. Carcone B, Pernes JM, Carcopino JM, Chollet D, Pouillard F.
Worsening of proximal thrombosis after insertion of a
temporary caval filter (3 cases). Rev Med Interne 1995;16:
351–3.
26. Stosslein F, Altmann E. A rare complication with an
Antheor vena cava filter. Cardiovasc Intervent Radiol
1998;21:165–7.
27. Tay KH, Martin ML, Fry PD, Webb JG, Machan LS.
Repeated Gunther tulip inferior vena cava filter reposition-
ing to prolong implantation time. J Vasc Interv Radiol
2002;13:509–12.
28. de Gregorio MA, Gamboa P, Gimeno MJ, Madariaga B,
Tobio R, Herrera M, et al. The Gunther Tulip retrievable
filter: prolonged temporary filtration by repositioning
within the inferior vena cava. J Vasc Interv Radiol
2003;14:1259–65.
29. Neuerburg JM, Gunther RW, Vorwerk D, Dondelinger RF,
Jager H, Lackner KJ, et al. Results of a multicenter study of
the retrievable tulip vena cava filter: early clinical experi-
ence. Cardiovasc Interv Radiol 1997;20:10–6.
30. Millward SF, Bhargava A, Aquino J Jr, Peterson RA, Veinot
JP, Bormanis J, et al. Gunther tulip filter: preliminary
clinical experience with retrieval. J Vasc Interv Radiol
2000;11:75–82.
31. Millward SF. Gunther tulip retrievable filter: why, when
and how? Can Assoc Radiol 2001;52:188–92.
32. Thery C, Bauchart JJ, Lesenne M, Asseman P, Flajollet JG,
Legghe R, et al. Predictive factors of effectiveness of
streptokinase in deep venous thrombosis. Am J Cardiol
1992;69:117–22.
33. Millward SF, Oliva VL, Bell SD, Valenti DA, Rasuli P, Asch
M, et al. Gunther tulip retrievable vena cava filter: results
from the registry of the Canadian interventional radiology
association. J Vasc Interv Radiol 2001;12:1053–8.
34. Yamagami T, Kato T, Iida S, Hirota T, Nishimura T.
Gunther tulip inferior vena cava filter placement during
treatment for deep venous thrombosis of the lower
extremity. Cardiovasc Interv Radiol 2005;28:442–53.
35. Owen RJT, Krarup KC. The successful use and removal of
the Gunther tulip inferior vena caval filter in pregnancy.
Clin Radiol 1997;52:241–3.
36. Lin M, Soo TB, Horn LC. Successful retrieval of infected
Gunther tulip IVC filter. J Vasc Interv Radiol 2000;11:
1341–3.
37. Terhaar OA, Lyon SM, Given MF, Foster AE, McGrath F,
Lee MJ. Extended interval for retrieval of Gunther tulip
filters. J Vasc Interv Radiol 2004;15:1257–62.
38. Grande WJ, Trerotola SO, Reilly PM, Clark TWI, Soulen
MC, Patel A, et al. Experience with the recovery filter as a
retrievable inferior vena cava filter. J Vasc Interv Radiol
2005;16:1189–93.
39. Bochenek KM, Aruny JE, Tal MG. Right atrial migration
and percutaneous retrieval of Günther tulip inferior vena
cava filter. J Vasc Interv Radiol 2003;14:1207–9.
591
 The British Journal of Radiology, 79 (2006), 592–596

A survey of MRI quality assurance programmes

C J KOLLER, MSc, J P EATOUGH, PhD, P J MOUNTFORD, PhD and G FRAIN, MMath

Medical Physics Directorate, University Hospital of North Staffordshire, Princes Road, Hartshill,
Stoke on Trent ST4 7LN, UK

ABSTRACT. There are currently no national guidelines on appropriate quality assurance

(QA) test frequencies for MRI equipment in clinical use. From a random selection of 45
hospitals in England, who were contacted by phone, 35 hospitals agreed to participate
in a survey of MRI QA and were sent a questionnaire requesting information on the
range and frequency of QA tests, as well as the staff groups who conduct these tests.
Twenty-four completed replies were received, representing a 68% response rate from
the distributed questionnaires. Of these, 79% undertook some form of QA, typically
conducted by the radiographic staff. Tests were most often undertaken on the head
coil, but there was a considerable variation in the frequency and range of tests Received 13 July 2005
Revised 7 October 2005
undertaken at different hospitals. For example, exactly half of the respondents Accepted 24 October 2005
conducted signal to noise ratio (SNR) tests on both head and body coils, but only 13%
of centres extended this test to other coils. Results from this survey should inform DOI: 10.1259/bjr/67655734
radiology departments regarding practice at other hospitals and should assist in
’ 2006 The British Institute of
formulating the frequency and scope of appropriate MRI QA programmes.
Radiology

The use of MRI has more than doubled over the last
decade and, on average, 3000 scans per day were
undertaken in England in 2004 [1]. There will be a
variation in the quality of images produced in different
centres, due to differences in equipment and scan
parameters. Within a given centre, it is important to
ensure that all medical images produced are consistent
and of sufficient quality to answer the diagnostic
question. A change in sensitivity of the system or an
increase in the noise present in an image may cause fine
detail to be lost. Image artefacts may affect the region of
interest. These effects may be subtle, leaving the observer
unaware that they are happening, yet can result in loss of
confidence in the diagnosis or even misdiagnosis. It may
be argued that the skilled observer would detect any
decrease in quality of the radiological image. However, a
case was reported in the national press several years ago
where more than 1000 patients had to be recalled as a
result of undetected image quality problems with an
MRI scanner (‘‘Hospital’s brain scanner fails’’ Daily
Telegraph, 6th June 2002).
Quality assurance (QA) is a process to ensure that any
product or service meets a required standard. This is
particularly important for diagnostic imaging equip-
ment, as it may not be immediately obvious that there
have been any changes in the performance of the
equipment. A QA programme for a MRI scanner must
be able to detect changes in system performance,
allowing equipment faults to be identified and rectified
before they become clinically significant. This goal can
only be achieved if the tests are of an appropriate type
and range, and they are carried out at an appropriate
frequency. Data from QA tests can be used to identify
trends and hence anticipate deterioration in perfor-
mance, and in some cases provide evidence to support
equipment replacement business cases. The results also
enable system drift to be monitored and quantified. This
is particularly important when comparing follow-up
scans on a patient, as it is necessary to ensure that any
change in the appearance of the image is not due to
changes in the equipment performance.
The importance of undertaking appropriate measure-
ments on new systems should also not be overlooked.
Acceptance and commissioning tests allow a judgement
to be made on whether the equipment meets the
manufacturer’s specifications, as well as providing
baseline performance data for future QA testing.
McRobbie et al [2] found considerable value in under-
taking acceptance tests on new MR installations, with
signal to noise ratio and geometric linearity being the
most common parameters to fail acceptance tests, and
these were successfully corrected and improved in the
majority of instances.
For radiological equipment where ionizing radiation is
used, not only are there national standards for QA tests
to be undertaken [3, 4], but also the requirement to carry
out these tests has been enshrined in legislation [5]. For
MRI the situation is somewhat different. Comprehensive
guidelines exist that detail a range of appropriate QA
tests and their methodology [6]. However, there appears
to be no consensus as to how frequently these tests
should be undertaken in clinical practice, and currently
there are no statutory requirements for these tests to be
carried out.
The purpose of this survey was to identify which MRI
QA tests are currently undertaken in England, who
undertakes the tests, and the frequency at which these
tests are carried out. The results of this survey should
assist radiology departments in the formulation of their
own appropriate QA programme and contribute towards
establishing which tests to undertake and their appro-
priate frequency.

592 The British Journal of Radiology, July 2006

Survey of QA programmes

Method
Within the National Health Service (NHS) there are
over 200 MRI scanners in England, located in around 180
hospitals. From an alphabetical list of hospital trusts with
MR scanners, 45 were selected by contacting every fourth
hospital, enquiring if they would participate in this
study. Of these, 35 hospitals agreed to take part in this
survey, and a questionnaire was sent to each. Where
hospitals had more than one MRI scanner, only one
questionnaire was returned which reflected their typical
QA programme. However, in practice this was similar
for all the scanners within a particular hospital. For those
hospitals that received medical physics support, a copy
of the questionnaire was also sent to the relevant medical
physics department. Follow-up calls were made to those
hospitals that had not returned the survey to maximize
the return rate.
Information was requested on the nature and fre-
quency of QA tests undertaken on the MRI equipment,
as well as the type, field strength and age of the MRI
equipment and also the staff group (radiographers,
medical physicists, clinical engineering or service engi-
neers) who undertook these tests.

Results
From 35 hospitals who received the questionnaire,
completed replies were received from 24, corresponding
to a response of just over 68%. Responses were received
in two further cases where the MRI scanner was
undergoing replacement and, as routine QA was under
review, the users felt unable to complete the question-
naire. Overall, this corresponded to a 74% response rate.
The responses covered a broad range of equipment
manufacturers, field strengths and ages (as shown in
Figure 1) and all four major MRI manufacturers were
represented to varying degrees.
Out of the 24 hospitals, 79% undertook some form of
QA (Figure 2a). Most routine QA was undertaken by the
radiographers who operated the equipment on a regular
basis. Service engineers have been excluded from this
graph as it is understood that they will invariably
undertake some QA during each service visit.
Figure 2b compares the different tests performed by
radiographers and medical physicists. Both staff groups
performed a range of tests, with signal to noise ratio Figure 1. Number of hospitals responding to the survey by
(SNR) measurements most commonly undertaken by (a) manufacturer, (b) field strength and (c) equipment age.
radiographic staff. However, there was a considerable
difference in the range of tests undertaken in different
hospitals with only four hospitals carrying out all eight Figure 2e,f detail how often the other QA tests (i.e.
of the QA tests listed. non-SNR) are performed by radiographers and medical
Test frequencies for head coils and for other coils are physicists, respectively. Image uniformity testing was
compared between radiographers and medical physicists carried out most often by both groups of staff. Slice
in Figure 2c. The head coil was the most frequently position was the least frequent test to be carried out. The
tested of all the coils, typically tested on either a daily or time interval between tests undertaken by the two
weekly basis by radiographic staff. The frequency of groups of staff generally mirrored that of the SNR tests,
testing of the other coils was considerably less and this with medical physicists tending to carry out a wider
was further highlighted in Figure 2d which shows that range of tests less often. Once again, there was a marked
where SNR measurements were performed, they were variation in practice between hospitals.
always undertaken on the head coil. Half of the hospitals The time interval between service visits varied
extended this test to the body coil, but less than 13% between 1 month and 6 months, with 3 months being
included any further coils. the most common (Figure 3a). At each service visit some

The British Journal of Radiology, July 2006 593

 C J Koller, J P Eatough, P J Mountford and G Frain

80
80

Percentage of hospitals
Percentage of hospitals

60 60

40 40

20 20

0
Any Staff Radiographers Medical Physics Clinical Engineering 0

Resolution
Geometric

Thickness
Uniformity

Distortion

Ghosting
Linearity

Position
Spatial
Image

Image

Slice

Slice
SNR
(a) (b)

45 80
40
35 60
30
Percentage of hospitals

Percentage of hospitals
25
20 40

15
10 20
5
0
6 Monthly +
Fortnightly

2 Monthly

3 Monthly

4 Monthly

0
Monthly
Weekly
Daily

Extremity
Shoulder

Surface
Spine
Head

Body
Frequency Coil

(c) (d)

30 10

25 8

20
Percentage of hospitals

Percentage of hospitals

6
15
4
10

5 2

0 0
6 Monthly +
Fortnightly

2 Monthly

3 Monthly

4 Monthly
Monthly
Weekly
Daily

6 Monthly +
Fortnightly

2 Monthly

3 Monthly

4 Monthly
Monthly
Weekly
Daily

Frequency Frequency

(e) (f)

Figure 2. The percentage of hospitals that: (a) undertake quality assurance (QA), by staff group; (b) undertake QA by type of
test, and by staff group; radiographers &medical physicists; (c) undertake signal to noise ratio (SNR) tests by frequency, by coil
and by staff group; head coil-radiographers & head coil-medical physicists % other coils-radiographers other coils-medical
physicists; (d) undertake SNR tests, by coil type; (e) undertake other QA tests, by type and by frequency – radiographers. Image
Uniformity % Geometric Distortion & Spatial Resolution Ghosting Slice Thickness Slice Position. (f) undertake QA tests by
type and by frequency – medical physicists. Image Uniformity % Geometric Distortion & Spatial Resolution Ghosting Slice
Thickness Slice Position.

QA was routinely undertaken on both the head and body [2, 7, 8]. Hence a significant change in SNR will indicate a
coils. However, the testing of the other coils was not as potential problem, but more specific tests are required to
regular (Figure 3b). locate the exact cause. It should be noted, however, that
testing the SNR on one coil does not itself provide
information on the performance of any of the other coils.
Discussion The other tests undertaken regularly, such as image
uniformity, spatial resolution and slice thickness exam-
One goal of undertaking QA is to maintain confidence ine, amongst other things, field gradient strength and
in the performance of imaging equipment and to ensure selection. Since many MRI systems have integral
that it is operating within specification. To that end, the gradient coils, this aspect of the system may be
measurement of SNR for an MRI unit gives a good, all sufficiently tested using just one receiver coil.
round indication of the performance of the whole SNR was the test most commonly performed in a QA
system. The SNR is determined by many different program, typically on a daily to weekly basis. This test
parameters, but for a particular sequence and coil was always performed on the head coil at hospitals
combination, should remain stable over the long term where QA was undertaken. Only half the hospitals

594 The British Journal of Radiology, July 2006

Survey of QA programmes

30 addition, the responsibility for equipment performance

and safety rests primarily with the user and not with the
25
engineers.
Percentage of hospitals

20 This survey was designed, in part, to capture a

representative sample of practice of MRI QA across
15
hospitals in England. It should, of course, be borne in
10 mind that the final response of 24 completed question-
naires from the random selection of 45 trusts might allow
5 the possibility of bias if the response, be it positive or
0
negative, was related to the extent of QA undertaken.
Monthly 2 Monthly 3 Monthly 4 Monthly 6 Monthly There was no evidence, however, that this was the case;
Frequency and, indeed, it is possible to argue both for and against a
bias towards hospitals that undertake greater QA.
(a)
Moreover, the possibility of bias would not affect the
30
major finding of this study, namely that the range and
25 frequency of MRI QA varies substantially between
different hospitals.
20
Percentage of hospitals

Quality is a core principle of both clinical governance

15 and the NHS Plan [10]. In terms of medical imaging, this
means the assurance that equipment is always working
10 optimally in order to enable the quality and confidence
5
of diagnosis to be maintained. However effective, QA is
also about maintaining a balance between undertaking
0 sufficient meaningful tests to ensure the equipment is
2 Monthly

3 Monthly

4 Monthly

6 Monthly
Monthly

Annual

operating optimally, whilst ensuring that the resource is

used effectively in scanning patients. In the current
Frequency climate, patient throughput is essential in maintaining
(b) waiting lists. Hence too much QA may also be
deleterious to the diagnostic service.
Figure 3. Percentage of hospitals where service engineers
undertake quality assurance (QA) tests: (a) by frequency;
(b) by coil tested. , Head coil; &, body coil; %, other coils.
Conclusion
extended this test to the body coil, and even fewer This national survey indicated that 21% of hospitals
hospitals (, 13%) tested the other coils available. Hence did not undertake any form of in-house QA on their MR
these hospitals are unable to make any judgment of the scanners, and as such may be out of step with national
long-term performance of these coils. practice. However, there was a wide variation in the
Another goal of QA is to maintain equipment number and frequency of QA tests carried out, and even
calibrations, to ensure that there is no image distortion, for the most common QA test (SNR measurements) there
and that any numerical quantities measured are accurate. was a wide variation in the frequency and range of coils
The importance of this will depend to a great extent on tested. A national standard would help users to achieve
how the diagnostic images are to be interpreted, and the the balance between maintaining quality and effective
type of diagnostic information required. Hence certain patient throughput.
specific QA tests will be more appropriate to some
hospitals [9].
Due to the complexity of MRI equipment, all hospitals Acknowledgments
received regular service visits on a 1–6 monthly basis.
During these service visits, the service engineer may The authors wish to thank the MR staff and the service
undertake some form of QA tests, but this will depend engineers who took part in this survey.
upon the work required and the availability of the MRI
equipment. The particular tests undertaken will vary
between manufacturers, engineers and between different
References
systems, and may well not be appropriate for the 1. Imaging and radiodiagnostics (KH12). http://www.perfor-
required diagnostic information. mance.doh.gov.uk/hospitalactivity. Department of Health.
Of note is that 21% of hospitals did not perform any October 2004.
QA measurements themselves, but left these tests to the 2. McRobbie DW, Quest RA. Effectiveness and relevance of
service engineers. These hospitals did not receive service MR acceptance testing: results of an 8 year audit. Br J Radiol
2002;75:523–31.
visits with any greater frequency than any of the other
3. The Institute of Physics and Engineering in Medicine.
hospitals. This approach should only be undertaken with Recommended standards for the routine performance
due caution as the users and the service engineers may testing of diagnostic X-ray imaging systems. IPEM Report
have different goals. The user may wish to observe and 91. York: IPEM, 2005.
minimize any drift in equipment performance whereas 4. The Institute of Physics and Engineering in Medicine.
the engineer is primarily concerned with ensuring that Quality control of gamma cameras and associated computer
the equipment is just operating within specification. In systems. IPEM Report 66. York: IPEM, 1997.

The British Journal of Radiology, July 2006 595


5. The Ionising Radiations Regulations 1999. London: The
Stationery Office, 2000.
6. The Institute of Physics and Engineering in Medicine.
Quality control in magnetic resonance imaging. IPEM
Report 80. York: IPEM, 2000.
7. Colombo P, Baldassarri A, Del Corona M, Mascara L, Strocchi
S. Multicenter trial for the set-up of a MRI quality assurance
programme. Magn Reson Imaging 2004;22:93–101.
596
C J Koller, J P Eatough, P J Mountford and G Frain
8. Lerski RA, De Certaines JD. Performance assessment and
quality control in MRI by eurospin test objects and
protocols. Magn Reson Imaging 1993;11:817–33.
9. Barker GJ, Tofts PS. Semiautomated quality assurance for
quantitative magnetic resonance imaging. Magn Reson
Imaging 1992;10:585–95.
10. The NHS Plan: a plan for investment, a plan for reform.
London: The Stationery Office, 2000.
The British Journal of Radiology, July 2006
The British Journal of Radiology, 79 (2006), 597–602

Influence of menopausal status and use of hormone replacement

therapy on radiation dose from mammography in routine breast
screening
C J WHITAKER, BSc, PhD, C M KELLY, BA, K FAULKNER, FIPEM, FInstP, FSRP and E C STAMP, MMathStat

Quality Assurance Reference Centre, 9 Kingfisher Way, Silverlink Business Park, Wallsend NE28 9ND,
UK

ABSTRACT. Menopausal status and hormone replacement therapy (HRT) cause

alterations in breast structure which can affect mammographic image quality. Here we
present the results of a study to discover the effect of menopausal status and HRT use
on breast dose. Women attending routine screening completed questionnaires which
included questions regarding menopausal status and HRT use. Details of the
radiographic technique factors were recorded, from which the mean glandular dose
(MGD) per film for each woman was calculated. MGD values were analysed with regard
to the woman’s menopausal status and HRT use. The data from 516 women were
analysed. Among the women who had never used HRT, women who had not
undergone the menopause had a mean MGD of 2.94 mGy per film, whereas post-
menopausal women had a lower mean MGD of 2.52 mGy per film: a difference which
was found to be highly significant (p50.0045). Post-menopausal women who had never
Received 1 September
used HRT and those who had previously used HRT, but had ceased using it, had 2005
identical mean MGDs (2.54 mGy per film), whereas current HRT users had a significantly Revised 6 October 2005
greater mean MGD (2.89 mGy per film, p50.003). Women currently using HRT receive a Accepted 25 October 2005
statistically significantly larger radiation dose from routine breast screening than other
DOI: 10.1259/bjr/68819456
women. However, this effect is small and only occurs during the period of HRT use.
Women who have ceased using HRT show no difference in MGD compared with ’ 2006 The British Institute of
women who have never taken HRT. Radiology

In recent years, there has been increasing interest in
the effects of hormone replacement therapy (HRT) on
women’s health [1–6]. HRT use has been increasing
rapidly since its widespread introduction in the early
1970s [7, 8], with an estimated 50% of women currently
aged 50 years and over having taken HRT at some time
[9]. Two of the largest studies encompassing research
into the effects of HRT on health are the Million Women
Study (MWS) (recruitment: 1996–2001) in the UK [4, 9]
and the Women’s Health Initiative (WHI) (recruitment
1993–1998) in the USA [3, 5]. One of the main differences
between these studies is that the WHI actively enrolled
women into a clinical trial of HRT, whereas the MWS
simply recorded the woman’s HRT use, as prescribed by
her GP.
The MWS was designed to investigate relationships
between the patterns of use of HRT and certain medical
conditions, including breast cancer, in women attending
routine screening as part of the NHS Breast Screening
Programme (NHSBSP) [4, 10]. Since the MWS began in
1996, it has recorded the personal details of over one
Address correspondence to: Dr Keith Faulkner.
This study was partially supported by the European Commission’s
Radiation Protection Research Programme, project DIMOND III
(Measures for Optimising Radiological Information and Dose in
Digital Imaging and Interventional Radiology), contract number
FIGM-CT-2000-00061.
million women [9], including specific details about
previous medical conditions, family history of cancer,
menopausal status and HRT use, by use of a self-
administered questionnaire.
Although data analysis for the MWS is still ongoing,
the study has already described some of the effects of
HRT on breast screening and breast cancer [9, 11, 12]. As
part of the MWS, women were tracked through the
NHSBSP to determine rates of breast cancer, which were
then analysed with regard to their original questionnaire
details. It was found that current users of HRT were
more likely to develop breast cancer than women who
had never taken HRT and women who were past users
of HRT [9]. Many other studies have also found a direct
link between HRT use and increased incidence of breast
cancer [1, 5, 13–15]. It was this HRT-induced increase in
breast cancer which prompted the premature end to the
WHI trial of combined oestrogen and progesterone HRT
in 2002 [5].
With specific regard to the population studied in the
MWS, i.e. women attending for mammograms as part of
the NHSBSP, the direct effects of HRT on mammography
need to be considered. This is because the clarity of a
mammogram is affected by the tissue composition of the
breast [16]. Fatty tissue has low absorption of X-rays,
whereas ductal and glandular tissues absorb radiation
and are, therefore, radiologically dense [17]. It is this
ratio between fatty and dense breast tissue which

The British Journal of Radiology, July 2006 597

 C J Whitaker, C M Kelly, K Faulkner and E C Stamp

determines overall breast density and which can be Every woman who attended a particular screening
strongly influenced by both menopausal status and HRT session was given the opportunity to take part in the
use [16–19]. Despite this, the impact of menopausal study. If the woman agreed, she was given an informa-
status and HRT use on the radiation dose received by tion sheet explaining the study and asked to complete a
women during routine breast screening has not been consent form and a modified version of the MWS
established. questionnaire [20] whilst at the screening centre. In total,
The objectives of this study were to: 520 women across the region gave their consent to the
1. Determine the radiation dose received by 500 women study and completed the questionnaire. The question-
during routine breast screening across five Breast naire was based on the MWS questionnaire [20] and
Screening Units; included detailed questions on the woman’s menopausal
2. Investigate the effect that a woman’s menopausal status and HRT use. However, in order that additional
status has on MGD from mammography; data on the woman could be obtained, the woman’s
3. Establish whether taking HRT affects radiation dose individual NHS Screening Number (Sx number) was
and whether this effect continues after the cessation of also recorded on each questionnaire.
HRT use; Following the woman’s routine mammogram, details
4. Compare the main types of HRT preparation to see of the radiographic technique used for each film (such as
what influence, if any, current HRT type has on projection, tube potential, tube current exposure time
mammographic radiation dose. product and breast thickness) were recorded. The radio-
graphic data, along with the X-ray tube output measure-
ments from the mammography machine, were inputted
into the NHSBSP Breast Dose Calculator program [21], to
Methods and materials estimate the mean glandular dose (MGD) for each film.
However, since the women had varying numbers of
Structure of the study films taken each due to various factors including the
Women were recruited from five Breast Screening introduction of two-view screening at three out of five of
Units across the North East, Yorkshire and the Humber the Units in this study, the average MGD for each
Government Office Regions of the UK between woman was calculated on a ‘‘per film’’ basis, giving each
November 2002 and September 2003, with Units con- woman a mean MGD per film (mean MGD). There were
tributing between 13.2% and 31.6% of the women in the only four women whose radiographic data was not
study (Table 1). Ethical approval was obtained from the sufficiently complete to calculate their MGD, leaving 516
Northern and Yorkshire Multi-Centre Research Ethics women for inclusion in the data analysis.
Committee, as well as from each specific local Research
Ethics Committee for each Unit.
Definition of groups for data analysis
Table 1. Characteristics of women in the study as deter-
mined from the questionnaire The mean MGD value for each woman was linked to
her questionnaire responses via her individual Sx
Criterion Number of Percentage number. This then allowed mean MGD values to be
women
analysed with regard to the woman’s menopausal status
Number of women in studya 516 100.0 and details of her HRT use.
Age (years) ,55 215 41.7 Initially, women were divided into three different
55–59 160 31.0 menopausal groups, pre-, peri- and post-menopausal,
60–64 84 16.3 based mainly on whether or not their periods had
64+ 57 11.0
stopped at the time of screening. In addition, women
Menopausal statusb Pre-menopausal 30 5.8
Peri-menopausal 30 5.8
aged 53 years and over who had had either a
Post-menopausal 454 88.0 hysterectomy without oophorectomy (removal of both
Unknown 2 0.4 ovaries) or who had started taking HRT before their
HRT use Never 272 52.7 natural menopause were defined as post-menopausal,
Previous 126 24.4 along with women of any age who had had a bilateral
Current 106 20.5 oophorectomy. This categorisation was in line with
Unknown 12 2.3 guidance from the Million Women Study [12].
HRT typec Combined 45 42.5 Subsequently, the pre- and peri-menopausal women
Oestrogen only 31 29.2 were combined for comparison against the mean MGD
Other/Unknown 30 28.3
of post-menopausal women. All data were analysed
Unit 1 68 13.2
2 91 17.6 using a one-way ANOVA to assess variation and to
3 99 19.2 establish statistical significance.
4 163 31.6 As part of the questionnaire, women were asked
5 95 18.4 whether they had ever taken HRT and, if so, whether
a they were currently taking HRT and which particular
Only women with a completed questionnaire, consent form
and full mammographic screening data were included in type. Their answers led to the post-menopausal women
data analysis. being categorised as never having taken HRT (never),
b
Definitions of menopausal status are included in the previously having taken HRT (previous) or currently
Methods and Materials. taking HRT (current). Mean MGD values were compared
c
Current hormone replacement therapy (HRT) users only. between the three groups and also for a combination of

598 The British Journal of Radiology, July 2006

Influence of menopausal status and HRT on breast dose
never and previous users, against current HRT users.
Finally, current HRT users were divided according to the
type of HRT they were taking; combined oestrogen and
progesterone, oestrogen-only or another type of prepara-
tion. Again, the mean MGDs for the women in each
group were compared using a one-way ANOVA.
Results
The characteristics of the women who took part in the
study are outlined in Table 1. The women ranged in age
from 42 years to 82 years, with a mean age of 57.0 years,
the majority being aged less than 60 years (72.7%). The
vast majority of the women’s questionnaires indicated
that they were post-menopausal (88.0%), with the
remainder being equally spilt between pre- and peri-
menopausal (5.8% each). Only two women (0.4%) who
completed the questionnaires failed to give details of
their menopausal status.
Just under half of the women had taken HRT (44.9%)
at some time, with 20.5% indicating that they were
current users. As with menopausal status, only a small
proportion of women (2.3%) did not specify whether
they had ever taken HRT. For current HRT users,
combined oestrogen and progesterone was taken by
42.5%, with oestrogen-only HRT comprising 29.2%.
However, there was a large percentage of women
(28.3%) who either used other forms of HRT (such as
progesterone-only) or did not state the type of HRT on
the questionnaire. This lack of knowledge about the type
of HRT used may have led to difficulties in ascertaining
any specific influence of the type of HRT on radiation
dose (see Discussion).
The mean MGD data from 516 women were analysed
using a one-way ANOVA. The mean MGD per film for
each woman was analysed by assigning the women to a
group depending on their menopausal status and details
of their HRT use, as obtained from their questionnaires
and defined in the Methods and Materials section.
Influence of menopausal status
Pre- and peri-menopausal women had similar mean
MGD values of 2.84 mGy and 2.91 mGy, respectively
The British Journal of Radiology, July 2006
(Figure 1). Post-menopausal women had a lower mean
MGD of 2.62 mGy, although this difference was not
statistically significant (p50.08). If, however, post-meno-
pausal women who have never used HRT are compared
directly with pre- and peri-menopausal women who
have also never used HRT (Figure 2), then the difference
in MGDs becomes statistically significant (p50.0045).
These non-HRT post-menopausal women have a mean
MGD of 2.52 mGy, compared with a mean MGD of
2.94 mGy for non-HRT pre- and peri-menopausal
women.
Influence of HRT use
In post-menopausal women, women who had never
used HRT and those women who had previously used
HRT, but now did not, had practically identical mean
MGDs of 2.54 mGy (Figure 2). However, current users of
HRT had a greater mean MGD of 2.89 mGy, which was
similar to the MGD for pre- and peri-menopausal
women who had never taken HRT (2.91 mGy). This
difference in MGD between current users of HRT and
women not currently taking HRT was found to be highly
statistically significantly (p50.003; Figure 3).
Influence of HRT type
As described previously, the large proportion of
women who did not specifically identify their current
type of HRT in the questionnaire may have adversely
affected the investigation as to whether the type of HRT
influences MGD. For current HRT users who specified
HRT type, there was no statistically significant difference
in mean MGDs between combined (2.73 mGy) and
oestrogen-only (2.88 mGy) HRT (Figure 4).
Discussion
The data obtained from the women’s questionnaires
was generally complete, with only 14 women (2.7%) not
providing enough details to classify either their meno-
pausal status or history of HRT use (Table 1). Overall,
only four women (0.8%) were excluded from the data
analysis due to lack of complete radiographic data.
Figure 1. Effect of menopausal sta-
tus on mean glandular dose (MGD)
per film. Error bars indicate 95%
confidence levels from one-way
ANOVA analysis.
599

The classification of women into three menopausal
groups was carried out following the criteria used by the
MWS [12]. The percentages of women in each category,
pre-, peri- and post-menopausal, were 5.8%, 5.8% and
88% in the study reported here (Table 1) and 6.3%, 6.5%
and 87% in the MWS [12], respectively. Classification of
women based on HRT use was more straightforward
than menopausal status and led to 45% of women being
classified as a previous or current user of HRT. This
figure is similar to the MWS where questionnaires
indicated that almost half of women (47%) had used
HRT at some time [4]. There is obviously close agreement
between our study and the MWS with regard to the
woman’s menopausal details and HRT use.
In our study, there was a large proportion of current
HRT users who were unable to accurately recall the
name or type of HRT preparation they were currently
taking. Combined oestrogen and progesterone accounted
for 43% of current users, with oestrogen-only prepara-
tions accounting for 29% (Table 1). However, there was a
large number of current HRT users classified as using
either another or unknown type of HRT (28%). These
figures are comparable with those obtained by the MWS.
In the MWS, the majority of current HRT users took
combined oestrogen and progesterone HRT (60%), with
around a third taking oestrogen-only HRT (31%) and
only 8.7% of women classed as other or unknown HRT
type [9]. This large difference in the proportion of current
HRT users classified as ‘‘unknown HRT type’’ between
600
C J Whitaker, C M Kelly, K Faulkner and E C Stamp
Figure 2. Effect of hormone repla-
cement therapy (HRT) use on mean
glandular dose (MGD) per film of
post-menopausal women (grey bars)
and pre- and peri-menopausal
women combined (dark grey bar).
‘‘Never’’ indicates women who have
never used HRT, ‘‘Previous’’ indi-
cates women who have taken HRT
in the past, but have now ceased
HRT use, and ‘‘Current’’ indicates
women currently taking HRT. Error
bars indicate 95% confidence levels
from one-way ANOVA analysis.
our study and the MWS is likely to be due to the
difference in questionnaire protocol.
In the MWS, the women completed their questionnaire
at home, where they were likely to have their current
HRT to hand and so were able to identify the type
accurately. However, since our questionnaire was com-
pleted at the screening unit, the women were less likely
to be able to remember the exact name or type of
preparation they were using. This led to difficulty in
interpreting the radiation dose results with regard to
HRT type for current HRT users. Although our results
indicated a small (5%) difference in dose received
between those women taking combined oestrogen and
progesterone and those taking oestrogen-only HRT
preparations (Figure 4), the lack of specific details about
HRT type resulted in too few women to analyse the data
meaningfully.
Radiation dose
The mean radiation dose received by the 516 women in
this study was 2.66 mGy. This compares with an average
MGD of 2.23 mGy determined by Young et al [22] for
women attending routine breast screening across the UK
during 2001 and 2002. The small difference between the
two mean MGDs is likely to be due to the use of Lorad
systems (Lorad M-IV) in four out of five of the Units
which took part in our study. Lorad systems are known
Figure 3. Effect of current hormone
replacement therapy (HRT) use on
the mean glandular dose (MGD) per
film of post-menopausal women.
Error bars indicate 95% confidence
levels and *** indicates highly sig-
nificant difference (p,0.005) from
one-way ANOVA analysis.
The British Journal of Radiology, July 2006
Influence of menopausal status and HRT on breast dose
to give a higher MGD than other systems and, according
to the Young et al study, account for only 15% of the
systems used nationally [22].
Ideally, the best approach to estimating the glandular
dose is to apply an accurate glandularity factor, specific
to each woman. In the absence of detailed knowledge of
breast glandularity for each woman, an average glandu-
larity factor has been applied. As the glandularity
between groups analysed here may vary, part of the
observed differences between groups will be due to this
effect. It is anticipated that the actual difference between
groups will be smaller.
Influence of menopausal status on radiation dose
Although the mean MGD for post-menopausal women
was around 9% lower than those for pre- and peri-
menopausal women together, this difference was not
statistically significant (p50.08) when HRT status was
not taken into account, perhaps due to the HRT masking
any real effect the menopause has on MGD. However,
when women who had never taken HRT were analysed
with regard to their menopausal status, the difference
between this group of women and the post-menopausal
women did reach statistical significance, with post-
menopausal women being found to have received a
statistically significantly smaller radiation dose
(2.52 mGy) than women who had not yet undergone
the menopause (2.94 mGy: p50.0045). The difference in
radiation dose between these groups may reflect the
changes which occur to the structure of a woman’s
breasts after the menopause [18].
Influence of HRT use on radiation dose
It was decided that only post-menopausal women
would be analysed with regard to HRT use, since the
numbers of pre- and peri-menopausal women were so
small (60 in total) that it would severely limit any
statistical analysis. It was also likely that the vast
majority of non-menopausal women would never have
taken HRT.
In post-menopausal women, current users of HRT
received an approximately 12% larger radiation dose
The British Journal of Radiology, July 2006
Figure 4. Effect of type of hormone
replacement therapy (HRT) on the
mean glandular dose (MGD) per film
of current HRT users. ‘‘Combined’’
indicates combined oestrogen and
progesterone users. Error bars indi-
cate 95% confidence levels from
one-way ANOVA analysis.
than women who were not taking HRT at present
(Figure 2). This difference was highly statistically
significant (p50.003). There was no statistical difference
in radiation dose received between women who had
stopped taking HRT and those who had never taken it
(Figure 2). This temporary nature of the effect of HRT on
radiation dose is supported by research which has
shown that the effects of HRT use, both on the incidence
of breast cancer [9, 14] and on the specificity of
mammography [11], decline rapidly after initial cessation
and return to the levels of women who have never used
HRT. These results indicate that the influence of HRT on
radiation dose, breast cancer and mammograms only
lasts for the length of time that the woman is taking HRT,
with little long term effect.
However, for current HRT users, the increased
radiation dose received may have important implica-
tions. The difference between radiation dose received by
a current user of HRT, versus a non-user is approxi-
mately 0.35 mGy per film, which equates to 0.7 mGy per
routine two-view screen (four films). Although this
additional amount of radiation received is relatively
small, there may still be an impact when it comes to
radiation-induced breast cancer. Law and Faulkner
calculated the effect of specific radiation doses on the
rates of induction of breast cancer [23]. For women aged
50–64 years, this equates to approximately 11.4 breast
cancers per million women for each mGy of radiation.
The small amount of additional radiation incurred as a
result of taking HRT is outweighed by the substantial
benefit of screening women on HRT, since this group of
women is known to have greater background incidence
of breast cancer than women not taking HRT [1, 5, 9, 11–
15]. Other factors, such as lack of specificity and
selectivity of mammograms with HRT use [24, 25], are
more likely to affect the benefit/risk ratio.
Conclusions
Women currently using HRT receive a statistically
significantly larger radiation dose from routine breast
screening than other women. However, this effect is
small and only occurs during the period of HRT use. The
increase in mammography radiation dose for women
taking HRT may result in a small increase in the number
601

of radiation-induced cancers, but this is probably
justified due to the higher incidence of breast cancer in
these women. Women who have ceased using HRT show
no difference in radiation dose received compared with
women who have never taken HRT.
Acknowledgments
We wish to thank all the women who agreed to take
part in the study and completed questionnaires, along
with the staff at the five Breast Screening Programmes in
the North East, Yorkshire and The Humber Government
Office Region who were involved in this study. Thanks
to Prof. V Beral, Ms B Crossley and Ms A Brown from the
Cancer Research UK Epidemiology Unit, Oxford, for
their initial advice about the questionnaire and subse-
quent computerized data extraction from the question-
naires. The authors are also grateful to Mrs C Clayton,
Mr G McGill, Mr C Moore and Dr K J Robson for their
help with the Medical Physics data.
References
1. Goddard MK. Hormone replacement therapy and breast
cancer; endometrial cancer and cardiovascular disease:
risks and benefits. Br J Gen Prac 1992;42:120–5.
2. Hulley S, Grady D, Bush T, Furberg C, Herrington D, Riggs
B, et al. Randomized trial of estrogen plus progestin for
secondary prevention of coronary heart disease in post-
menopausal women. JAMA 1998;280:605–13.
3. Women’s Health Initiative Study Group. Design of the
Women’s Health Initiative Clinical Trial and Observational
Study. Control Clin Trials 1998;19:61–109.
4. The Million Women Study Collaborative Group: The
Million Women Study: design and characteristics of the
study population [peer-reviewed research].
5. Writing Group for the Women’s Health Initiative
Investigators. Risks and benefits of estrogen plus progestin
in healthy postmenopausal women. Principal results from
the Women’s Health Initiative randomized controlled trial.
JAMA 2002;288:321–33.
6. Li C, Wilawan K, Samsioe G, Lidfeldt J, Agardh C-D,
Nerbrand C. Health profile of middle-aged women: the
Women’s Health in the Lund Area (WHILA) study. Human
Repro 2002;17:1379–85.
7. Gambrell RD Jr. Studies of endometrial and breast disease
with hormone replacement therapy. In: Studd JWW,
Whitehead MI, editors. The menopause. Oxford, UK:
Blackwell Scientific Publications, 1988:247–61.
8. Jolleys JV, Olesen F. A comparative study of prescribing of
hormone replacement therapy in USA and Europe.
Maturitas 1996;23:47–53.
602
C J Whitaker, C M Kelly, K Faulkner and E C Stamp
9. Million Women Study Collaborators. Breast cancer and
hormone-replacement therapy in the Million Women
Study. Lancet 2003;362:419–27.
10. http://cancerscreening.org.uk/breastscreen/ [Accessed 2
February 2006].
11. Banks E, Reeves G, Beral V, Bull D, Crossley B, Simmonds
M, et al. Impact of use of hormone replacement therapy on
false positive recall in the NHS breast screening pro-
gramme: results from the Million Women Study. BMJ
2004;328:1291–2.
12. Banks E, Reeves G, Beral V, Bull D, Crossley B, Simmonds
M, et al. Influence of personal characteristics of individual
women on sensitivity and specificity of mammography in
the Million Women Study: cohort study. BMJ 2004;329:477.
13. Marsden J. Hormone-replacement therapy and breast
cancer. Lancet Oncol 2002;3:303–11.
14. Coombs NJ, Taylor R, Wilcken N, Boyages J. Hormone
replacement therapy and breast cancer: estimate of risk.
BMJ 2005;331:347–9.
15. Ewertz M, Mellemkjaer L, Poulsen AH, Friis S, Sorensen
HT, Pedersen L, et al. Hormone use for menopausal
symptoms and risk of breast cancer. A Danish cohort
study. Br J Cancer 2005;92:1293–7.
16. Warren R. Hormones and mammographic breast density.
Maturitas 2004;49:67–78.
17. Sterns EE, Zee B. Mammographic density changes in
perimenopausal and postmenopausal women: is effect of
hormone replacement therapy predictable? Breast Cancer
Res Treat 2000;59:125–32.
18. Miller WR, Anderson TJ. Oestrogens, progestogens and the
breast. In: Studd JWW, Whitehead MI, editors. The
menopause. Oxford, UK: Blackwell Scientific Publications,
1988:234–46.
19. Greendale GA, Reboussin BA, Sie A, Singh R, Olson LK,
Gatewood O, et al. Effects of estrogen and estrogen-
progestin on mammographic parenchymal density. Ann
Intern Med 1999;130:262–9.
20. http://www.millionwomenstudy.org/[Accessed 2 February
2006].
21. http://www.cancerscreening.nhs.uk/breastscreen/publica-
tions/mammography-equipment.html#me-iqad [Accessed
2 February 2006].
22. Young KC, Burch A, Oduko JM. Radiation doses received in
the UK Breast Screening Programme in 2001 and 2002.
Br J Radiol 2005;78:207–18.
23. Law J, Faulkner K. Cancers detected and induced, and
associated risk and benefit, in a breast screening pro-
gramme. Br J Radiol 2001;74:1121–7.
24. Evans A. Hormone replacement therapy and mammo-
graphic screening. Clin Radiol 2002;57:563–4.
25. Carney PA, Miglioretti DL, Yankaskas BC, Lerlikowske K,
Rosenberg R, Rutter CM, et al. Individual and combined
affects of age, breast density, and hormone replacement
therapy use on the accuracy of screening mammography.
Ann Intern Med 2003;138:168–75.
The British Journal of Radiology, July 2006
The British Journal of Radiology, 79 (2006), 603–607

Kodak EDR2 film for patient skin dose assessment in cardiac

catheterization procedures
R E MORRELL, MSc, MIPEM and A T ROGERS, MSc, MIPEM

Medical Physics Directorate, Nottingham City Hospital NHS Trust, Hucknall Road, Nottingham NG5
1PB, UK

ABSTRACT. Patient skin doses were measured using Kodak EDR2 film for 20 coronary
angiography (CA) and 32 percutaneous transluminal coronary angioplasty (PTCA) Received 22 April 2005
procedures. For CA, all skin doses were well below 1 Gy. However, 23% of PTCA patients Revised 7 October 2005
Accepted 25 October 2005
received skin doses of 1 Gy or more. Dose–area product (DAP) was also recorded and was
found to be an inadequate indicator of maximum skin dose. Practical compliance with DOI: 10.1259/bjr/78359708
ICRP recommendations requires a robust method for skin dosimetry that is more accurate
than DAP and is applicable over a wider dose range than EDR2 film. ’ 2006 The British Institute of
Radiology

Cardiac catheterization procedures can involve pro-
longed fluoroscopic imaging and large numbers of
acquired images. As a result, patient skin doses may
approach or exceed the threshold for deterministic skin
effects [1–6].
The US Food and Drug Administration [7], and more
recently the International Commission on Radiological
Protection (ICRP), have published guidelines for dose
minimization and for recording skin doses to patients
who are suspected to be at risk. ICRP Report 85 [8]
recommends that the magnitude and position of the
maximum skin dose should be recorded in the patient’s
notes if it exceeds 1 Gy for procedures that are likely to
be repeated, or 3 Gy for all procedures. Patients who are
considered to be at risk should receive appropriate
information and clinical follow-up.
Implementation of these guidelines requires a robust
method for assessment of patient skin doses. Modern
cardiac X-ray units are fitted with dose–area product
(DAP) meters, which indicate the total amount of radiation
incident on the patient’s skin. However, there is no simple
relationship between DAP and maximum skin dose. The
dose distribution depends on which imaging projections
are used, and for what proportion of the procedure. This
can vary greatly from one patient to the next depending on
operator preference, the anatomy of the patient’s disease,
and the complexity of the procedure. Whilst some authors
have proposed DAP values to alert the operator to the
potential for deterministic effects [9–12], others have
reported poor correlations between DAP and maximum
skin dose [13–15].
Slow radiographic film can be used to record a map of
the skin dose distribution over a large area. As long as
the film is not saturated, absolute dose measurements
can be made directly from the film and any regions of
high dose can be easily identified. Of the films that are
compatible with standard radiology processors, the wide
latitude films developed for portal imaging and quality
control applications in radiotherapy are responsive to the
highest radiation doses, and are thus most suitable for
dosimetry in high dose diagnostic and interventional
procedures [13, 16–18].
The dosemeter selected for this study was Kodak
EDR2 film (Eastman Kodak Company, Rochester, NY),
which currently has the widest available dose range.
Guibelalde et al [19] have successfully used it for skin
dosimetry during interventional cardiology procedures.
They reported its saturation point at 1.4 Gy, and found
saturation to occur in about 1% of cases.
The performance of the film has been characterized in
detail, as described in a previous paper by the authors
[20]. It is available in 35 cm 6 43 cm sheets, which are
large enough to capture most of the radiation fields on
the patient’s back. It is pre-wrapped in light-proof paper,
ready for use. It can be processed in a standard radiology
processor using non-glutaraldehyde chemicals.
The purpose of the study was to determine typical skin
doses for patients undergoing coronary angiography
(CA) and percutaneous transluminal coronary angio-
plasty (PTCA) in our cardiac catheterization laboratory,
and to estimate the percentage of patients receiving
doses of 1 Gy or more. The film measurements were
compared with DAP to determine whether DAP could
be used as a predictor of maximum skin dose.
Method
The film was previously calibrated over the range of
exposure conditions typically encountered in our cardiac
catheterization laboratory [20]. Its response was char-
acterized across the dose range 20–1000 mGy. The effects
of beam energy and filtration, field size, exposure rate,
film batch and processing conditions were quantified.
The relationship between dose (D) and optical density
(OD) was found to be:
 
1 ODmax {OD
D~{ ln ð1Þ
a ODmax {ODmin
The constant a had a value of 0.0027 mGy21 when a 20 cm
polymethylmethacrylate (PMMA) phantom was used to
simulate the backscatter from a patient and with the

The British Journal of Radiology, July 2006 603

 R E Morrell and A T Rogers

processing conditions used in our department. The mean

optical density of unexposed film (ODmin) was 0.21, whilst
the mean density of fully exposed film (ODmax) was 3.92.
The film saturated at around 1 Gy. The uncertainty in dose
per optical density was estimated to be 229 mGy to
+62 mGy at a fixed dose of 160 mGy, across the full range
of exposure and processing conditions employed. At other
dose levels, this interval was assumed to scale linearly
with the gradient of the calibration curve, given by:

dD eaD
~ ð2Þ
d ðODÞ aðODmax {ODmin Þ

Skin doses were measured for 20 CA and 32 PTCA

procedures, performed on an Integris H5000F C-arm
imaging unit (Philips Medical Systems, Best,
Netherlands). The imaging equipment was subject to
monthly and annual quality control checks, as recom-
mended by the Institute of Physics and Engineering in
Medicine [21]. Patients were selected sequentially, and
the study included only those procedures performed by
our in-house consultant cardiologists and the registrars
working under their supervision. Fluoroscopy was
performed using the ‘‘low continuous’’ factory setting,
which has a nominal input dose rate at the detector of
740 nGy s21 and employs 0.4 mm copper filtration. All
acquisition runs were performed on the ‘‘12.5 FPS
Coronary’’ setting, which has a nominal detector input
dose rate of 870 nGy s21 and has no copper filter.
Before commencing each procedure, a sheet of 35 cm
6 43 cm EDR2 film was positioned on the imaging table,
underneath the mattress. The dotted rectangle in
Figure 1 demonstrates the position and orientation of
the film. Its long axis was perpendicular to the long axis
of the table, and its top edge was approximately level
with the patient’s shoulders. Each film was labelled to
indicate which side was face-up, and which edge was
closest to the patient’s head. Following exposure, a
pinhole was made in the corner of the film packet
corresponding to the patient’s left shoulder, to identify
the orientation of the processed film.
All films were stored in their packets overnight before
processing, since the response of EDR2 film takes some
time to stabilize following exposure [20, 22]. They were
then processed in a Kodak X-OMAT M6B processor
(Eastman Kodak Company, Rochester, NY), with
Photosol developer and fixer (Photosol Limited,
Basildon, UK). The maximum optical density of each
film was determined manually, using a Pehamed
Densoquick 2 densitometer (Pehamed, Sulzbach,
Germany).
The DAP for each procedure was measured using an
integral PTW-DIAMENTOR-M1 DAP meter (PTW- Figure 1. Film position and orientation.
FREIBURG, Freiburg, Germany). This had previously
been calibrated over the same range of exposure
conditions as the film. The uncertainty in its
response was estimated at ¡13%. The Pearson correla- Results
tion coefficient between DAP and peak skin dose
was calculated for each procedure type, for those Figure 2 shows a dosimetry film from a PTCA
procedures where no film saturation occurred. The procedure, viewed as if looking at the patient’s back.
significance of the correlations was determined using The patient’s left shoulder is indicated by the black spot
Student’s t-test. in the top left-hand corner of the image. The region of

604 The British Journal of Radiology, July 2006

Kodak EDR2 film for patient skin dose assessment
Figure 2. A dosimetry film from a coronary angioplasty,
viewed as if looking at the patient’s back. The black spot in
the top left-hand corner indicates the patient’s left shoulder.
maximum dose can be readily identified by the darkest
patch, in the top right-hand quadrant.
One film from each procedure type showed abnor-
mally large numbers of radiation fields, with extensive
areas of film saturation. Since the DAP from these
examinations was not particularly high, it seemed likely
that these films had been left on the imaging table for
more than one procedure. Both were excluded from the
data analysis.
Figure 3 shows the distribution of maximum skin
doses for coronary angiography and angioplasty proce-
dures. Skin doses for the angiograms were all less than
600 mGy. Seven films from angioplasty procedures were
saturated in at least one region, implying a skin dose of
1 Gy or more.
Figures 4 and 5 are scatter plots of maximum skin dose
against DAP for the two procedure types, for films that
demonstrated no saturation. A linear trend has been
fitted to each. The error bars indicate the expected
uncertainties in the measurements, as described in the
method.
The Pearson correlation coefficients between max-
imum skin dose and DAP were 0.76 for coronary
The British Journal of Radiology, July 2006
angiography, and 0.61 for coronary angioplasty. These
correlations were both significant (p 5 0.05). However,
Figures 4 and 5 show that some of the points deviate
considerably from the trend lines.
Discussion
For coronary angiography, all skin doses were well
below 1 Gy. This is in agreement with other published
studies using thermoluminescence dosimetry or film
[13–15, 23, 24]. The mean peak skin dose was 195 mGy,
with a range from 70 mGy to 520 mGy. Patients under-
going these diagnostic investigations are unlikely to
receive doses sufficient to cause deterministic effects.
For coronary angioplasty, 23% of patients received
skin doses of 1 Gy or more, sufficient to saturate the film.
It must be assumed that a similar proportion of our
patients approach or exceed this level in routine clinical
practice. Since it is fairly common for patients to undergo
more than one procedure, the dose to each of these
patients should be assessed and recorded.
Thermoluminescence dosimetry studies by Van de
Putte et al, Waite and Fitzgerald, and Verdun et al all
found some skin doses approaching or exceeding 1 Gy,
for PTCA patients [14, 15, 24]. The increased incidence of
film saturation compared with that reported by
Guibelalde et al may be at least partly explained by
our lower film saturation point of 1 Gy. Even if this is
increased to 1.5 Gy using a dedicated processor as
Guibelalde did, the film still cannot measure doses up
to the 2 Gy threshold for deterministic effects.
Given the large proportion of patients receiving skin
doses of at least 1 Gy, it seems likely that some of them
exceed the 2 Gy threshold for deterministic effects [25].
These patients should be monitored for skin effects, and
informed about potential symptoms and appropriate
action to take should any skin changes occur.
The study clearly identifies a need for routine assess-
ment of patient skin doses for coronary angioplasty
procedures in our cardiac catheterization laboratory.
EDR2 film identifies those patients whose doses may
exceed 1 Gy, and who may therefore be at risk of
deterministic skin effects. However, since this film
saturates at 1 Gy, it cannot be used to assess these
higher doses. It is labour-intensive as a dosimetry
Figure 3. Maximum skin doses for
coronary angiography (CA) and per-
cutaneous transluminal angioplasty
(PTCA) procedures.
605

method, because each film must be individually posi-
tioned and labelled, processed and analysed to deter-
mine the maximum dose. A further limitation of film
dosimetry is that contributions from lateral and very
wide oblique views are not measured.
DAP was found to be a poor indicator of maximum
skin dose, because clinical practice varies so much from
one procedure to the next. Although there was a
significant correlation between DAP and maximum skin
dose for both procedure types, it is evident from
Figures 4 and 5 that DAP alone cannot not reliably
predict high skin doses.
Alternative dosemeters used in cardiac catheterization
procedures include thermoluminescent dosemeters and
scintillation detectors [14, 15, 24, 26]. A major disadvan-
tage of such detectors is their small area. The location of
the maximum skin dose is not usually known in advance
and, if there is no detector at this exact location, skin dose
can be grossly underestimated. There is now a growing
range of ‘‘Gafchromic’’ films, produced by International
Speciality Products (Wayne, New Jersey). These are
designed specifically for patient dosimetry applications,
have higher dose ranges than EDR2 film, and do not
require any processing. At present however, they are
prohibitively expensive for routine dosimetric use.
A potential solution would be a mathematical model,
to calculate skin dose distribution from the exposure
parameters recorded in the DICOM image files for each
606
R E Morrell and A T Rogers
Figure 4. Maximum skin dose ver-
sus dose–area product (DAP), for
coronary angiography (CA).
individual acquisition run. As well as enabling assess-
ment of doses above the film’s saturation point, this
would require less staff involvement than film dosime-
try. The model could be applied routinely for all patients
or run for individual patients whose DAP exceeds a
certain trigger level.
Conclusion
Skin doses in our cardiac catheterization laboratory are
unlikely to exceed 1 Gy for CA. However, around 23% of
our patients receive maximum skin doses of at least 1 Gy
during PTCA procedures. DAP is not an adequate
indicator of patient skin dose. Practical compliance with
ICRP recommendations requires a method for routine
assessment of skin doses that is more accurate than DAP,
and is applicable over a wider dose range than EDR2
film. This may be achievable by means of a mathematical
model.
Acknowledgments
We would like to thank our cardiologists for agreeing
to participate in this study, and our catheterization
laboratory radiographers and nurses for their invaluable
assistance with data collection. We are grateful to Prof.
Alan Perkins for helpful discussions.
Figure 5. Maximum skin dose versus
dose–area product (DAP), for cor-
onary angioplasty (PTCA) proce-
dures for which no film saturation
occurred.
The British Journal of Radiology, July 2006
Kodak EDR2 film for patient skin dose assessment
References
1. Kawakami T, Saito R, Miyazaki S. Chronic radiodermatitis
following repeated percutaneous transluminal coronary
angioplasty. Br J Dermatol 1999;141:150–3.
2. D’Incan M, Roger H, le Boudec MCF, Souteyrand P.
Radiodermatitis following cardiac catheterization. Arch
Dermatol 1997;133:242–3.
3. Dehen L, Vilmer C, Humiliere C, Corcos T, Pentousis D,
Ollivaud L, et al. Chronic radiodermatitis following cardiac
catheterisation: a report of two cases and a brief review of
the literature. Heart 1999;81:308–12.
4. Shope TB. Radiation-induced skin injuries from fluoro-
scopy. Radiographics 1996;16:1195–9.
5. Vaño E, Goicolea J, Galvan C, Gonzalez L, Meiggs L, Ten JI,
et al. Skin radiation injuries in patients following repeated
coronary angioplasty procedures. Br J Radiol
2001;74:1023–31.
6. Vaño E, Arranz L, Sastre JM, Moro C, Ledo A, Garate MT,
et al. Dosimetric and radiation protection considerations
based on some cases of patient skin injuries in interven-
tional cardiology. Br J Radiol 1998;71:510–6.
7. Food and Drug Administration. Recording information in
the patient’s medical record that identifies the potential for
serious X-ray induced skin injuries following fluoroscopi-
cally guided procedures. Rockville, MD: Center for Devices
and Radiological Health, 1995.
8. International Commission on Radiological Protection.
Report 85: Avoidance of radiation injuries from medical
interventional procedures. Ann ICRP 2000;30:45–7.
9. Hansson B, Karambatsakidou A. Relationships between
entrance skin dose, effective dose and dose area product for
patients in diagnostic and interventional cardiac proce-
dures. Radiat Prot Dosim 2000;90:141–4.
10. Skinner CL. Implementation of IR(ME)R in cardiac angio-
graphy. In: Proceedings of UK Radiological Congress 2002;
2002 June 9–11; Birmingham, UK. London, UK: British
Institute of Radiology, 2002.
11. Neofotistou V, Vaño E, Padovani R, Kotre J, Dowling A,
Toivonen M, et al. Preliminary reference levels in interven-
tional cardiology. Eur Radiol 2003;13:2259–63.
12. McFadden SL, Mooney RB, Shepherd PH. X-ray dose and
associated risks from radiofrequency catheter ablation
procedures. Br J Radiol 2002;75:253–65.
13. Vaño E, Gonzalez L, Ten JI, Fernandez JM, Guibelalde E,
Macaya C. Skin dose and dose-area product values for
The British Journal of Radiology, July 2006
interventional cardiology procedures. Br J Radiol
2001;74:48–55.
14. van de Putte S, Verhaegen F, Taeymans Y, Thierens H.
Correlation of patient skin doses in cardiac interventional
radiology with dose-area product. Br J Radiol
2000;73:504–13.
15. Waite JC, Fitzgerald M. An assessment of methods for
monitoring entrance surface dose in fluoroscopically
guided interventional procedures. Radiat Prot Dosim
2001;94:89–92.
16. Geise RA, Ansel HJ. Radiotherapy verification film for
estimating cumulative entrance skin exposure for fluoro-
scopic examinations. Health Phys 1990;59:295–8.
17. Fajardo LC, Geise RA, Ritenoure RA. A survey of films for
use as dosimeters in interventional radiology. Health Phys
1995;68:595–9.
18. Vaño E, Guibelalde E, Fernandez JM, Gonzalez L, Ten JI.
Patient dosimetry in interventional radiology using slow
films. Br J Radiol 1997;70:195–200.
19. Guibelalde E, Vaño E, Gonzalez L, Prieto C, Fernandez JM,
Ten JI. Practical aspects for the evaluation of skin doses in
interventional cardiology using a new slow film. Br J Radiol
2003;76:332–6.
20. Morrell RE, Rogers A. Calibration of Kodak EDR2 film for
patient skin dose assessment in cardiac catheterization
procedures. Phys Med Biol 2004;49:5559–70.
21. Institute of Physics and Engineering in Medicine. Report 77:
Recommended standards for routine testing of diagnostic
X-ray imaging systems. York, UK: IPEM, 1997.
22. Childress NL, Rosen II. Effect of processing time delay on
the dose response of Kodak EDR2 film. Med Phys
2004;31:2284–8.
23. Delichas MG, Psarrakos K, Giannoglou G, Molyvda-
Athanasopoulou E, Hatziioannou K, Papanastassiou E.
Skin doses to patients undergoing coronary angiography
in a Greek hospital. Radiat Prot Dosim 2005;113:449–52.
24. Verdun FR, Capasso P, Valley JF, Schnyder P. Dose
evaluation in fluoroscopy. Radiat Prot Dosim 1998;
80:139–41.
25. Wagner LK, Eifel PJ, Geise RA. Potential biological effects
following high x-ray dose interventional procedures. J Vasc
Interv Radiol 1994;5:71–84.
26. Hwang E, Gaxiola E, Vlietstra RE, Brenner A, Ebersole D,
Browne K. Real-time measurement of skin radiation during
cardiac catheterization. Cathet Cardiovasc Diagn 1998;
43:367–70.
607
 The British Journal of Radiology, 79 (2006), 608–613

Randomized phase II study of GM-CSF to reduce mucositis caused

by accelerated radiotherapy of laryngeal cancer
1 2 2
J J MCALEESE, FRCR, K M BISHOP, BSc, RGN, R A’HERN, PhD and 2J M HENK, FRCR

1
Belvoir Park Hospital, Hospital Road, Belfast BT8 8JR and 2Head and Neck Unit, Royal Marsden
Hospital, Fulham Road, London SW3 6JJ, UK

ABSTRACT. Acute mucositis is dose-limiting in many accelerated radiotherapy schedules

for head and neck cancer. Cytokines may be one means of reducing the severity of
mucositis. A study was designed to assess the effect of subcutaneous molgramostin
(granulocyte-macrophage colony stimulating factor; GM-CSF) injections on acute
radiation morbidity in patients undergoing accelerated radiotherapy for laryngeal
cancer. A prospective, randomized, observer-blind, controlled trial was conducted in 29
patients who were to receive radical radiotherapy over 3 weeks for early stage
laryngeal cancer. Patients were randomized to receive 150 mg (,2 mg kg21) GM-CSF
subcutaneously once daily for 14 days after the second week of radiotherapy, or no
GM-CSF. Patients were assessed weekly for grade of mucositis, skin reactions and
related parameters. The severity of mucositis was reduced in the GM-CSF arm (p,0.05).
Received 8 March 2005
No other end-points reached statistical significance. Two patients failed to complete Revised 9 September 2005
their courses of GM-CSF. Three developed influenza type symptoms and in one an Accepted 16 September
allergic reaction was noted. There was no difference in tumour control rates. 2005
Subcutaneous GM-CSF reduced the severity of mucositis in patients undergoing
DOI: 10.1259/bjr/55190439
accelerated radiotherapy. Injections were well tolerated. Further studies of cytokines
are warranted, to assess the feasibility of increasing the total doses of accelerated ’ 2006 The British Institute of
radiotherapy given, with the aim of improving tumour cure rates. Radiology

When head and neck cancers are treated with
accelerated radiotherapy, acute normal tissue effects
can be dose limiting, especially when concurrent
chemotherapy is given. This is because severe mucositis
can lead to consequential late damage [1]. Reduction of
the severity of mucositis could permit dose escalation
with an increased chance of tumour cure.
The pathogenesis of mucositis is death of the actively
dividing cells of the mucosal epithelium. The first clinical
signs usually appear on the 12th or 13th day of
radiotherapy. Healing takes on average 3 weeks after
conventionally fractionated radiotherapy, but is pro-
longed in cases of severe mucositis. Various methods
have been tried to reduce mucosal damage. Antiseptic
and anti-inflammatory mouthwashes have no effect in
reducing the severity of mucositis [2, 3]. Overgrowth of
yeasts and aerobic gram negative cocci has been
demonstrated, but studies of the use of local anti-
fungals and antibiotics have given conflicting results;
some suggest a small benefit, while others are negative
[4–6].
Another approach is the use of growth factors to
stimulate repair and proliferation of surviving mucosal
cells. Granulocyte-macrophage colony stimulating factor
(GM-CSF) influences the proliferation and differentiation
of stem cells and regulates several functions in mature
leukocytes, macrophages and dendritic cells of the
submucosa and dermis [7, 8]. GM-CSF administered to
Address correspondence to: Dr J J McAleese, 46 Onslow Gardens,
Belfast BT6 0AQ, UK.
patients suffering cuts and burns decreases the healing
period significantly [9]. GM-CSF has been successfully
used to promote healing of chemotherapy-induced
mucositis [10, 11]. It is therefore appropriate to investi-
gate the use of GM-CSF to reduce the severity and
increase the rate of healing of mucositis in patients with
head and neck cancer. Several studies carried out to date
have used GM-CSF with conventionally fractionated
radiotherapy [12, 13].
Accelerated radiotherapy is used routinely for a
variety of sites and stages of head and neck cancer, one
of which is early laryngeal cancer. A large multicentre
trial compared short (3–4 weeks) and long (5–6 weeks)
treatment times for radiotherapy of laryngeal carcinoma.
The short times gave equal local control and less late
morbidity [14]. Since then, a 3-week 16-fraction regimen
has been adopted as standard at the Royal Marsden
Hospital. The majority of patients treated with this
regimen develop symptomatic mucositis; grade 2 is
observed in most, and grade 3 in about 10%. This group
of patients was therefore deemed to be especially
suitable for a study of the effect of GM-CSF with
accelerated radiotherapy.
A possible disadvantage of growth factors is risk of
stimulation of cell division in tumour cells, leading to
accelerated repopulation during radiotherapy and a
reduced chance of tumour cure. In one study another
cytokine, G-CSF, was associated with reduced tumour
control [15] when given from days 15 to 19 of a 38-day
course of radiotherapy, i.e. during the first half of the
course. We decided to try the effect of GM-CSF daily for

608 The British Journal of Radiology, July 2006

GM-CSF to reduce mucositis

2 weeks starting on day 14 of a 21-day course of
radiotherapy. Accordingly, the drug was not given until
two-thirds of the radiation dose had been accumulated,
so that the risk of accelerated repopulation should be
minimized.
The optimal dose of GM-CSF is not known. The dose
most often used to treat oral mucositis is 4 mg kg21, but
other studies suggest 1 mg kg21 may be effective [13, 16].
Commercially available vials of GM-CSF contain 150 mg,
sufficient for a dose of 2 mg kg21 for most patients, so
this latter dose was chosen for our study.
Materials and methods
Study design
The study was a prospective, randomized, observer-
blind phase II trial. Approval was granted by the Royal
Marsden Ethics Committee. Patients treated by radio-
therapy for early glottic carcinoma were randomly
allocated to receive or not daily injections of GM-CSF
for 2 weeks, beginning on day 15 of the course of
radiotherapy. A placebo injection was not used because
it was not considered ethically justifiable. Accordingly,
assessments of radiation effects were made and recorded
by a blinded observer.
Eligibility criteria
Patients were eligible for the study if they had
histologically proven T1 N0 or T2 N0 glottic carcinoma
and were to be treated with radiotherapy using a 16-
fraction 3-week regimen. They were required to be of
WHO (World Health Organization) performance status
grade 0 or 1. Patients were excluded if they had evidence
of renal or hepatic impairment, serious infections
requiring antibiotic therapy, were taking or likely to
need corticosteroids, or were known to be allergic to
GM-CSF.
Radiotherapy
External beam radiotherapy was given with a linear
accelerator using 6 MV photons. The technique was to
use lateral parallel-opposed fields, except in patients
with a short neck, in whom anterior oblique fields with
wedge filters were used. In most cases the fields were
5 cm square centred on the vocal cord, but in T2 cases the
fields were extended by up to 2 cm either above or below
the vocal cord to cover supraglottic or subglottic
extension, respectively. All doses were prescribed at
the ICRU intersection point. Once-daily fractions of
3.125 Gy were delivered to a total dose of 50 Gy in 16
fractions in 21 days.
GM-CSF
GM-CSF was administered at a dose of 150 mg by
subcutaneous injection once daily for 14 days, beginning
at the end of the second week of radiotherapy. It was
therefore given during the final week of radiotherapy,
and for 1 week after the end of radiotherapy.
Patients
All patients meeting the eligibility criteria treated at
the Royal Marsden Hospital between September 1997
and October 2000 were offered entry to the study. 29
agreed to enter, and were randomly assigned to the
active or control arms. The characteristics of the two
groups are compared in Table 1. There was an imbalance
in stage distribution, with more T2 patients in the GM-
CSF arm. Consequently more of this group were treated
with larger fields. The mean field sizes were 32.7 cm2 in
the GM-CSF group and 30.6 cm2 in the control group.
The two groups were well balanced for all other
variables.
Assessment and follow-up
Patients were seen weekly during treatment and for
3 weeks after treatment or until acute reactions settled.
At each visit one of two independent observers, blinded
to group allocation, scored mucositis by the Radiation
Therapy Oncology Group (RTOG) system reproduced in
Table 2 [17] using indirect laryngoscopy or flexible
nasendoscopy to provide adequate visualization; data on
skin erythema, moist and dry desquamation (as % of
treatment field), pain on swallowing, severity of dyspha-
gia, analgesic usage, evidence of candida infection and
laryngeal oedema were collected at each visit. Weight
was measured before starting and 2 weeks after the end
of radiotherapy.
Statistical methods
It was anticipated that 60% of patients receiving the
3 week course of radiotherapy would develop one of the
more severe grades of mucositis. We decided that in
order for the routine use of GM-CSF to be worthwhile,
this figure needed to be reduced to 10%. In order to
demonstrate this difference at 90% power and 5%
significance, 17 patients in each group would be needed.
Accordingly, the intention was to recruit 34 patients to
Table 1. Pre-treatment characteristics of the study
population
Variable GM-CSF arm Control arm
Gender
Male 14 12
Female 1 2
Age range (years) 48–79 32–70
Median age (years) 60 65
Smoking
Current smoker 9 9
Ex-smoker 5 4
Unknown 1 1
Stage I 6 11
Stage II 9 3
Total 15 14
GM-CSF, granulocyte-macrophage colony stimulating factor.

The British Journal of Radiology, July 2006 609


the study. Data were analysed on an intention to treat
basis. A ranking method (Mann Whitney U-test) was
chosen to analyse mucositis scores, as the grading system
employed uses an ordinal scale. Patients were ranked on
the basis of the frequency of maximal mucositis score
over the study period. The other data were also analysed
by non-parametric methods. Local control, disease-free
survival and overall survival rates were computed by the
Kaplan-Meier method and analysed by the log-rank
method.
Results
Recruitment to the study proved to be disappointingly
slow. There was a high refusal rate, because many
patients were unwilling to receive a course of injections
for a possible benefit only in terms of the side-effects of
treatment. Eventually the study was terminated when 29
patients had been randomized.
15 patients were allocated to the GM-CSF treatment
arm. Two patients failed to complete their prescribed
course of injections. One patient discontinued after his
second injection because of influenza-like symptoms.
The second patient was admitted to a local hospital
because of a chest infection and had his GM-CSF
discontinued. All 29 patients completed their radio-
therapy course as planned.
Figures 1 and 2 show the proportions of patients in
each group with each grade of mucositis at each week
after starting radiotherapy. The overall areas of the
graphs are similar, but the arms differ in the relative
contributions of each grade. The number of patients in
the GM-CSF arm with maximum mucositis scores of
grade 0, I, II and III were 1, 4, 10 and 0, respectively. The
corresponding numbers for the control arm were 1, 0, 12
and 1, respectively. There were a greater proportion of
maximum grade I scores in the GM-CSF arm compared
with the control arm (27% vs 0%) but decreased grade II
Figure 1. Proportion of patients with mucositis grades for
control arm.
610
J J McAleese, K M Bishop, R A’Hern and J M Henk
Figure 2. Proportion of patients with mucositis grades for
granulocyte-macrophage colony stimulating factor (GM-CSF)
arm.
scores (67% vs 86%), and the only patient with a grade III
score was in the control arm. The difference between the
mucositis grades of the two groups was significant
(ranking by Mann Whitney U556.5, p,0.05) The mean
time to healing of mucositis in the GM-CSF group was
approximately 60% that of the control arm, but this did
not reach significance on the log rank test (p50.25).
One patient in the control arm required tube feeding.
Otherwise, no patients had dysphagia at grade 3 (fluids
only) or greater at 5 weeks. No differences were detected
in dysphagia, odynophagia, analgesic usage, candida
infection or laryngeal oedema. There was a difference
noted in the average erythema grade in favour of the
GM-CSF group (0.83 vs 1.33, p,0.05). No difference in
moist desquamation or dry desquamation was noted. No
statistically significant difference in weight change was
noted between the groups.
Side effects of GM-CSF
12 patients who received GM-CSF had elevated white
cell counts (WCC). The range of maximal WCC was
7.2–30.5 (median 19.7). All WCC had returned to normal
within 3 weeks of completing injections (median
2 weeks). Three patients developed influenza-like symp-
toms with the GM-CSF and in one patient the injections
were stopped because of this symptom. One patient
developed an erythematous rash at his injection sites
after completing his course of 14 injections (Figure 3). He
had a past history of allergy to radiographic contrast
medium.
Tumour outcomes
The median follow-up is 28 months (range 6–45
months) during which time there have been three
relapses in the control arm. One occurred at the primary
site at 36 months and was successfully salvaged with
The British Journal of Radiology, July 2006
GM-CSF to reduce mucositis
Figure 3. Allergic reaction at granulocyte-macrophage
colony stimulating factor (GM-CSF) injection sites.
laryngectomy. The other two patients had cervical node
metastases, one at 10 months post-radiotherapy, the
other at 13 months, and both underwent successful neck
dissection. One patient in the GM-CSF arm developed a
local recurrence and required a salvage laryngectomy at
12 months post-treatment. One patient in the GM-CSF
group died of a myocardial infarction 5 months after his
radiotherapy. The study is too small to draw conclusions
about relative relapse rates, but there is no suggestion
that GM-CSF adversely affected outcome.
Three patients in the control arm were diagnosed with
a second malignancy. Two developed non-small cell
lung cancer, one of which died within 3 months; the
other received a course of radical radiotherapy. The third
patient was diagnosed with small cell lung cancer and
has received second line chemotherapy. One patient in
the GM-CSF group developed a carcinoma of the rectum,
which was treated with surgery alone.
Discussion
A meta-analysis of prophylactic GM-CSF in che-
motherapy and radiotherapy induced mucositis [18]
has shown a benefit with a relative risk of 0.51 (95%
confidence intervals 0.29–0.91). Several studies have
previously examined the ability of GM-CSF to modify
radiation-induced mucositis (Table 3) but using differing
radiotherapy regimens, doses of growth factor, and
scoring systems for mucosal reactions. Five prospective
randomized controlled trials of GM-CSF have been
reported; in two, the agent was administered subcuta-
neously, and in the other three as a mouthwash.
Throuvalas et al [13] reported a randomized trial in 10
patients receiving conventionally fractionated radiother-
apy. GM-CSF was administered subcutaneously at a
dose of 1 mg kg21 daily, beginning at the start of the
third week of radiotherapy. The GM-CSF arm showed
better pain control (p50.004) and less erythema (p50.09).
Makonnen et al [12] conducted a randomized controlled
trial comparing subcutaneous GM-CSF 150–300 mg per
day plus sucralfate mouthwash versus sucralfate
mouthwash control, in 40 patients receiving either
The British Journal of Radiology, July 2006
Table 2. RTOG mucositis scores
RTOG Grade Descriptor
Grade 0 No change over baseline
Grade 1 Injection / may experience mild pain not
requiring analgesia
Grade 2 Patchy mucositis that may produce an inflam-
matory serosanguinous discharge / may
experience moderate pain requiring analgesia
Grade 3 Confluent mucositis / may include severe pain
requiring narcotic
Grade 4 Ulceration, haemorrhage or necrosis
conventionally fractionated or hyperfractionated radio-
therapy with break. GM-CSF was begun at the end of the
first week of radiotherapy and continued throughout the
course of treatment. No difference was found in degree or
frequency of mucositis in the overall group or by type of
fractionation. Skin reactions, fever and bone pain were the
most common side effects in the GM-CSF arm.
Sprinzl et al [19] conducted a randomized control trial
in 35 patients using GM-CSF mouthwash. All patients had
advanced head and neck cancer and were treated with
two cycles of radiotherapy (each 30 Gy in 15 fractions)
and concomitant mitomycin C and 5FU chemotherapy.
Data were collected during the first radiotherapy cycle.
No differences in mucositis or oral pain were noted.
Saarilathi et al [20] compared GM-CSF with sucralfate
mouthwash in a randomized study in 40 patients with
oral cancer receiving post-operative radiotherapy.
Mucositis tended to be less severe and heal more quickly
in the GM-CSF treated group (p50.07) and there were
fewer requirements for opioid pain relief (p,0.05).
Two case-control studies of subcutaneous GM-CSF
have been reported. Kannan et al [16] compared the
results of 10 patients treated with 1 mg kg21 body weight
GM-CSF and conventionally fractionated radiotherapy
with historic controls. None of the 10 patients developed
grade III mucositis compared with 49% of controls.
Itching in 20% and body aches in 40% were the main side
effects, but both were successfully treated with simple
oral medications. Wagner et al [21] compared the use of
subcutaneous GM-CSF 5 mg kg21 body weight in 16
patients receiving conventionally fractionated radiother-
apy with historical controls. Pain relief was improved in
the GM-CSF group (p50.011). A trend towards a
decreased degree of mucositis was noted. Only mild
reactions to GM-CSF were reported.
All of the above studies were too small to provide a
conclusive result, but overall there is a trend towards a
reduction in severity of radiation mucositis with GM-
CSF. Our study also showed a reduction in the severity
of mucositis experienced by patients receiving acceler-
ated radiotherapy. This was despite an imbalance in the
two arms so that more patients in the GM-CSF arm were
treated with larger fields.
Side effects to GM-CSF seem to be generally mild.
Kannan et al [16] reported itching in 20% and body aches
in 40% of patients receiving the drug. Wagner et al [21]
observed bone pain in 3% and pyrexia in 3%. Makonnen
et al [12] reported local skin reaction in 13%, fever in 6%
and bone pain in 5%. In our study, 20% had myalgia/flu-
like symptoms and one patient had an allergic reaction.
611

Table 3. Published studies of GM-CSF in radiation mucositis
Author Year Number Random
Throuvalis [13] 1995 10 Yes
Kannan [16] 1997 10 No
Wagner [21] 1999 32 No
Makonnen [12] 2000 40 Yes
Sprinzl [19] 2001 35 Yes
Saarilhati [20] 2002 40 Yes
GM-CSF, granulocyte-macrophage colony stimulating factor.
Subcutaneous GM-CSF therefore appears to be well
tolerated.
Maurer-Schulze et al [22] suggest that low doses of
GM-CSF stimulate cell proliferation in human tumour
xenografts in nude mice, but inhibit proliferation at
higher doses. The data in Makonnen’s paper [12] show a
non-significant trend for decreased survival in the
GM-CSF arm. Because of the possible potential for
accelerating tumour cell re-population [23, 24], GM-CSF
was not administered until the final week of radio-
therapy in our study. We did not observe a difference in
local control, but larger studies would be needed to
confirm this finding.
Conclusions
GM-CSF was well tolerated and no decrease in local
control was seen. A decrease in severity of mucositis,
consistent with the biological actions of GM-CSF as a
promoter of wound healing, has been shown. GM-CSF
appears promising as a potential modifier of radio-
therapy-induced mucositis, but larger studies are
required to establish its value. More specific growth
factors are now under investigation: for example, Dörr
and colleagues have demonstrated a radio-protective
effect of keratinocyte growth factor on mouse oral
mucosa [25]. Cytokines are a promising method of
improving therapeutic ratio in radiotherapy of head
and neck cancer, especially with accelerated regimes in
which acute normal tissue effects are dose-limiting.
References
1. Dore W, Hendry JH. Consequential late effects in normal
tissues. Radiother Oncol 2001;61:223–31.
2. Foote RL, Loprinzi CL, Frank AR, O’Fallon JR, Gulavita S,
Tewfik HH, et al. Randomised trial of a chlorhexidine
mouthwash for alleviation of radiation induced mucositis. J
Clin Oncol 1994;12:2630–3.
3. Samaranyake LP, Robertson AG, MacFarlane TW, Hunter
IP, MacFarlane G, Soutar DS, et al. The effect of chlorhex-
idine and benzdamine mouthwashes on mucositis induced
by therapeutic irradiation. Clin Radiol 1988;39:291–4.
4. El-Sayed S, Nabid A, Shelley W, Hay J, Balogh J, Gelinas M,
et al. Prophylaxis of radiation-associated mucositis in
conventionally treated patients with head and neck cancer:
a double-blind, phase III, randomised, controlled trial
evaluating the clinical efficacy of an antimicrobial lozenge
using a validated mucositis scoring system. J Clin Oncol
2002;20:3956–63.
5. Symonds RP, McIlroy P, Khorrami J, Paul J, Pyper E, Alcock
SR, et al. The reduction of radiation mucositis by selective
decontamination antibiotic pastilles: a placebo controlled
double-blind trial. Br J Cancer 1996;73:312–7.
612
J J McAleese, K M Bishop, R A’Hern and J M Henk
Difference Dose
Pain p50.004 1 mg kg21
Mucositis 1 mg kg21
Pain p50.0011 5 mg kg21
No 150–300 mg
No Mouthwash
Opiod use p50.042 Mouthwash
6. Wijers OB, Levenda PC, Harms ER, et al. Mucositis
reduction by selective elimination of oral flora in irradiated
cancers of the head and neck: a placebo-controlled double-
blind randomised study. Int J Radiat Oncol Biol Phys
2001;50:341–52.
7. Berdel WE, Danhauser-Riedl S, Steinhauser C, Winton EF.
Various human haemopoietic growth factors (IL-3, GM-CSF,
G-CSF) stimulate clonal growth of hamatopoietic tumour
cells. Blood 1990;73:80–3.
8. Dexter M. Haemopoietic growth factors. Review of biology
and clinical potential. Macclesfield, Cheshire: Gardiner-
Caldwell Ltd., 1990.
9. Masucci G. New clinical applications of granulocyte-
macrophage colony-stimulating factor. Med Oncol 1996;13:
149–54.
10. Chi KH, Chen CH, Chan WK, Chow KC, Chen SY, Yen SH,
et al. Effect of granulocyte-macrophage colony-stimulating
factor on mucositis in head and neck cancer patients after
cisplatin, fluorouracil and leucovorin chemotherapy. J Clin
Oncol 1995;13:2620–8.
11. Hejna M, Kostler WJ, Raderer M, Steger GG, Brodowicz T,
Scheithauer W, et al. Decrease of duration and symptoms in
chemotherapy-induced oral mucositis by topical GM-CSF:
results of a prospective randomised trial. Eur J Cancer
2001;37:1994–2002.
12. Makkonen TA, Minn H, Jekunen A, Vilja P, Tuominen J,
Joensuu H, et al. Granulocyte macrophage-colony stimulat-
ing factor (GM-CSF) and sucralfate in prevention of
radiation-induced mucositis: a prospective randomised
study. Int J Radiat Oncol Biol Phys 2000;46:525–34.
13. Throuvalis N, Antonadou D, Pulizzi M, Sarris G. Evaluation
of the efficacy and safety of GM-CSF in the prophylaxis
of mucositis in patients with head and neck cancer
treated with radiotherapy. Eur J Cancer 1995;Suppl. 5:
Abstract 431.
14. Wiernik G, Alcock CJ, Bates TD, Brindle JM, Fowler JF,
Gajek WR, et al. Final report of the second British Institute
of Radiology fractionation study: short vs long overall
treatment times for radiotherapy of carcinoma of the
laryngopharnx. Br J Radiol 1991;64:232–41.
15. Staar S, Rudat V, Stuetzer H, Dietz A, Volling P, Schroeder
M, et al. Intensified hyperfractionation accelerated radio-
therapy limits the additional benefit of simultaneous
chemotherapy – results of a multicentric randomised
German trial in advanced head and neck cancer. Int J
Radiat Oncol Biol Phys 2001;50:1161–71.
16. Kannan V, Bapsy PP, Anantha N, Doval DC, Vaithianathan
H, Banumarthy G, et al. Efficacy and safety of granulocyte
macrophage-colony stimulating factor (GM-CSF) on the
frequency and severity of radiation mucositis in patients
with head and neck carcinoma. Int J Radiat Oncol Biol Phys
1997;37:1005–10.
17. Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation
Therapy Oncology Group (RTOG) and the European
Organization for Research and Treatment of Cancer
(EORTC). Int J Radiat Oncol Biol Phys 1995;31:1341–8.
18. Clarkson JE, Worthington HV, Eden OB. Interventions for
preventing oral mucositis for patients with cancer receiving
The British Journal of Radiology, July 2006
GM-CSF to reduce mucositis
treatment (Cochrane Review). In: The Cochrane Library,
Issue 2, 2004.
19. Sprinzl GM, Galvan O, de Vries A, Ulmer H, Gunkel AR,
Lukas P, et al. Local application of granulocyte-macrophage
colony stimulating factor (GM-CSF) for the treatment of
oral mucositis. Eur J Cancer 2001;37:2003–9.
20. Saarilahti K, Kajanti M, Joensuu T, Kouri M, Joensuu H.
Comparison of granulocyte-macrophage colony-stimulat-
ing factor and sucralfate mouthwashes in the prevention of
radiation-induced mucositis: a double-blind prospective
randomised phase III study. Int J Radiat Oncol Biol Phys
2002;54:479–85.
21. Wagner W, Alfrink M, Haus U, Matt J. Treatment of
irradiation-induced mucositis with growth factors (rhGM-
CSF) in patients with head and neck cancer. Anticancer Res
1999;19:799–804.
The British Journal of Radiology, July 2006
22. Maurer-Schultze B, Bassukas ID, Hofmockel G. Effect of
biological response modifiers on growth and cell prolifera-
tion of human tumour xenografts in nude mice. Cell Mol
Biol 1995;41:65–78.
23. Foulke R, Marshal M, Trotta P, Van Hoff D. In vitro
assessment of the effects of granulocyte-macrophage
colony-stimulating factor on primary human tumours and
derived lines. Cancer Res 1990;50:6264–7.
24. De Riese W, Allhoff EP, Werner M, Strief CG, Liedke S,
Kircher H, et al. Effects of cytokines on clonogenic growth
in vitro of primary human renal cell carcinoma. Urological
Res 1992;20:369–73.
25. Dörr W, Spekl K, Farrell CL. Amelioration of acute oral
mucositis by keratinocyte growth factor: fractionated
irradiation. Int J Radiat Oncol Biol Phys 2002;54:
245–51.
613
 The British Journal of Radiology, 79 (2006), 614–626

PICTORIAL REVIEW

Magnetic resonance urography: a pictorial overview

1 2 1 1
R GARCÍA-VALTUILLE, MD, A I GARCÍA-VALTUILLE, MD, F ABASCAL, MD, L CEREZAL, MD and
3
M C ARGÜELLO, MD

1
Instituto Radiológico Cántabro, Clı́nica Mompı́a, Avenida de los Condes, s/n. 39108 Santa Cruz de
Bezana (Cantabria), 2Department of Pathology, Clı́nica Mompı́a, Santa Cruz de Bezana (Cantabria)
and 3Department of Oncology, Clı́nica Mompı́a, Santa Cruz de Bezana (Cantabria), Spain

ABSTRACT. Magnetic resonance urography (MRU) can be performed on the basis of

two different imaging strategies: static-fluid MRU, based on heavily T2 weighted turbo Received 6 February 2005
spin echo (TSE) sequences, and gadolinium-enhanced excretory MRU. Both MR Revised 10 May 2005
Accepted 23 May 2005
urographic techniques in combination with standard MRI permit a comprehensive
examination of the entire urinary tract. This pictorial review illustrates the MRU DOI: 10.1259/bjr/21075982
features of the a wide spectrum of pathological conditions affecting the urinary tract.
’ 2006 The British Institute of
Radiology

Introduction Before the acquisition of the excretory MR urographic

Magnetic resonance urography (MRU) is an emerging
technique of MRI which provides a non-invasive
visualization of urinary tract. Most of previous studies
have used the unenhanced, heavily T2 weighted pulse
sequences to obtain images in which static fluid exhibits
a higher signal intensity relative to background (static
MRU) [1–4]. Clinical urography requires both morpho-
logical and functional information about the kidneys and
the collecting system. However, these sequences do not
provide information about the renal excretory function.
MRU performed with contrast material can meet all
demands of clinical urography and, in some cases, could
replace conventional X-ray urography [1, 4, 5].
This pictorial essay reviews the MRU features of the
major urinary tract disorders in which static or excretory
MRU provides information of diagnostic value.
Technique
The images were acquired by a 1 T superconducting
magnet (New technology [NT] Gyroscan; Philips
Medical Systems, Best, The Netherlands) using a body
coil. The MR sequence protocol started with localizing T1
weighted gradient-echo sequence (repetition time (TR)
18 ms/echo time (TE) 6.9 ms; flip angle 30 ˚; scan time
3 min 30 s) and T2 weighted turbo spin echo (TSE)
sequence (TR 4200 ms/TE 100 ms; flip angle 90 ˚; scan
time 3 min 12 s) in axial and coronal planes.
In static MRU, heavily T2 weighted TSE pulse
sequences are used to obtain water images of the urinary
tract (three-dimensional; respiratory-triggering; TR
2000 ms/TE 700 ms; flip angle 90 ˚; TSE-factor 101;
matrix size 2566256; number of excitations 2; field of
view (FOV) 360–390 mm; number of slices 40–50; slice
thickness 2 mm; scan time 3 min 30 s to 4 min 20 s).
sequences, the patients received an intravenous dose of
0.1 mg kg21 of furosemide and 0.1 mmol kg21 of
GdDTPA-BMA (Gadodiamide). A delay of 1–5 min
between the administration of both drugs is necessary
for achieving optimal contrast enhancement of the
urinary tract. Excretory MRU was performed at our
institution using a respiratory gating, three-dimensional,
T1 weighted gradient-echo sequence (TR 15 ms/TE 5 ms;
flip angle 70 ˚; matrix size 2566256; number of excita-
tions 2; FOV 360–390 mm; scan time 3 min) with an
anteriorly located pre-saturation slab. 60 sections,
2.2 mm thick, were obtained in coronal plane 5 min,
10 min and 20 min after diuretic and contrast material
injection. In selected cases, additional transverse planes
were performed to optimize visualization of anatomic
structures.
For the examination of children, we reduce the FOV of
the sequences and adjust furosemide and gadolinium
dosages (0.05 mg kg21 of furosemide and 0.05 mmol kg21
of gadolinium).
The source images of static and excretory MRU were
then post-processed by the use of a maximum intensity
projection (MIP) algorithm.
When no dilatation of the urinary tract is visible on
the initial T2 weighted images we use excretory
MRU. With the use of a diuretic in MRU within the
dose range of 4–10 mg of furosemide, the induced
distention of the urinary tract was mild and did not
result in false-positive diagnosis of substantial
dilatation. In patients with mild dilatation, both techni-
ques (static and excretory MRU) are employed. In
cases of marked dilatation of the urinary tract and
impaired excretory function, static MRU is used. Static
MRU is also used for the visualization of urinary tract
disorders in women during pregnancy (Figures 1 and 2)
[5–7].

614 The British Journal of Radiology, July 2006

Pictorial review: MR urography

(a) (b)

Figure 1. A 52-year-old woman with an extrinsic ureteral obstruction caused by a metastasis of an ovarian carcinoma.
(a) Maximum intensity projection (MIP) image from an unenhanced T2 weighted MR urograph (MRU) shows a left ureteral
obstruction (arrow). Note the changes of chronic hydronephrosis and hydroureter. (b) The axial standard T2 weighted turbo spin
echo (TSE) image visualizes a soft tissue mass with heterogeneous signal intensity surrounding the ureter (arrowheads).

Normal variants and congenital anomalies
The main indications in children of MRU are con-
genital anomalies of the kidneys and collecting system
[6]. Normal variants and congenital anomalies of the
collecting system can be accurately identified with this
technique [5]. Knowledge of the myriad appearances of
congenital renal and collecting system anomalies and
minor anatomic variants is essential for the correct
interpretation of urograms.
Congenital ureteropelvic junction (UPJ) obstruction is
sharply defined UPJ narrowing with dilatation of the
pelvocalyceal system, which persists even when patient
is placed in a position favouring gravity drainage of the
pelvis (Figure 3) [1]. Large extrarenal pelves may
simulate hydronephrosis when they are stressed by
diuresis.
MRU can also accurately detect complete and incom-
plete ureteral duplication by locating the level of fusion.
In cases of complete duplication, the insertion of the
superior collecting system is usually ectopic [8].
Common congenital anomalies of the fusion variety
have characteristic MR appearances. True congenital
hypoplasia is distinctly rare or very difficult to docu-
ment. Hypoplastic kidneys usually are caused by
trauma, infection or ischaemic or obstructive insult
during the growth phase. Renal agenesis with contra-
lateral solitary kidney usually associates with Müllerian
duct abnormalities (Figure 4) [1].
Filling defects in the ureter or in the
pelvocalyceal system
Filling defects are demonstrated on MRU as signal-
void areas outlined by the hyperintense surrounding
urine, except when they are impacted or filling the entire
lumen of ureter. We sometimes perform complementary

The British Journal of Radiology, July 2006 615

 R Garcı́a-Valtuille, A I Garcı́a-Valtuille, F Abascal et al

(a) (b)

Figure 2. Staghorn calculus and chronic hydronephrosis in a 32-year-old pregnant patient. (a) Coronal T2 weighted turbo spin
echo (TSE) image and (b) urogram from static MR urography show diffuse cortical atrophy, pyelocaliectasis and a voluminous
pyelocaliceal filling defect (arrows). Note also the gestational sac (arrowheads) and a left corpus luteum cyst (white arrow).

616 The British Journal of Radiology, July 2006

Pictorial review: MR urography
Figure 3. A 20-year-old man with ureteropelvic junction
narrowing (arrow). Coronal maximum intensity projection
(MIP) excretory MR urography. The renal pelvis shows typical
dilatation and convex inferior border.
axial images because the small filling defects are better
visualized in this plane. Instead of MIP images, the
source images must always be reviewed because small
defects may be obscured by the surrounding urine on
MIP projections [5].
The acute stone colic should not be a primary
indication for MRU. However, it is important to be
aware of the findings of stones in MRU because most
common filling defects are the calculi (Figures 2 and 5);
round or oval filling defects that tend to become
impacted in areas of normal anatomic narrowing –
ureteropelvic and ureterovesical junctions, and the site
where the ureter crosses the sacrum and the iliac vessels
– and cause a variable degree of dilatation of the urinary
tract [1].
Blood clots are single or multiple filling defects of
various sizes and shapes that may cause temporary
ureteral obstruction (Figure 6). They are usually hyper-
intense on T1 weighted MR images, do not enhance with
gadolinium and become much smaller or disappear
within several weeks [8].
Transitional cell carcinoma appears as smooth or
irregular shaggy filling defects (Figure 7). The signal
intensity of transitional cell carcinoma usually differs
sufficiently from that of other causes of ureteral filling
defects, on conventional T1 and T2 weighted images, to
suggest the diagnosis. There is often localized dilatation
of the ureter below the level of the expanding intralum-
inal tumour, in contrast to ureteral collapse distal to an
The British Journal of Radiology, July 2006
obstructing stone. The ‘‘globet sign’’ and the ‘‘sipple
sign’’ are also useful in differentiation with other entities.
However, the morphological differentiation between a
small calculus and a small early intrinsic tumour is
difficult in some cases, especially if the clinical symp-
toms are non-specific [1, 5, 8].
Mimickers of filling defects are flow artefacts (usually
with central location within the ureter) [9], vessels that
can cause an extrinsic impression on the ureter
(Figure 8), and ureteral spasm and peristalsis.
Obstruction of the ureter
The differential diagnosis of ureteral obstruction
include intraluminal (calculi, blood clots, papillary
necrosis with sloughed papilla), intramural (tumour,
infection diseases, post-surgery/instrumentation
trauma, lesions after radiotherapy, ureterocele, mega-
ureter) and extrinsic abnormalities (retroperitoneal
fibrosis, invasion or compression by extrinsic malig-
nancy, lymphadenopathy, inflammatory diseases) [1, 3,
8].
MRU allows the precise depiction of the site of the
obstruction and the degree of ureterectasis, and may
demonstrate the underlying pathology with the help of
conventional T1 and T2 weighted sequences (Figures 1, 6
and 9).
Filling defects in the urinary bladder
MRU provides a non-invasive mean to detect filling
defects in the urinary bladder – calculus, blood clot, air
bubble, neoplasm, prostatic enlargement, ureterocele or
foreign body [5].
The transitional cell carcinoma of the urinary bladder
is a single or multiple polypoid defect that arises from
the bladder wall and is fixed in position – unlike a
calculus, blood clot or air. Sometimes they may produce
only focal bladder wall thickening and rigidity
(Figure 10).
Prostatic enlargement causes a smooth or irregular
extrinsic filling defect of varying size at the base of the
bladder (Figure 11). If a chronic process, there is
trabeculation of the bladder wall and diverticula forma-
tion. The distal ureters often have a fishhook deformity
due to elevation of the trigone.
Post-operative changes
The role of MRU in patients undergoing urinary
diversion (ureteroileal by-passes, ureterosigmoidostomy,
skin ureterostomy, orthotopic neobladder reconstruc-
tion) or after renal transplantation is emerging. MRU
allows visualization of anastomoses, as well as of
associated complications such as strictures (Figure 12),
ureteral compression by lymphocele or haematoma,
urine leaks, fistulae (Figure 13), stones or signs of
infection. Signal-void within urinary tract in post-
operative patients does not always correspond to stones,
but may be due to air bubbles or susceptibility artefacts
caused by surgical material [1, 10, 11].
617
 R Garcı́a-Valtuille, A I Garcı́a-Valtuille, F Abascal et al

(a) (b)

(c)

Figure 4. Left renal agenesia in a 14-year-old woman with didelphic uterus and a vaginal septum. Maximum intensity
projection (MIP) image from (a) excretory MR urography demonstrates a normal right kidney with no evidence of left renal
tissue. (b) Axial and (c) coronal T2 weighted turbo spin echo (TSE) images show a bicornuate uterus (arrows) with two cervix
(arrowheads).

618 The British Journal of Radiology, July 2006

Pictorial review: MR urography

(a) (b)

Figure 5. A 55-year-old man with left-sided ureteral stone. (a) Coronal maximum intensity projection (MIP) excretory MR
urograph shows filling defect (arrow) in left distal ureter that is causing mild pyeloureterectasis. (b) Enhanced axial T1 weighted
gradient-echo image shows a round dependent filling defect (arrow) in left ureter.

The British Journal of Radiology, July 2006 619

 R Garcı́a-Valtuille, A I Garcı́a-Valtuille, F Abascal et al

(a) (b)

(c) (d)

Figure 6. A 52-year-old woman with temporary ureteral obstruction caused by blood clots. (a) Coronal maximum intensity
projection (MIP) excretory MR urography and (b) complementary retrograde pyelography show complete proximal ureteral
obstruction (arrow) and mild dilatation of the collecting system. (c) Axial gadolinium-enhanced T1 weighted gradient-echo and
(d) T2-weighted turbo spin echo (TSE) images demonstrate hypointense tissue filling completely a mildly dilated ureter
(arrowhead). (Continued)

620 The British Journal of Radiology, July 2006

Pictorial review: MR urography

(e)

Figure 6. (Cont.) (e) After several days, excretory urogram from conventional intravenous pyelography demonstrates patency of
previously occluded ureter (arrowheads).

The British Journal of Radiology, July 2006 621

 R Garcı́a-Valtuille, A I Garcı́a-Valtuille, F Abascal et al

(a) (b)

(c)

Figure 7. Transitional cell carcinoma of the midureter in a 68-year-old man. (a) Coronal maximum intensity projection (MIP) and
(b) source images from excretory MR urography demonstrate a large mass inside the midureter (arrows) with proximal ureteral
and pelvocalyceal dilatation. (c) The axial standard T2 weighted turbo spin echo (TSE) sequence confirms the diagnosis by
demonstrating a soft-tissue mass (arrowhead) with heterogeneous signal intensity.

622 The British Journal of Radiology, July 2006

Pictorial review: MR urography
(a)
Figure 8. A 73-year-old woman with mild narrowing of the midureter (arrow) caused by left common iliac artery. (a) Coronal
maximum intensity projection (MIP) image from excretory MR urography and (b) composite coronal MIPs of both urogram an
MR angiography.
Conclusions
Static and excretory MRU are complementary methods
for morphological and functional evaluation of the urinary
system, which can be alternatively employed according to
the degree of urinary tract dilatation and renal function.
These techniques have some advantages over ultrasound,
conventional urography and CT urography in the diag-
nosis of urological diseases. The three-dimensional nature
of the data permits reformation into any plane, and thus
virtually eliminates the potential of projection related
errors in the diagnosis of different pathological conditions.
There are also the safety advantages of eliminating
ionizing radiation and the risk of medical complications
due to iodinated contrast agents, and is even suitable for
assessing transplanted kidneys because of the low
nephrotoxicity of gadolinium.
MRU, due to its non-use of ionizing radiation, is the
most important tool in the diagnostic work-up of
The British Journal of Radiology, July 2006
(b)
genitourinary pathologies in infants, small children and
in women during pregnancy.
The major drawback of MRU is its low sensitivity in
detecting calcifications and subtle urothelial lesions, the
latter due to the reduced spatial resolution compared
with conventional excretory urography. However, MRU
can be offered as an alternative to conventional urogra-
phy and CT urography to avoid repetitive radiation
exposure in patients with chronic urolithiasis.
In conclusion, static and contrast-enhanced excretory
MRU provide high-quality imaging of the urinary tract
and are an accurate and safe diagnostic alternative to
other urological diagnostic procedures. These techni-
ques, combined with conventional MR images, func-
tional MR sequences or MR angiography, in a single
session yields a rapid and complete diagnostic evalua-
tion of the entire urinary tract, and have the potential to
provide the same information as can be obtained with
multiple separate diagnostic studies.
623
 R Garcı́a-Valtuille, A I Garcı́a-Valtuille, F Abascal et al

(a) (b)

(c) (d)

Figure 9. A 77-year-old man with transitional cell carcinoma of the right ureter. (a) Maximum intensity projection (MIP) image
from excretory MR urography (MRU) demonstrates right ureteral obstruction (arrow), hydronephrosis and hydroureter.
(b) Original source image from excretory MRU shows a large hypointense filling defect inside distal ureter (arrows). (c) Axial T1
weighted image shows a hypointense soft-tissue mass (arrowhead) in the pelvis. (d) An area of subtle enhancement (arrowhead)
is demonstrated on the axial section of a contrast-enhanced T1 weighted sequence.

624 The British Journal of Radiology, July 2006

Pictorial review: MR urography

(a) (b)

Figure 10. A 61-year-old man with transitional cell carcinoma of the bladder. (a) Axial T2 weighted turbo spin echo (TSE) image
shows an irregular wall thickening at the left-side of the bladder (arrows). (b) Maximum intensity projection (MIP) image from
excretory MR urography confirms large irregular filling defect (arrows) on the floor and left-sided wall of the bladder. The
tumour does not produce obstruction at the ureterovesical junction.

Figure 12. A 56-year-old woman with ileal loop urinary

diversion. Maximum intensity projection (MIP) image from
Figure 11. An excretory MR urograph in a 78-year-old man excretory MR urography shows the post-operative urinary
with benign prostatic hypertrophy. Large, smooth filling tract anatomy. Both sides are dilated because of stenosis
defect at the base of the bladder (arrowheads). (arrowheads) close to the ureteroenteric implantation site.

The British Journal of Radiology, July 2006 625


Figure 13. Vesicovaginal fistula (arrowheads) formation
caused by inadvertent injury to the bladder during surgery
in a 48-year-old woman. Sagittal maximum intensity projec-
tion (MIP) excretory MR urography.
References
1. Nolte-Ernsting C, Adam G, Bücker A. MR urography:
examination techniques and clinical applications. Eur
Radiol 2001;11:355–72.
626
R Garcı́a-Valtuille, A I Garcı́a-Valtuille, F Abascal et al
2. O’Malley ME, Soto JA, Yucel EK, Hussain S. MR urography:
evaluation of a three-dimensional fast spin-echo technique
in patients with hydronephrosis. AJR Am J Roentgenol
1997;168:387–92.
3. Regan F, Bohlman ME, Khazan R, Rodriguez R, Schultze-
Haakh H. MR urography using HASTE imaging in the
assessment of ureteric obstruction. AJR Am J Roentgenol
1996;167:1115–20.
4. Rohrschneider WK, Haufe S, Wiesel M, Tonshoff B,
Wunsch R, Darge K, et al. Functional and morphologic
evaluation of congenital urinary tract dilatation by using
combined static-dynamic MR urography: findings in
kidneys with a single collecting system. Radiology
2002;224:683–94.
5. Nolte-Ernsting C, Bücker A, Adam G, Neuerburg JM, Jung
P, Hunter DW, et al. Gadolinium-enhanced excretory MR
urography after low-dose diuretic injection: comparison
with conventional excretory urography. Radiology
1998;209:147–57.
6. Nolte-Ernsting C, Staatz G, Tacke J, Günther RW. MR
urography today. Abdom Imaging 2003;28:191–209.
7. El-Diasty T, Mansour O, Farouk A. Diuretic contrast-
enhanced magnetic resonance urography versus intrave-
nous urography for depiction of nondilated urinary tracts.
Abdom Imaging 2003;28:135–45.
8. Blandino A, Gaeta M, Minutoli F, Salamone I, Magno C,
Scribano E, et al. MR urography of the ureter. AJR Am J
Roentgenol 2002;179:1307–14.
9. Girish G, Chooi WK, Morcos SK. Filling defect artefacts in
magnetic resonance urography. Eur Radiol 2004;14:145–50.
10. Schubert RA, Göckeritz S, Mentzel HJ, Rzanny R, Schubert
J, Kaiser WA. Imaging in ureteral complications of renal
transplantation: value of static fluid MR urography. Eur
Radiol 2000;10:1152–7.
11. Zielonko J, Studniarek M, Markuszewski M. MR urography
of obstructive uropathy: diagnostic value of the method in
selected clinical groups. Eur Radiol 2003;13:802–9.
The British Journal of Radiology, July 2006
The British Journal of Radiology, 79 (2006), 627–629
CASE OF THE MONTH
Focal pancreatic lesion: can a neoplasm be confidently excluded?
J HAGUE, FRCR and Z AMIN, FRCR
Department of Radiology, Middlesex Hospital, Mortimer Street, London W1T 3AA, UK
A 56-year-old female patient was referred to our
institution for further investigation of diarrhoea and
weight loss over the previous 3–4 months. Extensive
biochemical investigation had revealed no abnormality.
The patient underwent a CT scan of the pancreas
(a)
(c)
The British Journal of Radiology, July 2006
Received 18 April 2005
Revised 11 July 2005
Accepted 25 July 2005
DOI: 10.1259/bjr/19356841
’ 2006 The British Institute of
Radiology
(pre-contrast, pancreatic phase and portal venous phase),
which demonstrated a low density focus in the head of
the pancreas. Selected images from this study are
presented below (Figure 1). What is the diagnosis?
What further tests should be done?
(b)
Figure 1. (a) Unenhanced axial image
through the head and neck of the
pancreas demonstrating a region of
hypoattenuation in the pancreatic head.
(b) 5 mm reconstructed contrast
enhanced axial section through the head
of pancreas. (c) 5 mm reconstructed con-
trast enhanced axial section through the
body and tail of pancreas.
627

(a)
Figure 2. (a) In phase T1 weighted gradient-echo image through the head and neck of pancreas. (b) Opposed phase T1
weighted gradient-echo image through the head and neck of pancreas demonstrating signal drop in part of the head
corresponding to the CT abnormality.
The patient subsequently underwent MRI. The in and
opposed phase axial images are presented (Figure 2).
This demonstrates uniformly high signal in the entire
pancreas on the in phase T1, with signal drop in the
anterior head/neck, body and tail on the opposed phase
images. This is in keeping with uneven fatty infiltration
in part of the head and uncinate process, which has been
described as a normal variant.
Discussion
Fatty change in the pancreas can be diffuse, or may be
uneven, sparing certain regions (focal fatty sparing, FFS).
Alternatively, fat infiltration can be confined to one
region of the pancreas (focal fatty infiltration, FFI) [1, 2].
Diffuse mild fatty infiltration of the pancreas fre-
quently occurs in elderly and obese persons, and is of
little clinical importance [1–3].
Focal areas of FFS in the pancreas are analogous to
focal sparing in the liver [4]. FFS can mimic a mass, in a
pancreas with diffuse fatty infiltration. Areas of FFS are
usually within the head or uncinate [1, 2]. FFS has been
associated with a pancreas divisum. FFS can appear as a
hypoechoic mass on ultrasound, and appears as an
enhancing mass relative to the normal pancreas on
contrast enhanced CT [4].
FFI is associated with insulin-dependent diabetes
mellitus (IDDM), chronic pancreatitis, hepatic disease,
628
J Hague and Z Amin
(b)
dietary deficiency, viral infection and steroid therapy.
Distribution of FFI is variable, dominant in the body and
tail. FFI of the head or uncinate or both have been
described [3].
FFI can appear as a hypoechoic mass on ultrasound,
and a hypoattenuating region on non-enhanced CT and
contrast enhanced CT. A pancreatic neoplasm may be
diagnosed on this basis. There will, however, be no
dilatation of the pancreatic duct, and no contour
deformation, but the appearances may be progressive
on serial scans.
MRI using chemical shift can detect and characterize
focal fatty infiltration of the pancreas and exclude a
diagnosis of neoplasm [5]. The loss of signal intensity on
an opposed phase T1 weighted gradient echo image
compared with a corresponding in phase image estab-
lishes the lipid content of the focal abnormality and
excludes a diagnosis of a pancreatic adenocarcinoma or a
neuroendocrine tumour, both of which do not contain
fat.
Focal lesions in the pancreas detected on CT (or
ultrasound) that do not have any associated features of
malignancy (for example, no venous attenuation, or duct
dilatation) and do not deform the contour, may be
caused by uneven pancreatic lipomatosis (FFI or FFS)
and can be further evaluated with in and opposed phase
MRI [5].
The patient’s symptoms resolved without treatment
and she remained well on follow up 1 year later.
The British Journal of Radiology, July 2006
Case of the month: Focal pancreatic lesion
References
1. Marchal G, Verbeken E, Van Steenbergen W, et al. Uneven
lipomatosis; a pitfall in pancreatic sonography. Gastrointest
Radiol 1989;14:233–7.
2. Donald JJ, Shorvon PJ, Lees WR. A hypo echoic area within
the head of the pancreas; a normal variant. Clin Radiol
1990;41:337–8.
The British Journal of Radiology, July 2006
3. Matsumoto S, Mori H, Miyake H, Takaki H, Maeda T,
Yamada Y, et al. Uneven fatty replacement of the pancreas:
evaluation with CT. Radiology 1995;194:453–8.
4. Jacobs JE, Coleman BG, Arger PH, Langer JE. Pancreatic
sparing of focal fatty infiltration. Radiology 1994;190:437–9.
5. Isserow JA, Siegelman ES, Mammone J. Focal fatty infiltra-
tion of the pancreas: MR characterization with chemical shift
imaging. AJR Am J Roentgenol 1999;173:1263–5.
629
BJR
The British Journal
of Radiology

August
2006
Volume 79
Issue 944
August 2006, Volume 79, Issue 944

● Diffusion-weighted MRI: a new functional clinical technique for

tumour imaging

● In vitro evaluation of stent patency and in-stent stenoses in 10

metallic stents using MR angiography

● Occupational exposure in the electrophysiology laboratory:

quantifying and minimizing radiation burden

● Usefulness of diffusion/perfusion-weighted MRI in patients with

non-enhancing supratentorial brain gliomas: a valuable tool to
predict tumour grading?

● Imaging well-differentiated hepatocellular carcinoma with

dynamic triple-phase helical computed tomography

● Building an anonymized catalogued radiology museum in PACS:

a feasibility study

● Optimizing localization accuracy in head and neck, and brain

radiotherapy

● Evaluation of the larynx for tumour recurrence by diffusionweighted

MRI after radiotherapy: initial experience in four cases

● Non-invasive measurement of perfusion: a critical review of

arterial spin labelling techniques

● Book review

● A young female with galactorrhoea and sudden onset chest pain

The British Journal of Radiology, 79 (2006), 633–635
COMMENTARY
Diffusion-weighted MRI: a new functional clinical technique for
tumour imaging
1
D-M KOH, MRCP, FRCR and 2A R PADHANI, FRCR, FRCP
1
Department of Radiology, Royal Marsden Hospital, Downs Road, Sutton, Surrey SM2 9PT and
2
Mount Vernon Hospital, The Paul Strickland Scanner Centre, Rickmansworth Road, Northwood,
Middlesex HA6 2RN, UK
One of the key aims of oncological imaging is to
differentiate between malignant and non-malignant
tissues at all stages of the patient’s cancer care.
Accurate staging and precise delineation of the extent
of malignancy influences therapeutic decisions, therapy
outcomes and, ultimately, patient prognosis.
Conventional imaging using ultrasound, CT or MRI
detects cancer by identifying anatomical distortion or
altered tissue appearances. Tumour tissue conspicuity
may be increased after the administration of intravenous
contrast medium, thus enhancing detection and delinea-
tion. However, identification of small volume active
tumour, either at presentation or at early disease relapse
remains challenging because small volume disease may
not result in detectable structural or morphological
change on conventional imaging. Furthermore, the
effects of therapy and complications thereof may obscure
or mimic recurrent disease.
Functional imaging techniques using CT, MRI and
positron emission tomography (PET) are increasingly
being applied to the evaluation of tumours. These
techniques exploit as their contrast mechanism unique
pathophysiological changes that occur within tumours;
such as altered blood flow, increased glucose meta-
bolism, hypoxia and cellularity. Such functional tech-
niques are increasingly used for tumour detection, for
the monitoring of treatment response and to detect
relapsed disease. Clinical experience has shown that
functional techniques have their own unique strengths
and limitations.
A new, emerging functional technique that is now
finding a role in cancer imaging is diffusion-weighted
MRI (DWI or DW-MRI), which produces information
about tissue cellularity and the integrity of cellular
membranes. This technique may not be well appreciated
by general radiologists. DWI can be performed on most
The British Journal of Radiology, August 2006
Received 30 March 2006
Revised 19 May 2006
Accepted 30 May 2006
DOI: 10.1259/bjr/29739265
’ 2006 The British Institute of
Radiology
modern MRI machines with relative ease, in a short
period of time and without the need for contrast medium
administration. The potential for this technique to
evaluate the larynx for tumour recurrence after prior
radiotherapy is demonstrated in a short communication
from Vandecaveye et al in this issue [1].
At a fundamental level, DWI provides information on
the random (Brownian) motion of water molecules in
tissues. The Brownian displacements of millions of water
molecules over time are normally distributed with a
mean final value of zero for all time periods measured,
but with a standard deviation that is proportional to the
diffusion coefficient and time measured. This was the
basis for Einstein’s diffusion equation published in 1905,
which subsequently helped to earn him the 1921 Physics
Nobel Prize.
In tissues, DWI probes the movement of water
molecules, which occurs largely in the extracellular
space. However, the movement of water molecules in
the extracellular space is not entirely free, but is modified
by interactions with hydrophobic cellular membranes
and macromolecules. Hence, diffusion in biological
tissue is often referred to as ‘‘apparent diffusion’’. By
comparing differences in the apparent diffusion between
tissues, tissue characterization becomes possible. For
example, a tumour would exhibit more restricted
apparent diffusion compared with a cyst because intact
cellular membranes in a tumour would hinder the free
movement of water molecules.
One of the simplest methods of obtaining DWI images
is to apply pairs of opposing and balanced magnetic field
gradients (but of differing durations and amplitudes)
around a spin-echo refocusing pulse of a T2 weighted
sequence. Stationary water molecules are unaffected by
the paired gradients, and thus retain their signal. Non-
stationary water molecules acquire phase information
633

from the first gradient, but are not rephased by the
second gradient, leading to an overall loss of the MR
signal. The signal reduction on the DWI image is
proportional to the amount of diffusion water motion
occurring during the pulse sequence. Hence, on DWI,
there is usually less signal attenuation (i.e. higher signal
intensity) of tumour compared with normal tissue due to
the restricted diffusion of water molecules in tumours,
which is presumed to be due to an increased cellular
density.
At some anatomical locations (e.g. the brain), DWI is
usually performed in three or more gradient directions
because of the unequal limitations to diffusion in some
directions imposed by tissue organization (e.g. white
matter tracts). This phenomenon is termed anisotropy
and may be also observed in some visceral organs. For
example, in the normal prostate gland, diffusion is
greater along the line of the ducts than across the ducts.
However, anisotropy is usually not seen in tumours
since cancers typically grow in a disorganized fashion.
The degree of diffusion-weighting applied is indicated
by the b-value (measured in s mm22), which indicates
the magnitude and duration of the applied gradients and
time between the paired gradients. By varying the
amplitudes, lengths and intervals between the diffusion
gradients, the sensitivity to the degree of diffusion
motion can be altered and the data processed to provide
information about actual diffusion distances. Hence,
DWI using a larger b-value (e.g. b5500 s mm22) is more
sensitive to the slower motion of water molecules and
smaller diffusion distances, whereas the converse is true
with a smaller b-value (e.g. b550 s mm22). It is important
to remember that the phenomenon of water molecule
movement detected by DWI occurs at a length scale
(typically micrometres) that is significantly larger than
intracellular distances, but significantly smaller than
the pixel dimensions of typical MR images (typically
millimetres).
When performing DWI cancer studies, images are
typically acquired using different b-values (typically 0–
1000 s mm22). The images obtained at different b-values
allow the calculation of the apparent diffusion coefficient
(ADC; unit mm2 s21), which is usually presented as a
quantitative parametric map. From the discussions
above, it is not surprising that on ADC maps, tumours
usually demonstrate low ADC values and appear as low
signal intensity area compared with normal tissue. This
appearance is the inverse of that observed on ‘‘raw’’ DWI
images obtained at high b-values. The quantitative ADC
values can aid in lesion characterization, and can also be
applied to evaluate the treatment response of tumours.
Diffusion-weighted MRI is an established tool for the
evaluation of intracranial diseases. The technique has
been applied successfully to detect early cerebral infarc-
tion and for the characterization of brain tumours. More
recently, DWI has been used to demonstrate early
response of brain tumours to radiation treatment.
However, motion-related artefacts, which degrade image
quality, have limited the clinical application of DWI to
extracranial sites. These motion-related artefacts can now
be substantially reduced by the use of parallel imaging,
combined with breath-hold, single-shot, echo-planar MRI
techniques. Such techniques are useful for the evaluation
of the larynx, where respiratory and swallowing
634
D-M Koh and A R Padhani
movement, together with susceptibility effects from
air within the larynx, can significantly degrade image
quality.
In this issue of BJR, Vandecaveye et al describe their
experience of using DWI for the detection of post-
treatment recurrent laryngeal tumours in a small number
of patients [1]. All cases were validated by histopa-
thology. Thin section (48 slices at 4 mm thickness) axial
DWI of the larynx was performed using six b-values,
which ranged from b50 s mm22 to 1000 s mm22.
Although parallel imaging was not employed, each
DWI study was completed in less than 6 min. The four
examples presented elegantly demonstrate the potential
role of DWI in distinguishing tumour from post-
treatment change. Furthermore, the exquisite radiologi-
cal-pathological comparison enhances our understand-
ing of the pathological basis for their imaging
appearances at DWI. In the study, two cases with
recurrent tumour were detected at DWI as focal areas
of restricted diffusion returning low ADC values. By
comparison, two cases demonstrated focal asymmetry of
the larynx due to inflammatory change. These areas
returned high ADC values at DWI and were due to
laryngeal necrosis and oedema at histopathology.
Intriguingly, one case with inflammatory change was
also evaluated using 18FDG-PET imaging, which
revealed a moderately hypermetabolic focus in the
larynx (false-positive). As the PET imaging was per-
formed within 6 months of radiotherapy, this case
emphasises the potential pitfall of using 18FDG-PET
imaging to distinguish between tumour and inflamma-
tion in the early post-treatment period.
There are challenges to the use of DWI for tumour
evaluation. First, the dichotomy of identifying tumour
versus post-treatment change may not always be
straightforward. There can be substantial overlap in the
ADC values between malignant and non-malignant
tissue making it difficult to determine disease status.
For example, a predominantly necrotic tumour may
potentially be confused with necrosis arising from
radiation treatment since both would result in higher
ADC values. Thus, as with all functional imaging
techniques, DWI image information should be inter-
preted with information from conventional imaging to
improve disease assessment. Second, the averaged ADC
values derived from regions of interest drawn around
tumours may not sufficiently characterize tumour
heterogeneity. More sophisticated methods of analysis
are needed to adequately account for regional variations,
which is particularly important for the evaluation of
treatment response in tumours. Third, the analysis of
quantitative ADC on commercial platforms lacks stan-
dardization. Some commercial software does not allow
drawing of free forms to encompass tumour regions.
Most commercial software also does not allow image
registration and noise filtration, which can significantly
impact on the quality of the quantitative data. Clearly,
further collaborative work in this area would be
welcomed.
Despite some of the above limitations, DWI is
emerging as a powerful, new diagnostic tool which will
be increasingly applied to the evaluation of tumours, as
has been demonstrated in the accompanying paper by
Vandecaveye et al [1]. Potential applications include
The British Journal of Radiology, August 2006
Commentary: Diffusion-weighted MRI

distinguishing tumour from non-tumour tissue, assess- Reference

ing of treatment response and for the prediction of
1. Vandecaveye V, De Keyzer F, Poorten VV, Deraedt K,
treatment outcome. As the examination is quick and can Alaerts H, Landuyt W, et al. Evaluation of the larynx for
be conveniently incorporated into existing protocols, tumour recurrence by diffusion-weighted MRI after radio-
assessment of its role in everyday clinical practice could therapy: initial experience in four cases. Br J Radiol
be expediently achieved. 2006;79:681–7.

The British Journal of Radiology, August 2006 635

 The British Journal of Radiology, 79 (2006), 636–643

In vitro evaluation of stent patency and in-stent stenoses in 10

metallic stents using MR angiography
O W HAMER, MD, I BORISCH, MD, C PAETZEL, MD, W R NITZ, PhD, J SEITZ, MD, S FEUERBACH, MD and
N ZORGER, MD

Department of Radiology, University Hospital of Regensburg, Franz-Josef-Strauss-Allee 11, 93042

Regensburg, Germany

ABSTRACT. In vitro study to investigate the suitability of contrast enhanced magnetic

resonance angiography (CEMRA) for determination of stent patency and grading of in-
stent stenoses in 10 metallic stents. The Acculink carotid, DynaLink, Easy Wallstent,
JostentSelfX XF, Luminexx, Omnilink, sinus-SuperFlex, SMART, Symphony and ZA stent
were separately placed in a vascular phantom. Dedicated stenoses inside the stents
generated a concentric lumen narrowing of 50%. CEMRA was performed for each
stent. Signal loss inside the stents and artificial lumen narrowing were assessed
objectively using the evaluation software of the MR imager. Moreover, three blinded
observers determined visibility of stent patency and in-stent stenoses subjectively on a
3-point scale and graded in-stent stenoses. Loss of signal intensity within the stent
lumen ranged between 90% (Wallstent) and 5% (ZA), artificial lumen narrowing
between 56% (Symphony) and 22% (ZA). For the Symphony and Wallstent, visibility of Received 23 August 2005
patency and in-stent stenoses was impaired and the observers’ grading exaggerated Revised 9 January 2006
the degree of stenoses (by 23% and 33%, respectively). For the remainder of stents, Accepted 30 January 2006
patency and stenoses were visible and stenoses were graded accurately (less than 10%
DOI: 10.1259/bjr/57301879
discrepancy from reference standard). In this in vitro study, eight of 10 stents presented
with MRI characteristics which enabled determination of stent patency and accurate ’ 2006 The British Institute of
grading of clinically relevant in-stent stenoses. Radiology

Contrast-enhanced three-dimensional magnetic reso-
nance angiography (CEMRA) is a non-invasive alter-
native to digital subtraction angiography (DSA) and has
meanwhile gained wide acceptance regarding the assess-
ment of arterial occlusive disease in virtually all arterial
territories [1–5]. Patient monitoring after stent placement
is still a challenge for CEMRA due to image distortion by
artefacts emanating from the stent. Computed tomogra-
phy angiography (CTA) has been advocated as a suitable
technique for follow-up after stent implantation.
However, CTA is associated with the need for ionizing
radiation and potentially harmful iodinated contrast
material [6, 7]. Moreover, it has been demonstrated that
stent visualization on CTA images can be hampered by
markedly artificial lumen narrowing and pseudo-
enhancement within the stent lumen [8]. Colour duplex
ultrasound is another non-invasive technique which is
capable of assessing stent patency and eventual in-stent
stenoses. Performance of this technique, however, can be
very difficult in patients with obesity and gaseous
distension of the bowel when iliac stents are the target.
However, monitoring of endoprostheses in iliac arteries
is of particular interest because iliac stenting is a
frequently performed procedure.
Several experimental studies have been performed to
evaluate the nature and amount of artefacts of a variety
of stent designs and materials on MR images [9–16].
Some of these studies evaluated whether stent patency
can be determined, reporting promising results for a
subgroup of stents [17–24]. However, for a clinically
valuable follow-up technique, it is crucial to detect not
only stent patency, but also significant in-stent stenoses.
The purpose of this in vitro study simulating iliac
anatomy was to investigate the suitability of CEMRA for
determining stent patency as well as detecting clinically
relevant in-stent stenoses in 10 metallic stents.
Materials and methods
Pre-study to determine the MR background signal
MR background signal in vivo
Within a pre-study, the MR signal of perivascular
pelvic soft tissue was retrospectively determined for 27
patients who had received a CEMRA of the iliac
vasculature for clinical purposes. The CEMRA sequence
was identical to the sequence used in the main in vitro
study (see below). The coronal source image best
showing the aortic bifurcation was determined for each
patient. Using the MR evaluation software (NUMARIS
3.5 a1.1b; Siemens AG, Medical Solutions, Erlangen,
Germany), one of the authors placed a circular user-
defined region of interest (ROI) with a diameter of 3–
4 cm in the soft tissues beneath the aortic bifurcation of
each patient. The mean signal intensity as well as
standard deviation (SD) was registered for each patient.
The mean signal intensity¡SD for all 27 patients was
calculated.

636 The British Journal of Radiology, August 2006

Detection of stent patency and in-stent stenoses with MRA

MR background signal in vitro aorta, renal arteries, common iliac arteries and external
Blankets which had been soaked in water at room iliac arteries of both sides. The common and external iliac
temperature were positioned in the centre of the magnet arteries were made out of the same tube and thus were of
and imaged with the identical CEMRA sequence as identical dimensions without noticeable transition. All
applied in the main study (see below). The same author stents were separately implanted in the right common/
who had performed the in vivo measurements for the 27 external iliac artery of the phantom, which had an inner
patients placed ROIs with a diameter of 3–4 cm on a diameter of 10 mm and a wall thickness of 2 mm.
coronal source image of the blankets and determined the Because the tubes were transparent, complete stent
mean signal intensity¡SD. opening and attachment to the tube wall was assured
visually. The phantom was filled with a 2.00 mmol l21
gadopentetate dimeglumine in saline solution at room
temperature: 1 l of saline solution was mixed with 4 ml
Main in vitro study to determine the MRI
of gadopentetate dimeglumine (0.5 mol l21; Magnevist;
characteristics of 10 metallic stents Schering GmbH, Berlin, Germany). This dilution was
chosen according to settings used by other authors who
Stents applied similar MR sequences [11, 13, 25]. The MR signal
Ten metallic stents were evaluated (Acculink carotid, value of this dilution had been shown to be similar to the
DynaLink, JostentSelfX XF, Luminexx vascular, signal of contrast filled vessels on CEMRA images
Omnilink, sinus-SuperFlex, SMART, Symphony, Easy in vivo. The phantom was embedded in water soaked
Wallstent, and ZA-stent). The dimensions, material blankets to simulate perivascular pelvic soft tissue and to
composition and manufacturers are listed in Table 1. enable proper shimming.
Nine of these 10 stents were self-expanding and one was
balloon-expanding (Omnilink). All stents had a dedi-
Artificial in-stent stenosis
cated diameter of 10 mm and thus were suitable for the
A defined lumen narrowing was created by placing a
treatment of iliac stenoses (the Acculink carotid is
dedicated stenosis made from rigid urethane (model #
intended for the treatment of carotid stenoses; the stent’s
1802-2; Sawbones, Malmö, Sweden) in each stent. The
dimensions, however, render an implantation in the iliac
artificial stenoses were 20 mm in length and cylindrically
vasculature possible also). The stents’ lengths ranged
shaped, showing a central lumen and thus simulating
from 38 mm to 60 mm (mean 45.0 mm¡7.6 mm).
concentric stenoses. The lumen narrowing for all
employed stenoses was set to be 50%.
Vascular phantom
The stents were placed in a vascular phantom
MRA imaging
(endovascular model # 1803; Sawbones, Malmö,
All images were acquired at a commercial 1.5 T MRI
Sweden). The phantom consisted of tubes made from
unit (Magnetom Sonata; Siemens AG, Medical Solutions,
flexible urethane polymer and simulated the abdominal Erlangen, Germany). The vascular phantom was posi-
tioned in the centre of the magnet. The amount of
Table 1. Technical stent parameters and manufacturers susceptibility artefacts caused by the stent depends on
the orientation of the stent compared with the main
Stent Diameter Length Material Manufacturer
(mm) (mm) magnetic field B0. The distortion of the magnetic field by
a cylindrical object is minimal when the cylinder’s long
Acculink 10 40 nitinol Guidant, Diegem, axis is aligned with the direction of the main magnetic
carotid Belgium field [9, 12, 14]. The artefacts grow when the object is
DynaLink 10 56 nitinol Guidant, Diegem,
rotated toward a direction in which its long axis is
Belgium
JostentSelfX 10 44 nitinol Abbott
perpendicular to B0. This dynamic was demonstrated in
XF Laboratories, several experimental studies for numerous stents [13,
Chicago, IL 15]. With the purpose of comparing the stent specific
Luminexx 10 40 nitinol Bard/Angiomed, artefacts rather than the orientational dependence, it was
Vascular Karlsruhe, decided to choose one orientation of the stent containing
Germany tube to B0, which is representative for the common/
Omnilink 10 38 stainless Guidant, Diegem, external iliac arteries. The image distortion by all of the
steel Belgium stents would be more severe for an angle setting more
(316L) perpendicular to B0 and less severe for an angulation
sinus- 10 60 nitinol Optimed,
more parallel to B0. However, the relative severity of
SuperFlex Ettlingen,
Germany image distortion and thus the ranking regarding suit-
SMART 10 60 nitinol Cordis, ability for CEMRA imaging seen for the 10 evaluated
Miami Lakes, FL stents would be identical, independent of the specific
Symphony 10 60 nitinol Boston Scientific, angle. Thus, in order to identify those stents which are
Natick, MA best suitable for CEMRA and which should be chosen if
Easy 10 42 cobalt alloy Boston Scientific, follow-up with CEMRA is intended, one angle setting is
Wallstent Natick, MA sufficient as long as the setting is identical for all stents.
ZA 10 60 nitinol Cook, Therefore, the tube which contained the stent was
Bjaeverskov,
orientated at a lateral deviation of 30 ˚ as well as an
Denmark
upward deviation of 45 ˚ to the z-axis and main magnetic

The British Journal of Radiology, August 2006 637


field (B0), respectively, to simulate the course of the
common/external iliac artery. A phased-array body coil
(circularly polarized) served as a receiver coil. Oblique
coronal source images (parallel to the orientation of the
stent containing tube) were obtained. The applied
CEMRA sequence was a three-dimensional gradient-
echo sequence with RF spoiling (fast low-angle shot,
repetition time (TR) 3.37 ms, echo time (TE) 1.24 ms, flip
angle 25 ˚, matrix 246 6 512, read-out bandwidth 390 Hz
pixel21, 1.1 mm partition thickness (after Fourier inter-
polation), voxel size 1.3 mm 6 0.7 mm 6 1.1 mm,
frequency encoding parallel to the main magnetic field)
identical to the sequence usually used for CEMRA of the
pelvis in the authors’ department. For each stent, one
angiographic data set was acquired. The data sets were
post-processed by application of a standard maximum
intensity projection (MIP) algorithm (22.5 ˚ rotational
intervals around the craniocaudal axis covering 180 ˚).
Finally, the images were sent to a MagicView work-
station (Siemens AG).
Subjective data analysis
Three radiologists, each with 4 years of experience in
MR angiogram interpretation, evaluated the images
independently on softcopy displays at a MagicView
workstation. The observers were blinded regarding the
type of stent and degree of in-stent stenosis. Each
observer evaluated each stent once; there were no
repeated measurements. They were instructed to rank
their overall impression of the quality of each CEMRA
study (source and MIP images were evaluated together
for this analysis) on a 4-point scale (15good, 25minor
limitations, 35major limitations, 45not diagnostic).
Furthermore, both stent patency and visibility of in-stent
stenoses had to be ranked on a 3-point scale (15visible,
25limited visibility, 35not visible). This was done
separately for source and MIP images. If ‘‘limited
visibility’’ or ‘‘visible’’, the degree of stenosis had to be
determined at its narrowest location on the source
images. The reference diameter was defined as the
diameter of the tube segment adjacent to the proximal
stent end. The severity of stenosis had to be expressed as
a percentage of the reference diameter. The mean¡SD of
the three observers’ gradings and the difference of this
mean to 50% (reference standard) were computed for
every stent.
Objective data analysis
For each stent, the author who had performed the
measurements of the pre-study assessed (1) the back-
ground signal intensity on CEMRA images, (2) the
apparent diameter of the stent lumen outside the
artificial stenosis and (3) the loss of signal intensity
within the stent outside the artificial stenosis. The author
was blinded to the type of stent and acquired one data
set per stent. Measurements were made on source
images by using the evaluation software of the MR
imager (NUMARIS 3.5 a1.1b; Siemens AG, Medical
Solutions, Erlangen, Germany).
Ad (1): The background signal intensity was deter-
mined as follows: ROIs with a diameter of 3–4 cm were
placed beneath the ‘‘aortic bifurcation’’ of the vascular
phantom on coronal source images. The mean signal
638
O W Hamer, I Borisch, C Paetzel et al
intensity¡SD was determined for each of the 10 studies.
The mean signal value¡SD of these 10 signal measure-
ments was calculated.
Ad (2): The apparent stent lumen and artificial lumen
narrowing, respectively, were determined as follows:
signal intensity plots were drawn perpendicular to the
long axis of the stent. The plots were placed within
the visible stent lumen outside the in-stent stenosis. The
location of the artificial in-stent stenosis was easily
visible for eight stents (Acculink, DynaLink,
JostentSelfX XF, Luminexx, Omnilink, sinus-SuperFlex,
SMART and ZA). For two stents (Symphony and
Wallstent), the margins of the in-stent stenoses could
not be determined unequivocally on MR images.
However, the examiner was aware of the exact location
because he knew the experimental setup. The outer
contour of the visible stent lumen was determined using
a method similar to that used by other groups [13, 26]:
Those two points (left and right margin of the stent
lumen) on the signal intensity plot were identified where
the signal exceeded twice the signal of the background.
The apparent stent lumen diameter was defined as the
distance between these two points (Figure 1). The ratio of
apparent stent lumen diameter and the respective value
of the reference tube, which was the tube segment
adjacent to the proximal stent end, was calculated and
represented the artificial lumen narrowing (the higher
the value, the less artificial lumen narrowing, and vice
versa).
Ad (3): The signal intensity within the stent was
determined as follows: a circular ROI was placed
(a) within the visible stent lumen without covering parts
of the stent struts or the in-stent stenosis and (b) within
the reference tube. The mean signal values within both
ROIs were measured. The loss of signal intensity within
the stent was determined by calculating the ratio of the
signal within the stent and the signal in the reference
tube for each stent (the higher the value, the less signal
loss, and vice versa).
Figure 1. Schematic of signal intensity plot (dotted line)
perpendicular to long axis of stent. The apparent stent
lumen was defined to be the distance between the two
points on the plot where the signal exceeded twice the
background signal.
The British Journal of Radiology, August 2006
Detection of stent patency and in-stent stenoses with MRA
To compare the suitability of the stents for CEMRA,
three groups were established on the basis of the degree
of artificial lumen narrowing and loss of in-stent signal
intensity. The thresholds were chosen according to the
strategy published by other authors [13]. Group 1 stents
were stents with a minor reduction of signal intensity
within the stent and a minor artificial lumen narrowing
(signal intensity and apparent stent lumen 66–100%
relative to the reference tube). These stents were
considered to be well suited for CEMRA. Group 2 stents
were considered to be partially suited for CEMRA: one of
the two parameters (signal intensity or apparent stent
lumen) was less than 66%. Group 3 stents were
considered to be not well suited for CEMRA: both
parameters were less than 66%.
Statistics
Interobserver variability for categorical data (ranking
of visibility stent patency and in-stent stenoses) was
determined using kappa statistics. In case kappa
statistics was not applicable because the cross tabulated
scores were not symmetrical, Spearman’s correlation
coefficient was calculated. For continuous data (severity
of in-stent stenoses), kappa statistics was not
applicable because calculation of kappa requires that
all observers use the same rating categories. In the
present study, however, the observers’ grading of
stenoses was not classified, but resulted in a large
number of possible responses. Thus, interobserver
variability regarding determination of severity of in-
stent stenoses was evaluated by calculating Pearson’s
correlation coefficient. A p-value # 0.05 was considered
to be statistically significant.
The British Journal of Radiology, August 2006
Results
MR background signal
The mean background signal intensity, as retrospec-
tively assessed for the 27 in vivo studies, was 10.4¡3.9.
The background signal intensity determined within the
pre-study for the water soaked blankets was 13.3¡3.5.
Because the difference between these two signal values
was negligible, the water soaked blankets were consid-
ered to be suitable to simulate perivascular pelvic soft
tissue, and thus were applied in the main study. The
mean background signal intensity for the 10 experimen-
tal stent studies was 14.8¡3.9.
Subjective data analysis
The source and MIP images of all 10 stents are given
side by side in Figure 2.
The quality of CEMRA images regarding the overall
diagnostic value was ranked to be good (score 1) for all
stents and observers.
The observers’ performance for visibility of stent
patency and in-stent stenoses is summarized in Table 2.
The scores for the Acculink, DynaLink, JostentSelfX XF,
Luminexx, Omnilink, sinus-SuperFlex, SMART, and ZA
were equally excellent on both CEMRA source images
and MIP images. All observers agreed that visibility of
stent patency and in-stent stenoses was worse for the
Symphony and Wallstent. Moreover, visibility of patency
and in-stent stenoses in these two stents was ranked
lower on MIP images compared with the respective
source images for 21/24 (88%) of corresponding score
pairs. Kappa was 0.51–1.00 (p50.027 to p ,0.0005) and
Figure 2. Source images
(upper row) and maximum
intensity projection (MIP)
images (lower row) obtained
from oblique coronal 3D con-
trast enhanced MR angiogra-
phy (CEMRA) images of 10
metallic stents. The stents from
left to right: group 1: DynaLink,
ZA; group 2: Acculink carotid,
JostentSelfX XF, Luminexx,
Omnilink, SMART, sinus-
SuperFlex; group 3: Symphony,
Wallstent. The signal loss
within the stents and the
degree of artificial lumen nar-
rowing differed considerably
between the stents. Imaging
characteristics of group 1 and
2 stents (for classification see
Table 1) enabled determina-
tion of stent patency and accu-
rate delineation of in-stent
stenoses by the observers. The
MIP algorithm can cause loss of
contrast which might artificially
worsen the severity of a steno-
sis or mimic an occlusion. This
phenomenon is best appre-
ciated for the group 3 stents.
639
 O W Hamer, I Borisch, C Paetzel et al

Table 2. Subjective and objective evaluation of contrast enhanced MR angiography (CEMRA) images
Visibility of stent Visibility of in-stent Accuracy of Intraluminal Apparent Group
patencya stenosisa assessment of signal stent lumen classificatione
in-stent intensityc diameterd
stenosisb

Source MIP images Source MIP images [%] [%] [%]

images images

Acculink carotid 1-1-1 1-1-1 1-1-1 1-1-1 (+) 1.1 64.8 70.1 2
DynaLink 1-1-1 1-1-1 1-1-1 1-1-1 (–) 1.7 75.6 76.2 1
JostentSelfX XF 1-1-1 1-1-1 1-1-1 1-1-1 (+) 3.5 84.9 62.3 2
Luminexx Vascular 1-1-1 1-1-1 1-1-1 1-1-1 (+) 4.3 62.8 76.7 2
Omnilink 1-1-1 1-1-1 1-2-1 1-2-1 (–) 2.8 47.0 73.4 2
sinus-SuperFlex 1-1-1 1-1-1 1-1-1 1-1-1 (–) 6.3 64.1 75.8 2
SMART 1-1-1 1-1-1 1-1-1 1-1-1 (+) 8.2 40.3 74.6 2
Symphony 1-1-1 2-3-2 1-2-2 2-3-3 (+) 22.6 34.4 43.9 3
Easy Wallstent 2-3-2 3-3-3 2-3-3 3-3-3 (+) 33.3 10.1 65.0 3
ZA 1-1-1 1-1-1 1-1-1 1-1-1 (+) 3.0 95.8 77.7 1
a
Three observers ranked the visibility according to the score: 1: visible, 2: limited visibility, 3: not visible. Each number refers to
one observer.
b
Expressed as the difference between the reference standard (stenosis of 50%) and the mean of severity of stenoses, as assessed
by the three observers on CEMRA; (+): degree of stenosis was overestimated by the observers, (2): degree of stenosis was
underestimated by the observers.
c
Expressed as percentage of signal intensity in reference tube.
d
Expressed as percentage of diameter of reference tube.
e
Classification on the basis of signal intensity and apparent stent lumen diameter; group 1: both parameters >66%, group 2: one
of the two parameters ,66%, group 3: both parameters ,66%.
MIP, maximum intensity projection.

Spearman’s correlation coefficient 0.65–1,00 (p50.044 to
p ,0.0005), indicating high interobserver agreement.
The observers’ performance regarding accuracy of
grading of in-stent stenosis is given in detail in Table 2.
For eight (Acculink, DynaLink, JostentSelfX XF,
Luminexx, Omnilink, sinus-SuperFlex, SMART, and
ZA) out of 10 evaluated stents, the observer’s grading
differed by less than 10% from the reference standard.
For five (Acculink, JostentSelfX XF, Luminexx, SMART,
and ZA) of these eight stents, the degree of stenosis was
overestimated; for three stents (DynaLink, Omnilink,
sinus-SuperFlex) it was underestimated. For the
Symphony and Wallstent, the discrepancy between the
observers’ grading and the reference standard was 22.6%
and 33.3%, respectively. For both stents, in-stent stenoses
were overestimated. Pearson’s correlation coefficients
ranged between 0.62 and 0.85 (p50.002 to p50.059)
indicating high interobserver agreement.
Objective data analysis
The magnitude of the in-stent signal intensities and
degree of artificial lumen narrowing and, based on these
criteria, the group classification are summarized in
Table 2. The measurements differed considerably
between the 10 stents. Eight of the 10 stents (Acculink,
DynaLink, JostentSelfX XF, Luminexx, Omnilink,
sinus-SuperFlex, SMART, and ZA) exerted less severe
artefacts and were consecutively ranked to be partially or
well suited for CEMRA. Two stents (Symphony and
Wallstent) caused marked image distortion and thus
were found to be not well suited for CEMRA.
Interestingly, the measurements for the subgroup of
eight stents made from nitinol (Acculink, DynaLink,
JostentSelfX XF, Luminexx, sinus-SuperFlex, SMART,
Symphony, and ZA) were inhomogeneous with the in-
stent signal intensity ranging between 34.4%
(Symphony) and 95.8% (ZA) and the apparent stent
lumen diameter between 43.9% (Symphony) and 77.7%
(ZA) relative to the reference tube. The magnitude of
artefacts exerted by the stent made from stainless steel
(Omnilink: signal intensity within the stent 47.0%,
apparent stent lumen diameter 73.6%) was found to be
less extensive than that seen in some nitinol stents.
Discussion
MRA is an alternative to digital subtraction angiogra-
phy providing several advantages. It is a non-invasive
technique without the need for ionizing radiation and
potentially harmful iodinated contrast material [6, 7]. In
recent years, there have been rapid developments
regarding the optimization of MRA sequences. With
the introduction of CEMRA, the limitations of time-of-
flight and phase contrast techniques in terms of long
imaging times and artefacts caused by slow and
turbulent flow have widely been overcome [2, 4].
CEMRA is meanwhile well accepted for assessing
arterial occlusive disease. However, the monitoring of
vascular endoprostheses is still a challenge because of
image distortion due to artefacts caused by the stents.
According to data in the literature, stent patency can be
assessed by CEMRA for a subgroup of stents [20, 21, 23,
24, 27, 28]. A clinically valuable follow-up technique,
however, should also enable a reliable detection of
critical in-stent stenoses which cause a significant
compromise to blood flow. In general, this occurs for
stenoses generating a diameter reduction of 50% or more

640 The British Journal of Radiology, August 2006

Detection of stent patency and in-stent stenoses with MRA
[29, 30]. Consecutively, these stenoses are considered to
be relevant for surgical or interventional treatment.
In this in vitro study, 10 stents suitable for the
implantation in iliac arteries were evaluated with regard
to their imaging characteristics on CEMRA images.
Emphasis was placed on subjective criteria (visibility of
stent patency and artificial in-stent stenoses, as well as
grading of lumen narrowing) and on objective para-
meters (signal loss within the stent and artificial lumen
narrowing). The stents were made from nitinol, stainless
steel and a cobalt-based alloy. The quality of images
regarding the overall diagnostic value was ranked to be
good for all studies. Thus, there was no bias due to a
potentially differing image quality.
There are three main types of artefacts associated with
metallic vascular implants causing image distortion:
(1) radiofrequency (RF) artefacts originating from RF-
induced eddy currents in the stent; (2) susceptibility
artefacts caused by differences between the magnetic
susceptibility of the stent material and the surrounding
tissue, leading to local field inhomogeneities; and
(3) flow-related artefacts caused by turbulences and
consecutive signal loss due to dephasing [9, 12, 14, 16,
27]. The absence of flow and pulsatility in the presented
study is a limitation and may reduce the comparability to
an in vivo study. However, CEMRA depends on the T1
shortening effects of gadopentetate dimeglumine and is
relatively insensitive to flow-related artefacts because of
extremely short echo times and thin slices [31]. Thus, the
presented setup can be assumed to obtain valid results
regarding interstent visibility.
In order to quantify RF induced and susceptibility
artefacts, the degree of signal loss inside the stent and
artificial lumen narrowing was assessed objectively.
According to these measurements, the stents were
classified into three groups representing their suitability
for MRI. This classification correlated well with the
observers’ performance regarding the subjective grading
of in-stent stenoses. Group 1 and 2 stents (Acculink,
DynaLink, JostentSelfX XF, Luminexx, Omnilink, sinus-
SuperFlex, SMART, and ZA) were considered to be at
least partially suited for CEMRA imaging. For these
stents, the observers’ accuracy of grading of in-stent
stenoses did meet the needs of clinical routine (difference
to the standard of reference ,10%). For group 3 stents
(Symphony and Wallstent), the number of artefacts was
more pronounced. Consecutively, the grading of in-stent
stenoses was markedly worse, resulting in an over-
estimation of 23–33%. Hence, implantation of these two
stents seems not be favourable if follow-up with CEMRA
is intended.
Published data regarding the evaluation of type and
amount of artefacts of various stent materials revealed
nitinol to be among the favourable alloys and stainless
steel to be less suited regarding MR compatibility
[12, 13, 15, 16, 32–34]. In the present study, substantial
differences between the imaging characteristics within
the subgroup of nitinol stents (Acculink, DynaLink,
JostentSelfX XF, Luminexx, sinus-SuperFlex, SMART,
Symphony and ZA) were noted. This fact emphasises
that the magnitude of artefacts not only depends on
material composition, but also on other features like wire
thickness, total weight and geometry of mesh or slotted
tubes. The observation that the Omnilink stent, made
The British Journal of Radiology, August 2006
from stainless steel, presented in this in vitro study with
less severe artefacts than some nitinol stents confirmed
that an advantageous design may be able to compensate
for disadvantages due to an unfavourable stent material.
For the Symphony and Wallstent, the visibility of stent
patency and in-stent stenoses was ranked worse on MIP
images than on source images. This was most likely
caused by an artificial loss of contrast on MIP images.
Loss of contrast is a result of the characteristics of the
MIP algorithm which projects noise or distant high
signal into the image [26]. MIP images can mimic a
stenosis or even a stent or vessel occlusion when the
original signal is low, as seen within the Symphony and
Wallstent. Therefore, it should be emphasised that for a
thorough evaluation of CEMRA studies, interpretation of
source and MIP images is mandatory.
In contradiction to the remainder of stents evaluated in
this study, the Wallstent presented with a complex
morphology of the in-stent signal in terms of bandlike
artefacts at both stent endings, a circumscribed region of
high signal intensity adjacent to the bandlike artefacts
and a severe signal reduction in the central part of the
stent. This morphology is the same observed in experi-
mental studies performed by other authors who did not
apply in-stent stenoses but evaluated the imaging
characteristics of the stent itself [13, 14, 32]. The length
of the central region of severe signal reduction observed
in the present study exceeded the length of the artificial
in-stent stenosis. The signal intensity plots for determi-
nation of the apparent in-stent lumen and the region of
interest for assessment of in-stent signal intensity were
placed within the area of severe signal reduction, but
outside the in-stent stenosis.
To our knowledge, there are two studies in literature
which evaluated in-stent stenoses on CEMRA images
with, however, different setups: on the basis of objective
measurements, Letourneau-Guillon et al determined the
degree of artificial lumen narrowing and the delineation
of a 50% in-stent stenosis in a Luminexx stent (diameter
8 mm), revealing comparable results as observed in our
study for this stent [25]. An observer based subjective
evaluation was not performed in this study. Maintz et al
employed a vascular phantom with an inner tube
diameter of 8 mm and evaluated, among others, the
Symphony, SMART, Wallstent and ZA stent [35]. The
stent-containing tubes were oriented along the main
magnetic field B0. The signal loss inside the stent was
assessed objectively, the degree of artificial lumen
narrowing and the visibility of in-stent stenoses were
determined subjectively in a consensus decision.
Grading of in-stent stenoses was not performed. As far
as comparisons are appropriate, the results are consistent
with the data evaluated in the present study. However,
the authors observed differing imaging characteristics
for the Wallstent in terms of no signal loss inside the
stent and a good visibility of the in-stent stenosis. This
discrepancy is most likely due to the different orientation
of the stent relative to the main magnetic field because it
is well documented that the artefact size increases with a
larger angle to B0 [9, 12, 13, 15, 25, 36]. This hypothesis is
confirmed by another study performed by the same
group of authors evaluating the Wallstent at vary-
ing angles relative to B0 [14]. According to this latter
analysis, the Wallstent caused severe artefacts when the
641

longitudinal axis of the stent was diagonal to B0.
Furthermore, two other groups observed similar results
as seen in our study in terms of a subtotal to total signal
void within the Wallstent even though the stent’s
orientation was parallel to the main magnetic field [13, 32].
There are limitations to our study. For the sake of
comparability, all evaluated stents had a dedicated
diameter of 10 mm. Thus, the observed results apply
for this specific dimension only. It is possible that the
relative artefact size changes with different stent dia-
meters, which might be expected for smaller diameters in
particular. However, considering the already outlined
data published by Letourneau-Guillon et al and Maintz
et al, the imaging characteristics of the Luminexx,
Symphony, SMART, and ZA stent were similar for the
10 mm and 8 mm diameter models, indicating low
impact of the stent diameter on artefact size [25, 35].
Although the vascular phantom and the experimental
setting were designed to simulate the conditions of the
iliac vasculature, clinical studies are warranted to
investigate if the observations are confirmed in an
in vivo setting. First clinical experiences in 27 patients
regarding the detection and grading of in-stent stenoses
in the JostentSelfX (Abbott Laboratories, Chicago, IL) by
CEMRA confirmed the in vitro results of the present
study because this stent, which is very similar to the
JostentSelfX XF, revealed to be suitable for MR [37].
However, in single cases susceptibility artefacts at the
stent’s ends simulated significant in-stent stenoses. This
phenomenon, which limited the reliability of CEMRA
significantly, was not observed under in vitro conditions.
Hence, in vitro studies seem to be suitable to roughly
classify stents with regard to their MR suitability.
However, the final decision has to be made in vivo.
In conclusion, the amount of stent related artefacts
differed considerably for the 10 evaluated stents. Two
out of 10 tested stents were not suitable for CEMRA
follow-up because stent-related artefacts were severe and
in-stent stenoses could not reliably be determined. In
eight out of 10 stents, artefacts were less pronounced and
the observers’ performance regarding the determination
of stent patency and grading of in-stent stenoses on
CEMRA images did meet the requirements of clinical
routine. Radiologists should be familiar with this
subgroup of stents if follow-up with CEMRA is intended.
However, confirmation of the evaluated data under
in vivo conditions is mandatory.
References
1. Borisch I, Horn M, Butz B, Zorger N, Draganski B,
Hoelscher T, et al. Preoperative evaluation of carotid artery
stenosis: comparison of contrast-enhanced MR angiography
and duplex sonography with digital subtraction angiogra-
phy. AJNR Am J Neuroradiol 2003;24:1117–22.
2. Koelemay MJ, Lijmer JG, Stoker J, Legemate DA, Bossuyt
PM. Magnetic resonance angiography for the evaluation of
lower extremity arterial disease: a meta-analysis. JAMA
2001;285:1338–45.
3. Lenhart M, Framme N, Volk M, Strotzer M, Manke C, Nitz
WR, et al. Time-resolved contrast-enhanced magnetic
resonance angiography of the carotid arteries: diagnostic
accuracy and inter-observer variability compared with
selective catheter angiography. Invest Radiol 2002;37:
535–41.
642
O W Hamer, I Borisch, C Paetzel et al
4. Prince MR, Yucel EK, Kaufman JA, Harrison DC, Geller SC.
Dynamic gadolinium-enhanced three-dimensional abdom-
inal MR arteriography. J Magn Reson Imaging
1993;3:877–81.
5. Sueyoshi E, Sakamoto I, Matsuoka Y, Ogawa Y, Hayashi H,
Hashmi R, et al. Aortoiliac and lower extremity arteries:
comparison of three-dimensional dynamic contrast-
enhanced subtraction MR angiography and conventional
angiography. Radiology 1999;210:683–8.
6. Cochran ST, Bomyea K, Sayre JW. Trends in adverse events
after IV administration of contrast media. AJR Am J
Roentgenol 2001;176:1385–8.
7. Laroche D, Namour F, Lefrancois C, Aimone-Gastin I,
Romano A, Sainte-Laudy J, et al. Anaphylactoid and
anaphylactic reactions to iodinated contrast material.
Allergy 1999;54 Suppl. 58:13–6.
8. Strotzer M, Lenhart M, Butz B, Volk M, Manke C,
Feuerbach S. Appearance of vascular stents in computed
tomographic angiography: in vitro examination of 14
different stent types. Invest Radiol 2001;36:652–8.
9. Bartels LW, Smits HF, Bakker CJ, Viergever MA. MR
imaging of vascular stents: effects of susceptibility, flow,
and radiofrequency eddy currents. J Vasc Interv Radiol
2001;12:365–71.
10. Bartels LW, Bakker CJ, Viergever MA. Improved lumen
visualization in metallic vascular implants by reducing RF
artifacts. Magn Reson Med 2002;47:171–80.
11. Graf H, Klemm T, Lauer UA, Duda S, Claussen CD, Schick
F. [Systematics of imaging artifacts in MRT caused by
metallic vascular implants (stents)]. Rofo Fortschr Geb
Rontgenstr Neuen Bildgeb Verfahr 2003;175:1711–9.
12. Klemm T, Duda S, Machann J, Seekamp-Rahn K, Schnieder
L, Claussen CD, et al. MR imaging in the presence of
vascular stents: A systematic assessment of artifacts for
various stent orientations, sequence types, and field
strengths. J Magn Reson Imaging 2000;12:606–15.
13. Lenhart M, Volk M, Manke C, Nitz WR, Strotzer M,
Feuerbach S, et al. Stent appearance at contrast-enhanced
MR angiography: in vitro examination with 14 stents.
Radiology 2000;217:173–8.
14. Maintz D, Kugel H, Schellhammer F, Landwehr P. In vitro
evaluation of intravascular stent artifacts in three-dimen-
sional MR angiography. Invest Radiol 2001;36:218–24.
15. Meyer JM, Buecker A, Schuermann K, Ruebben A,
Guenther RW. MR evaluation of stent patency: in vitro test
of 22 metallic stents and the possibility of determining their
patency by MR angiography. Invest Radiol 2000;35:739–46.
16. Wang Y, Truong TN, Yen C, Bilecen D, Watts R, Trost DW,
et al. Quantitative evaluation of susceptibility and shielding
effects of nitinol, platinum, cobalt-alloy, and stainless steel
stents. Magn Reson Med 2003;49:972–6.
17. Hagspiel KD, Leung DA, Nandalur KR, Angle JF, Dulai HS,
Spinosa DJ, et al. Contrast-enhanced MR angiography at
1.5 T after implantation of platinum stents: in vitro and in
vivo comparison with conventional stent designs. AJR Am J
Roentgenol 2005;184:288–94.
18. Amano Y, Gemma K, Kawamata H, Kumazaki T. Fat-
suppressed gadolinium-enhanced three-dimensional mag-
netic resonance angiography adequately depicts the status
of iliac arteries following atherectomy and stent placement.
Cardiovasc Intervent Radiol 1998;21:345–7.
19. Cavagna E, Berletti R, Schiavon F. In vivo evaluation of
intravascular stents at three-dimensional MR angiography.
Eur Radiol 2001;11:2531–5.
20. Juergens KU, Tombach B, Reimer P, Vestring T, Heindel W.
Three-dimensional contrast-enhanced MR angiography of
endovascular covered stents in patients with peripheral
arterial occlusive disease. AJR Am J Roentgenol 2001;176:
1299–303.
The British Journal of Radiology, August 2006
Detection of stent patency and in-stent stenoses with MRA
21. Link J, Steffens JC, Brossmann J, Graessner J, Hackethal S,
Heller M. Iliofemoral arterial occlusive disease: contrast-
enhanced MR angiography for preinterventional evaluation
and follow-up after stent placement. Radiology
1999;212:371–7.
22. Matsumoto AH, Teitelbaum GP, Barth KH, Carvlin MJ,
Savin MA, Strecker EP. Tantalum vascular stents: in vivo
evaluation with MR imaging. Radiology 1989;170:753–5.
23. Schurmann K, Vorwerk D, Bucker A, Neuerburg J,
Grosskortenhaus S, Haage P, et al. Magnetic resonance
angiography of nonferromagnetic iliac artery stents and
stent-grafts: a comparative study in sheep. Cardiovasc
Intervent Radiol 1999;22:394–402.
24. Tello R, Thomson KR, Witte D, Becker GJ, Tress BM.
Dynamic gadolinium DTPA-enhanced magnetic resonance
of intravascular stents. Invest Radiol 1998;33:411–4.
25. Letourneau-Guillon L, Soulez G, Beaudoin G, Oliva VL,
Giroux MF, Qin Z, et al. CT and MR imaging of nitinol
stents with radiopaque distal markers. J Vasc Interv Radiol
2004;15:615–24.
26. Anderson CM, Saloner D, Tsuruda JS, Shapeero LG, Lee RE.
Artifacts in maximum-intensity-projection display of MR
angiograms. AJR Am J Roentgenol 1990;154:623–9.
27. Amano Y, Ishihara M, Hayashi H, Gemma K, Kawamata H,
Amano M, et al. Metallic artifacts of coronary and iliac
arteries stents in MR angiography and contrast-enhanced
CT. Clin Imaging 1999;23:85–9.
28. Matsumoto AH, Teitelbaum GP, Carvlin MJ, Barth KH,
Savin MA, Strecker EP. Gadolinium enhanced MR imaging
of vascular stents. J Comput Assist Tomogr 1990;14:357–61.
The British Journal of Radiology, August 2006
29. May AG, van de Berg L, Deweese JA, Rob CG. Critical
arterial stenosis. Surgery 1963;54:250–9.
30. Rose SC. Noninvasive vascular laboratory for evaluation of
peripheral arterial occlusive disease: Part I--hemodynamic
principles and tools of the trade. J Vasc Interv Radiol
2000;11:1107–14.
31. Prince MR. Gadolinium-enhanced MR aortography.
Radiology 1994;191:155–64.
32. Hilfiker PR, Quick HH, Debatin JF. Plain and covered stent-
grafts: in vitro evaluation of characteristics at three-
dimensional MR angiography. Radiology 1999;211:693–7.
33. Barras CD, Myers KA. Nitinol - its use in vascular surgery
and other applications. Eur J Vasc Endovasc Surg
2000;19:564–9.
34. Duerig T, Tolomeo D, Wholey M. An overview of super-
elastic stent design. Min Invas Ther Allied Technol
2000;9:235–46.
35. Maintz D, Tombach B, Juergens KU, Weigel S, Heindel W,
Fischbach R. Revealing in-stent stenoses of the iliac arteries:
comparison of multidetector CT with MR angiography and
digital radiographic angiography in a Phantom model. AJR
Am J Roentgenol 2002;179:1319–22.
36. Trost DW, Zhang HL, Prince MR, Winchester PA, Wang Y,
Watts R, et al. Three-dimensional MR angiography in
imaging platinum alloy stents. J Magn Reson Imaging
2004;20:975–80.
37. Hamer OW, Finkenzeller T, Borisch I, Paetzel C, Zorger N,
et al. In vivo evaluation of patency and in-stent stenoses
after implantation of nitinol stents in iliac arteries using MR
angiography. AJR Am J Roentgenol 2005;185:1282–8.
643
 The British Journal of Radiology, 79 (2006), 644–651

Occupational exposure in the electrophysiology laboratory:

quantifying and minimizing radiation burden
1,2
N THEOCHAROPOULOS, MSc, 1J DAMILAKIS, PhD, 1K PERISINAKIS, PhD, 3E MANIOS, MD, 3P VARDAS,
4
MD, PhD, FESC, FACC and N GOURTSOYIANNIS, MD, PhD

1
Departments of Medical Physics, 3Cardiology and 4Radiology, Faculty of Medicine, University of
Crete, P.O. Box 2208, Iraklion 71003, Crete and 2Department of Natural Sciences, Technological
Education Institute of Crete, P.O. Box 140, Iraklion 71004, Crete, Greece

ABSTRACT. Fluoroscopically guided procedures in the electrophysiology room, such as

radiofrequency catheter ablation and implantation of cardiac resynchronization
devices, may result in high radiation exposure of electrophysiologists and assisting
staff. Our aim was to provide accurate and applicable data on occupational doses to
the electrophysiology laboratory personnel. We exposed fluoroscopically an
anthropomorphic phantom at three projections common in electrophysiology studies.
For each exposure, scattered radiation was measured at 182 sites of the cardiology
room at four body levels. Effective dose values, eye lens, skin and gonadal doses to the
laboratory staff were calculated. Our study has shown that a procedure requiring
40 min of fluoroscopy yields a maximum effective dose of 129 mSv and a maximum
value of gonadal dose of 56.8 mSv to staff using a 0.35 mm lead-equivalent apron. A
conservative estimate of the electrophysiologist’s annual maximum permissible Received 7 June 2005
workload is 155 procedures. Staff effective dose values vary by a factor of 40 due to Revised 9 December 2005
positioning during fluoroscopy and by a factor of 11 due to radiation protection Accepted 3 January 2006
equipment. Undercouch protective shields may reduce gonadal doses up to 98% and
DOI: 10.1259/bjr/76128583
effective dose up to 25%. Consequently, radiation levels in the electrophysiology room
are not negligible. Mitigation of occupational exposure is feasible through good ’ 2006 The British Institute of
fluoroscopy and working practices. Radiology

In recent years, due to technological advances, a ple-
thora of new fluoroscopically guided procedures has
been introduced in the electrophysiology suite [1–9].
Radiofrequency catheter ablation and implantation of
rhythm devices and cardioverter defibrillators have been
proven to supersede the effectiveness of other therapeu-
tic approaches and, as a result, their application is
increasing [7, 8]. Hence, electrophysiologists and support
personnel may be exposed to considerable levels of
radiation, depending on the laboratory workload
and complexity of the procedures [9–11]. The need for
quantifying the risk of radiation detriment to laboratory
staff is imperative.
Because of the different radiosensitivities of the various
body organs and tissues, for non-homogeneous irradia-
tions, radiation detriment is assessed by a radiation
quantity termed effective dose, introduced by the
International Commission on Radiological Protection
(ICRP). The effective dose E is the weighted average of
the mean absorbed dose DT to 22 organs/tissues, where
the organ/tissue-specific weighting factor wT is the
fractional organ/tissue contribution to the total body
detriment [12]:
P (a) posteroanterior (PA), (b) left anterior oblique 45 ˚
E~ wT DT
T (LAO) and (c) right anterior oblique 25 ˚ (RAO). A mobile
Address correspondence to: J Damilakis, Assistant Professor,
Department of Medical Physics, Faculty of Medicine, University of
Crete, P.O. Box 2208, Iraklion 71003, Crete, Greece.
The risk of genetic effects, i.e. impairment on progeny,
depends on the preconceptual radiation dose delivered
to staff gonads. Other organ/tissue doses of interest are
the skin and the eye lens dose, which together with
effective dose, are controlled by regulatory annual limits
[12].
To the best of our knowledge, there is no reported
experience on occupational effective and gonadal doses
as well as maximum permissible workloads regarding
the electrophysiology suite. The present study aims to
provide accurate and applicable data on occupational
doses to the electrophysiology laboratory personnel.
Methods and materials
Simulated exposures and scatter measurements
An anthropomorphic tissue-equivalent phantom
(RANDO; Alderson Research Labs, Stanford, CA) simu-
lating an adult was positioned supine on the surgical
table and fluoroscopically exposed at three distinct
projections common in interventional cardiology:
undercouch C-arm fluoroscopic unit was used (Philips
BV300-R2; Philips Medical Systems, Best, The
Netherlands) at a focus to image intensifier distance of

644 The British Journal of Radiology, August 2006

Occupational exposure in the electrophysiology laboratory
100 cm. The half value layer of the X-ray tube was
4.7 mm aluminium at 70 kVp and the diameter of the
input field size was 23 cm.
Scatter air kerma rates were measured using a hand-
held ionization chamber (LB 1236, EG&G BERTTHOLD )
connected to a portable digital meter (UMo LB 123,
EG&G BERTTHOLD). The operating theatre floor was
divided into two grids relative to the long sides of the
table. The grids were 3 m61.5 m in dimensions and
consisted of 25 cm625 cm square cells. 91 measuring
points were thus defined on each grid. Measurements
were performed at the height of the gonads (80 cm above
the floor), the waist (100 cm), the neck (150 cm) and the
face (165 cm). The air kerma readings were divided by
the dose–area product (DAP) rate of each exposure to
obtain data independent of exposure parameters. All
measurements were repeated with two removable
flexible undercouch 0.5 mm lead-equivalent protective
shields (60 cm670 cm) attached to the table’s long sides
(Mavig, Muenchen, Germany).
Calculation of occupational doses
Using coefficients provided by ICRP [13], the face level
air kerma measurements were converted into eye lens
dose and to personal dose equivalent Hp (0.07)F [14]. The
latter corresponds to the dose received by the facial skin.
The neck level air kerma measurements were converted
into personal dose equivalent Hp (0.07)N while the waist
level measurements were converted to personal dose
equivalent Hp (10)W. The genital levels measurements
were converted to male and female staff gonadal dose
with use of coefficients provided by ICRP [13].
Previously published attenuation data were used for
the calculation of gonadal and deep doses under a
0.35 mm and 0.50 mm lead-equivalent apron, and of the
eye lens doses under 0.35 mm lead-equivalent goggles
[15]. The polychromatic nature of X-rays was accounted
for by weighting the available monoenergetic dose
conversion coefficients on the basis of the characteristics
of the utilized spectra.
The values of Hp W under apron protection and the
Hs N values at neck level were used for the calculation of
staff effective dose according to the following equations
[16]:
E~0:06|(Hp (0:07) N
Hp (10) W )zHp (10) W ð1Þ
when a thyroid shield is not used, and
E~0:02|(Hp (0:07) N
Hp (10) W )zHp (10) W ð2Þ
when a thyroid shield is used.
Equation (1) was also used for the calculation of
effective dose to unshielded personnel by means of
attenuation-free Hp (10)W values.
The neck level air kerma measurements were also
converted into to personal dose equivalent Hp (10)N, to
enable the correlation of the reading of an over-apron
chest dosemeter with the actual value of effective dose to
monitored staff. To investigate the effect of beam quality
on staff doses, the Hp (0.07)N, Hp (10)W and Hp (10)N
values were calculated for four tube voltage settings:
70 kVp, 85 kVp, 100 kVp and 120 kVp.
The British Journal of Radiology, August 2006
Calculation of radiation burden from a complex
procedure
Derived occupational doses are projection specific and
DAP normalized since they originate from the initial
DAP normalized scatter radiation measurements. This
normalization eliminates the dependence on exposure
technique factors (kilovoltage and milliamperage) or
instrumentation and enables occupational dose calcula-
tions from any complex procedure on the basis of the
individual DAP value of each projection involved [17].
The fluoroscopy time required per patient strongly
depends on the type of electrophysiological study.
Reported fluoroscopy course durations range from
3.54 min per procedure for conventional cardiac rhythm
device implantation [18] to 77 min per patient for
biventricular pacing [10]. For the estimation of the staff
radiation burden per patient, a complex procedure
requiring 40 min of fluoroscopy was considered. The
total DAP of the examination was assumed to be
4300 cGy cm2 and the relative contribution of the PA,
RAO and LAO projection was 58%, 15% and 27%,
respectively [3]. Adopted fluoroscopy time and total
DAP value are similar to those derived from patient
studies performed at our institution regarding biventri-
cular pacing (35.2 min and 4765 cGy cm2) [19] and
radiofrequency catheter ablation (41 min) [3].
For the hypothesised complex procedure, staff doses
and workloads were calculated at three positions at the
left hand side of the patient and at the symmetrical
positions at the right hand side of the patient. Two sites
of interest adjacent to (12 cm from) the surgical table
were considered, at the level of the patient’s groin
(femoral area) and heart (subclavian position), while the
third position was selected 1 m away from the operating
table, also at the level of the patient’s heart.
The effect of projection ratios constituting a complex
procedure on staff effective dose values was investigated
by considering an alternative procedure in which the
roles between the LAO and RAO projections were
interchanged, i.e. the RAO contribution was increased
by a factor of 1.8 and the LAO contribution was
decreased by a factor of 0.56.
Calculation of workloads
Maximum annual permissible workloads were calcu-
lated on the basis of the ICRP recommendations regard-
ing effective dose (20 mSv per year), eye lens dose
(150 mSv per year) and skin dose (500 mSv per year) [12].
Results
Distribution of effective dose values per projection
The highest occupational exposure during a PA or
RAO projection is delivered next to the patient’s chest
at the left hand side (subclavian position), while during
a LAO projection the symmetrical site at the right
hand side of the patient receives the highest radia-
tion burden (Figure 1). The maximum values of effec-
tive dose to the electrophysiology laboratory staff,
645

protected by 0.5 mm lead-equivalent apron and thyroid
collar, are 14 nSv cGy21, 14.8 nSv cGy21 and
21
10.5 nSv cGy for the PA, RAO and LAO projection,
respectively.
Radiation burden from complex procedure
During an electrophysiological study, the highest
radiation levels occur adjacent to the patient’s heart. A
complex 40 min procedure yields between 54.2 mSv and
129 mSv depending on radiation protection measures
(Table 1). Effective dose values are considerably lower at
the right hand side of the patient where, at 1 m from the
table, they range from 3.38 mSv to 8.47 mSv, depending
on protective equipment used. The annual permissible
workload derived for the effective dose constraint is at
minimum for a staff member using only a 0.35 mm lead-
equivalent apron while occupying the site of highest
exposure, and equals 155 procedures. Since cancer risk
increases with effective dose and workloads decrease
with dose, their variation with staff positioning and
protection devices can be readily appreciated.
Interchanging roles between the LAO and RAO
projections, as in the alternative complex procedure
considered, would increase effective doses to staff
working adjacent to the tableside at the left hand side
of the patient by 10.1%, and reduce effective dose values
at the right hand side of the patient by 10.6%.
By analogy to effective dose, eye lens and face skin
dose present their peak values, 389 mSv and 446 mSv per
procedure, respectively, at the left hand side of the
patient (Table 2). Maximum permissible workloads
derived from the skin dose limit always exceed that
derived from the eye lens dose constraint, unless eye
protection is used.
Gonadal doses are also higher at the left hand side of
the patient. Moreover, male gonads receive approxi-
mately twice the dose delivered to female gonads. The
Figure 1. Distribution of effective dose in the electrophysiology laboratory to 0.5 mm lead-equivalent apron and collar
protected staff from (a) posteroanterior (PA), (b) right anterior oblique 25 ˚ (RAO) and (c) left anterior oblique 45 ˚ (LAO) heart
fluoroscopy. Provided values are dose–area product (DAP)-normalized and expressed in nSv cGy cm22.
646
N Theocharopoulos, J Damilakis, K Perisinakis et al
maximum estimated gonadal dose was 56.8 mSv per
procedure for the male staff using a 0.35 mm lead-
equivalent apron (Table 3).
Correlation between effective dose and dosemeter
reading
An over-apron dosemeter worn at chest height records
an Hp value that is from 3.6 to 27 times higher than the
effective dose actually received by the monitored staff
member in the electrophysiology laboratory, depending
on radiation protection equipment used and the tube
voltage (Table 4).
Effective dose and tube voltage
The penetrating ability of radiation through protective
apron and tissue increases with beam energy, and hence
effective dose values depend also on tube voltage. The
correction factors of Table 5 show the effect of voltage on
staff effective dose and can be used as multipliers to
adjust the data of Table 1 or Figure 1, which were
derived for 100 kVp.
The effect of removable protective shields
For the complex 3-projection procedure considered,
the effect of the removable under-apron protective
drapes on gonadal doses is markedly different from that
on effective dose values, within the confines of the
electrophysiology laboratory (Figure 2). Effective dose is
reduced up to 20% for staff members working at the left
hand side of the patient while for those working at the
contralateral side, reduction up to 40% occurs in the
vicinity of the patient’s groin and up to 50% at 0.5 m
from the couchside. The reduction in gonadal doses is
The British Journal of Radiology, August 2006
Occupational exposure in the electrophysiology laboratory

Table 1. Effective dose and maximum permissible annual workloads for electrophysiology laboratory staff, calculated for four
radiation protection conditions: (a) 0.35 mm lead-equivalent apron, (b) 0.35 mm lead-equivalent apron and 0.5 mm lead-
equivalent collar, (c) 0.5 mm lead-equivalent apron and (d) 0.5 mm lead-equivalent apron and 0.5 mm lead-equivalent collar
Side of the patient Position (coordinates)* Radiation protection Effective dose mSv Maximum permissible
working from measures per procedure{ workloads{

Procedures Gy m2

Left hand side Subclavian position (0,0) a 129 155 0.7

b 106 189 0.8
c 79.1 253 1.1
d 54.2 369 1.6
Femoral area (20.5,0) a 19.7 1015 4.4
b 13.7 1460 4.8
c 14.1 1418 6.2
d 7.82 2558 11.0
1 m from table (0,1) a 21.2 943 4.1
b 14.8 1353 5.8
c 15.1 1325 5.7
d 8.42 2375 10.3
Right hand side Subclavian position (0,0) a 67.8 295 1.3
b 54.2 369 1.6
c 42.9 466 2.0
d 28.2 709 3.1
Femoral area (20.5,0) a 18.4 1087 4.7
b 13.7 1460 6.3
c 12.4 1616 7.0
d 7.47 2677 11.6
1 m from table (0,1) a 8.47 2361 10.2
b 5.99 3339 14.5
c 5.98 3344 14.5
d 3.38 5917 25.6
*As in Figures 1 and 2. {4300 cGy cm2 at 100 kVp, 40 min of fluoroscopy. {Based on the 20 mSv annual limit on effective dose.

Table 2. Eye lens dose, face skin dose and corresponding maximum permissible annual workloads for the electrophysiology
laboratory staff
Side of the patient Position (coordinates) Eye lens Face skin
staff are working
from Eye protection Dose mSv per Maximum Dose mSv Maximum
procedure* permissible per procedure permissible
workloads{ workloads{
procedures procedures

Left hand side Subclavian position (0,0) Unprotected 389 386 446 1121

Goggles 19.4 7721

Femoral area (20.5,0) Unprotected 153 979 176 2844

Goggles 7.7 19587

1 m from table (0,1) Unprotected 125 1199 144 3481

Goggles 6.3 23972

Right hand side Subclavian position (0,0) Unprotected 194 773 223 2246

Goggles 9.7 15466

Femoral area (20.5,0) Unprotected 107 1402 123 4072

Goggles 5.3 28046

1 m from table (0,1) Unprotected 31.2 4809 35.8 13964

Goggles 1.6 96175

*4300 cGy cm2 at 100 kVp, 40 min of fluoroscopy. {Based on the 150 mSv annual limit on eye lens dose. {Based on the 500 mSv
annual limit on skin dose.

The British Journal of Radiology, August 2006 647

 N Theocharopoulos, J Damilakis, K Perisinakis et al

Table 3. Female and male gonadal doses for the electrophysiology laboratory staff protected by (a) 0.35 mm lead-equivalent
apron and (b) 0.5 mm lead-equivalent apron
Side of the patient Position (coordinates) Radiation protection Female staff Male staff
working from measures
Gonadal dose mSv Gonadal dose mSv
per procedure* per procedure

Left hand side Subclavian position (0,0) a 56.8 119

b 25.0 52.2
Femoral area (20.5,0) a 6.4 13.3
b 2.8 5.8
1 m from table (0,1) a 3.6 7.4
b 1.6 3.3
Right hand side Subclavian position (0,0) a 34.6 72.3
b 15.2 31.8
Femoral area (20.5,0) a 8.6 17.9
b 3.8 7.9
1 m from table (0,1) a 2.5 5.1
b 1.1 2.3
*4300 cGy cm2 at 100 kVp, 40 min of fluoroscopy.

also asymmetrical. However, protection is more effective impact on doses to tableside workers. However, if more
since staff gonadal doses are reduced by 90–98% at both accurate calculations are required, the isodose maps of
sides of the patient’s groin. It is evident that the under- Figure 1 and the individual DAP values of the constitu-
couch shields protect mostly the lower part of the body. ent projections may be used. Effective dose estimation
using the above methods is demonstrated in the
following example. A cardiologist has performed an
Discussion electrophysiology study positioned next to the patient’s
right hip (femoral area, x520.5, y50). The total DAP
value recorded during fluoroscopy at 85 kVp was
Applicability of provided dosimetric data
5000 cGy cm2 and the estimated relative contribution of
A recent study has shown that scattered radiation PA, LAO and RAO projections is 60%, 25% and 15%,
patterns from fluoroscopically exposed patients are respectively. From the data of Table 1 and the high
similar to those obtained from simulated exposures of voltage correction factors of Table 5, the effective dose
a RANDO phantom and that variations in the body (E) received by the cardiologist may be obtained as
habitus have a minor impact on scatter air kerma rates follows (1st method):
[20]. Hence, presented data allow the accurate estimation 7:47 mSv
of staff occupational doses and radiogenic risks in the E~ |5000 cGy cm2 |0:58~5:04 mSv
4300 cGy cm2
electrophysiology laboratory. Although not explicit,
doses of Tables 1–3 are in effect DAP-normalized, since More precisely (2nd method), the isodose curves of
a strictly defined complex procedure of 4300 cGy cm2 Figure 1 can be combined with Table 5 correction factors
was considered. Hence, staff doses in any institution to yield E:
from procedures similar in geometry to the procedure 1:5 nSv
adopted in the current study may be estimated on the E~( |5000 cGy cm2 |0:60|0:58)
cGy cm2
basis of the tabulated data, corrected for total DAP value.
Even major alterations in the ratios of the constituent 1:5 nSv
z( |5000 cGy cm2 |0:25|0:58)
components of fluoroscopic imaging will have minor cGy cm2
3 nSv
z( |5000 cGy cm2 |0:15|0:58)~5:00 mSv
cGy cm2
Table 4. Ratios of dosemeter reading* to effective dose, for
an over-apron chest dosemeter
Table 5. Correction factors (multipliers) for effective dose
Radiation protection measures Tube voltage (kVp)
calculation relative to data of Table 1 and Figure 1
70 85 100 120
Radiation protection measures Tube voltage (kVp)
No protection 0.39 0.47 0.53 0.56 70 85 120
0.35 mm lead-equivalent apron 9.4 7.6 5.6 3.6
0.35 mm lead-equivalent apron 15 10 6.8 4.7 0.35 mm lead-equivalent apron 0.52 0.72 1.57
and 0.5 mm lead-equivalent 0.35 mm lead-equivalent apron 0.36 0.62 1.84
collar and 0.5 mm lead-equivalent
0.5 mm lead-equivalent apron 13 11 9.1 6.2 collar
0.5 mm lead-equivalent apron 27 19 13 7.7 0.5 mm lead-quivalent apron 0.44 0.65 1.66
and 0.5 mm lead-equivalent 0.5 mm lead-equivalent apron 0.34 0.58 1.55
collar and 0.5 mm lead-equivalent
collar
*Hp (10)N.

648 The British Journal of Radiology, August 2006

Occupational exposure in the electrophysiology laboratory
A third, retrospective, method proposed for effective
dose estimation is based on personal dosemeter reading
and the conversion coefficients of Table 4. Suppose that
during an electrophysiological study performed at an
average voltage of 70 kVp, the over-apron dosemeter
positioned over a 0.35 mm lead-equivalent apron records
an Hp N value of 4.7 mSv. The actual effective dose
delivered will be:
dosemeter reading Hp N 4:7 mSv
E~ ~ ~0:5 mSv
9:4 9:4
The actual radiation risks
Occupational exposure in the electrophysiology labo-
ratory is most unlikely to incur deterministic effects to
staff. Not only are dose limits well below the threshold
of the induction of such effects, but also the fluoro-
scopy times required for the accumulation of thre-
shold doses are extremely high. The 2 Sv-threshold for
the induction of erythema [12] will be reached after at
least 3000 h of fluoroscopy, while cataract formation
occurs after at least 1700 h of fluoroscopy or 1 Sv. The
threshold dose for temporary sterility under prolonged
exposure is 0.4 Gy per year to male gonads [12], which
corresponds to 2240 h of fluoroscopy. Hence, the actual
concern for the laboratory staff is the induction of
stochastic effects, such as cancer and impairment to
progeny.
Factors affecting occupational doses
The quantity directly related to the risk of cancer
induction and used for the expression of occupational
dose limits is effective dose. Genetic risk, on the other
hand, calls for the estimation of gonadal doses to staff.
Our study has shown that, within a 1.5 m radius from
the tableside, effective dose varies by a factor of 40 due to
staff positioning and by a factor of 11 due to differences
in radiation protection measures. Gonadal doses present
even higher discrepancies. Hence, providing limited data
regarding scatter dose levels, or even effective doses,
without exact information on beam geometry, location or
The British Journal of Radiology, August 2006
Figure 2. Percent reduction of
(a) effective dose and (b) gonadal
dose delivered to electrophysiology
laboratory staff from the complex
three-projection procedure consi-
dered, accomplished with use of
removable undercouch protective
drapes.
protective devices, is of little value in developing a
radiation protection strategy. Mapping the electrophy-
siology laboratory in terms of effective dose and gonadal
dose for several beam geometries and different protec-
tive equipment is essential for the precise calculation of
staff radiogenic risk. Moreover, the use of any available
removable radiation barriers should be accurately
duplicated in a separate set of measurements, since there
is no other reliable method of predicting their efficiency
in dose reduction [21, 22].
Maximum permissible workloads
For each working site of interest, the overall maximum
permissible workload is the minimum of the three partial
workloads imposed by the constraints on eye lens, face
skin and effective dose. Since workloads depend on
positioning and protection measures, a simplified and
safe approach is to adopt the workload derived for the
most burdened position with the minimum protection of
a 35 mm lead-equivalent apron. Hence, no dose limit
will be violated if the maximum DAP value of fluoro-
scopy performed annually is less than 0.7 Gy m2. This
corresponds to 6200 min of fluoroscopy per year or 155
procedures.
Comparison with previous studies
Calkins et al [24] measured radiation exposure to
medical personnel during radiofrequency catheter abla-
tion using thermoluminescent dosimetry. Using the
mean exposure, they recorded per case at waist and
thyroid level (532 mSv and 156 mSv, respectively) and by
assuming a 0.5 lead-equivalent apron and collar, the
current study methodology yields a mean effective dose
value of 15.2 mSv for 44 min of fluoroscopy. Lindsay et al
[25] estimated an effective dose equivalent, which is the
predecessor of effective dose, of 18 mSv per case or
55 min of fluoroscopy when a thyroid shield is used, and
of 28 mSv per case without thyroid protection, for the
physician located at the femoral area. The effective dose
equivalent was approximately twice as high near the
649

subclavian position. Regarding eye lens dose, Calcins
et al [24] reported 281 mSv per case, Vano et al [26]
294 mSv and Kuon et al [21] more than 165 mSv per year
for an annual workload of 1000 interventions. Given
the multiplicity and complexity of factors affecting
occupational doses, reported data are in broad agree-
ment with the current study estimates. It should be
stressed, however, that the vast majority of previously
reported occupational doses are based on fluoroscopy
duration observations and not on DAP. Hence, they
are strongly dependent on fluoroscopy technique factors
and equipment used, and cannot serve the purpose of
accurate dose estimation in other electrophysiology
laboratories.
Mitigation of radiation risks
Two easy-to-apply practices are recommended for
minimizing personnel exposure in the electrophysiology
laboratory. Approaching the patient from the right hand
side, specifically at the groin level, is preferable since the
radiation backscatter effect is less pronounced. Also,
reducing tube voltage and milliamperage is beneficial for
staff doses in two ways: the total DAP per patient is
diminished and staff effective dose values per unit DAP
fall as the beam energy decreases. Personal protection is
also extremely important. A 0.35 mm lead-equivalent
apron reduces effective dose to staff by a factor of 10,
compared with the unshielded values. Additional use of
a 0.5 mm thyroid collar provides a further 1.5-fold
decrease. Use of both 0.5 mm lead-equivalent apron
and collar provides a protection factor of 26.
Furthermore, operators should avoid or limit the use of
LAO projection when positioned on the right side of the
patient and of RAO projection when positioned on the
left [22]. Other good fluoroscopy practices, such as strict
beam collimation use of pulsed fluoroscopy and last
frame hold feature, act to reduce radiation administered
per patient and keep staff radiation doses as low as
reasonably achievable [21, 27, 28]. However, the degree
of radiation reduction depends on specific character-
istics of the equipment used. Therefore, optimization of
imaging protocols should be based on a comprehen-
sive evaluation of the dosimetric characteristics and
the performance of automatic exposure control of the
employed fluoroscopic unit, preferably in terms of DAP
rate [29].
Limitations of the study
The major limitations of the present study are that not
all of the beam geometries possibly encountered in
electrophysiological studies were simulated, and that no
ceiling-suspended radiation barriers were available in
our institution. However, the selected projections are the
major constituents of any cardiac study. Moreover, the
effect of commercially available removable radiation
barriers should be individually examined at each
institution, according to their specific use, with extensive
radiation measurements regarding beam angulation and
staff/barrier positioning. Furthermore, since the distri-
bution and attenuation of scatter is likely to be
650
N Theocharopoulos, J Damilakis, K Perisinakis et al
influenced by the volume of tissue lateral to the
scattering site, real patients, who may differ considerably
in size compared with the RANDO phantom, may
produce a somewhat different scatter pattern.
Conclusions
Radiation hazards in the electrophysiology laboratory
should not be overstated nor ignored. Continuous risk
assessment and minimization is required. Good fluoro-
scopy and radiation protection practices to mitigate
occupational exposure to radiation should be proposed
and implemented.
References
1. Calkins H, Canby R, Weiss R, Taylor G, Wells P, Chinitz L,
et al. 100W Atakr II Investigator Group. Results of catheter
ablation of typical atrial flutter. Am J Cardiol
2004;94:437–42.
2. Lickfett L, Mahesh M, Vasamreddy C, Bradley D, Jayam V,
Eldadah Z, et al. Radiation exposure during catheter
ablation of atrial fibrillation. Circulation 2004;110:3003–10.
3. Perisinakis K, Damilakis J, Theocharopoulos N, Manios E,
Vardas P, Gourtsoyiannis N. Accurate assessment of patient
effective radiation dose and associated detriment risk from
radiofrequency catheter ablation procedures. Circulation
2001;104:58–62.
4. Leclercq C, Hare JM. Ventricular resynchronization: current
state of the art. Circulation 2004;109:296–9.
5. Roguin A, Bomma CS, Nasir K, Tandri H, Tichnell C, James
C, et al. Implantable cardioverter-defibrillators in patients
with arrhythmogenic right ventricular dysplasia/cardio-
myopathy. J Am Coll Cardiol 2004;3:1843–52.
6. Leclercq C, Kass DA. Retiming the failing heart: principles
and current clinical status of cardiac resynchronization. J
Am Coll Cardiol 2002;39:194–201.
7. Boehmer JP. Device therapy for heart failure. Am J Cardiol
2003;91:53D–9D.
8. Lee KL, Hafley G, Fisher JD, Gold MR, Prystowsky EN,
Talajic M, et al. Multicenter Unsustained Tachycardia Trial
Investigators. Effect of implantable defibrillators on
arrhythmic events and mortality in the multicenter unsus-
tained tachycardia trial. Circulation 2002;106:233–8.
9. Izutani H, Quan KJ, Biblo LA, Gill IS. Biventricular pacing
for congestive heart failure: early experience in surgical
epicardial versus coronary sinus lead placement. Heart
Surg Forum 2002;6:E1–6.
10. Balter S. Stray radiation in the cardiac catheterisation
laboratory. Radiat Prot Dosim 2001;94:183–8.
11. Vano E. Radiation exposure to cardiologists: how it could
be reduced. Heart 2003;89:1123–4.
12. International Commission on Radiological Protection.
Recommendations of the International Commission on
Radiological Protection, ICRP Publication 60. Ann ICRP
1991;21(1/3).
13. International Commission on Radiological Protection.
Conversion coefficients for use in radiological protection
against external radiation, ICRP Publication 74. Ann ICRP
1996;26(3/4).
14. International Commission on Radiation Units and
Measurements. Measurement of dose equivalents from
external photon and electron radiations, ICRU Report 47.
Bethesda, MD: ICRU, 1992.
15. Rawlings DJ, Faulkner K, Harrison RM. Broad-beam
transmission data in lead for scattered radiation produced
at diagnostic energies. Br J Radiol 1991;64:69–71.
The British Journal of Radiology, August 2006
Occupational exposure in the electrophysiology laboratory
16. Niklason LT, Marx MV, Chan HP. The estimation of
occupational effective dose in diagnostic radiology with
two dosimeters. Health Phys 1994;67:611–5.
17. Theocharopoulos N, Perisinakis K, Damilakis J,
Papadokostakis G, Hadjipavlou A, Gourtsoyiannis N.
Occupational exposure from common fluoroscopic projec-
tions used in orthopaedic surgery. J Bone Joint Surg Am
2003;85:1698–703.
18. Wiegand UKH, Bode F, Bonnemeier H, Eberhard F, Schlei
M, Peters W, et al. Long-term complication rates in
ventricular single lead VDD and dual chamber pacing.
Pacing Clin Electrophysiol 2003;26:1961–9.
19. Perisinakis K, Theocharopoulos N, Damilakis J, Manios E,
Vardas P, Gourtsoyiannis N. Fluoroscopically guided
implantation of modern cardiac resynchronization devices:
radiation burden to the patient and associated risks. J Am
Coll Cardiol 2005;46:2335–9.
20. Damilakis J Perisinakis K, Theocharopoulos N, Manios E,
Vardas P, Gourtsoyiannis N. Anticipation of radiation
dose to the conceptus from occupational exposure of
pregnant staff during fluoroscopically guided electrophy-
siological procedures. J Cardiovasc Electrophysiol 2005;16:
773–80.
21. Kuon E, Schmitt M, Dahm JB. Significant reduction of
radiation exposure to operator and staff during cardiac
interventions by analysis of radiation leakage and
improved lead shielding. Am J Cardiol 2002;9:44–9.
The British Journal of Radiology, August 2006
22. Kuon E, Dahm JB, Empen K, Robinson DM, Reuter G,
Wucherer M. Identification of less-irradiating tube angula-
tions in invasive cardiology. J Am Coll Cardiol
2004;44:1420–8.
23. Hall EJ. Radiobiology for the radiologist. Philadelphia, PA:
JB Lippincott Co, 1994:363–78.
24. Calkins H, Niklason L, Sousa J, el-Atassi R, Langberg J,
Morady F. Radiation exposure during radiofrequency
catheter ablation of accessory atrioventricular connections.
Circulation 1991;84:2376–82.
25. Lindsay BD, Eichling JO, Ambos HD, Cain ME. Radiation
exposure to patients and medical personnel during radio-
frequency catheter ablation for supraventricular tachycar-
dia. Am J Cardiol 1992;70:218–23.
26. Vano E, Gonzalez L, Guibelalde E, Fernandez JM, Ten JI.
Radiation exposure to medical staff in interventional and
cardiac radiology. Br J Radiol 1998;71:954–60.
27. Balter S. Radiation safety in the cardiac catheterization
laboratory: operational radiation safety. Catheter
Cardiovasc Interv 1999;47:347–53.
28. Kuon E, Birkel J, Schmitt M, Dahm JB. Radiation exposure
benefit of a lead cap in invasive cardiology. Heart
2003;89:1205–10.
29. Kuon E, Dorn C, Schmitt M, Dahm JB. Radiation dose
reduction in invasive cardiology by restriction to adequate
instead of optimized picture quality. Health Phys
2003;84:626–31.
651
 The British Journal of Radiology, 79 (2006), 652–658

Usefulness of diffusion/perfusion-weighted MRI in patients with

non-enhancing supratentorial brain gliomas: a valuable tool to
predict tumour grading?
G G FAN, PhD, MD, Q L DENG, MD, Z H WU, MD and Q Y GUO, PhD, MD

Department of Radiology, Second Hospital of China Medical University, No.36 Sanhao St., Heping
Dist., Shenyang, Liaoning,110004, People’s Republic of China

ABSTRACT. 22 patients with non-enhancing supratentorial gliomas on contrast-

enhanced MRI underwent both diffusion- and perfusion-weighted MRI (DWI/PWI)
before surgical resection. 14 low-grade gliomas (WHO Grade I and II) and 8 anaplastic
gliomas were verified histologically. Both apparent diffusion coefficient (ADC) values
and relative cerebral blood volume (rCBV) ratios were calculated on the solid portion of
the tumour, on peritumoural area, as well as on the contralateral normal white matter,
respectively. The results showed that lower ADC values were present in the solid
portions of anaplastic gliomas, but not in low grade (p,0.01). All ADC values in
peritumoural regions of tumours were decreased compared with the contralateral
normal white matter. However, there was no significant difference between anaplastic
gliomas and low-grade gliomas. Meanwhile, higher rCBV ratios were present in both Received 9 January 2006
solid portions and peritumoural regions of anaplastic gliomas, but not in low grade Revised 27 January 2006
gliomas (p,0.01). In conclusion, non-enhancing brain gliomas with lower ADC values in Accepted 2 February 2006
the solid portions and higher rCBV ratios in both solid portions and peritumoural
DOI: 10.1259/bjr/25349497
regions of tumours are significantly correlated with anaplasia. Therefore, DWI and PWI
should be integrated in the diagnostic work-up of non-enhancing gliomas in order to ’ 2006 The British Institute of
predict grading. Radiology

Gliomas are the most common primary neoplasms of
the brain, varying histologically from low grade to high
grade [1]. MRI plays a crucial role in the evaluation of
patients with gliomas [2]. The use of gadolinium-based
contrast agents yields further improvement in the
demonstration and detection of cerebral gliomas.
Patterns and extent of contrast enhancement have been
suggestive of a malignant potential [3]. However, this
approach is limited because 14–45% of non-enhancing
supratentorial gliomas are malignant (especially in older
patients) and some enhancing gliomas (i.e. pilocytic
astrocytoma) are benign [4, 5]. Moreover, large cerebral
gliomas are often histopathologically heterogeneous and
may have components of varying grades of malignancy
within them. Hence, accurate pre-operative grading of
gliomas and planning of adequate treatment strategies
are often difficult with conventional MRI [5, 6].
Diffusion-weighted MRI (DWI), which is sensitive to
the molecular diffusion of water, has been well estab-
lished as a reliable non-invasive method for the early
detection of cerebral ischaemic stroke, and DWI has been
reported to be helpful in differentiating necrotic cavities
associated with malignant gliomas from the benign ones
[7]. The use of DWI to better characterize enhancing
tumours and vasogenic oedema has been explored, but
the results obtained have been conflicting [7, 8]. Recent
developments in perfusion-weighted MRI (PWI) techni-
ques have permitted the creation of relative cerebral
blood volume (rCBV) maps, leading to the qualitative
and quantitative assessment of tumour vascularity.
These maps have helped in the assessment of tumour
grade and in targeting the site of biopsy [8, 9].
Although DWI and PWI have been widely used in pre-
operative grading of gliomas, to the best of our knowl-
edge, few studies evaluating the usefulness of diffusion
and perfusion MRI solely in non-enhancing gliomas have
been reported, although the utility of perfusion MRI in
gliomas without enhancement on conventional contrast-
enhanced MRI was performed by Maia et al [10]. In
addition, the efficiency of peritumour region in the grade
assessment of cerebral gliomas is still being investigated.
Our hypothesis was that DWI/PWI could provide
additional useful information in the assessment and
tumour grading of supratentorial glial neoplasms,
which lacked contrast enhancement on pre-operative
neuroimaging.
Materials and methods
Patient selection and clinical data collection
22 patients (10 women and 12 men; median age
48.3 years, age range 38–68 years) with non-enhancing
supratentorial gliomas on contrast-enhanced MRI under-
went both diffusion/perfusion-weighted MRI (DWI/
PWI) before surgical resection. Informed consent was
obtained from all patients prior to the investigation, and
all procedures were performed under the approval of
our institutional review board for clinical studies. 14 low

652 The British Journal of Radiology, August 2006

Diffusion/perfusion weighted MRI and supratentorial gliomas

grade gliomas (WHO Grade I and II) and 8 high grade
(WHO Grade III) were verified histologically. Low grade
gliomas consisted of low grade astrocytomas (Grade I,
n57), low grade oligodendrogliomas (Grade II, n54) and
low grade mixed oligoastrocytomas (Grade II, n53);
while high grade gliomas consisted of anaplastic astro-
cytomas (Grade III, n55) and anaplastic oligodendro-
gliomas (Grade III, n53).
MRI examination
MRI examinations were performed on 1.5 T super-
conduction whole-body MR system (Intera Gyroscan;
Philips Medical Systems, Best, The Netherlands) and an
eight-channel Sensitivity Encoding (SENSE) head coil.
After scout view MRI, the examination protocol consists
of pre-contrast conventional MRI followed by DWI, PWI,
and finally post-contrast T1 weighted images.
Conventional MR images were obtained with T1 (spin
echo (SE), 442/15) and turbo T2 (turbo spin echo (TSE),
3235/100) weighted spin echo sequences (both with a
2566192 matrix, 6 mm slice thickness and 2 averages).
DWIs were acquired using single-shot echo-planar
imaging (EPI) sequence at multiple levels. 18 slices of
6 mm thickness were obtained (repetition time (TR)
5000 ms, echo time (TE) 104 ms, field of view 40620,
matrix size 2566128, b values of 0 and 1000 mm2 s21) in
3 orthogonal directions.
For susceptibility-based PWI, the transitory signal loss
during the bolus passage was detected with a T2*
weighted fast field echo (FFE) EPI sequence (TR
232 ms, TE 25 ms, 6 slices with 6 mm slice thickness;
matrix, 896128; and 1 average). 40 dynamic scans with a
time resolution of 1.6 s per image were performed after
intravenous bolus injection of 20 ml Gd-DTPA
(Magnevist; Schering AG) at a flow rate of 4 ml s21
and a 20 ml saline flush.
cover the region of expected maximum rCBV on maps of
rCBV. Additionally, we measured the ratio of maximum
rCBV in either the tumour or peritumoural area of that in
the contralateral normal white matter in order to
standardize variations in each examination.
All data obtained were summarized as the mean ¡
standard deviation (SD). The Student t-test was used to
determine if there were statistically significant differ-
ences in both averaged ADC value and maximum rCBV
ratio between anaplastic gliomas (Grade III) and low
grade (Grade I and Grade II) gliomas. A p-value of
less than 0.05 was considered to indicate statistical
significance.
Results
Conventional MRI revealed heterogeneous signal
intensity of tumours with clear evidence of central
necrosis in all 22 patients. Of all these patients, the solid
tumour tissue exhibited typical hyperintense on T2
weighted images (T2WI) and hypointense on T1
weighted images (T1WI) (Figure 1). Six of the seven
patients with Grade II tumours and all eight patients
with Grade III tumours (Figure 2) were found to have
peritumoural oedema and mass effect. Infiltration of the
tumour producing nodular thickening of the grey matter
was seen in three patients with Grade II tumours and
five patients with Grade III tumours (Figures 3 and 4).
All tumours, however, did not produce significant

MR data analysis and statistics

The apparent diffusion coefficient (ADC) maps and
values were calculated on a separate workstation (Easy
Vision Intera workstation, release 8.1.3; Philips Medical
Systems). We recorded the ADC values from both the
solid portion of the tumour (seen as the highest signal
intensity lesion at b value 0 and 1000) and peritumoural
area. The ADC values in our study represented averaged
ADCs of three to five regions of interest (ROIs). A ROI,
varying from 40 mm3 to 60 mm3, was positioned care-
fully to avoid contamination from adjacent different
tissues. The ROI was drawn as large as possible using a
circular ROI on the workstation. As a control, the ADC
was obtained from contralateral normal white matter.
Raw PWIs were transferred to a PC workstation (Easy Figure 1. MRI and diffusion weighted imaging (DWI) from a
Vision Intera workstation, release 8.1.3; Philips Medical patient with anaplastic astrocytoma. (a) T2 weighted image.
(b) T1 weighted MR image with contrast material. (c) DWI
Systems) for post-processing. With the aid of the
(b51000). (d) Apparent diffusion coefficient (ADC). The
implemented software, the rCBV could be calculated tumour exhibited typical hyperintense on T2 weighted
on the basis of the indicator dilution method and were images; peritumoural oedema and mass effect was present;
displayed as spectral colour images. ROI analyses were no contrast enhancement. On DWIs (b51000), the signal
performed on the solid portion of the tumour, on intensity in the solid portion of the tumour was hyperintense
peritumoural area, as well as on the contralateral normal with respect to the white matter; lower ADC values were
white matter, respectively. ROI was placed carefully to present in the solid portions of high grade gliomas.

The British Journal of Radiology, August 2006 653


contrast enhancement on visual assessment, suggesting
lack of blood–brain barrier breakdown.
On DWIs (b50, 1000), the signal intensity in the solid
portion of the tumour was hyperintense with respect to
the white matter (Figures 1, 3 and 5). The averaged ADC
values for the solid tumour component ranged from
Figure 3. MRI and diffusion weighted imaging (DWI) from a
patient of anaplastic astrocytoma. (a) T2 weighted image.
(b) T1 weighted MR image with contrast material. (c) DWI
(b51000). (d) Apparent diffusion coefficient (ADC). The
tumour exhibited typical hyperintensity on T2 weighted
images; infiltration of the tumour producing nodular
thickening of the grey matter was observed. The tumour
did not produce significant contrast enhancement on visual
assessment, On DWIs (b51000), the signal intensity in the
solid portion of the tumour was hyperintense with respect to
the white matter; lower ADC values were present in the solid
portions of high grade gliomas.
654
G G Fan, Q L Deng, Z H Wu and Q Y Guo
Figure 2. Perfusion weighted ima-
ging (PWI) and histopathological
photograph from a patient with
anaplastic astrocytoma (same
patient as in Figure 1). (a) Relative
cerebral blood volume (rCBV) colour
map. (b) Signal-intensity time-curve.
(c) Histopathological photograph.
rCBV maps were inhomogeneous
with various increases of signal
intensity in both solid portion and
peritumoural region of tumour.
Histopathological photograph of
tumour confirmed the diagnosis of
anaplastic astrocytoma.
0.4361023 mm2 s21to 1.3961023 mm2 s21 (mean
0.89¡0.3661023 mm2 s21) and for white matter from
1.0261023 mm2 s21to 1.7761023 mm2 s21 (mean
1.40¡0.3561023 mm2 s21) (Table 1).
Lower ADC values were present in the solid portions
of high grade gliomas (Grade III) (0.52¡0.1161023 mm2
s21), but not in low grade (Grade I and Grade II)
(1.15¡0.1661023 mm2 s21); the difference was statisti-
cally significant (p,0.01). All ADC values in peri-
tumoural regions of tumours were decreased compared
with the contralateral normal white matter; however,
there was no significant difference between high grade
(0.69¡0.1361023 mm2 s21) and low grade gliomas
(0.79¡0.0761023 mm2 s21) (p.0.05).
The signal intensity in the peritumoural region was
homogeneous on rCBV map in all 14 cases of low-grade
(Grade I and Grade II) tumours (Figure 6). However, a
slight increase of signal intensity of the tumour was
observed in three cases of Grade II tumours and in one
case of Grade I tumour. In contrast, all rCBV maps for
Grade III tumours were inhomogeneous with various
increases of signal intensity in solid portions of tumour
(Figures 2 and 4).
Table 2 summarizes the measurements of maximum
relative rCBV ratios of all non-enhancing gliomas.
Measured maximum rCBV ratio in the solid portion of
tumour varied from 0.76 to 5.01, with a mean of
2.02¡0.73 (¡SD); whereas in peritumoural region,
maximum rCBV ratio varied from 0.92 to 3.55, with a
mean of 1.79¡1.03 (¡SD). Higher maximum rCBV
ratios were present in both solid portions and peritu-
moural regions of anaplastic gliomas (3.27¡1.54;
2.92¡0.44), but not in low grade (Grade I and Grade
II) (1.52¡0.49; 1.36¡0.83) (p,0.01).
Discussion
MRI plays a critical role in the pre-operative assess-
ment of brain gliomas. Mass effect, cyst formation and
necrosis on MRI studies do correlate significantly with
The British Journal of Radiology, August 2006
Diffusion/perfusion weighted MRI and supratentorial gliomas
malignant behaviour [2, 3, 4, 6]. Traditionally, the extent
of contrast enhancement has been used as a mark of
malignancy: most high-grade gliomas on post-contrast T1
weighted images generally show moderate to strong
enhancement; on the other hand, the low-grade gliomas
have minimal or no enhancement [6, 10]. However, lack
of contrast enhancement on MRI studies does not equate
with low tumour grade. In fact, a broad spectrum of
histological types may present as non-enhancing lesions
[5]. Even after contrast administration, up to 25% of high-
grade gliomas may show faint or no detectable enhance-
ment, such as our cases (8 of 22 cases of anaplastic
gliomas have been confirmed) [6, 11]. In addition, the
Figure 5. MRI and diffusion weighted imaging (DWI) from a
patient with low-grade oligodendrogliomas. (a) T1 weighted
image. (b) T 2 weighted image. (c) DWI (b51000).
(d) Apparent diffusion coefficient (ADC). The tumour
exhibited inhomogeneous with various increase of signal
intensity on both T2 weighted image and DWI.
The British Journal of Radiology, August 2006
Figure 4. Perfusion weighted ima-
ging (PWI) and histopathological
photograph from a patient with
anaplastic astrocytoma (same
patient as in Figure 3). (a) Relative
cerebral blood volume (rCBV) colour
map. (b) Signal-intensity time-curve.
(c) Histopathological photograph.
rCBV maps were inhomogeneous
with various increases of signal
intensity in solid portion of tumour.
Histopathological photograph of
tumour confirmed the diagnosis of
anaplastic astrocytoma.
risk of anaplasia in non-enhancing lesions on MRI
increases significantly with the patient’s age [12].
Therefore, in the case of tumours that do not enhance,
pre-operative evaluation of tumour grade is often
difficult on conventional MRI.
DWI has been used by some to evaluate intra-axial
tumours [7, 8, 13]. DWI and calculation of ADC values
have been used to distinguish the normal white matter
areas from necrosis, cyst formation, oedema, and solid
tumour by measuring differences in ADC values caused
by water proton mobility alterations [13, 14]. These
differences are thought to result from both changes in the
balance between intracellular and extracellular water
and changes in the structure of the two compartments
[14]. Our result showed that calculated ADC values from
tumoural core added more information to MRI in the
differentiation and grading of non-enhancing gliomas;
these results suggested different ADC values due to
different tumour grades [15]. Although ADC values of
biological tissue are determined by many factors, results
of previous studies have already confirmed that the
lower ADC values in tumour core were mainly affected
by tumour cellularity. With higher diffusivity found in
the extracellular volume, the increase of intracellular
space due to highly cellular tissue is coupled with a
decrease of the ADC [8, 15, 16]. Therefore, higher
cellularity in anaplastic neoplasm would contribute to
the lower ADC values.
Contrary to most previous studies, averaged ADC
values in peritumoural area were also analysed in our
Table 1. Apparent diffusion coefficient values comparison
of non-enhancing gliomas with contralateral white matter
(WM) (61023 mm2 s21)
Grade Cases Solid tumoural Peritumoural Contralateral
region region WM
I 7 1.17¡0.14 0.83¡0.16 1.44¡0.32
II 7 1.01¡0.17 0.74¡0.09 1.35¡0.28
III 8 0.52¡0.11* 0.69¡0.13 1.40¡0.47
Data are the mean¡SD; * is statistical significance (p,0.01).
655

study. The results showed that ADC values in peri-
tumoural regions were decreased compared with the
contralateral normal white matter. It is assumed that
malignant gliomas are not strictly focal lesions, but
rather are characterized by intracerebral dissemination of
malignant glial cells along the myelinized axons and
blood vessels, or through the subarachnoid space [8]. We
therefore speculated that there might be a difference in
the ADC values of the peritumoural region between low
grade and high grade gliomas. However, no significant
difference was found. We postulate that the ADC value
in the peritumoural region is unreliable as a means of
differentiating between high grade and low grade
gliomas, because a partial volume effect contaminated
by surrounding oedema may affect the accurate outcome
[8, 17]. Recently, usefulness of diffusion tensor imaging
(DTI) in the study of peritumoural region of gliomas has
been investigated and the result seems promising [18].
DTI, which is capable of visualizing the anisotropy of
proton motion, may make it possible to eliminate the
partial volume effect of peritumoural oedema, and the
difference between high-grade and low-grade gliomas in
this respect seems worthy of evaluation.
For planning the optimal treatment strategy, accurate
determination of tumour grade is critical, and in most
histological grading systems, vascular proliferation of
gliomas is a diagnostic criterion for malignancy [19].
New blood vessel growth is a critical phase of solid
tumour growth. The growth of a solid tumour mass at 1–
2 mm3 depends upon simple diffusion of oxygen,
Table 2. Maximum relative cerebral blood volume (rCBV)
ratio in both tumoural and peritumoural regions of non-
enhancing gliomas
Grade Cases Solid tumoural Peritumoural
region region
I 7 1.09¡0.26 1.27¡0.33
II 7 1.91¡0.78 1.54¡0.60
III 8 3.27¡1.54* 2.92¡0.44*
Data are the mean¡SD; * is statistical significance (p,0.01).
656
G G Fan, Q L Deng, Z H Wu and Q Y Guo
Figure 6. Perfusion weighted ima-
ging (PWI) and histopathological
photograph from a patient with
low-grade oligodendrogliomas (same
patient as in Figure 5). (a) Relative
cerebral blood volume (rCBV) colour
map. (b) Signal-intensity time-curve.
(c) Histopathological photograph.
The signal intensity in both peri-
tumoural region and solid portion of
tumour was homogeneous on rCBV
map. Histopathological photograph
of tumour confirmed the diagnosis of
oligodendrogliomas (Grade II).
nutrients and other essential materials. However, tumour
mass growth over 1–2 mm3 can not occur and metastasise
without angiogenesis [20]. Although conventional MRI
with gadolinium-based contrast enhancement has been
useful for grading gliomas, contrast enhancement itself
reflects disruption of the blood–brain barrier, not tumour
angiogenesis [3, 5, 6]. PWI techniques now have been used
for the assessment of tumour vascularity in vivo. rCBV
maps and measurements have been shown to correlate
reliably with tumour grade and histological findings of
tumoural microvascular density, the current standard for
assessing the degree of angiogenesis [8–11].
The correlation between the histopathological grade of
cerebral gliomas and rCBV has already been evaluated
by various groups. Although these studies showed a
wide range of rCBV ratios and overlapping between
tumours of different grades, there were statistically
significant differences between high grade and low
grade gliomas [8–11], even in the study of non-enhancing
gliomas, as our results confirm. In our study, non-
enhancing anaplastic gliomas often demonstrated sig-
nificant heterogeneity and areas of high rCBV. The
presence of contrast enhancement on conventional
MRI only represents a pathological alteration in the
blood–brain barrier (with or without concomitant angio-
genesis), whereas the degree of perfusion MR abnor-
mality can truly reflect the degree of angiogenesis (with
or without destruction of the blood–brain barrier) [6, 21].
Therefore, the advantage of perfusion MRI over contrast-
enhanced MRI is in depicting tumour angiogenesis and
hence in pre-operative grading. Moreover, because large
cerebral gliomas are often histopathologically hetero-
geneous, areas with higher rCBV values, which may be
regarded as greater tumour vascularity, can be selec-
tively targeted by stereotactic biopsies to reduce tumour
under-grading [6, 22]. This is especially true for non-
enhancing gliomas with relative intact blood–brain
barrier [10]. On the contrary, for high grade enhancing
gliomas with concomitant breakdown of the blood–brain
barrier, the first pass of contrast material may leak
into extravascular space, and thus the produced
The British Journal of Radiology, August 2006
Diffusion/perfusion weighted MRI and supratentorial gliomas
susceptibility effects may be decreased between intra-
vascular and extravascular space near the disrupted
blood–brain barrier, which is considered to cause the
underestimation of the true tumour vascularity [8].
In contrast to measuring average rCBV within tumoural
core alone, the measurement of maximum rCBV in
peritumoural regions was also performed in our study.
The results showed that elevated rCBV ratios were also
present in peritumoural brain regions in high grade
gliomas, suggests increased peritumoural perfusion due
to tumour infiltration. In anaplastic tumours, peritumoural
areas demonstrate not only altered capillary morphologic
findings but also scattered tumour cells infiltrating along
newly formed or pre-existing but dilated vascular chan-
nels [8, 22, 23]. In low grade gliomas, on the other hand, the
peritumoural region contains less infiltrating tumour cells.
This interpretation is consistent with elevated blood
volume preceding the appearance of enhancement, which
reflects blood–brain barrier breakdown [22]. The fact that
we found comparable elevated rCBV in the peritumoural
area of grade III gliomas is both novel and significant:
Information regarding heterogeneity of peritumoural
region in terms of vascularity as depicted by PWI can be
effectively used to best estimate of the true brain tumour
size pre-operatively [24].
In PWI study, we chose spin-echo echo-planar images
because of their presumed higher sensitivity in detecting
tumour vascularity at capillary level than at large vessel
level [10]. Two different sequences, including spin echo
and gradient-echo echo-planar sequences are generally
used in first-pass perfusion MR study [21]. Because the
gradient-echo echo-planar technique is sensitive to
susceptibility effects from the total volume of blood
contained in both capillaries and large vessels, we prefer
spin echo echo-planar sequences in PWI study of non-
enhancing gliomas in order to eliminate the interference
of susceptibility artefact [25].
In conclusion, usefulness of DWI and PWI in non-
enhancing cerebral gliomas is not only feasible, but also
offers clinically relevant physiological data not obtain-
able by conventional MRI. Non-enhancing brain gliomas
with lower ADC values in the solid portions and higher
rCBV ratios in both solid portions and peritumoural
regions of tumours are significantly correlated with
anaplasia. Therefore, DWI and PWI should be integrated
in the diagnostic work-up of non-enhancing gliomas in
order to predict grading. However, with the advent of
advanced MR techniques, a more sophisticated study
using a larger sample is needed in the near future.
Acknowledgments
We wish to thank Dr Bing Yu and Dr Songmin Quan
for their expert neuroradiological opinion and assistance,
and Prof. Liying Chen for helpful advice and discussion.
The authors also thank Department of Neurosurgery
of China Medical University for their fellowship.
References
1. Daumas-Duport C, Scheithauer B, O’Fallon J, Kelly P.
Grading of astrocytomas: a simple and reproducible
method. Cancer 1988;62:2152–65.
The British Journal of Radiology, August 2006
2. Rees J. Advances in magnetic resonance imaging of brain
tumours. Curr Opin Neurol 2003;16:643–50.
3. Kurki T, Lundbom N, Kalimo H, Valtonen S. MR
classification of brain gliomas: value of magnetization
transfer and conventional imaging. Magn Reson Imaging
1995;13:501–11.
4. Mihara F, Numaguchi Y, Rothman M, Kristt D, Fiandaca M,
Swallow L. Non-enhancing supratentorial malignant astro-
cytomas: MR features and possible mechanisms. Radiat
Med 1995;13:11–7.
5. Scott JN, Brasher PM, Sevick RJ, Rewcastle NB, Forsyth PA.
How often are nonenhancing supratentorial gliomas malig-
nant? A population study. Neurology 2002;59:947–9.
6. Batra A, Tripathi RP, Singh AK. Perfusion magnetic
resonance imaging and magnetic resonance spectroscopy
of cerebral gliomas showing imperceptible contrast
enhancement on conventional magnetic resonance imaging.
Australas Radiol 2004;48:324–32.
7. Tien RD, Felsberg GJ, Friedman H, Brown M, MacFall J. MR
imaging of high-grade cerebral gliomas: value of diffusion-
weighted echoplanar pulse sequences. AJR Am J
Roentgenol 1994;162:671–7.
8. Fan G, Zang P, Jing F, Wu Z, Guo Q. Usefulness of
diffusion/perfusion-weighted MRI in rat gliomas: correla-
tion with histopathology. Acad Radiol 2005;12:640–51.
9. Law M, Yang S, Babb JS, et al. Comparison of cerebral blood
volume and vascular permeability from dynamic suscept-
ibility contrast-enhanced perfusion MR imaging with
glioma grade. AJNR Am J Neuroradiol 2004;25:746–55.
10. Maia AC Jr, Malheiros SM, da Rocha AJ, et al. MR cerebral
blood volume maps correlated with vascular endothelial
growth factor expression and tumor grade in nonenhancing
gliomas. AJNR Am J Neuroradiol 2005;26:777–83.
11. Law M, Yang S, Wang H, et al. Glioma grading: sensitivity,
specificity, and predictive values of perfusion MR imaging
and proton MR spectroscopic imaging compared with
conventional MR imaging. AJNR Am J Neuroradiol
2003;24:1989–98.
12. Bruner JM. Neuropathology of malignant gliomas. Semin
Oncol 1994;21:126–38.
13. Bulakbasi N, Guvenc I, Onguru O, Erdogan E, Tayfun C,
Ucoz T. The added value of the apparent diffusion
coefficient calculation to magnetic resonance imaging in
the differentiation and grading of malignant brain tumors. J
Comput Assist Tomogr 2004;28:735–46.
14. Bulakbasi N, Kocaoglu M, Ors F, Tayfun C, Ucoz T.
Combination of single-voxel proton MR spectroscopy and
apparent diffusion coefficient calculation in the evaluation
of common brain tumors. AJNR Am J Neuroradiol
2003;24:225–33.
15. Hein PA, Eskey CJ, Dunn JF, Hug EB. Diffusion-weighted
imaging in the follow-up of treated high-grade gliomas:
tumor recurrence versus radiation injury. AJNR Am J
Neuroradiol 2004;25:201–9.
16. Sugahara T, Korogi Y, Kochi M, et al. Usefulness of
diffusion-weighted MRI with echo-planar technique in the
evaluation of cellularity in gliomas. J Magn Reson Imaging
1999;9:53–60.
17. Pauleit D, Langen KJ, Floeth F, et al. Can the apparent
diffusion coefficient be used as a noninvasive parameter to
distinguish tumor tissue from peritumoral tissue in cerebral
gliomas? J Magn Reson Imaging 2004;20:758–64.
18. Lu S, Ahn D, Johnson G, Law M, Zagzag D, Grossman RI.
Diffusion-tensor MR imaging of intracranial neoplasia and
associated peritumoral edema: introduction of the tumor
infiltration index. Radiology 2004;232:221–8.
19. Zhen HN, Zhang X, Hu PZ, et al. Survivin expression and
its relation with proliferation, apoptosis, and angiogenesis
in brain gliomas. Cancer 2005;104:2775–83.
657

20. Bello L, Giussani C, Carrabba G, Pluderi M, Costa F,
Bikfalvi A. Angiogenesis and invasion in gliomas. Cancer
Treat Res 2004;117:263–84.
21. Hakyemez B, Erdogan C, Ercan I, Ergin N, Uysal S, Atahan
S. High-grade and low-grade gliomas: differentiation by
using perfusion MR imaging. Clin Radiol 2005;60:
493–502.
22. Lupo JM, Cha S, Chang SM, Nelson SJ. Dynamic suscepti-
bility-weighted perfusion imaging of high-grade gliomas:
characterization of spatial heterogeneity. AJNR Am J
Neuroradiol 2005;26:1446–54.
658
G G Fan, Q L Deng, Z H Wu and Q Y Guo
23. Fan G, Sun B, Wu Z, Guo Q, Guo Y. In vivo single-voxel
proton MR spectroscopy in the differentiation of high-grade
gliomas and solitary metastases. Clin Radiol 2004;59:77–85.
24. Bulakbasi N, Kocaoglu M, Farzaliyev A, Tayfun C, Ucoz T,
Somuncu I. Assessment of diagnostic accuracy of perfusion
MR imaging in primary and metastatic solitary malignant
brain tumors. AJNR Am J Neuroradiol 2005;26:2187–99.
25. Sugahara T, Korogi Y, Kochi M, Ushio Y, Takahashi M.
Perfusion-sensitive MR imaging of gliomas: comparison
between gradient-echo and spin-echo echo-planar imaging
techniques. AJNR Am J Neuroradiol 2001;22:1306–15.
The British Journal of Radiology, August 2006
The British Journal of Radiology, 79 (2006), 659–665

Imaging well-differentiated hepatocellular carcinoma with

dynamic triple-phase helical computed tomography
1,3
C-S LI, 1,3,4R-C CHEN, 1H-Y TU, 2L-S SHIH, 2T-A ZHANG, 1J-M LII, 1W-T CHEN, 1S-J DUH and
1
L-C CHIANG

Departments of 1Radiology, 2Pathology, Renai Branch, Taipei City Hospital, Taipei, 3Department of
Radiology, School of Medicine, National Yang-Ming University, Taipei and 4Department of
Radiology, School of Medicine, Taipei Medical University, Taipei, Taiwan

ABSTRACT. To investigate the imaging appearance of well-differentiated

hepatocellular carcinoma (HCC) on dynamic CT, a total of 38 histopathologically
proven well-differentiated HCC were included in a retrospective study. We
reviewed the contrast-enhanced dynamic CT of all 38 tumours for attenuation of each
tumour in unenhanced scan, arterial-dominant and delayed portal venous phases. Our
results showed that dynamic CT identified 26 (68.4%) out of the 38 lesions. The
remaining 12 lesions were isodense compared with surrounding liver parenchyma in
each dynamic CT phase. There was no statistically significant difference between the
mean size of tumours detected by dynamic CT and that of tumours not detected
by dynamic CT (p50.1). Of a total of 38 tumours, most were isodense (n519) or
hypodense (n516) in unenhanced scan, mostly hyperdense (n518) or isodense Received 7 November 2005
(n515) in arterial-dominant phase and mostly isodense (n522) or hypodense (n515) Revised 26 January 2006
in delayed portal venous phase. Enhancement of tumour was observed in 19 Accepted 13 February 2006
(50.0%) of 38 lesions. In conclusion, the ability of dynamic CT to detect well-
DOI: 10.1259/bjr/12699987
differentiated HCC is poor, and negative CT findings cannot exclude the presence
of well-differentiated HCC, especially if there is well-grounded clinical suspicion for ’ 2006 The British Institute of
HCC. Radiology

Hepatocellular carcinoma (HCC) is one of the most Materials and methods

common malignancies in the world. Globally, there is an
increasing incidence of HCC in both eastern and western
countries [1, 2]. HCC usually develops in the setting of
chronic liver disease and cirrhosis [3]. Screening with
ultrasound and a-fetoprotein levels to detect HCC in
patients with chronic liver disease has become common
practice [4, 5]. Consequently, an increasing number of
nodular hepatic lesions has been detected on screening
ultrasound. It is important to distinguish well-differen-
tiated HCC, an early form of HCC in hepatocarcino-
genesis, from dysplastic hepatic nodule [3, 6]. Treatment
should be applied at this stage with local ablation
therapy, surgical resection or transplantation [7]. The
biological behaviour of well-differentiated HCC is
uncertain, but it is thought to have relatively low
malignant potential and rarely invade vessels or metas-
tasise to other sites [8, 9].
When hepatic nodules are identified on ultrasound,
CT or MRI is usually used for further characterization
because ultrasound appearance of a hepatic nodule may
be non-specific [10]. However, imaging characteristics of
well-differentiated HCCs have been described in only a
few papers in the English literature [11–15]. In this
retrospective study, we would like to investigate the
imaging appearance of well-differentiated HCC on
dynamic dual-phase helical CT.
Patients
From January 2000 to July 2004, we retrospectively
reviewed the histopathological reports of ultrasound-
guided liver tumour biopsy in our hospital. A total of 68
patients with 71 well-differentiated HCCs were identified.
All patients received abdominal ultrasound for screening
of HCC or follow-up after treatment for HCC. Among this
group, 37 patients with 38 well-differentiated HCCs were
referred for dynamic CT scan because of hepatic nodules
detected on abdominal ultrasound, and they were
included in our study. They comprised 25 men and 12
women, with ages ranging from 31 years to 82 years (mean
age 59.8 years). All patients suffered from chronic liver
disease including hepatitis B in 22 patients, hepatitis C in
12 patients, alcoholic liver disease in 1 patient and
cryptogenic liver cirrhosis in 2 patients. The dynamic CT
images of each patient were reviewed. Written informed
consent was obtained from each patient.
All 37 patients underwent ultrasound-guided biopsy
of hepatic nodules before or soon after (within 1 month)
CT scan. All biopsies were performed with a 3.5 MHz
guiding probe with a Logiq 400 unit (GE pro series). The
biopsy needle was either a 1.2 mm Surecut needle (Top
Surgical Manufacturing, Tokyo, Japan) or a 0.8 mm

The British Journal of Radiology, August 2006 659


Majima needle (Top Surgical Manufacturing, Tokyo,
Japan).
The needle biopsy specimens were fixed in formalin
immediately and processed according to standard
procedures in the department of pathology. Both
haematoxylin and eosin (H&E) and reticulin stains were
routinely performed on each case. The specimens were
interpreted by one of the two authors (LSS and TAZ);
both of them were experienced pathologists in the
diagnosis of differentiation of HCC. The histopatho-
logical criteria of well-differentiated HCC included:
(1) increased cell density (more than 1.8 times non-
neoplastic hepatic tissue of the same patient);
(2) increased nuclear/cytoplasmic ratio of individual
tumour cells; and (3) disorganized reticulin framework
with irregular thin-trabecular pattern [16]. In most cases,
the cancer cells of well-differentiated hepatocellular
carcinoma were smaller in size with increased cytoplas-
mic eosinophilia. In addition, fatty changes of tumour
tissue were occasionally observed [17]. Correlation
between imaging characteristics and histopathological
evidence of fatty change and vascular density was not a
focus of our study, because the histopathological
diagnosis in our study was based on liver biopsy.
Image acquisition
Hepatic CT scan was performed with a helical CT
scanner (HiSpeed Advantage; General Electric Medical
Systems, Milwaukee, WI). All patients received oral
contrast material before the CT examination. The
unenhanced scans of entire liver were performed with
slice thickness of 7 mm and space of 10 mm. The
enhanced CT scans were then performed by intravenous
injection of 100–120 ml iopromide (Ultravist 370;
Schering, Berlin, Germany), depending on patients’ body
weight. A mechanical power injector was used. The rate
of injection was 2.5–3.5 ml s21 depending on accessi-
bility to patients’ peripheral veins. The length of delay
between intravenous contrast material administration
and scanning was 30–35 s for the arterial-dominant
phase, according to cardiac function estimated by
patient’s age and general medical condition. The delayed
portal venous phase was obtained at 110 s after contrast
material administration. The delayed portal venous
phase, in-between the usual portal venous phase and
delayed phase, was the standard dynamic CT protocol
for evaluation of nodular liver lesions in our institution.
The dual-phase enhanced CT scan was performed with a
slice thickness of 7 mm and a pitch of 1.2:1. Images
acquired during each phase of contrast were obtained
using a breath-hold technique and required two split
acquisitions to cover the entire liver. For most cases, a
total of 15 slices were obtained in each acquisition and
there was a 6 s interval between the two acquisitions.
Image interpretation
The ultrasound features of each patient in this study
were carefully correlated with the CT scan findings in
respect of tumour location. The tumour sizes were
determined by the maximal diameter of each tumour
660
C-S Li, R-C Chen, H-Y Tu et al
on the basis of ultrasound. The attenuations of these
lesions in unenhanced scan, arterial-dominant phase and
delayed portal venous phase of dynamic CT scan were
independently analysed by two experienced radiologists
specialized in liver imaging. The attenuations of these
lesions were compared with surrounding liver and were
categorized as hyperdense (or high-density), isodense (or
iso-density) and hypodense (or low-density) by visual
inspection. Identification of extremely low attenuation
representing the fat component within these lesions was
also noted. Positive enhancement of tumour was defined
as when a hypodense tumour in unenhanced scan
becomes isodense or hyperdense in arterial-dominant
phase, or when an isodense tumour in unenhanced scan
turns hyperdense in arterial-dominant phase. The win-
dow levels of images were kept constant during
interpretation. Disagreements of images interpretation
were resolved by consensus.
Statistical analysis
All 38 tumours in the study were subdivided into two
groups, one that could be visualized by CT scan and
another group that could not. The size of tumour in each
group was expressed as mean ¡ standard error (SE).
Comparison of sizes between different groups was
performed using the Student’s t-test. A threshold p-value
of 0.05 was chosen for statistical significance.
Results
The sizes of these 38 well-differentiated HCCs
detected by ultrasound ranged from 0.5 cm to 2.7 cm
(mean: 1.7¡0.6 cm). 27 (71.1%) out of 38 tumours were
small HCCs with the sizes less than 2 cm. Dynamic CT
scan depicted 26 (68.4%) out of the 38 nodules. The
remaining 12 nodules were isodense in unenhanced,
arterial-dominant and delayed portal venous phases of
dynamic CT scan (we coded this as ‘‘iso-iso-iso’’ pattern)
and, as a result, could not be identified (Figure 1).
The mean size of the 26 nodules detected by dynamic
CT scan was 1.8¡0.6 cm; and that of the remaining 12
nodules not identified by dynamic CT scan was 1.5
¡0.4 cm. There was no statistically significant difference
in size between the two groups of nodules (p50.1).
Besides that, there was no statistically significant
difference in size between all 38 nodules and the group
of 26 nodules that could be visualized by CT scan
(p50.4), as well as the group of 12 nodules that could not
be visualized by CT scan (p50.3).
In unenhanced CT scans, these 38 well-differentiated
HCCs were mostly isodense (n519, 50.0%) or hypodense
(n516, 42.1%); in arterial-dominant phase, they were
mostly hyperdense (n518, 47.4%) or isodense (n515,
39.5%); in delayed portal venous phase, they were
mostly isodense (n522, 57.9%) or hypodense (n515,
39.5%) (Table 1). Positive enhancement in arterial-domi-
nant phase was observed in 19 (50.0%) of 38 tumours
(Figure 2). One of the lesions was hypodense and
enhanced to be isodense on the arterial-dominant phase
images. All of the 19 tumours demonstrated washout of
enhancement in delayed portal venous phase.
The British Journal of Radiology, August 2006
Dynamic CT of well-differentiated HCC

(a) (b)

(c) (d)

Figure 1. Images of a 1.1 cm-diameter well-differentiated hepatocellular carcinoma in segment IV of the liver in a 73-year-old
man. (a) Abdominal ultrasound disclosed a hepatic tumour with high echogenicity (arrow). (b,c) The tumour was isodense to
surrounding liver parenchyma in unenhanced (not shown), arterial-dominant phase (b) and portal venous phase (c) of dynamic
CT and, as a result, could not be identified. (d) Photomicrograph of the needle biopsy specimen showed a well-differentiated
hepatocellular carcinoma. Haematoxylin and eosin, original magnification 6100.

In our study, the most common appearance of well-
differentiated HCC in each phase of dynamic CT was the
‘‘iso-iso-iso’’ pattern (n512, 31.6%). These tumours could
not be identified on dynamic CT. The other well-
differentiated HCC in our study presented varying
imaging appearances, and common patterns included:
a ‘‘low-high-low’’ pattern (n57, 18.4%) (Figure 3), an
‘‘iso-high-iso’’ pattern (n55, 13.2%), a ‘‘low-high-iso’’
pattern (n54, 10.5%), and a ‘‘low-low-low’’ pattern (n54,
10.5%).
Discussion
Early detection of HCC is becoming feasible owing to
the wide use of ultrasound for screening [18]. High rates
of ultrasound detection for small HCCs, ranging from
89% to 98.6%, have been reported [19, 20]. However, the
differentiation of a hepatic nodule between early
malignant tumour and pre-malignant nodule is often
difficult. The distinction between well-differentiated
HCC and dysplastic nodule is clinically important
because of the difference in their management [10].
Therefore, advanced imaging modalities such as CT and
MRI play an important role in characterizing these
hepatic nodules in a non-invasive fashion.
The appearance of well-differentiated HCC on CT scan
has been studied before, but these prior studies either
had too small a case number of well-differentiated HCCs
or were lacking in dynamic CT protocols [5, 13, 14, 21].
Takayasu et al reported a poor detection rate of 56% for
well-differentiated HCC, which was based on mixed
table incremental CT and helical CT protocols [5]. Our
study presented a 68.4% detection rate on the basis of
dynamic helical CT protocol. The detection rate in our
study was still low. Our study revealed that the mean
size of well-differentiated HCC undetected by dynamic
CT was not statistically different from tumours detected
by dynamic CT. The reason behind the low detection
rate was that well-differentiated HCC was commonly

The British Journal of Radiology, August 2006 661

 C-S Li, R-C Chen, H-Y Tu et al

(a) (b)

(c) (d)

Figure 2. Dynamic CT images of a 1.5 cm-diameter well-differentiated HCC in segment II of the liver in a 43-year-old man.
(a) The tumour was hypodense in unenhanced scan (arrow), (b) hyperdense in arterial-dominant phase (arrow) and (c) isodense
in portal venous phase. (d) Photomicrograph of the needle biopsy specimen with reticulin stain showed a disorganized reticulin
framework with irregular thin-trabecular pattern. Original magnification 6100.

isodense in unenhanced, arterial-dominant and delayed
portal venous phases of dynamic CT scan.
Only half of our cases showed tumour enhancement in
arterial-dominant phase of dynamic CT scan, which
means the other half were hypovascular hepatocellular
carcinoma. Well-differentiated HCC had less neoplastic
angiogenesis and incomplete vascularization of the
sinusoid-like blood spaces of the tumour, and therefore
frequently presented as a hypovascular tumour [22–25].
In contrast, most moderately- and poorly-differentiated
HCCs present as hypervascular lesions, and are easier to
detect on dynamic CT [21].
Choi et al considered that combination of arterial and
portal venous phases on dynamic CT was enough for
detecting HCC, and delayed phase images could be
omitted to decrease scanning time and radiation hazard
[26]. However, other authors suggested that delayed
phase CT images were important in detecting early or
hypovascular hepatocellular carcinoma [5, 27, 28]. The
CT protocols used in these studies were composed of
arterial-dominant phase of 30–35 s, portal venous phase
of 60 s and delayed phase of 180 s after commencing
intravenous contrast medium administration [27, 28]. In
our institution, we used arterial-dominant phase to
detect hypervascular lesions, followed by delayed portal
venous phase at 110 s after contrast medium admini-
stration. The purpose of our dual-phase CT protocol was
to avoid excessive radiation exposure of triple-phase CT;
and our study showed that the CT protocol could offer a
comparable (68.4%) detection rate for well-differentiated

662 The British Journal of Radiology, August 2006

Dynamic CT of well-differentiated HCC

(a) (b)

(c)

Figure 3. Dynamic CT images of a 2.1 cm-diameter well-differentiated HCC in segment III of the liver in a 54-year-old man.
(a) The tumour was hypodense in unenhanced scan (arrow). (b) Positive enhancement of the tumour was demonstrated in
arterial-dominant phase (arrow). (c) In portal venous phase, the tumour was again hypodense (arrow), representing the ‘‘low-
high-low’’ pattern on dynamic CT scan.

HCC. According to the results of our study, roughly half
of well-differentiated HCCs are hypovascular; we
hypothesized that dynamic CT scan with additional
delayed-phase images might increase the detection rate
of well-differentiated HCC. However, the hazard of
increased radiation exposure should also be a concern.
Further study with the triple-phase dynamic CT protocol
to detect well-differentiated HCCs may be needed.
It is clinically important to distinguish pre-malignant
hepatic dysplastic nodules from well-differentiated
HCCs. However, our study showed some overlapping
dynamic CT appearances between the well-differentiated
HCCs and hepatic dysplastic nodules. Choi et al had
reported that hepatic dysplastic nodules were mostly
isodense or hypodense in arterial-phase and portal-
phase CT images; most of them were relatively avascular
[7]. Compared with radiographic features demonstrated
in our study, high density or positive enhancement of a
small hepatic nodule in arterial-dominant phase of
dynamic CT could be an important feature pointing to
the diagnosis of a HCC rather than a pre-malignant
dysplastic nodule. However, in the appropriate clinical
settings (chronic B or C viral hepatitis, elevation of
serum a-fetoprotein level, positive ultrasound findings),

The British Journal of Radiology, August 2006 663


negative CT finding may not be sufficient to exclude a
well-differentiated HCC. Further imaging study or
biopsy should be considered in such situations.
Lee et al reported a 96% arterial-phase enhancement
rate of HCC, using triple-phase multidetector CT
(MDCT) as well as rapid injection of contrast material
at 5 ml s21 [29]. However, only a small number of well-
differentiated HCCs were presented in their study.
MDCT, with its rapid scanning time, should have better
detection ability than single-detector scanner; but the
role of MDCT in diagnosis of well-differentiated HCC
still needs further investigation.
Dynamic gadolinium-enhanced MRI has been consi-
dered the method of choice for HCC diagnosis [30–32]. In
cirrhotic patients, dynamic gadolinium-enhanced MRI is
also a useful modality for detection and characterization
of regenerative or dysplastic nodules [33]. Through its
superior tissue contrast, MRI might detect more well-
differentiated HCCs based on signal intensity change on
pre-contrast T1 weighted or T2 weighted images than CT
[11, 13, 34]. Amano et al studied the CT and MRI patterns
of HCC and concluded that they were useful in
predicting the degree of histological differentiation of
cancer cells in HCC [35]. However, no specific CT
patterns or MR signals can offer accurate diagnosis of
well-differentiated HCC, and absence of arterial phase
enhancement cannot exclude the possibility of early or
‘‘borderline’’ malignancy [32]. In recent years, super-
paramagnetic iron oxide (SPIO)-enhanced MRI has been
reported as useful for the detection of hepatic tumours
[36, 37]. SPIO-enhanced MRI provided a fundamentally
different approach for HCC diagnosis: SPIO would be
taken up by Kupffer cells in normal liver parenchyma,
but not in hepatic tumours which generally lacked
reticulo-endothelial cells. However, the diagnosis of
well-differentiated HCC could still be problematic
because well-differentiated HCC may contain Kupffer
cells as in normal liver parenchyma. Consequently, it
may show similar SPIO uptake comparable with sur-
rounding normal liver parenchyma and would be poorly
delineated on SPIO-enhanced MRI [30, 38, 39]. Contrast-
enhanced MRI could be a better imaging modality for
diagnosis of well-differentiated HCC than CT, but
further investigation would be needed to achieve better
sensitivity and specificity.
We recognized two limitations in our study. First, this
is a retrospective study in which there might be case
selection bias, with respect to referral and patients
undergone CT scanning. Second, there is a limitation in
the correlation between CT appearances and histopatho-
logical findings in our study. All of our cases only had
needle biopsy specimen for histopathological diagnosis,
and underwent percutaneous ethanol injection therapy
(PEIT) or transcatheter arterial chemoembolisation
(TACE) for subsequent treatment. HCC may have
different grades of differentiation within the tumour
and the biopsy sample may not be representative of the
whole tumour. In a study from Japan [24], about 20% of
small HCCs (less than 2 cm in diameter) of the distinctly
nodular type were composed of varying mixtures of
well-differentiated and moderately-differentiated cancer-
ous tissues. Needle biopsy of a tumour cannot fully reflect
the differentiation of HCC. Better imaging-pathology
664
C-S Li, R-C Chen, H-Y Tu et al
correlation could be obtained if we were able to
examine the entire tumours.
Conclusion
In conclusion, the detecting ability of dynamic dual-
phase helical CT for well-differentiated HCC is low. This
type of tumour is frequently isodense to surrounding
liver parenchyma in unenhanced, arterial-dominant or
portal venous phase of CT scan, and positive enhance-
ment is seen only in 50% of them. It is difficult to
diagnose well-differentiated HCCs based solely on the
density changes on CT scans. Hepatic nodules with high
density or positive enhancement in arterial-dominant
phase of dynamic CT scans suggest a diagnosis of a HCC
rather than a hepatic dysplastic nodule in cirrhotic liver.
However, in the appropriate clinical settings, a negative
CT scan cannot be relied upon in exclusion of well-
differentiated HCCs. Further imaging study, including
contrast enhanced-MDCT or MRI, and liver biopsy
should be considered in this situation.
References
1. Deuffic S, Poynard T, Buffat L, Valleron AJ. Trends in
primary liver cancer. Lancet 1998;351:214–5.
2. El-Serag HB, Mason AC. Rising incidence of hepatocellular
carcinoma in the United States. N Engl J Med
1999;340:745–50.
3. Sakamoto M, Hirohashi S, Shimosato Y. Early stages of
multistep hepatocarcinogenesis: adenomatous hyperplasia
and early hepatocellular carcinoma. Hum Pathol
1991;22:172–8.
4. Okazaki N, Yoshida T, Yoshino M, Matue H. Screening of
patients with chronic liver disease for hepatocellular
carcinoma by ultrasonography. Clin Oncol 1984;10:241–6.
5. Takayasu K, Furukawa H, Wakao F, Muramatsu Y, Abe H,
Terauchi T, et al. CT diagnosis of early hepatocellular
carcinoma: sensitivity, findings, and CT-pathologic correla-
tion. AJR Am J Roentgenol 1995;164:885–90.
6. Choi BI, Takayasu K, Han MC. Small hepatocellular
carcinomas and associated nodular lesions of the liver:
pathology, pathogenesis, and imaging findings. AJR Am J
Roentgenol 1993;160:1177–87.
7. Choi BI, Han JK, Hong SH, Kim TK, Song CS, Kim KW, et
al. Dysplastic nodules of the liver: imaging findings.
Abdom Imaging 1999;24:250–7.
8. Sugihara S, Nakashima O, Kojiro M, Majima Y, Tanaka M,
Tanikawa K. The morphologic transition in hepatocellular
carcinoma. A comparison of the individual histologic
features disclosed by ultrasound-guided fine-needle biopsy
with those of autopsy. Cancer 1992;70:1488–92.
9. Winter TC 3rd, Takayasu K, Muramatsu Y, Furukawa H,
Wakao F, Koga H, et al. Early advanced hepatocellular
carcinoma: evaluation of CT and MR appearance with
pathologic correlation. Radiology 1994;192:379–87.
10. Lencioni R, Mascalchi M, Caramella D, Bartolozzi C. Small
hepatocellular carcinoma: differentiation from adenoma-
tous hyperplasia with color Doppler US and dynamic Gd-
DTPA-enhanced MR imaging. Abdom Imaging
1996;21:41–8.
11. Muramatsu Y, Nawano S, Takayasu K, Moriyama N,
Yamada T, Yamasaki S, et al. Early hepatocellular carci-
noma: MR imaging. Radiology 1991;181:209–13.
12. Sadek AG, Mitchell DG, Siegelman ES, Outwater EK,
Matteucci T, Hann HW. Early hepatocellular carcinoma
that develops within macroregenerative nodules: growth
The British Journal of Radiology, August 2006
Dynamic CT of well-differentiated HCC
rate depicted at serial MR imaging. Radiology
1995;195:753–6.
13. Monzawa S, Omata K, Shimazu N, Yagawa A, Hosoda K,
Araki T. Well-differentiated hepatocellular carcinoma:
findings of US, CT, and MR imaging. Abdom Imaging
1999;24:392–7.
14. Karahan OI, Yikilmaz A, Isin S, Orhan S. Characterization
of hepatocellular carcinomas with triphasic CT and correla-
tion with histopathologic findings. Acta Radiol
2003;44:566–71.
15. Takayasu K, Muramatsu Y, Mizuguchi Y, Moriyama N,
Ojima H. Imaging of early hepatocellular carcinoma and
adenomatous hyperplasia (dysplastic nodules) with
dynamic CT and a combination of CT and angiography:
experience with resected liver specimens. Intervirology
2004;47:199–208.
16. Nakashima O, Kojiro M. [Pathomorphologic characteristics
and differential diagnosis of small hepatocellular carci-
noma]. Gan To Kagaku Ryoho 1996;23:827–34, [In
Japanese].
17. Kutami R, Nakashima Y, Nakashima O, Shiota K, Kojiro M.
Pathomorphologic study on the mechanism of fatty change
in small hepatocellular carcinoma of humans. J Hepatol
2000;33:282–9.
18. Oka H, Kurioka N, Kim K, Kanno T, Kuroki T, Mizoguchi
Y, et al. Prospective study of early detection of hepatocel-
lular carcinoma in patients with cirrhosis. Hepatology
1990;12:680–7.
19. Teefey SA, Hildeboldt CC, Dehdashti F, Siegel BA, Peters
MG, Heiken JP, et al. Detection of primary hepatic
malignancy in liver transplant candidates: prospective
comparison of CT, MR imaging, US, and PET. Radiology
2003;226:533–42.
20. Horigome H, Nomura T, Saso K, Itoh M, Joh T, Ohara H.
Limitations of imaging diagnosis for small hepatocellular
carcinoma: comparison with histological findings. J
Gastroenterol Hepatol 1999;14:559–65.
21. Sakabe K, Yamamoto T, Kubo S, Hirohashi K, Hamuro M,
Nakamura K, et al. Correlation between dynamic computed
tomographic and histopathological findings in the diag-
nosis of small hepatocellular carcinoma. Dig Surg
2004;21:413–20.
22. Ohashi I, Hanafusa K, Yoshida T. Small hepatocellular
carcinomas: two-phase dynamic incremental CT in detec-
tion and evaluation. Radiology 1993;189:851–5.
23. Takayasu K, Muramatsu Y, Furukawa H, Wakao F,
Moriyama N, Takayama T, et al. Early hepatocellular
carcinoma: appearance at CT during arterial portography
and CT arteriography with pathologic correlation.
Radiology 1995;194:101–5.
24. Kojiro M. Focus on dysplastic nodules and early hepato-
cellular carcinoma: an Eastern point of view. Liver Transpl
2004;10:S3–8.
25. Kim SR, Kim KI, Maekawa Y, Imoto S, Ninomiya T, Mita K,
et al. Well-differentiated HCC manifesting hyperattenua-
tion on CT during arterial portography. Hepato-
gastroenterology 2005;52:1559–62.
The British Journal of Radiology, August 2006
26. Choi BI, Lee HJ, Han JK, Choi DS, Seo JB, Han MC.
Detection of hypervascular nodular hepatocellular carcino-
mas: value of triphasic helical CT compared with iodized-
oil CT. AJR Am J Roentgenol 1997;168:219–24.
27. Lim JH, Choi D, Kim SH, Lee SJ, Lee WJ, Lim HK, et al.
Detection of hepatocellular carcinoma: value of adding
delayed phase imaging to dual-phase helical CT. AJR Am J
Roentgenol 2002;179:67–73.
28. Iannaccone R, Laghi A, Catalano C, Rossi P, Mangiapane F,
Murakami T, et al. Hepatocellular carcinoma: role of
unenhanced and delayed phase multi-detector row helical
CT in patients with cirrhosis. Radiology 2005;234:460–7.
29. Lee KH, O’Malley ME, Haider MA, Hanbidge A. Triple-
phase MDCT of hepatocellular carcinoma. AJR Am J
Roentgenol 2004;182:643–9.
30. Simon G, Link TM, Wortler K, Doebereiner F, Schulte-
Frohlinde E, Daldrup-Link H, et al. Detection of hepatocel-
lular carcinoma: comparison of Gd-DTPA- and ferumox-
ides-enhanced MR imaging. Eur Radiol 2005;15:895–903.
31. Peterson M, Baron R, Murakami T. Hepatic malignancies:
usefulness of acquisition of multiple arterial and portal
venous phase images at dynamic gadolinium-enhanced MR
imaging. Radiology 1996;201:337–45.
32. Rode A, Bancel B, Douek P, Chevallier M, Vilgrain V,
Picaud G, et al. Small nodule detection in cirrhotic livers:
evaluation with US, spiral CT, and MRI and correlation
with pathologic examination of explanted liver. J Comput
Assist Tomogr 2001;25:327–36.
33. Martin DR, Semelka RC. Magnetic resonance imaging of the
liver: review of techniques and approach to common
diseases. Semin Ultrasound CT MR 2005;26:116–31.
34. Kajiwara M. MR imaging of small hepatocellular carcinoma
(, or 5 20 mm)--correlation with vascularity and histolo-
gical features. Kurume Med J 1997;44:327–8.
35. Amano S, Ebara M, Yajima T, Fukuda H, Yoshikawa M,
Sugiura N, et al. Assessment of cancer cell differentiation in
small hepatocellular carcinoma by computed tomography
and magnetic resonance imaging. J Gastroenterol Hepatol
2003;18:273–9.
36. Paley MR, Mergo PJ, Torres GM, Ros PR. Characterization
of focal hepatic lesions with ferumoxides-enhanced T2-
weighted MR imaging. AJR Am J Roentgenol
2000;175:159–63.
37. Tang Y, Yamashita Y, Arakawa A, Namimoto T, Mitsuzaki
K, Abe Y, et al. Detection of hepatocellular carcinoma
arising in cirrhotic livers: comparison of gadolinium- and
ferumoxides-enhanced MR imaging. AJR Am J Roentgenol
1999;172:1547–54.
38. Lim JH, Choi D, Cho SK, Kim SH, Lee WJ, Lim HK, et al.
Conspicuity of hepatocellular nodular lesions in cirrhotic
livers at ferumoxides-enhanced MR imaging: importance of
Kupffer cell number. Radiology 2001;220:669–76.
39. Kato H, Kanematsu M, Kondo H, Goshima S, Matsuo M,
Hoshi H, et al. Ferumoxide-enhanced MR imaging of
hepatocellular carcinoma: correlation with histologic tumor
grade and tumor vascularity. J Magn Reson Imaging
2004;19:76–81.
665
 The British Journal of Radiology, 79 (2006), 666–671

Building an anonymized catalogued radiology museum in PACS:

a feasibility study
A P TOMS, FRCS, FRCR, B KASMAI, MSc, MIPEM, S WILLIAMS, MRCP, FRCR and P WILSON, MRCP, FRCR

Department of Radiology, Norfolk & Norwich University Hospital NHS Trust, Colney Lane, Norwich,
Norfolk NR4 7UY, UK

ABSTRACT. The aim of this study was to test the feasibility of a software application
that would allow the anonymization and cataloguing of whole DICOM datasets in
order to build searchable radiology museums within PACS. The application was
developed on a dedicated networked PC, using C# and HL7 coding. Whole DICOM
datasets were pushed from PACS to a networked PC on which the application, Museum
Builder, was developed. Museum Builder works by replacing the patient specific data
(the forename, surname and hospital number) within each header of each DICOM file
with terms from anatomical and surgical sieve menus. The date of birth is anonymized
to 1 January of the same year. Whole DICOM datasets comprising hundreds of images
can be anonymized and catalogued in a single episode. Museum Builder primes PACS
with an HL7 script to receive a ‘‘new’’ patient. DICOM datasets are then pushed back to
PACS where they are added to the database as ‘‘new’’ cases. The museum cases can
then be searched for, on PACS, by any combination of terms that correspond to
appropriate anatomical units, surgical sieve headings or radiological specialty. New
radiology reports containing clinical histories, radiological descriptions, differential
diagnoses and discussion can be added through the report window. Our institution has Received 22 June 2005
developed and used this tool to generate a PACS based radiology museum containing Revised 6 December 2005
not only full DICOM datasets, but also relevant histological and clinical photographs. In Accepted 9 December 2005
conclusion, this technique offers a mechanism for generating anonymized catalogued
DOI: 10.1259/bjr/11861604
radiology museums in PACS. Museum Builder represents a working prototype that
demonstrates some of the archiving functions that are expected by teaching ’ 2006 The British Institute of
institutions from PACS. Radiology

The radiology museum is an integral component of
every radiological training scheme. Over the past 5 years,
the practice of radiology has moved from film to PACS,
but the ability to build radiology museums has not kept
pace [1].
For many of us, the hardcopy ACR collection provided
hours of study and exam practice material. Those cases we
saw in our formative years of radiological training often
become ‘‘index’’ cases against which those that followed
were measured. These museums, which are becoming
increasingly obsolete [2], often comprised cupboard-like
rooms filled with shelf upon shelf of ageing radiographs in
various states of disorganization. Radiological museums
have now diversified into multiple digital formats.
DICOM files can easily be converted and saved in a
number of manageable formats [3]. Large institutional
collections can be acquired on CD-ROM [4, 5]. Personal
teaching collections can be created in any number of
readily available image databases [6–10]. Some of these
databases are specifically designed for archiving radio-
logical teaching cases and sometimes for storage on
servers [4, 11, 12] for sharing access across networks or
the World Wide Web [13]. Online database applica-
tions allow free text searches across thousands of cases,
This projected was supported by a Kodak Radiology Fund
Scholarship awarded by the Royal College of Radiologists in 2003.
sometimes in multiple institutions [13]. These archives are
used for research, teaching [2, 9, 10, 14] and for assessment
of radiological expertise [15]. With this digital diversity has
come a subtle change in the look and feel of the radiology
museum. Hardcopy museum cases must be read in the
same way that hardcopy radiology is practiced; with a
light box. Digital museum cases are read from personal
computers and not in the PACS environment in which
many of us now work. One of the obstacles to replicating
the PACS environment on a PC is the prohibitive size of
the DICOM files. A solution to these problems is not to
replicate PACS in order to build a work-like radiology
museum, but to build a radiology museum within PACS.
To the best of our knowledge, the major PACS manufac-
turers provide only limited tools for archiving radiology
teaching cases, whereas most radiologists consider this
sort of functionality important or even essential when
considering the purchase of PACS [16]. Some provide a
system of academic folders that require a system admin-
istrator to set up. These provide inadequate archiving and
retrieval mechanisms for generating usable databases
within PACS [1]. Neither does there appear to be any
third party solutions that meet these criteria. PACS
manufacturers prohibit access to their databases other
than by their employees and, therefore, a novel approach
is required for third parties to generate teaching cases on
PACS. The aim of this study was to see if it was feasible to

666 The British Journal of Radiology, August 2006

Building anonymized radiology museums in PACS
develop a software application that could edit the DICOM
headers to anonymize and catalogue teaching cases in
order to build a radiology museum within PACS.
Materials and methods
Principle
Radiological studies stored on PACS can be identified
and retrieved using a number of search fields common to
all PACS – namely, the patient’s surname, forename,
middle names and unique hospital number. This
information is stored within a header in every DICOM
file (Figure 1). This usually means that every image, in
every series, in every radiological study contains this
information embedded within it. Multiframe images,
generated with ultrasound, contain the same information
in a single header. Both of these types of DICOM file can
be handled in the same way. After DICOM data has been
generated by a radiological investigation, it is pushed to
the PACS server to be archived. As PACS receives the
DICOM data, it reads the DICOM header and stores the
patient specific data in a database. When PACS is
queried to search for a particular patient, it is this
database that is searched and not the DICOM archive
itself. This entry in the database, however, points to the
DICOM dataset within the archive, which can then be
retrieved and opened. Our application, called Museum
Builder, exploits this process by replacing the patient
specific data within the header of DICOM files that have
been exported from PACS. When the DICOM files are
returned to PACS, the new information within the header
is added to the PACS database as a ‘‘new’’ patient. The
DICOM files for the museum case are effectively dupli-
cated, but with a new DICOM header. In effect, PACS sees
Museum Builder as any other radiological modality
contributing to the local PACS archive.
Hardware and software
Our institution is a film-free hospital with a GE
Centricity PACS (General Electric, Milwaukee, WI).
Images can be viewed from a mixture of dedicated
reporting workstations and PC-based web-browsers,
which cover the whole hospital on a network with a
2 Gb s21 backbone and a 100 Mb s21 link to workstations.
Museum Builder was developed using .NET technology
and C# as the preferred language (Microsoft .Net
Framework to run and Visual Studio to compile the C#
source code). Museum Builder was installed on a
networked PC (Pentium 4 CPU 2.80 Ghz with 1.0 Gb of
Figure 1. A screenshot of a readout from a hexadecimal
editor of the header from a DICOM file demonstrating the
ASCII values of the binary data. The patient specific binary
data can be identified and edited with Museum Builder.
The British Journal of Radiology, August 2006
RAM running Microsoft Windows XP Professional, ver-
sion 2002, with Service Pack 1). An academic licensed
version of eFilm Workstation 1.9.4 [17] was installed as the
helper application.
Results
Process
Museum Builder can work with most PC based DICOM
browsers, which for the purpose of this article will be
referred to as the helper application. After opening
Museum Builder, the helper application’s database can
be browsed or searched using search fields for the patient’s
name or hospital number. Cases are then identified for
museum archiving and highlighted. Once selected, the
patient’s surname, forename, hospital number and date of
birth are displayed in a row of text fields. Below this a
second row contains the text fields for the anonymized
museum case. The patient’s ‘‘new’’ surname and forename
are selected from pre-defined menus. Optional middle
names can also be entered. The new hospital number
comprises a three letter code, again selected from a menu,
followed by a unique four digit number generated by
Museum Builder (Figure 2). The drop-down menu
selections can be over-ridden by entering text into the
fields manually. The patient’s date of birth is automatically
anonymized to 1 January of the year of their birth, thus
preserving age-related information in the DICOM header.
Once the minimum number of fields has been completed
(surname, forename and hospital number), the case is
ready for anonymization (Figure 2). Clicking the ‘‘anon-
ymize’’ button will then replace the patient specific data in
each header of each DICOM file within each directory of
the chosen radiological study or studies and the helper
application database is modified accordingly. All other
identifiable labels, private or otherwise, are changed or
erased. This one click anonymizes every radiological study
in the eFilm database for that patient; there is no limit to
the number of types of examination, e.g. conventional
radiography, CT, ultrasound or MRI, that can be processed
at once. Even for large datasets with multiple examinations
this only takes a few seconds on the workstation described.
The process is simple and is currently performed in our
department by clerical staff. Every DICOM file generated
by a medical imaging device contains a unique identifier
(the study SOP instance UID). At this stage, Museum
Builder assigns each new case with a new unique
identifier, generated internally from a sub-delegate range
offered by Medical Connections [18], so that PACS does
not recognize it when it returns. PACS just sees another
new case arriving from a medical imaging device. When
coding and anonymization is complete, Museum Builder
composes an HL7 Radiology order message, which is sent
to the PACS HIS/RIS broker. On receipt of this acknowl-
edgment the new museum case is pushed from the helper
application back to PACS (Figure 3).
Catalogue headings
The American College of Radiologists (ACR) has
established a well-recognized classification system for
667
 A P Toms, B Kasmai, S Williams and P Wilson

Figure 2. A screenshot from Museum Builder’s anonymization window. The patient specific fields are replaced by catalogue
headings that are selected from drop down menus.

radiology [19]. This system allows high-level discrimina- import process, DICOM header fields are entered
tion of radiological diagnoses, and therefore accurate manually and new DICOM files are generated. This
retrieval of data, particularly for research. However, the allows non-radiological images to be added to the
system is not entirely intuitive, and therefore a different
classification system has been used in Museum Builder.
The catalogue headings are almost universal within
radiology. The patient’s surname is changed to a radi-
ological anatomical unit, which broadly defines the
sections of the human body that radiological investigations
cover. These consist of head, neck, chest, abdomen, pelvis,
extremities and breast. The patient’s forename is selected
from a surgical sieve consisting of normal, developmental,
trauma, infection, neoplasia, inflammation, vascular,
metabolic. Thereafter there is an option to add one, two
or more middle names from a selection of organ specific
titles such as liver, lung, brain, adrenal and so on. The
hospital number is replaced by a unique museum number
that comprises a three-letter code followed by a four-digit
number. The code reflects subspecialty interests within
radiology and include MSK for musculoskeletal, GIT for
gastrointestinal and H&N for head and neck (Figure 2).

Non-radiological images
Figure 3. A schematic diagram demonstrating data flow
Many DICOM browsers will import non-DICOM between Museum Builder, the helper application, PACS and
image files such as JPEG and TIFF files. During the RIS/HIS.

668 The British Journal of Radiology, August 2006

Building anonymized radiology museums in PACS
Figure 4. A montage of screenshots demonstrating material from a single museum case including computed radiography, MRI,
arthroscopy and histology.
museum case, including histology, arthroscopy, endo-
scopy and clinical photographs (Figure 4). The report
field of these ‘‘new studies’’ can then contain pathology
reports, operative notes and clinical findings.
Reports
Reports for the museum cases can be added to the PACS
RIS. These can be added by pasting text into a report
window in Museum Builder (Figure 5). These reports can
include copies of the original radiological report or can be
entirely new and contain clinical histories, updated
radiological reports, results of other special investigations,
differential diagnoses and discussion.
Discussion
Museum Builder is a fully working prototype that has
some of the functionality required by teaching institu-
tions to develop radiology museums within PACS. It
uses a novel approach to generating anonymized
searchable teaching cases without accessing the PACS
database directly. Teaching cases can then be read in the
PACS environment in which the radiology trainee and
his or her trainers work. Instead of being presented with
The British Journal of Radiology, August 2006
single or selected digital images from a teaching archive
or CD-ROM, the trainee has access to the whole DICOM
dataset. When reading cross-sectional investigations, the
trainee would have to interrogate the entire dataset,
including localizers and sequences repeated because of
technical problems, to gather the signs necessary to yield
a diagnosis. It is this process that cannot be replicated by
non-DICOM museums and teaching collections.
There are limitations to radiological museums created
by Museum Builder. There is no free-text search function
within the PACS browser window, which would allow
the user to search for a specific diagnosis. However, the
objective was to create a radiological museum that
functioned in a similar manner to the ACR hard copy
museum and therefore did not require the ability to
immediately recall specific cases. Trainees can search
through the database through catalogue headings based
on anatomical site, disease process and radiology
subspecialty.
Museum Builder can be used with any PACS and, in
theory, integrated with any PACS broker, but this has
only been tested with PACS in our institution. Whilst it
should work with any PACS broker, the concept of
Museum Builder does not allow for a ‘‘plug and play’’
solution. Museum Builder needs to be configured for
each PACS broker in the same way that any CT or
ultrasound machine must be configured to work with a
669

Figure 5. A screenshot demonstrating the ‘‘Report Window’’ where the text of the museum case has been pasted in prior to
sending to RIS in an HL7 script.
particular PACS. However, configuring Museum Builder
has been simplified by using variables in the HL7 code
that can be defined from within the Access database
according to the local PACS broker profiles.
The concept of Museum Builder is relatively simple
and the coding is mainstream. It currently works as a
bolt-on application to PC-based DICOM browsers, but
there are a number of options for further development.
Museum Builder could be coded to work as a DICOM
client and, therefore, could stand alone in its integration
with PACS and RIS. However, it does not make sense to
repeat the work done by many affordable or free, readily
available proprietary DICOM browsers. It would be
easier to add Museum Builder’s functionality to these
670
A P Toms, B Kasmai, S Williams and P Wilson
applications. The most elegant solution would be for the
PACS manufacturers to add this functionality to their
current systems. That way the radiologist could build his
or her DICOM radiology museum without leaving
PACS.
This sort of functionality within PACS has certain
implications for governance of the educational material
because other allied healthcare workers outside radio-
logy, and IT personnel, also have access to the database.
In our institution we developed a governance protocol
that was approved by the Caldicott Guardian to ensure
that the limitations of patients’ consent to procedures
and investigations were adhered to. Rather than being a
risk, the radiology museum is considered a valuable
The British Journal of Radiology, August 2006
Building anonymized radiology museums in PACS
hospital-wide resource. Generating validated case mate-
rial is always time consuming, and therefore the number
of museum cases has, so far, made a negligible impact on
the PACS archive capacity. In theory, however, duplicat-
ing large volumes of archive material could have serious
implications for storage and therefore needs to be
carefully controlled. In our institution, all museum
material must be approved by a Radiology Museum
Committee, which acts as a gatekeeper safeguarding the
quality of the PACS museum and controlling its impact
on the clinical archive.
Conclusion
Museum Builder demonstrates that it is feasible to
build anonymized catalogued radiology museums
within PACS, by editing the patient specific headers
within the DICOM files, and therefore without directly
accessing the PACS database. Teaching cases generated
with this tool allow the trainee to read the full DICOM
datasets within the normal PACS working environment.
By using PACS as the radiology museum repository, the
problems of storing and transmitting large image files
and directories can be overcome. Museum Builder
provides a model for the some of functionality that
many academic institutions would like to see added to
PACS by the PACS manufacturers.
References
1. Siegel E, Reiner B. Electronic teaching files: seven-year
experience using a commercial picture archiving and
communication system. J Digit Imaging 2001;14:125–7.
2. Wiggins RH, Davidson HC, Dilda P, Harnsberger HR,
Katzman GL. The evolution of filmless radiology teaching. J
Digit Imaging 2001;14:236–7.
3. Halsted MJ, Moskovitz J, Johnson N, Perry L. A simple
method of capturing PACS and other radiographic images
for digital teaching files or other image repositories. AJR
Am J Roentgenol 2002;178:817–9.
The British Journal of Radiology, August 2006
4. Rosset A, Muller H, Martins M, Dfouni N, Vallee JP, Ratib
O. Casimage project: a digital teaching files authoring
environment. J Thorac Imaging 2004;19:103–8.
5. Bernard J, Gerber S, Oppenheim C, Marsault C, Dormont D.
Creation of an educational CD-ROM using a PACS. J Radiol
2002;83:68–70.
6. Maldjian JA, Listerud J. Automated teaching file and slide
database for digital images. AJR Am J Roentgenol
2000;175:1249–51.
7. Macura KJ, Macura RT, Morstad BD. Digital case library: a
resource for teaching, learning, and diagnosis support in
radiology. Radiographics 1995;15:155–64.
8. Weinberger E, Jakobovits R, Halsted M. MyPACS.net: a
web-based teaching file authoring tool. AJR Am J
Roentgenol 2002;179:579–82.
9. Strickland NH, Gishen P. From Silver to Silicon: film
libraries of the future. Imaging 1994;6:143–6.
10. Strickland NH, Allison DJ, Gishen P. Technical note: a
radiological educational system – organization of an image
library. Br J Radiol 1995;68:524–7.
11. Rosset A, Ratib O, Geissbuhler A, Vallee JP. Integration of a
multimedia teaching and reference database in a PACS
environment. Radiographics 2002;22:1567–77.
12. Radfiler Website. Available at: www.radfiler.com [Accessed
19 April 2005].
13. Lim CC, Yang GL, Nowinski WL, Hui F. Medical Image
Resource Center--making electronic teaching files from
PACS. J Digit Imaging 2003;16:331–6.
14. Dugas M, Trumm C, Stabler A, et al. Case-oriented
computer-based-training in radiology: concept, implemen-
tation and evaluation. BMC Med Educ 2001;1:5.
15. Mullins ME, Will M, Mehta A, Novelline RA. Evaluating
medical students on radiology clerkships in a filmless
environment: use of an electronic test prepared from PACS
and digital teaching collection images. Acad Radiol
2001;8:514–9.
16. Blunt D, O’Regan D. Using PACS as a teaching resource. Br
J Radiol 2005;78:483–4.
17. Efilm Website. Available at www.merge.com [Accessed 22
June 2005].
18. Medical Connections Website. Available at www.medical-
connections.co.uk [Accessed 22 June 2005].
19. The American College of Radiology Website. Available at
www.acr.org [Accessed 22 June 2005].
671
 The British Journal of Radiology, 79 (2006), 672–680

Optimizing localization accuracy in head and neck, and brain

radiotherapy
1
M McJURY, PhD, 2K DYKER, MBChB, MRCP, FRCR, 3R NAKIELNY, MA(Cantab), BM BCh, FRCR, 1J CONWAY, PhD
and 2M H ROBINSON, MD, FRCP, FRCR

Departments of 1Radiotherapy Physics and 2YCR Clinical Oncology, Weston Park Hospital, Whitham
Road, Sheffield and 3Department of Radiology, Royal Hallamshire Hospital, Glossop Road,
Sheffield, UK

ABSTRACT. The purpose of this study was to investigate the impact on localization of
utilizing contrast-enhanced CT scans and the formal input of a radiologist in the
planning process. 25 head and neck/brain patients had pre- and post-contrast CT scans
in the treatment position. Radiotherapy treatment was planned on the unenhanced CT
images as per standard practice. Retrospectively, their scans (unenhanced and
enhanced) were re-contoured by two oncologists and a radiologist. These new contours
were compared with the original unenhanced treatment contours and differences in
contour volume, geographical isocentre position and tolerance coverage of the
associated planning target volumes (PTVs) were evaluated using the original plans. The
use of contrast enhanced CT data during localization by the oncologist shows little
change in gross tumour volumes (GTVs) or PTVs, geographical position or tolerance
coverage for the targets in the brain studied here. Larger changes in mean volume are
seen for the head and neck cases alone. Changes are greater and statistically significant
(p,0.05, Wilcoxon signed rank test) for localization by the radiologist. Furthermore,
when comparing the original PTV marked by the oncologist with a new PTV re-
contoured by the oncologist, but based on a GTV marked-up by the radiologist, again
statistically significant (p,0.01) changes in percentage volume are noted. Intraoperator
precision is good, percentage volume differences being of the order 3–6%. PTVs also
show improved standard deviations compared with GTVs. Geographic shifts are
generally within our departmental tolerance levels for daily patient setup. Comparing
precision of unenhanced data with enhanced, mean percentage volume changes are
smaller, but not statistically significant. The use of enhanced scan data for localization Received 15 December
has little effect on size, geographical position or tolerance coverage of PTVs marked up 2005
by the oncologists in this study. However, more important is the input from a Accepted 7 February 2006
radiologist. Statistically significant differences due to mark-up on enhanced scans by
DOI: 10.1259/bjr/14663755
the radiologist are shown. Furthermore, significant differences are also seen between
PTVs based on oncologist-generated GTVs, and those based on radiologist-generated ’ 2006 The British Institute of
GTVs. Radiology

The goal of modern three-dimensional (3D) conformal
radiotherapy is to accurately conform dose to the tumour
target whilst minimizing dose to nearby normal tissue.
For patients with brain and head and neck tumours, the
targets are in close proximity to many critical structures,
making tissue avoidance a high priority. Failure of
locoregional tumour control is also a particular issue
for these patients as brain tumours do not tend to
metastasise and nodal involvement in head and neck
tumours is one of the major prognostic factors. These
factors determine the quality and length of life in many
patients undergoing radiotherapy, and are known to be
due to biological and technical factors [1]. The inter-
Current address for Dr M McJury: Department of Medical Physics,
The Northern Ireland Cancer Centre, Belfast City Hospital, Lisburn
Road, Belfast BT9 7AB, UK.
The authors gratefully acknowledge support from Weston Park
Research Fund (MM) and Yorkshire Cancer Research (KD, MHR).
clinician variability is known to be high across all sites
[2–4]. Many methods to improve the consistency of
tumour and organ delineation have been tried, including
the development of volume delineation protocols [5, 6],
the application of additional imaging modalities [7–9],
the use of contrast-enhanced data [10] and publication of
anatomical maps [11].
The first step in the treatment process is localization of
the tumour, usually using radiographic films and/or CT
data. Tepper et al [12] showed that performing a
planning CT scan in addition to existing diagnostic
information enabled improvements in target localization
in 49% of patients. In the authors’ oncology centre,
intravenous (IV) contrast is used routinely in diagnostic
CT scanning of these patients, but not used routinely
when acquiring CT scans for treatment planning
purposes. The use of enhanced CT scans can offer
improved tumour visibility in many cases and may
enable improved localization for planning [13, 14].

672 The British Journal of Radiology, August 2006

Optimizing localization in radiotherapy
Whilst seemingly obvious, improvements may be avail-
able in marking-up visible gross tumour volumes
(GTVs), although the overall impact on the target or
planning target volume (PTV) and general plan quality is
unknown. The use of contrast-enhanced data has been
shown (for other sites, e.g. Zhou et al and Valcenti et al
[10, 15]) to provide improved tumour delineation. For
brain tumours and head and neck cancer, there have
been no previous reports that assess the impact of
contrast-enhanced CT on the delineation of the GTV and
the effect of such change on the PTV.
When contouring, the oncologist will mark the gross
tumour volume (GTV), which is the visible extent of the
tumour, with the help of diagnostic MRI images. A
margin is added to this to allow for non-visible tumour
infiltration, creating a clinical target volume (CTV). An
additional margin is added to the CTV to account for
patient movement and set-up inaccuracies, generating a
final planning target volume (PTV) to be treated [16]. For
this study, PTVs are not marked on directly, but are
always generated by adding a uniform 2D margin to the
initial GTV. Once generated, the initial PTV may be
edited by the clinician to achieve a final PTV contour.
This editing may be necessary, for example, if the
software-generated PTV extends beyond the patients’
anatomy. Although a radiologist is the recognized expert
in the interpretation of medical images, in many centres
(the authors’ included) definition of the GTV is
performed solely by the oncologist. By requiring the
GTV to be defined by a radiologist and the remainder of
the marking-up process (definition of the CTV and PTV)
to be done by the oncologist, improvements in planning
accuracy and outcome may be possible.
This study addresses two main questions:
(i) Does the use of IV contrast during the acquisition
of CT data for treatment planning significantly
alter tumour localization and plan quality?
and
(ii) Does the input of a radiologist in delineating the
tumour GTV significantly alter the volume or
position of the primary, or quality of the final
treatment plan?
Methods
A group of 13 head and neck and 12 brain cases were
included in the investigation. Table 1 gives a list of sites
and associated patient numbers. Patient selection criteria:
all patients having radical external beam treatment with
CT planning for head and neck or brain sites are
considered. Each patient must be physically able to
Table 1. Breakdown of the individual sites for the study
patient group
Primary site Number
Brain – surgical resection 5 initial GTV, a PTV was generated by the oncologist
Brain – no surgical resection 7 and tolerance coverage determined.
Larynx 2
Nasopharynx 4 In comparing contour pairs, three indices were used:
Tongue 1
Tonsil 6
The British Journal of Radiology, August 2006
undergo CT scanning (weight and girth can result in
some exclusions) and be able to consent. Ethics approval
was obtained for non-standard administration of con-
trast as part of the treatment planning process. They each
had pre- and post-contrast CT scans carried out in the
treatment position. Our standard CT protocols were used
for image acquisition: 5 mm slice thickness with 5 mm
interslice gap, acquired in helical mode with pitch of 1,
on our Picker PQS CT. After administering contrast, the
couch was returned to the original position without any
patient movement, enabling both scans (unenhanced
and enhanced) to be acquired with identical scanner
coordinates.
GTV and PTV contours were marked-up using the
AcQSIMTM virtual simulator (Philips Medical Systems,
Best, The Netherlands) on the unenhanced scan and the
patients were treated based on this scan. After a gap of
several weeks, each patient’s unenhanced and enhanced
scans were then retrospectively re-contoured by (i) the
original oncologist (A) [to generate data on intraobserver
precision], (ii) a second oncologist (B) [to generate
oncologist interobserver precision], and (iii) a radiologist.
In this way, several sets of GTVs and PTVs were
generated for each patient. At each contouring session,
all previous contours were removed from the image
display, so the clinician was blinded to all previous
work. Diagnostic films/images and patient notes were
made available to the clinician marking-up at each
contouring session.
The enhanced studies were intrinsically registered
with the unenhanced, such that contours marked on the
enhanced image would be automatically transferred and
stored with all previous others already marked on the
unenhanced study. For data sets with identical scanner
coordinates, image registration was performed automa-
tically by the AcQSIMTM software (see Figure 1). The
accuracy of this registration is dependent on negligible
patient movement during scanning. Diagnostic scan data
was not registered to planning CT data, but was
generally available as hardcopy films and viewed on a
light-box beside the AcQSIM work-station. All contours
were stored on the unenhanced data set and analysis
performed on this data set.
The data for each patient allow us to make a number of
comparisons of localization and planning:
1. To investigate the difference between marking-up by
the oncologist and radiologist: a comparison of the
original unenhanced and contrast enhanced scan
contours and the re-contour;
2. To investigate the influence of contrast: a comparison
of unenhanced and enhanced contours marked-up by
(i) oncologist (A) and (ii) the radiologist; changes in
contour data (volume, tolerance coverage and target
isocentre displacement) are determined;
3. To investigate the influence of the radiologist: a
comparison of GTV contours marked-up by oncolo-
gist (A) and the radiologist; also using the radiologists
Volume changes: contour pairs were analysed to
identify any changes in the volume of the GTV or PTV
673
 M McJury, K Dyker, R Nakielny et al

Figure 1. The image fusion workspace showing marked-up contours. If the CT coordinates are the same for both scans,
enhanced CT data on the left can be automatically fused to the unenhanced data on the right. Contours marked-up on the
enhanced scan are then automatically transferred to the unenhanced scan for storage with previous contours. Contours shown
are planning target volume (PTVunenh) (dark line) and gross tumour volume (GTVunenh) (light line).

contours. All unenhanced patient scans and contours
were imported into the CADPLANTM (Dosetek and
Varian Medical Systems) treatment planning system
(TPS) and dose–volume histograms (DVHs) were gener-
ated to yield values for GTV and PTV volume.
Geographical changes: pairs of GTVs were compared to
identify any geographical shift of the re-marked contours
from the position of the original GTV. On the AcQSIM
virtual simulator, shifts between the geometric centres of
the GTV contours were measured in three orthogonal
axes, defined as lateral (L), anterior/posterior (A/P) and
superior/inferior (S/I) shift. Using the Isocentre
Manager, the centre of gravity of each GTV was
identified automatically. Shifts between centres of
GTVs under comparison were then simply found by
subtraction of the coordinates in the orthogonal axes.
From these shifts, an overall 3D scalar value was
computed for each contour pair. The displacement of
the re-marked contours can be assessed in terms of
comparison with conventional treatment set-up toler-
ance. That is to say, we can note when the change in
geographical isocentre, due to the use of additional input
(contrast-enhanced data or radiologist input) is of the
same order as an alteration in patient geographical
set-up isocentre, which would conventionally require
action by staff to reposition the patient.
Area coverage changes: in the TPS, the original
treatment plan was applied to all subsequent sets of
contours. Pairs of PTVs (original and re-marked) were
then compared in terms of tolerance volumes (TV), i.e.
the percentage of the target which is either below 90%
prescribed dose (target under-coverage) or above 105%
prescribed dose (target over-coverage). If, for example,
the enhanced target contour is assumed to be the ‘‘true’’
target, the amount of under- or over-coverage the ‘‘true’’
target will experience can be measured, the original
treatment plan (based on the unenhanced target contour)
having been applied.
Results
Interclinician data
Table 2 shows a summary of the data for changes in
contour volumes. The unenhanced contours marked up
by the oncologist and radiologist show good agreement,
with percentage volume changes in the order of 1–6%.

674 The British Journal of Radiology, August 2006

Optimizing localization in radiotherapy

Table 2. Percentage volume changes for precision data Table 4. Tolerance coverage changes
Data Volume % GTV PTV Data comparison Over- Under-
comparison differences coverage coverage
Mean SD Mean SD
Mean SD Mean SD
Unenhanced – Intraoperator –4.1 62 1.2 24
unenhanced (Oncologist) Unenhanced – Intraoperator 0.2 3.2 5.6 5.9
Interoperator –3.2 52 5.9 33 unenhanced (Oncologist)
(Oncologist) Interoperator 0.8 4.8 6.0 12.2
Intraoperator –1.34 30.8 (Oncologist)
(Radiologist) Enhanced – Contrast 1.2 2.5 5.9 9.8
Radiologist- 39.1 131 unenhanced changes
Oncologist
SD, standard deviation.
Radiologist (GTV)+ 2.53 43.7
Oncologist (PTV)
Enhanced – Intraoperator – 1.7 7.4 required to mark-up GTVs only and an oncologist
unenhanced Oncologist generated a PTV based on this. Table 2 shows percentage
differences in contour volumes for the radiologist mark-
GTV, gross tumour volume; PTV, planning target volume; SD,
standard deviation. up precision and for the radiologist compared with the
oncologist. Mark-up precision for the radiologist is good,
and in fact slightly higher than that of the oncologist,
The mean percentage changes between pairs of GTVs
both in mean volume change and SD on the mean. There
and PTVs correlate well. The PTVs show smaller
is a marked difference in mean percentage GTV volume
standard deviations compared with the GTVs.
(unenhanced) between radiologist and oncologist, with
When re-contouring the contrast-enhanced images, the
the radiologist marking larger volumes. This is not
oncologist intraoperator GTV precision shows a signifi-
statistically significant for the group as a whole, but
cant (p,0.01, Wilcoxon signed rank test, two tailed)
when the data are split into head and neck, and brain
improvement with mean of 1.7 and SD of 7.4 compared
groups.
with mean 4.1 and SD 61.7.
For data split into the two groups (Table 6), the
Isocentre shifts between pairs of contours are shown in
precision of radiologist mark-up is lower in the brain
Table 3. For the oncologist, all shifts are within our daily
cases, although this is not significant. Whilst differences
setup tolerance of 5 mm for mould-immobilized
patients. Our standard protocol requires no corrective in (unenhanced) GTV mark-up between oncologist and
action by staff for errors in geographical isocentre set-up radiologist for the group as a whole are not significant,
of this level. Isocentre shift difference between contours when split into two, mark-up by the differing clinicians
marked by radiologist and oncologist are also higher is found to be significant at p50.01 (head and neck) and
(though not significantly) than precision data from either p,0.02 (brain). A similar trend significance is found for
operator, with value 5.8 mm, which is outside the limit PTV mark-up (p,0.05). Head and neck cases, in fact
for daily patient setup tolerance. show an improved precision for the radiologist com-
When the original plans were applied to the re- pared with the oncologist, with a lower intraoperator SD.
marked-up contours (PTVs), changes in target and Considering data where the oncologist marks-up a PTV
normal tissue coverage were calculated. Under-coverage based on a GTV delineated by the radiologist, again,
defined as less than 90% and over-coverage as greater significance change in unenhanced PTV percentage
than 105%. These are shown in Table 4. There is good volume change is only seen when the data is split into
agreement with plan coverage, with maximum percen- two groups, with head and neck cases showing a
tage changes being of the order of 5–6% for oncologist significant difference (p,0.02, Wilcoxon rank sum test).
mark-up precision.
Many consider a radiologist’s delineation of a GTV as
the gold standard, due to the radiologists’ specific Use of contrast
training [25, 26]. In this study, the radiologist was
Data for differences in volume between contrast-
enhanced and unenhanced scan contours is shown in
Table 3. Isocentre shifts for gross tumour volume (GTV) Table 5.
precision data For the group as a whole, the use of contrast in CT
Data comparison GTV isocentre shifts Mean SD scanning and subsequent mark-up by the oncologist,
leads to little (non-significant) change in GTVs and PTVs,
Unenhanced – Intraoperator 3.7 2.1
unenhanced (Oncologist)
similar in value to the typical precision differences seen
Interoperator 4.2 2.4 in Table 2. Neither of the percentage GTV volume
(Oncologist) changes (for oncologist or radiologist) was statistically
Intraoperator 4.2 1.6 significant. However, for the GTVs, a comparison of the
(Radiologist) impact of contrast for the radiologist and oncologist, is
Radiologist- 5.8 2.4 significant (p,0.01, Wilcoxon signed rank test).
Oncologist Obviously, contrast-enhanced data has a differing
Enhanced – Oncologist 4.8 2.5 impact on tumour delineation depending on which
enhanced Radiologist 5.2 2.6
clinician is using it. This result can also be compared
SD, standard deviation. with the difference between radiologist-oncologist

The British Journal of Radiology, August 2006 675


Table 5. Percentage volume changes for data involving
contrast-enhanced data or input from a radiologist
Data comparison Volume % GTV PTV
differences
Mean SD Mean SD
Enhanced – Oncologist –5.9 53 2.30 46
unenhanced Radiologist 98.8 172
Enhanced – Radioologist –4.1 24.2
enhanced GTV,
Oncologist PTV
GTV, gross tumour volume; PTV, planning target volume; SD,
standard deviation.
mark-up of unenhanced scans (see Table 2). The use of
contrast seems to have a greater difference in the
marking-up of the radiologist than that of the oncologist.
Isocentre shifts, which result from use of enhanced
data (Table 3), are similar to those of precision data.
Whilst there may be changes in volume, the geographical
position of the structure has not changed. In agreement
with data on volume changes, differences in coverage for
contrast-enhanced data also show little change to
precision values (see Table 4).
When comparing the difference between enhanced
and non-enhanced PTV mark-up, there is a statistical
difference (p,0.01) between these for the contours
generated from the radiologist’s GTV compared with
those generated solely by the oncologist, see Table 5.
Percentage volume changes for the radiologist mark-up
is significantly larger with contrast than without (p,0.01,
Wilcoxon signed rank test), so contrast is obviously a
much greater influence on mark-up for the radiologist
than for the oncologist.
When investigating the data split into two groups, the
mean percentage difference in GTV for head and neck
cases is larger, at mean –20.2% compared with 8.5% for
the brain group (although not statistically significant).
There is little difference in PTV changes between the two
groups, and similarly in isocentre shifts and tolerance
coverage (data not shown).
Discussion
These results highlight some difficulties inherent in
this study and in the marking-up process more generally.
Other authors have reported similar variations in
marking-up targets in radiotherapy [2, 3, 15, 17–24] due
to several factors.
Imaging
The first step in the process is the assimilation of
information from the available radiological images and
clinical data. Some authors have noted significant
variation amongst clinicians at this first step [4, 19].
When first marking-up for treatment, the oncologist
may have additional information, which he/she will not
have when re-contouring months later. They may have
seen and examined the patient recently or have
discussed the case with the surgeon regarding tumour
676
M McJury, K Dyker, R Nakielny et al
extent and geographic infiltration. Excluding the effect of
these factors is obviously difficult.
Clinicians when marking-up will use input from other
diagnostic scans which may not have been performed
with the patient in the same position as the planning
scan. The scans may be pre- or post-operative, and thus
be anatomically different. There may also be inconsis-
tencies in the timing and modalities of imaging available
for different patients. There is obviously a role here for
image registration methods. Image registration will
obviously increase the accuracy of combining diagnostic
data in the planning process. Furthermore, the applica-
tion of more advanced methods, such as elastic rather
than rigid-body registration methods, will improve
accuracy, especially for patients with data acquired on
different patient couches and patients in slightly differ-
ent positions.
It is also difficult, over the course of a long study such
as this, to monitor and control the use of notes and files
for additional information used in marking-up as they
are in constant clinical demand and so may not have
been available (this was not recorded).
Contouring
When contouring, the clinician will therefore have to
mentally translate visual information on certain planes
into contours in a different 3D plane. This complex
process leads to increased inaccuracies in the final
contour volume and may contribute to the large standard
deviations we see in this study. Other authors have
noted potential errors introduced in this step in the
overall process and report large variations in target
volumes [2, 4].
The CT images assist in delineating only the gross
tumour volume. Assessing microscopic tumour involve-
ment is difficult, highly subjective, and sometimes
controversial, as noted by some reports [23].
Subjectivity will also be introduced in margin growing
and editing when going from CTV to PTV [4, 22]. Indeed,
Yamamoto et al [2] report on reduced mark up precision
for CTVs compared with GTVs highlighting this
problem.
Although they mark up a GTV in a similar fashion,
different oncologists in an institution may have slightly
different philosophies when growing and editing mar-
gins. There can even be differences in interpretation of
the ICRU [16] criteria for marking between clinicians
[22].
PTV contouring precision should be better than GTV
precision (differences in marking-up at the upper and
lower extent of the target will have a smaller impact on
the larger PTV volume), but may not be due to large
variations in the size of margin added. The difference
could be in the order of 0.5–1.0 cm in 2D, making a large
difference to volume. Our results show little difference
between intraoperator and interoperator volume preci-
sion for oncologists. Volume differences are presented as
a percentage change from the original contour. In the
case of GTVs with small numbers of slices, adding or
removing a small numbers of slices will have a much
greater effect on the percentage change than for PTVs
which have 2–3 times the number of slices. At the
The British Journal of Radiology, August 2006
Optimizing localization in radiotherapy
Table 6. Data for the whole group split into head and neck,
and brain cases
Data comparison Head and neckBrain
Mean SD Mean SD
Unenhanced – Intraoperator 9.2 69 –16.5 54
unenhanced (Oncologist)
Interoperator –25.3 48 17.0 48
(Oncologist)
Intraoperator 5.4 20 49.7 136
(Radiologist)
Radiologist – –31.4 58 191 178
Oncologist
extremes of the tumour volume (most superior and
inferior positions) marking-up to include or exclude a
slice can often be very subjective and will impact on
precision results. This may explain the much improved
standard deviations for the precision of PTV contour
mark-up compared with GTV mark-up, which is in
agreement with other authors [20].
Size and sites
In this work, the data are recorded for a sample size of
25 patients, which will have an influence on the
uncertainties and statistical confidence. Although not
large in statistical terms, the sample size is certainly
larger then several similar studies in the literature [3, 4,
15, 19, 23]. Volumes were generated from full 3D CT data
using DVH algorithms on our TPS. Volume data will
therefore be accurate and should be an improvement on
comparative assessments used by other authors involv-
ing 2D assessment of maximal tumour extent on specific
CT slices [10], limited calculations from selected slices
from a volume set [3] or volumes manually calculated
from hardcopy films [22].
The cases in this work consist of small numbers from
several different sites as shown in Table 1. Analysis of
the data as a single group could obscure any benefit that
may exist for a particular site(s). In certain circum-
stances, therefore, the data have been further split into
smaller groups for consideration. With small numbers
for individual sites, it was considered statistically
prudent to merely split the data into two groups, namely
brain and head and neck. The difficulty in delineating
the target will depend on the site in question. Some
authors report a variation between different groups of
clinicians (radiologists and oncologists) when dealing
with more or less ‘‘difficult’’ cases [19].
In the case of some resections, the oncologist may
simply be marking-up a post-surgical cavity as a GTV,
although technically this should be a CTV. In others, they
can mark the site of the original tumour. Many of our
patients had only cavities remaining. For the brain
patients, the variation in marking was already so large
that any difference made by the contrast may be too
small to be detected. Some individuals seemed to be
marking up the tumour cavity only, and some were
marking a larger volume, more like a true CTV.
Consistency was lacking and it was not always clear if
The British Journal of Radiology, August 2006
the marked volumes were supposed to include a margin
for microscopic disease or not. Thus, we could have been
comparing unlike volumes for some patients. For these
patients, getting good agreement between oncologist and
radiologist mark-up may prove difficult and, indeed, we
do see poor agreement between these clinicians for brain
cases in particular. Other authors also report significant
differences in mark-up between oncologist and radiolo-
gist [4, 19]. Yamamoto et al [2] also note a variation in
precision between pre-operative and post-operative
cases. Without contrast, contouring is far more depen-
dent on the diagnostic scans. This leads to further
inaccuracies if tumour volume changes due to surgery
or chemotherapy have occurred.
The use of contrast may be very helpful for some sites,
and significantly less so for others. Unfortunately, with
low numbers and a large mix of sites, any large changes
or improvements for a particular site may not be
apparent when looking only at results for the entire
group and only by running a much larger study may
influences of this nature be more apparent. As an
example, consider the three cases shown in Figures 2–
4. Figure 2 shows pre- and post-contrast images for a
tonsil patient. The post-contrast scan offers little addi-
tional information about the GTV. The PTV was never
likely to show any significant change as it includes nodal
groups in the neck, and so is not solely dependent on
ascertaining the exact extent of the GTV. The head and
neck tumours in general were much easier to outline,
however, even using the diagnostic scan while volum-
ing, as vessels and nodes were more easily distinguish-
able from other soft tissues. The efficiency of contouring
is likely to improve with the use of contrast.
Figure 3 shows a set of pre- and post-contrast
images for a nasopharynx patient. This is helpful in
showing the extension into the cranial contents. This
illustrates that direct tumour enhancement is sometimes
useful in head and neck cases. Compare this with
Figure 4, which shows images of a patient with a
glioblastoma. Here, the periphery of the tumour,
between tissue and oedema, is well visualized in the
post-contrast image on the left, compared with the pre-
contrast scan on the right. We observed that contrast was
very helpful in identifying the GTV in patients who had
had a biopsy or minimal debulking only. For patients
who had maximal debulking of their brain tumours,
there was little enhancement.
There are several characteristics of the tumour itself
which can influence the impact of contrast on imaging.
Tumours in certain sites, e.g. oral cavity, are more likely
to enhance with contrast due to increased vascularity
compared with others, e.g. larynx. As mentioned above,
surgery not only changes the anatomy, but also the
vasculature and oedema can mimic tumour very well.
Contrast should make contouring more consistent, and
our results show that.
Finally, it must also be accepted, that although an
expert in interpretation of medical images, the radiolo-
gist is not as expert in radiotherapy treatment planning.
In agreement with other reports [4, 19], the input of the
radiologist has shown to lead to significant mark-up
changes, both in volume and isocentre shift, a stronger
influence than that of using contrast alone. In certain
circumstances, the radiologist may mark a considerably
677
 M McJury, K Dyker, R Nakielny et al

Figure 2. Pre- and post-contrast images of a tonsil patient. Enhanced scan is shown on the left. The gross tumour volume (GTV)
based on the unenhanced scan is shown in red.

different contour to the oncologist, but from our data it is
not possible to ascertain whether it is more or less
accurate.
In the study, all patients were treated using a plan
based on the unenhanced scans marked-up by the
oncologist. It is not possible, therefore, to compare the
outcome of patients treated on plans generated with and
without contrast and therefore it is not possible to say
clinically whether the use of contrast-enhanced scan data
or radiologist input resulted in more accurate, and
therefore, improved treatment. In essence, we cannot
say which of the clinicians’ mark-ups is the ‘‘true’’ one,
or most accurate.
Others also note differences between radiologists and
oncologists, with radiologists marking consistently
smaller volumes [4, 19], which is the opposite of what

Figure 3. Pre- and post-contrast images of a nasopharynx patient. Enhanced scan is on the left. The gross tumour volume (GTV)
based on unenhanced data is shown.

678 The British Journal of Radiology, August 2006

Optimizing localization in radiotherapy
Figure 4. Pre- and post- contrast images of a glioblastoma patient. Enhanced scan on the left. The gross tumour volume (GTV)
based on the unenhanced data is shown.
we see in this study. However, in this study we have
only a single radiologist compared with other reports
performed with a larger and perhaps more representa-
tive group of radiologists.
Conclusions
Radiotherapy has become more accurately targeted
over the last 10 years. This continues with the advent of
intensity-modulated radiotherapy (IMRT). As treatment
becomes more complex by requiring different dose levels
to be given to different areas, depending on level of risk,
our plans become more heterogeneous. This makes
accurate contouring a cornerstone of these advance-
ments, especially if we aim to dose escalate or alter
fractionation schedules.
It was reassuring to observe good intraclinician and
interclinician precision for the GTVs and PTVs, with
isocentre shifts within daily setup tolerance and plan
coverage changes also acceptably small. Brain cases
showed less precision than head and neck cases.
The use of contrast markedly improved the intraon-
cologist precision. The impact of using contrast caused
greater differences for the radiologist than the oncologist.
In fact, this difference was greater than the variation
between them. This was more marked for the head and
neck cases, although it was non-significant.
We found the radiologist marked significantly differ-
ent volumes for both GTV and generated PTV. Also, the
mean isocentre shifts for these contours were outside our
daily setup tolerance.
Contrast enhanced planning appears to offer benefit in
planning head and neck patients and those brain tumour
patients who have not had a maximal debulking
surgical procedure and so still have macroscopic tumour
remaining.
The British Journal of Radiology, August 2006
References
1. Le QTX, Fu KK, Kroll S, et al. Prognostic factors in adult
soft-tissue sarcomas of the head and neck. Int J Radiat
Oncol Biol Phys 1997;37:975–84.
2. Yamamoto M, Nagata Y, Okajima K, Ishigaki T, et al.
Differences in target outline delineation from CT scans of
brain tumours using different methods and different
observers. Radiother Oncol 1999;50:151–6.
3. Ketting CH, Austin-Seymour M, Kalet I, Unger J, Hummel
S, Jacky J. Consistency of three-dimensional planning target
volumes across physicians and institutions. Int J Radiat
Oncol Biol Phys 1997;37:445–53.
4. Logue JP, Sharrock CL, Cowan RA, Read G, Marrs J, Mott
D. Clinical variability of target volume description in
conformal radiotherapy planning. Int J Radiat Oncol Biol
Phys 1998;41:929–31.
5. Bowden P, Fisher R, MacManus M, Wirth A, Duchesne G,
Millward M, et al. Measurement of lung tumour volumes
using three-dimensional computer planning software. Int J
Radiat Oncol Biol Phys 2002;53:565–73.
6. Senan S, van Sornsen de Koste J, Samson M, Hankink H,
Jansen P, Nowak PJCM, et al. Evaluation of a target
contouring protocol for 3D conformal radiotherapy in
non-small cell lung cancer. Radiother Oncol 1999;53:
247–55.
7. Aoyama H, Shirato H, Nishioka T, Hashimoto S, Tsuchiya
K, Kagei K, et al. Magnetic resonance imaging system for
three-dimensional conformal radiotherapy and its impact
on gross tumour volume delineation of central nervous
system tumours. Int J Radiat Oncol Biol Phys 2001;50:821–7.
8. Vordermark D, Becker G, Flentje M, Richter S, Goerttler-
Krauspe I, Koelbl O. Transcranial sonography: integration
into target volume definition for glioblastoma multiforme.
Int J Radiat Oncol Biol Phys 2000;47:565–71.
9. Hawighorst H, Schreiber W, Knopp MV, Essig M,
Engenhart-Cabilic R, Brix G, et al. Macroscopic tumour
volume of malignant glioma determined by contrast-
enhanced magnetic resonance imaging with and without
magnetization transfer contrast. Magn Reson Imaging
1996;14:1119–26.
679

10. Zhou SM, Bental GC, Lee CG, Anscher MS. Differences in
gross target volumes on contrast vs. non-contrast CT scans
utilised for conformal radiation therapy treatment planning
for prostate carcinoma. Int J Radiat Oncol Biol Phys
1998;42:73–8.
11. Chao KS, Wippold FJ, Ozyigit G, Tran BN, Dempsey JF.
Determination and delineation of nodal target volumes for
head and neck cancer based on patterns of failure in
patients receiving definitive and postoperative IMRT. Int J
Radiat Oncol Biol Phys 2002;53:1174–84.
12. Tepper JE, Padikal TN. The role of computed tomography
in treatment planning, In: Bleehen NM, Glastein E,
Haybittle JL, editors. Radiation therapy planning. New
York, NY: Marcel Dekker, 1983:139–58.
13. McJury M, Nakielny R, Levy D, Lilley J, Conway J,
Robinson MH. Improving the localisation of radiotherapy
treatments in head and neck and brain cancer: some initial
findings. J Radiother Practice 2001;2:125–32.
14. Sharma R, Duclos M, Chuba PJ, Sharmsa F, Foreman JD.
Enhancement of prostate tumour volume definition
with intravesical contrast: a three-dimensional dosimetric
evaluation. Int J Radiat Oncol Biol Phys 1997;38:
575–8.
15. Valcenti RK, Sweet JW, Hauck WW, et al. Variation of
clinical target volume definition in three-dimensional
conformal radiation therapy for prostate cancer. Int J
Radiat Oncol Biol Phys 1999;44:931–5.
16. ICRU Report 50. Prescribing recording and reporting
photon beam radiotherapy. Bethesda, MD: ICRU, 1993.
17. Cazzanigna LF, Marinoni MA, Bossi A, et al. Interphysician
variability in defining the planning target volume in the
irradiation of prostate and seminal vesicles. Radiother
Oncol 1998;47:293–6.
18. Fiorino C, Reni M, Bolognesi A, Cattanero GM, Calandrino
R. Intra- and inter-observer variability in contouring
680
M McJury, K Dyker, R Nakielny et al
prostate and semial vesicles: implications for conformal
treatment planning. Radiother Oncol 1998;47:285–92.
19. Giraud P, Elles S, Helfre S, et al. Conformal radiotherapy
for lung cancer: different delineation of the gross tumor
volume (GTV) by radiologists and radiation oncologists.
Radiother Oncol 2002;62:27–36.
20. Foppiano Foppiano F, Fiorino C, Frezza G, Greco C,
Valdagni R. The impact of contouring uncertainty on rectal
3D dose-volume data: results of a dummy run in a
multicenter trial (AIROPROS01-02). Int J Radiat Oncol Biol
Phys 2003;57:573–9.
21. Senan S, Chapet O, Lagerwaard FJ, Ten Haken RK. Defining
target volumes for non-small cell lung carcinoma. Semin
Radiat Oncol 2004;14:308–14.
22. Tai P, Van Dyk J, Yu E, Battista J, Stitt L, Coad T. Variability
of target volume delineation in cervical oesophageal cancer.
Int J Radiat Oncol Biol Phys 1998;42:277–88.
23. Khoo VS, Adams EJ, Saran F, Bedford JL, Perks JR,
Warrington AP, et al. A comparison of clinical target
volumes determined by CT and MRI for the radiotherapy
planning of base of skull meningiomas. Int J Radiat Oncol
Biol Phys 2000;46:1309–17.
24. Geets X, Daisne JF, Arcangeli S, Coche E, De Poel M,
Duprez T, et al. Inter-observer variability in the delineation
of pharyngo-laryngeal tumor, parotid glands and cervical
spinal cord: comparison between CT-scan and MRI.
Radiother Oncol 2005;77:25–31.
25. Bowden P, Fisher R, MacManus M, Wirth A, Duchesne G,
Millward M, et al. Measurement of lung tumour volumes
using three-dimensional computer planning software. Int J
Radiat Oncol Biol Phys 2002;53:566–73.
26. Giraud P, Elles S, Helfre S, De Rycke Y, Servois V, Carette
M-F, et al. Conformal radiotherapy for lung cancer:
different delineation of the gross tumor volume (GTV) by
radiologists and radiation oncologists. Radiother Oncol
2002;62:27–36.
The British Journal of Radiology, August 2006
The British Journal of Radiology, 79 (2006), 681–687

SHORT COMMUNICATION

Evaluation of the larynx for tumour recurrence by diffusion-

weighted MRI after radiotherapy: initial experience in four cases
1
V VANDECAVEYE, MD, 1F DE KEYZER, MSc, 2V VANDER POORTEN, MD, PhD, 3K DERAEDT, MD,
3
H ALAERTS, MD, 4W LANDUYT, PhD, 5S NUYTS, MD, PhD and 1R HERMANS, MD, PhD

Departments of 1Radiology, 2Otorhinolaryngology, Head and Neck Surgery, 3Pathology and

5
Radiation Oncology, University Hospitals Leuven, Leuven and 4Department of Experimental
Radiobiology/LEO, Katholieke Universiteit Leuven, Belgium

ABSTRACT. Radiotherapy-induced changes in the soft tissues of the neck hamper the
early detection of persistent or recurrent tumour by clinical examination and imaging
procedures. Diffusion-weighted (DW) MRI is a non-invasive technique capable of
probing tissue properties by measuring the movement of water. The purpose of the
ongoing study is to examine the usefulness of DW-MRI for differentiation of persistent
or recurrent tumour from post-radiotherapeutic sequelae or complications. Four
patients, suspected of tumour recurrence after radiotherapy for laryngeal squamous
cell carcinoma, were examined using a DW-MRI sequence on a clinical 1.5 T MR system
prior to surgery. In two patients, the DW-MRI images showed an asymmetric
hyperintense lesion on b1000 images with low apparent diffusion coefficient (ADC)-
value, compatible with tumour on histopathology. All surrounding tissue presented
high ADC values and absent signal on the b1000 images, histopathologically correlating Received 12 October 2005
to post-radiotherapeutic changes. The images of the third and fourth patient showed Revised 9 February 2006
absent or minimal symmetric hyperintensity of the laryngeal soft tissues on the b1000 Accepted 20 February 2006
images and high ADC-values. In these cases, the histopathological diagnosis of
DOI: 10.1259/bjr/89661809
radionecrosis was made and no tumour was found. In all four cases, differentiation
of tumoral tissue from radiotherapy-induced tissue alterations was possible with ’ 2006 The British Institute of
DW-MRI. Radiology

The clinical and radiological detection of persistent or
recurrent head and neck cancer is difficult in the early
phase after radiotherapy (RT) [1]. Also, the differentia-
tion between tumour recurrence and laryngeal necrosis
can be challenging after radiotherapy.
Diffusion-weighted MRI (DW-MRI) is a non-invasive
technique capable of probing the micro-environment of
tissue by measuring water movement, and has not yet
been reported for evaluation of head and neck lesions in
a post-RT setting. Four patients are presented in whom
DW-MRI was used to evaluate possible tumour recur-
rence in the larynx after RT for squamous cell cancer
(SCC).
Imaging technique
All examinations were approved by the local ethics
committee.
The MRI study was performed on a 1.5 T SONATA
scanner (Siemens, Erlangen, Germany). A T2 weighted
Address correspondence to: Dr Robert Hermans, Department of
Radiology, Herestraat 49, 3000 Leuven, Belgium.
This work was partly financially supported by the research grant
‘‘Prof. em. A. L. Baert, Siemens Medical Solutions’’.
turbo spin-echo (TSE) sequence was performed in the
transverse plane, with parameters: 48 slices, 4 mm slice
thickness, 0.4 mm intersection gap, field of view (FOV)
of 20 cm625 cm, matrix of 2916512, repetition time
(TR)/echo time (TE)53080 ms/106 ms, 2 averages, an
echo train length of 9 and a resulting pixel resolution of
0.7 mm60.5 mm64.0 mm. The total acquisition time
was 5 min 42 s. Then, a T1 weighted TSE sequence in
the transverse plane was acquired with the following
parameters: 48 slices, 4 mm slice thickness, 0.4 mm
intersection gap, FOV of 20 cm625 cm, matrix of
2506512, TR/TE5775 ms/8.3 ms, 3 averages, an echo
train length of 19 and a resulting pixel resolution of
0.8 mm60.5 mm64.0 mm. The total acquisition time
was 5 min 35 s. This sequence was performed before
and after administration of 15 cm3 of gadolinium-BOPTA
(Multihance; BRACCO, Milan, Italy). In one patient, a fat
suppression pulse was added to the T1 weighted TSE
after contrast-administration, increasing the sequence
time to 6 min 59 s. Additional coronal or sagittal T1
sequences after contrast administration were used
depending on tumour localization.
Diffusion-weighted echo planar images (EPI) were
acquired with 48 slices in the transverse plane, band-
width of 1502 Hz/pixel, 4 mm slice thickness, 0.4 mm

The British Journal of Radiology, August 2006 681


intersection gap, FOV 20 cm625 cm, matrix of 1046128,
TR/TE57100 ms/84 ms, 3 averages and a resulting
pixel resolution of 2.0 mm62.0 mm64.0 mm. The
images were acquired using six different b-values (b50,
50, 100, 500, 750 and 1000 s mm22). The total acquisition
time for this sequence was 5 min 48 s. All diffusion-
sensitizing gradients were applied in three orthogonal
directions and combined to create a 3-scan trace. An
apparent diffusion coefficient (ADC) map was calculated
automatically using the built-in manufacturer’s software.
All sequences were acquired with identical geometry
to allow correlation of the DW images with the TSE
sequences.
Image analysis was performed on an off-line work-
station using dedicated software (BioMAP; Novartis
Pharma AG, Basel, Switzerland).
DW-MRI was analysed in a first step by visual inspec-
tion of DW images with b-value 1000 s mm22 (b1000
images) and ADC maps in correlation to the co-regis-
tered anatomical images. Afterwards, multiple regions of
interest (ROIs) were placed by two authors in consensus
(VV, 2 years of experience, and RH, 15 years of experi-
ence in head and neck radiology) over the larynx at
the supraglottic, glottic and infraglottic levels and
averaged, excluding the site of any suspect lesions.
Separately, the suspect sites were delineated. For all
ROIs the ADC values were calculated using all b-values.
More specifically, ADC-calculation was acquired from
the native DW-MR images. The ROIs were drawn on the
DW-MR images for b50 s mm22 and were then copied
to correct positions on all other images (b550, 100, 500,
750 and 1000 s mm22) automatically by the software
allowing for correct determination of signal intensity per
b-value and thus allowing for accurate ADC-calculation.
Both the qualitative and quantitative analyses were
correlated to the histological findings (localization of
tumour on DW-MRI and histological specimen, and
signal intensity versus histological tissue type).
Routine imaging procedures in three patients included
a CT study of the head and neck during intravenous
injection of a contrast agent, using a multidectector scan-
ner (Siemens Sensation 16). Collimation was 16 mm6
0.75 mm, feed/rotation 9.9 mm s21, tube voltage 120 kV
and mAseff 250. The effective slice thickness was 1.5 mm,
the reconstruction interval 0.75 mm; axial and coronal
slices were reformatted with a thickness of 2 mm parallel
to and perpendicular on the true vocal cords.
Whole-body FDG-PET studies were performed in
two patients on a CTI Siemens ECAT 931 (Siemens,
Knoxville, TN) with an in-plane spatial resolution of
8 mm and a transverse FOV of 10.1 cm for each bed
position. The emission scan was initiated 60 min after
the intravenous injection of 6.5 MBq kg21 FDG (to a
maximum of 555 MBq). The raw imaging data were
reconstructed in a 1286128 matrix with the use of an
iterative reconstruction algorithm.
Case presentations
Patient 1
3 months after completion of RT for a T1N0 SCC of the
right true vocal cord (25 daily fractions of 2.2 Gy
682
V Vandecaveye, F de Keyzer, V Vander Poorten et al
resulting in a total dose of 55 Gy), a 66-year old patient
presented with progressive hoarseness.
CT showed a contrast-enhancing nodular lesion in
the right true and false vocal cord, suggestive of tumour
recurrence (Figure 1).
The MR TSE sequences confirmed the irregular,
contrast-enhancing mass in the right hemilarynx. On
the b1000 images the lesion appeared hyperintense and
on the ADC map hypointense, with ADC value (in mm2
s21) of 0.8361023. The surrounding tissue at the
supraglottic, glottic and infraglottic levels showed no
clear signal on the b1000 images; this tissue was marked
by diffuse hyperintensity on the ADC maps, showing an
average ADC value (in mm2 s21) of 1.4361023 for all
measured ROIs.
A direct laryngoscopy under anaesthesia showed a
thickened right true vocal cord with adjacent granulation
tissue. Biopsies were taken and histological examination
confirmed the presence of tumour.
The patient underwent a total laryngectomy with
unilateral neck dissection. Histopathology confirmed
presence of SCC in the right vocal cord, characterized
by densely packed cells with large cytoplasmatic con-
tent and irregular nuclei, multiple mitoses and inter-
cellular bridging, anatomically corresponding to the
suspect lesion on DW-MRI. The surrounding tissue
showed post-radiotherapeutic changes, including vari-
able amount of inflammation, fibrosis and limited
necrosis.
Patient 2
6 months after completion of radiotherapy for a T2N0
SCC of the left true vocal cord (35 fractions of 2 Gy
resulting in a total dose of 70 Gy), a 54-year-old patient
showed on indirect laryngoscopy during routine follow-
up an irregular appearance of the left vocal cord. CT
showed an irregularly thickened, slightly enhancing left
true vocal cord, compatible with tumour recurrence; the
soft tissue infiltration extended into the false vocal cord,
subglottic region, anterior and posterior commissure,
and involvement of the cricoid cartilage was suspected
(Figure 2). The PET scan showed moderate FDG-uptake
at the level of the larynx; this finding was interpreted as
tumour recurrence.
On MRI, the TSE sequences confirmed the presence
of a contrast-enhancing mass in the left hemilarynx.
Despite the small dimensions, the b1000 image showed
an asymmetric hyperintense lesion corresponding to
the lesion seen on CT scan. On the ADC map, this lesion
was hypointense, with an ADC value (in mm2 s21) of
0.9661023. The signal of the surrounding tissue was
suppressed on the b1000 images and the ADC maps
showed diffuse hyperintensity, with an average ADC
value (in mm2 s21) of 1.4161023 for all measured ROIs.
Direct laryngoscopy was performed at the time of the
laryngectomy and showed the presence of a left sided
laryngeal mass. No biopsies were taken and surgery was
performed immediately. Histopathology confirmed SCC
in the left vocal cord corresponding to the suspect lesion
on DW-MRI (Figure 2). Typical neoplastic features
were present, showing densely packed cells with large
cytoplasmatic content and irregular nuclei, multiple
The British Journal of Radiology, August 2006
Short communication: Evaluation of the larynx using DW-MRI after radiotherapy

(a) (b) (c)

(d) (e) (f)

Figure 1. (a) Transverse CT image and (b) transverse contrast-enhanced T1 weighted MR image show an ulcerated lesion in the
right false vocal cord (arrows), corresponding to (c) a hyperintense rim on the b1000 image (arrows) and (d) a hypointense rim
on the apparent diffusion coefficient (ADC) map (arrows). (e) This rim corresponds histologically to hypercellular cancer tissue (T)
in the lateral part of the false vocal cord; medially, a necrotic ulcer (NU) is present. (f) A more detailed view of the tumour shows
multiple nests of densely packed squamous cell carcinoma (SCC) cells.

mitoses and intercellular bridging. The surrounding
tissue showed post-radiotherapeutic changes, including
variable amounts of inflammation, fibrosis and limited
necrosis.
Patient 3
value (in mm2 s21) of 1.8461023. A direct laryngoscopy
was not performed. Because of a worsening situation,
the patient underwent total laryngectomy. Histological
examination of the entire larynx showed severe radio-
therapy-induced changes with necrosis and inflamma-
tion, including purulent infiltration; no neoplastic tissue
was found.

A 45-year old patient presented with progressive pain,

dysphagia and dyspnoea, 4 months after completion of
RT for a T2N0 SCC of the right true vocal cord (35
fractions of 2 Gy resulting in a total dose of 70 Gy).
Clinical examination showed diffuse laryngeal oedema.
Progressive dyspnoea required placement of a tracheo-
stomy. CT showed diffuse and pronounced thickening
of the laryngeal soft tissues, with obliteration of the
laryngeal lumen. No focal soft tissue mass was dis-
cerned. The presence of some small gas bubbles in the
crico-arytenoid joints was interpreted as indicating
laryngeal necrosis (Figure 3). However, PET showed a
moderately hypermetabolic focus in the larynx; this
was reported as being suspect for tumour recurrence.
Similar to the CT findings, the TSE sequences on MRI
(Figure 3) showed diffuse soft tissue thickening in the
larynx, without a focal lesion. The b1000 images
showed no asymmetric hyperintense signal and the
ADC map showed diffuse hyperintensity with an ADC
Patient 4
A 53-year-old patient presented with progressive pain
and dysphagia 6 months after completion of chemo-
radiotherapy for a SCC of the left true vocal cord (35
fractions of 2 Gy resulting in a total dose of 70 Gy).
T-stage before treatment was unknown.
The TSE-sequences on MRI showed diffuse laryngeal
soft tissue thickening, more pronounced in the left true
vocal cord, and a soft tissue defect in the posterior part
of the right true vocal cord, suggesting laryngeal
necrosis. However, based on conventional MRI findings
a tumoral lesion could not be excluded on the left glottic
level.
Histological examination on multiple laryngeal biop-
sies taken during panendoscopy suggested the presence
of tumoral recurrence at the level of the true and false
vocal cords. DW-MRI showed only slight symmetric

The British Journal of Radiology, August 2006 683

 V Vandecaveye, F de Keyzer, V Vander Poorten et al

(a) (b) (c)

(d) (e) (f)

Figure 2. (a) Transverse CT image and (b) transverse contrast-enhanced T1 weighted MR image show thickening and increased
contrast enhancement of the left vocal cord (arrows), corresponding to a hypermetabolic spot on (c) coronal FDG-PET image
(arrow). The lesion is detected on (d) the b1000 DW-MR image (arrows) as a hyperintensity rim showing low intensity on (e) the
apparent diffusion coefficient (ADC) map (arrows). (f) Histological section shows corresponding tumoral infiltration (arrows).

hyperintensity on the b1000 images in the laryngeal soft
tissues, no focal lesion with low ADC-value could be
detected. The laryngeal soft tissues appeared hyper-
intense on the ADC map, with ADC value (in mm2 s21)
of 1.8761023.
The patient underwent total laryngectomy. Histo-
logical examination of the entire larynx showed severe
radiotherapy-induced changes with ulceration, necrosis
and inflammation but no neoplastic tissue was found.
Discussion
DW-MRI is a non-invasive technique able to depict the
extent of random movement of water protons in biological
tissues; the addition of two opposed magnetic field
gradients makes the signal intensity dependent on the
mobility of water molecules [2]. The amount of signal loss
over the range of b-values correlates with the mobility of
protons and is quantified by means of the ADC. The ADC
value is mainly influenced by the size of the extracellular
extravascular space (EES) containing free moving protons,
with additive contributions from bulk water movement
(e.g. intravascular flow) in the low b-value images and
minor contributions from intracellular diffusion and
transmembraneous transport [3]. Thus, any tissue archi-
tectural change causing structural barriers or influencing
the proportion of the extracellular versus the intracellular
compartment is expected to alter the ADC value.
The anatomic heterogeneity of the head and neck
region with numerous air–soft tissue interfaces makes
DW-MRI in this region prone to susceptibility artefacts,
ghosting and image distortion. However, recent technical
developments make EPI-based (including DW-MRI)
imaging feasible in this anatomically highly demanding
region. Application of parallel imaging decreases the
echo-train length, which in turn reduces off-resonance
and blurring artefacts [4], while application of high
bandwidth [5], thin slices and meticulous shimming
reduces image distortion and chemical shift artefacts.
The acquisition of images with a large range of b-values
allows a more accurate calculation of ADC and improves
the image quality of ADC-maps by reducing movement
artefacts and noise propagation.
Experimental and clinical data support the potential
use of DW-MRI for in vivo characterization of tissue.
Wang et al [5] showed a significantly smaller ADC for
malignant lesions, including SCC, than for benign lesions
in the head and neck. Furthermore, ADC measurements
provide reliable information on remaining viable tumour
tissue in the follow-up of human high-grade gliomas
after RT [6]. For SCC in an animal model, Herneth et al
show that DW-MRI differentiates viable from necrotic
tumour tissue [7].

684 The British Journal of Radiology, August 2006

Short communication: Evaluation of the larynx using DW-MRI after radiotherapy

(a) (b) (c)

(d) (e) (f)

Figure 3. (a) Transverse CT image and (b) transverse contrast-enhanced T1 weighted MR image show diffuse thickening and
contrast-enhancement of the false vocal cords, without focal nodular mass. Small air bubble adjacent to the right arytenoid can
be appreciated (a, arrow). (c) Coronal FDG-PET image shows tracer uptake at the laryngeal level. No asymmetric hyperintensity is
revealed by (d) b1000 and (e) the ADC map shows diffuse hyperintensity of the soft tissues at the same level. Histological
examination reveals necrosis, inflammation and purulent infiltration, without evidence for tumour recurrence. (f) Detailed
histological image shows granulation tissue and inflammatory infiltrate.

Recent progress in non-surgical treatment of head and
neck cancer, by combining multifractioned high-dose
radiotherapy (RT) with radiosensitizing measures [8] or
chemotherapy [9], allows us to obtain tumour control
even in advanced laryngeal SCC. The diagnostic accu-
racy of currently used clinical and imaging follow-up
procedures may be compromised when organ preserva-
tion is attempted in such advanced disease. Treatment-
induced tissue changes are anticipated to be more
pronounced in such circumstances, menacing the early
detection of persistent or recurrent tumour [10–12]. DW-
MRI may allow differentiation between neoplastic tissue
and post-radiotherapy inflammatory or necrotic tissue as
the differences in tissue microstructure are expected to
create differences in proton mobility.
In patients one and two, the appearance of recurrent
tumoral tissue on DW-MRI is illustrated. The lesions
were hyperintense on b1000 and hypointense on the
ADC map, contrasting with the surrounding tissue. His-
tologically, the recurrent SCC showed densely grouped
cells with large cytoplasmatic content and occasionally
intercellular bridging. These tumoral characteristics are
expected to restrict the movement of protons as the high
cellular index and large cytoplasmatic content increase
the cellular tissue fraction and reduce the EES [13].
The diffuse high ADC value and the absence of any
focal restrictive signal on the b1000 images in the
laryngeal soft tissues in the third and fourth patient
correlated with diffuse laryngeal necrosis and absence
of tumoral recurrence. The findings on DW-MRI were
contradictory to the FDG-PET findings in the third
patient. The increased laryngeal uptake of FDG in this
patient was presumably caused by granulation tissue
and metabolically active leukocytes [14]. Indeed, the
results reported on the value of FDG-PET in the post-
radiotherapy evaluation of laryngeal cancer are variable
[11, 12, 15]. The specificity of this technique is dimi-
nished by inflammatory tissue alterations present early
after radiotherapy.
In all four patients, the diffuse hyperintensity on the
ADC maps in the normal soft tissues of the larynx and
hypopharynx correlated with expected post-radiothera-
peutic alterations, such as inflammation and interstitial
oedema, promoting free movement of protons.
In all three patients examined with CT, a correct
diagnosis could be made. This technique has a high
accuracy for diagnosing recurrent laryngeal cancer after
radiotherapy, but false positive and false negative results
may occur [16]. Differentiation of tumour recurrence
from therapy-induced laryngeal necrosis based on

The British Journal of Radiology, August 2006 685

 V Vandecaveye, F de Keyzer, V Vander Poorten et al

(a) (b) (c)

(d) (e) (f)

Figure 4. (a) Transverse non-fat saturated T1 weighted MR image before and (b) fat-saturated image after administration of
contrast agent at the level of the vocal cords showing right-sided soft tissue ulceration (b, arrowhead) and tissue swelling on the
left side (b, arrows), not allowing exclusion of tumour recurrence. Both (c) b1000 image and (d) apparent diffusion coefficient
(ADC) map do not show a restrictive focal lesion, supporting the diagnosis of laryngeal necrosis without tumour recurrence.
(e) Five and (f) 10 times magnified histopathological sections show necrotic tissue (N), and stromal tissue with reactive changes
(S); multiple neutrophils and giant cells are visible, suggesting profound inflammatory reaction. No tumoral tissue was found.

anatomical findings may be problematic, although some
CT-findings allow the correct diagnosis to be made [17].
The accuracy of conventional MRI-techniques has shown
to be similar to CT in the post-radiotherapeutic neck [11].
As illustrated in the fourth patient, the asymmetric soft
tissue thickening, visible on conventional MRI, did not
allow exclusion of the presence of tumour recurrence.
However, DW-MRI showed no restrictive lesion in the
laryngeal soft tissues, supporting the diagnosis of post-
radiotherapeutic alterations or complications rather than
tumour recurrence. This specific ability of DW-MRI to
probe tissue microstructure is an interesting comple-
ment to the currently used imaging procedures in the
evaluation of the post-radiotherapeutic neck.
Conclusion
This report illustrates the ability of DW-MRI to
characterize the tissue changes observed in the post-
radiotherapeutic larynx. The use of DW-MRI may have
a complementary role in the non-invasive evaluation
of the larynx after RT. Early, and preferentially non-
invasive, differentiation of tumour recurrence from
a treatment-induced complication is desirable, as in a
number of patients with the latter condition the laryn-
geal function may be saved by conservative mea-
sures. Further studies in a large patient population are
ongoing to validate the reproducibility and diagnostic
accuracy of the technique in the post-radiotherapeutic
neck.
References
1. Parsons JT. The effect of radiation on normal tissues of
the head and neck. In: Million RR, Cassisi NJ, editors.
Management of head and neck cancer: a multidisciplinary
approach. Philadelphia, PA: Lippincott, 1994:245–89.
2. Stejskal EO, Tanner J. Spin diffusion measurements: spin-
echoes in the presence of a time-dependent field gradient. J
Chem Phys 1965;42:288–92.
3. Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J,
Laval-Jeantet M. Separation of diffusion and perfusion in
intravoxel incoherent motion MR imaging. Radiology
1988;168:497–505.
4. Taouli B, Martin AJ, Qayyum A, Merriman RB, Vigneron D,
Yeh BM, et al. Parallel imaging and diffusion tensor
imaging for diffusion-weighted MRI of the liver: pre-
liminary experience in healthy volunteers. AJR Am J
Roentgenol 2004;183:677–80.
5. Wang J, Takashima S, Takayama F, Kawakami S, Saito A,
Matsushita T, et al. Head and neck lesions: characterization

686 The British Journal of Radiology, August 2006

Short communication: Evaluation of the larynx using DW-MRI after radiotherapy
with diffusion-weighted echo-planar MR imaging.
Radiology 2001;220:621–30.
6. Hein PA, Eskey CJ, Dunn JF, Hug EB. Diffusion-weighted
imaging in the follow-up of treated high-grade gliomas:
tumor recurrence versus radiation injury. AJNR Am J
Neuroradiol 2004;25:201–9.
7. Herneth AM, Guccione S, Bednarski M. Apparent diffusion
coefficient: a quantitative parameter for in vivo tumor
characterization. Eur J Radiol 2003;45:208–13.
8. Kaanders JH, Pop LA, Marres HA, Bruaset I, van den
Hoogen FJ, Merkx MA, et al. ARCON: experience in 215
patients with advanced head-and-neck cancer. Int J Radiat
Oncol Biol Phys 2002;52:769–78.
9. Browman GP, Hodson DI, Mackenzie RJ, Bestic N, Zuraw
L, Cancer Care Ontario Practice Guideline Initiative Head
and Neck Cancer Disease Site Group. Choosing a con-
comitant chemotherapy and radiotherapy regimen for
squamous cell head and neck cancer: a systematic review
of the published literature with subgroup analysis. Head
Neck 2001;23:579–89.
10. Viani L, Stell PM, Dalby JE. Recurrence after radiotherapy
for glottic carcinoma. Cancer 1991;67:577–84.
11. Briggs RJ, Gallimore AP, Phelps PD, Howard DJ. Laryngeal
imaging by computerized tomography and magnetic
resonance following radiation therapy: a need for caution.
J Laryngol Otol 1993;107:565–8.
The British Journal of Radiology, August 2006
12. Terhaard CH, Bongers V, van Rijk PP, Hordijk GJ. F-18-
fluoro-deoxy-glucose positron-emission tomography scan-
ning in detection of local recurrence after radiotherapy
for laryngeal/pharyngeal cancer. Head Neck 2001;23:
933–41.
13. Lyng H, Haraldseth O, Rofstad EK. Measurement of cell
density and necrotic fraction in human melanoma xeno-
grafts by diffusion-weighted magnetic resonance imaging.
Magn Reson Med 2000;43:828–36.
14. Kubota R, Yamada S, Kubota K, Ishiwata K, Tamahashi N,
Ido T. Intratumoral distribution of fluorine-18-fluorodeox-
yglucose in vivo: high accumulation in macrophages and
granulation tissues studied by microautoradiography. J
Nucl Med 1992;33:1972–80.
15. McGuirt WF, Greven KM, Keyes JW Jr, Williams DW 3rd,
Watson NE Jr, Geisinger KR, et al. Positron emission
tomography in the evaluation of laryngeal carcinoma. Ann
Otol Rhinol Laryngol 1995;104:274–8.
16. Hermans R, Pameijer FA, Mancuso AA, Parsons JT,
Mendenhall WM. Laryngeal or hypopharyngeal squamous
cell carcinoma: can follow-up CT after definitive radio-
therapy be used to detect local failure earlier than clinical
examination alone? Radiology 2000;214:683–7.
17. Hermans R, Pameijer FA, Mancuso AA, Parsons JT,
Mendenhall WM. CT findings in chondroradionecrosis of
the larynx. AJNR Am J Neuroradiol 1998;19:711–8.
687
 The British Journal of Radiology, 79 (2006), 688–701

REVIEW ARTICLE

Non-invasive measurement of perfusion: a critical review of

arterial spin labelling techniques
1 1 1 1,2
E T PETERSEN, MSc, I ZIMINE, PhD, Y-C L HO, MSc and X GOLAY, PhD

1
Department of Neuroradiology, National Neuroscience Institute and 2Singapore Bioimaging
Consortium, A*STAR, Singapore

ABSTRACT. The non-invasive nature of arterial spin labelling (ASL) has opened a unique
window into human brain function and perfusion physiology. High spatial and
temporal resolution makes the technique very appealing not only for the diagnosis of
vascular diseases, but also in basic neuroscience where the aim is to develop a more Received 15 December
comprehensive picture of the physiological events accompanying neuronal activation. 2005
However, low signal-to-noise ratio and the complexity of flow quantification make ASL Accepted 29 March 2006
one of the more demanding disciplines within MRI. In this review, the theoretical
DOI: 10.1259/bjr/67705974
background and main implementations of ASL are revisited. In particular, the perfusion
quantification methods, including the problems and pitfalls involved, are thoroughly ’ 2006 The British Institute of
discussed in this article. Finally, a brief summary of applications is provided. Radiology

Perfusion, or the steady state nutritive delivery of
blood to the tissue capillary bed, is vital for the
homeostasis, and thereby survival, of an organ.
Accurate perfusion measurement can provide important
diagnostic information on pathological conditions, e.g.
whether an ischaemic organ is viable or not. Among
other methods, MRI has emerged as a powerful tool for
assessing tissue perfusion and possesses strong diag-
nostic and prognostic capabilities, especially when
combined with additional MRI modalities such as T1,
T2 and diffusion-weighted images.
Historically, radiology has principally offered mor-
phological imaging techniques; however, these new MRI
perfusion techniques will add routine physiological
imaging to its portfolio. This information is not only
valuable in clinical settings, but is of great importance for
basic physiological science and drug development,
where, for example, perfusion changes due to a
pharmacological stimulus might be monitored over time.
There exist two main MRI perfusion methods: bolus
tracking after the injection of an exogenous endovascular
tracer and arterial spin labelling (ASL), which uses
magnetically labelled water protons as an endogenous
tracer. The complete non-invasiveness of ASL makes it
very suitable for perfusion studies of healthy volunteers
and in patient groups requiring repetitive follow-ups.
This is especially important in patients with particular
conditions, such as kidney failure, or in paediatric
populations where the use of radioactive tracers or
exogenous contrast agents may be restricted. Recently,
ASL has gained more clinical acceptance partly due to
Address correspondence to: Xavier Golay, Singapore Bioimaging
Consortium, Agency for Science, Technology and Research
(A*STAR), Singapore, 138667. E-mail: xavier_golay@sbic.a-star.
edu.sg.
these advantages, but also as a result of the recently
introduced technique for imaging the perfusion territory
of individual blood vessels in the brain. In addition, the
increased availability of high-field (> 3 T) clinical
scanners has moved ASL from the research and devel-
opment stage towards the clinics.
The present article will focus on arterial spin labelling
approaches. Different implementations and their advan-
tages and disadvantages will be briefly reviewed. More
attention will be given to the different ways of modelling
the ASL data for perfusion quantification and the
problems and pitfalls involved. Finally, research and
clinical applications will be revisited.
Methodology
Terminology
The term ‘‘perfusion’’ refers to the process involved in
the nutritive blood delivery to the tissue’s capillary bed.
The physiology behind this blood delivery can be
described by many different parameters, such as the
blood flow itself, the volume of the blood vessels, the
time it takes a particle, such as a red blood cell to traverse
the vasculature. Finally, it can be described by the
velocity of these particles. The results of 2D and 3D brain
perfusion imaging techniques are commonly expressed
as cerebral blood flow (CBF), which has a typical unit of
[ml 100 ml21 min21]. As noted, this is a rate [s21] rather
than a volume-flux [m3s21] measurement, which one
would associate, for instance, with the flow from a pipe.
This is, however, convenient, as the average flow rate is
measured in a voxel of arbitrary volume. For historical
reasons, CBF is often stated as [ml 100 g21min21], a

688 The British Journal of Radiology, August 2006

Review article: Non-invasive measurement of perfusion
rather misleading unit because no information can be
obtained about the tissue mass of the individual voxels
and, in practice, a mean brain density of 1 g ml21 is
assigned to all voxels. The cerebral blood volume (CBV)
describes the fraction of a voxel that contains blood
vessels and is therefore dimensionless and usually
expressed as [%] or [ml 100 g21]. Finally, the mean
transit time (MTT) is the average time it takes a particle
to traverse the vasculature [s]. These are the main
parameters in general use, although measures such as
time to peak concentration of the tracer (TTP) and flow
heterogeneity (FH) can also be found in the perfusion
imaging literature. All of them provide information
about perfusion and are often referred to as perfusion
maps.
Quantitative perfusion
In order to measure tissue perfusion, one needs to
follow the course of blood flow through the organ, and
for this there exist two methods, one based on freely
diffusible and the other on intravascular tracers. As their
names suggest, freely diffusible tracers can leave the
intravascular space without restriction and be distri-
buted throughout the entire tissue volume, whereas
intravascular tracers remain in the vasculature, which
constitutes only a fraction of the full volume.
Early perfusion experiments carried out by the
pioneers Kety and Schmidt [1] used inhalation of free
diffusible nitrous oxide (N2O) as a tracer. This tracer
distributes throughout the entire tissue volume, having a
MTT in the order of minutes at normal physiological
flow values. The relationship between the flow, distribu-
tion volume and the MTT is described by the central
volume theorem, which states that the ratio between
volume and flow is equal to the MTT. This rather long
MTT allowed them to sample both arterial and venous
blood, as arterial and venous tracer concentrations
reached equilibrium after a few minutes. In this experi-
ment, the time to reach this equilibrium is directly
related to the cerebral blood flow using Fick’s principle:
dCt ðtÞ
~CBFðca ðtÞ{cv ðtÞÞ ð1Þ
dt
where Ct(t), ca(t) and cv(t) are the concentrations of tracer
in the tissue, arterial and venous blood, respectively. The
theory behind their method provides the basis for
arterial spin labelling, which uses magnetically labelled
water protons as the freely diffusible tracer.
However, when a freely diffusible tracer is used, there
is no access to the blood volume and an intravascular
tracer would be needed in order to obtain this volume
information. In this case, the distribution volume is equal
to the blood volume and using the terminology for the
brain, then according to the central volume theorem:
MTT 5 CBV/CBF. This principle is used in MRI bolus
tracking experiments, where an intravascular tracer such
as gadolinium-DTPA allows measurements of MTT, CBF
and CBV. Here, the first pass of the bolus is monitored,
an arterial input function (AIF) selected and based on the
indicator-dilution theory of Meier and Zierler [2] and, a
The British Journal of Radiology, August 2006
model-independent perfusion estimation can be
obtained [3]. Time-domain impulse functions are
employed to describe and compute the tissue perfusion
CBF, which is calculated using the deconvolution of the
tissue concentration curve Ct(t) by the measured arterial
input function Ca(t):
ðt
Ct ðtÞ~CBF:Ca ðtÞ6RðtÞ~CBF: Ca ðtÞRðt{tÞdt ð2Þ
0
R(t) is the residue function describing the fraction of
contrast remaining in the system after a given time, t. If,
alternatively, a steady state experiment is used, i.e.
letting the contrast distribute to the total body blood
volume, one can measure CBV by comparing images
from before and after contrast. For recent reviews on
these methodologies, see Østergaard et al, Grandin,
Grenier et al, and Barbier et al [4–7].
The theory behind the two methods described above
provides the basis for all applied perfusion measurement
techniques used today. They range from MRI to CT and
nuclear medicine imaging, all capable of measuring
perfusion and/or blood volume in different parts of the
body, depending on the selected tracer and imaging
technique.
Perfusion imaging encompasses physiological mass
transport and exchange mechanisms, where the system
is required to be stationary, linear and time invariant in
order to satisfy the underlying flow quantification
theory. This means that no major physiological altera-
tions are allowed during the acquisition scheme in order
to obtain quantitative perfusion. This should be kept in
mind when planning an experiment, especially when
dealing with functional studies for instance, where the
perfusion rate is manipulated over time.
Basic arterial spin labelling
The overall goal of all existing ASL techniques is to
produce a flow-sensitized image or ‘‘labelled’’ image and
a ‘‘control’’ image in which the static tissue signals are
identical, but where the magnetization of the inflowing
blood differs. The subtraction control-label yields a
signal difference DM that directly reflects local perfusion
because the signal from stationary tissue is completely
eliminated (Figure 1). The label is usually performed by
inverting or saturating the water molecules of the blood
supplying the imaged region. By adding a delay between
labelling and image acquisition, called inversion delay
(TI), the labelled blood spins are allowed to reach the
capillaries where they exchange with tissue water and
thereby give rise to the perfusion signal. The signal
difference, which is only 0.5–1.5% of the full signal,
depends on many parameters such as the flow, T1 of
blood and tissue, as well as the time it takes blood to
travel from the labelling to the imaging region. Multiple
repetitions are needed for ensuring sufficient signal-to-
noise, and a model of the perfusion signal is usually used
in order to quantify the perfusion.
There exist two main classes of ASL techniques:
continuous ASL (CASL) and pulsed ASL (PASL). In
689

Figure 1. Schematic description of a perfusion weighted
image (DM) obtained by subtraction of the labelled images
from the control images.
CASL, the supplying blood is continuously labelled
below the imaging slab, until the tissue magnetization
reaches a steady state (Figure 2a). The PASL approach
labels a thick slab of arterial blood at a single instance in
time, and the imaging is performed after a time long
enough for that spatially labelled blood to reach the
tissue and exchange at the region of interest (Figure 2b).
Both methods need a control experiment in order to
visualize and quantify the perfusion.
Continuous arterial spin labelling
The original ASL method proposed by Williams et al
[8, 9] in 1992 used a continuous flow-driven adiabatic
inversion scheme, a method that was previously used for
angiography [10]. This type of adiabatic inversion of the
arterial magnetization is realized using a 2–4 s contin-
uous radiofrequency (RF) pulse while applying a
magnetic field gradient in the flow direction. The moving
arterial spins will therefore experience a slow variation
of the resonance frequency, which will result in their
inversion, while static (tissue) spins will just be satu-
rated. Typically, the inversion ‘‘slice’’ will be selected
just proximal to the circle of Willis, near the medullosp-
inal junction or at the level of the common carotid, and
the spins in blood that flows through this plane will be
inverted. The inversion efficiency a, which is important
for further quantification, depends on factors like the
mean velocity of the blood, angulations of the vessels to
the plane and the selected RF amplitude and gradient
strength. Typical labelling efficiency is in the range of 80–
95% [8, 11–19].
Among the confounding factors of these long lasting
inversion pulses are magnetization transfer (MT) effects
[20]. When using a single coil for labelling and imaging,
the off-resonance labelling pulse (with respect to
imaging slice) will act as a powerful MT pulse in a
way similar to an MT-weighted technique. The resulting
saturation effect of the macromolecular pool will result
in a reduced signal of the free water pool from the tissue
of interest [21]. This is a very important issue, as the
perfusion weighted images are calculated by subtracting
690
E T Petersen, I Zimine, Y-C L Ho and X Golay
Figure 2. (a) Continuous arterial spin labelling (ASL) multi-
slice experiment, using double adiabatic inversion for the
control experiment, where labels get inverted during the
passage of the first plane and returned to equilibrium during
the subsequent passage of the second plane. (b) The EPISTAR
pulsed ASL sequence, which labels everything at once and
uses two 180 ˚ + 180 ˚ 5 0 ˚ pulses for the control images.
a labelled from a control acquisition, and if this MT effect
is present only during the labelling scheme, it will lead to
overestimated perfusion.
In the first implementation, these MT-effects were
compensated for by applying a distal labelling during
the control experiment. This produces identical satura-
tion effects but, due to the applied gradients during
labelling, this is unfortunately valid only for a single
slice. For multislice acquisition, Alsop et al [22] proposed
the use of two closely spaced inversion planes, also
called double adiabatic inversion (DAI). In the control
experiment, the magnetization gets inverted while
traversing the first plane and returns theoretically to its
original state during the passage through the second
plane (the CASL experiment is shown in Figure 2a).
Double inversion is achieved by applying a sinusoidal
modulation of the RF waveform. Global control of the
MT-effects is obtained by matching the RF power and the
location of the planes. Figure 3 shows a perfusion map
obtained using this method. Another method was also
proposed, called simultaneously proximal and distal RF
The British Journal of Radiology, August 2006
Review article: Non-invasive measurement of perfusion
irradiation (SPDI) for multislice acquisition [23]. In this
later scheme, the RF power on the control scan is
distributed evenly on both sides of the acquisition
volume. However, a big limitation of these approaches
is the doubled RF deposition, resulting in higher specific
absorption rates (SAR). This is particularly important at
higher field strength, and must be carefully considered
for human studies at 3 Tesla and above.
Using two coils is another way of avoiding MT-effects
and reducing the RF deposition [24–26]. In such
methods, a small dedicated coil is used for labelling the
carotid arteries and, due to the small physical extent of
the applied RF field, no saturation occurs in the imaging
region. Another advantage is that selective labelling of
each carotid artery is possible, allowing independent
mapping of the left- and right-internal carotid perfusion
territories [27–29]. The main disadvantage of this
approach is linked to the need for non-standard hard-
ware, such as a separate transmit channel and a detune-
able RF coil, which are not usually available on
commercial clinical scanners. Another disadvantage of
this approach is that the labelling takes place further
away from the imaging slices, resulting in increased
relaxation of the label before entering the imaging region.
This approach has recently been applied for full brain
coverage in humans at 3 Tesla without exceeding current
SAR limits [30].
Pulsed arterial spin labelling
In 1994, Edelman et al [31] proposed the first pulsed
ASL scheme. Contrary to CASL, the labelling is
performed once in a 10–15 cm slab proximal to the
image slices. For the PASL sequences, MT-effects have to
be considered as well, although these are much smaller
compared with CASL. In this first version of the ‘‘Echo-
Planar MR Imaging and Signal Targeting with
Alternating Radio frequency’’ (EPISTAR) sequence,
inversion was performed distal to the image slice during
the control experiment to induce identical MT effects in
both cases. Again, this truly compensates for a single
slice only and, therefore, the sequence was modified for
multislice acquisition using a single 180 ˚ adiabatic pulse
The British Journal of Radiology, August 2006
Figure 3. Full brain continuous
arterial spin labelling (CASL) DM
images acquired using double adia-
batic inversion [22] in a healthy 27-
year-old female. These images were
acquired on a 1.5 T scanner of the
F.M. Kirby Research Center for
Functional Brain Imaging at
Kennedy Krieger Institute.
for the label experiment and two 180 ˚ + 180 ˚ pulses of
half the power for the control experiment at the same
proximal location (Figure 2b) [31, 32]. Figure 4 shows a
multislice experiment using this technique.
Shortly afterwards, an alternative to this asymmetric
method of labelling was proposed by Kwong et al [33]
and independently by Kim [34], who named it flow
alternating inversion recovery (FAIR). Here, the label is
applied using a non-selective inversion pulse, while the
control employs a concomitant slice selective gradient
pulse. The symmetric nature of this sequence automati-
cally compensates for MT-effects.
PASL allows inversion of the arterial spins closer to the
image slices and the inversion efficiency a is improved;
however, due to imperfect inversion profiles, a gap
between the labelling region and the image slices of
typically 1–2 cm is needed, depending on the type of RF
pulse used. This increases transit time ta from the
labelling slab to the imaging slices leading to decreased
efficiency. In addition, T1 relaxation of all the inverted
spins will also result in a theoretically lower signal
difference. Nevertheless, the ease of implementation and
reduced practical problems, as compared with CASL,
have made PASL a popular choice for perfusion imaging.
This is reflected in the wide range of sequences available
today, the common ones being listed in Table 1. For
further in depth explanation of these sequences, see
recent ASL reviews [4, 35].
ASL perfusion quantification
Having acquired the data using either technique, the
subtracted control-label images will be perfusion
weighted (Figure 1). The relationship between the DM
signal and the actual CBF depends mainly on proton
density and T1 relaxation rates of tissue and inflowing
blood, the and their respective differences. In addition,
the label transit time from the inversion slab to the
observed region in the images is also an important factor.
Traditionally, quantitative CBF estimation is carried out
using the tracer clearance theory originally proposed by
Kety and Schmidt [1], which was first adapted to ASL
experiments by Detre and Williams et al [8, 9]. In the
691
 E T Petersen, I Zimine, Y-C L Ho and X Golay

Figure 4. Full brain pulsed arterial

spin labelling (PASL) using EPISTAR/
PULSAR [32]. Images are the average
of 30 control-label pairs acquired in
3 min using a TR of 3 s and TI of
1.7 s.

original model, it is assumed that the labelled arterial
blood water is a free diffusible tracer, implying that the
exchange of blood water with tissue water happens
instantaneously upon its arrival to the parenchyma.
Therefore, this model corresponds to a single compart-
ment tracer kinetic, described by a mono-exponential
tissue response function. The modified Bloch equation,
including the flow dependent exchange term, becomes:


dMt ðtÞ Mt,0 {Mt ðtÞ Mt ðtÞ
~ zCBF Ma ðtÞ{ ð3Þ
dt T1t l
where Mt, Mt,0 and Ma are the tissue-, equilibrium- and
arterial-magnetizations, respectively, l is the blood–brain
partition coefficient, and T1t is the longitudinal relaxation
rate of the tissue. In the original quantification model,
further assumptions about uniform plug flow and equal
T1 relaxation of both tissue and arterial blood were made
[8, 9].
Many derivatives and improved versions of this first
solution exist for both CASL and PASL experiments.
Calamante et al [36] took the difference in R1 relaxation
rate for tissue and arterial blood into account as well as

Table 1. Common arterial spin labelling sequences

Method Sequence name Reference

CASL (Asymmetric) The original continuous arterial spin labelling [8, 9, 120]
DAI (Double Adiabatic Inversion) [22]
(Symmetric) SPDI (Simultaneously Proximal and Distal RF Irradiation) [23]
(Two coil) Two coil methods [24, 25]
PASL (Asymmetric) EPISTAR (Echo-Planar MR Imaging and Signal Targeting with Alternating Radio frequency) [31]
PICORE (Proximal Inversion with Control for Off-Resonance Effects) [121]
TILT (Transfer Insensitive Labelling Technique) [44, 52]
DIPLOMA (Double Inversion with Proximal Labelling of bOth tagged and control iMAges) [122]
STAR-HASTE (Signal Targeting with Alternating Radio frequency - HAlf-fourier Single [123]
shot Turbo spin-Echo)
PULSAR (PULsed Star labelling of Arterial Regions) [32]
QUASAR (QUAntitative Star labelling of Arterial Regions) [124]
(Symmetric) FAIR (Flow Alternating Inversion Recovery) [33, 34, 125]
UNFAIR (UNinverted Flow Alternating Inversion Recovery) [126]
FAIRER (Flow Alternating Inversion Recovery Extra Radiofrequency pulse) [127, 128]
FAIRER (Flow Alternating Inversion Recovery Excluding Radiation damping) [129, 130]
BASE (unprepared BAsis and SElective inversion) [52, 131]
CASL, continuous arterial spin labelling; PASL, pulsed arterial spin labelling.

692 The British Journal of Radiology, August 2006

Review article: Non-invasive measurement of perfusion

the trailing edge td of the bolus. In the work of Kwong

et al [33], the transit time ta was also considered.
Buxton et al [37] proposed a general kinetic model
where all the above mentioned parameters can be taken
into account. Here, the magnetization difference between
labelled and control measurements is described using
the convolution integral in a way similar to Equation (2):

ðt
DM~2 Ma,0 CBF cðtÞ:rðt{tÞ:mðt{tÞdt
: : : ð4Þ
0

where M0,a is the equilibrium magnetization in a blood

filled arterial voxel, c(t) is the delivery function or
fractional arterial input function (AIF). The residue
function r(t2t) describes the washout of labelled spins
from a voxel, and m(t2t) includes the longitudinal
magnetization relaxation effects. The possibility to
choose a particular arterial input function and to Figure 5. Diagram on an arterial spin labelling (ASL) time
course. In this graph, an example of two voxels having the
consider a certain exchange mechanism (single or
same flow (60 ml 100 g21 min21) and relaxation character-
multicompartment) allows greater flexibility for data istics, but different arrival time, is presented. The measured
analysis of both CASL and PASL experiments. signal at a single inversion time of 1.5 s would result in a
However, analytical solutions are only possible with 12% lower DM signal in a voxel with delayed arrival of
simple assumptions, and the most widely used 500 ms (open square) as compared with a voxel with an
‘‘Standard Model’’ can be summarized using the follow- arrival time of 200 ms (closed square).
ing terms:
8 Different issues relating to the actual perfusion
>
> 0, tvta
>
> quantification, using a single or multiple inversion time
>
< a:e{t:R1a ,ðPASLÞ ta ƒtvtd points, and the possible errors introduced using the
cðtÞ~ standard model vs more advanced methods, will be
> {ta :R1a
>
> a:e ,ðCASLÞ ta ƒtvtd
>
> ð5Þ discussed next.
:
0, t§td
:
CBF t
r(t)~e{ l

:
Quantification errors
m(t)~e{t R1t
Transit time
The major source of error in the quantitative estimate
Again, the assumptions are: uniform plug flow and of cerebral perfusion is the arterial transit time, ta, which
fast exchange, which is equivalent to single compartment even in healthy subjects differs across the brain, being
kinetics. Solving Equation (5) using Equation (4) in the longest in distal branches, especially in the regions
case of a PASL experiment gives a stepwise defined between perfusion territories also known as border-zone
equation: areas. In most PASL approaches, information on perfu-
sion is assessed at a single inversion time point
DM ðtÞ~ (Figure 5), and therefore without information about the
8
> 0, tvta transit time. Quantification is then based on the second
>
< {2:a:M   ð6Þ
a,0
:CBF
{R1a :t
e 1
e
dR:ðt
ta Þ
, ta ƒtvtd step in Equation (6), either where ta is assumed equal all
dR
>
>   (6) over, or simply set to zero. The problem in doing this is
: {2:a:Ma,0 :CBF
R1a :td {dR:ðt
ta Þ : {R1app :ðt
td Þ
dR e 1{e e , t§td that not only is the quantification of CBF impossible, but
relative perfusion values between regions are not valid
where, dR 5 R1a – R1app and R1app 5 R1t + CBF/l, also either. Sequences like QUIPSS II and Q2-TIPS [38, 39]
called the apparent tissue relaxation rate. A similar set of were developed to render ASL more transit-time
equations can be obtained for the CASL experiments, insensitive. Constant bolus duration is assured by
with the only difference being that c(t) in Equation (5) is saturating the part of the label that remains within the
constant for CASL, whereas it is subject to T1a decay for labelling slab at a time delay short enough to be able to
PASL, since the latter is not a steady-state method. As intersect the trailing edge of the fastest blood.
can be seen, various parameters like the transit time ta, Appropriate timing of this saturation and the following
blood–tissue partition coefficient l, Ma,0, R1a and R1t need image acquisition makes these sequences insensitive to
to be estimated or measured in order to obtain the transit time. This is really an advantage in volunteers
quantitative CBF values. The difference between the and patients without vascular diseases where the
many ASL sequences is mainly in the measurement (or difference in transit time is small (,1.5 s) [40], as well
not) of these parameters. as for quantitative functional studies where it is known

The British Journal of Radiology, August 2006 693


that the transit time changes are small (0.1–0.2 s)
between baseline and activation conditions [41, 42].
However, these methods will fail in patients with
atherosclerosis, where the transit time can be long in
affected areas (.2.0 s) due to low perfusion velocity and,
in some cases, extensive collateral perfusion. In these
cases, the problem can be solved by acquiring images at
multiple inversion times and therefore measuring the
entire DM curve (Figure 6). The standard sequences, in
which a single time point is acquired at a time (2–4 min
per time point) are generally too lengthy to be suitable
for clinical examinations [37, 41]. Günther et al [43]
introduced an elegant solution to this problem using a
Look-Locker-like readout to measure the ASL signal at
multiple inversion times in a single scan (Figure 6).
Because of multiple low flip angle readouts, the general
model needs to be modified by substituting R1app in
Equation (6) with R1app,eff 5 R1t + CBF/l - ln(cosw)/DTI,
where w is the flip angle and DTI is the interval between
the excitation pulses. A similar scheme was recently
implemented using the transfer insensitive labelling
technique (TILT) [44] by Hendrikse et al [45].
CASL is also sensitive to the transit time, but because
of the steady state behaviour of this type of sequences,
the effect is smaller than in PASL. Further improvements
can be achieved using a pre-delay of typically 0.9–1.5 s
between the continuous labelling and the readout, which
renders CASL methods almost insensitive to transit time
differences [46]. However, the duration of the delay
should be chosen according to the subject’s condition; in
a healthy volunteer, a delay of 1 s would be suitable,
Figure 6. Dynamic perfusion characterization. In this time-
course diagram, the black triangles show an example of
repeated acquisition at multiple inversion time points.
Usually, a three-parameter fit model is applied [37] to
estimate cerebral blood flow (CBF) when multiple inversion
times arterial spin labelling (ASL) sequences are used, which
reveal additional information about arrival time and bolus
duration. The second time-course (white triangles) has been
calculated with identical perfusion parameters, while taking
into account a Look-Locker readout method [43, 45, 124]
(flip angle 5 30 ˚). The advantage of this method is a higher
signal to noise ratio (SNR) as per acquisition time.
694
E T Petersen, I Zimine, Y-C L Ho and X Golay
while in patients with cerebrovascular diseases, longer
delays are necessary [46].
Vascular artefacts
Vascular artefacts, associated with the inflow of
labelled arterial blood into the arteries, can introduce
important errors in CBF quantification. Strictly speaking,
for Equation (5) to be valid, none of the DM signal should
originate from within the arterial vasculature. This
assumption is often violated in voxels containing feeding
vessels or traversing arteries, resulting in a substantial
overestimation of perfusion values. Ye et al [47]
proposed the use of bipolar crusher gradients to
eliminate the signal from the large feeding arteries in
CASL experiments. This was also adapted for PASL
sequences. Another solution is to choose a sufficiently
long inversion time, allowing the feeding vessels to
empty before acquisition [40, 46–48].
Inversion pulse shape and efficiency
In PASL sequences, where a spatially defined label is
used, the shape of the inversion pulse is of great
importance. Ideally, the profile should be truly rectan-
gular, allowing zero spacing between the labelling slab
and the imaging region; due to finite duration of the RF
pulse, this is not realizable in practice. Imperfect profile
can reduce labelling efficiency, but, more importantly,
there can be contamination in the imaging region from
the labelling slab. To avoid this, a gap between the
inversion and image plane is often introduced, with
increased transit times as a result. In order to minimize
this gap, longer adiabatic pulses like the hyperbolic
secant (HS) [49], frequency offset corrected inversion
(FOCI) [50] or bandwidth-modulated adiabatic selective
saturation and inversion (BASSI) [51] pulses are often
used. Another approach, as implemented in the TILT
sequence [44], is the self-refocusing concatenated 90 ˚
Shinar-Leroux pulse combination that maintains the
profile efficiency of the 90 ˚ pulses [52].
For CASL labelling, where labelling is performed in
general at a distance from the imaging slices, the
‘‘profile’’ is less of a concern. It is the fulfilment of the
adiabatic condition ensuring proper inversion that can be
problematic [9, 15, 16, 22, 53].
Signal to noise issues
At normal perfusion rates of 40–100 ml 100 g21 min21
the signal change DM is in the order 0.5–1.5% of the full
signal. Therefore, an average from typically 30 to 40 pairs
of subtracted control and labelled images are required to
get the desired signal to noise ratio (SNR) in the
perfusion-weighted maps. A total scan time of 3–4 min
is needed in order to acquire these data, which makes the
technique very sensitive to motion artefacts. Proper head
fixation and collaboration of the subject are necessary in
order to obtain good results. Fast imaging techniques like
single shot EPI and spiral sequences are often used to
reduce the scan time between successive control-label
The British Journal of Radiology, August 2006
Review article: Non-invasive measurement of perfusion
pairs as well as overall scan time. Also, prior saturation
of the image plane reduces the sensitivity to motion. In
case of limited motion, realignment of the images can
improve quantification, but in general it is compromised
because of low resolution, small number of slices, and
small DM signal.
Finally, the running state of the scanner is important.
Having a perfusion signal DM at only a fraction of the
full signal, a slight drop in the scanner performance can
preclude acquisition of clinically relevant ASL images,
even if no effects can be detected in the standard clinical
images.
Fast vs intermediate water exchange
The complex nature of the brain vasculature might not
act like a single compartment to water based tracers. In
particular, the blood–brain barrier seals the vasculature
from the extravascular space, allowing a dynamic inter-
face that protects the brain from toxic substances while
allowing nutrients and other essential compounds to
pass and thereby maintain its homeostasis. The ‘‘free
diffusible’’ water is known to be limited to ‘‘free’’
diffusion through the lipid membrane of the endothelial
cells as well as through dedicated water channels [54,
55], whereas diffusion in between adjacent endothelial
cells is impossible due to tight junctional complexes [56].
The speed of the exchange in comparison with the 2–
5 s duration of the ‘‘control’’ or ‘‘label’’ experiment still
remains controversial. At higher flow rates, the exchange
appears restricted [54], as though the water channels
have been saturated. In order to deal with these effects, a
few groups have suggested more elaborate multicom-
partmental approaches [57–59]. In general, the conclu-
sion is that these effects are negligible at normal human
perfusion rates at 1.5 Tesla, whereas the effect might be
more pronounced when going toward higher field
strengths. However, the intrinsic low SNR of the ASL
techniques combined with the problems related to fitting
all these additional exchange parameters will keep these
models within the animal experimental world for at least
another few years.
Bolus dispersion
Even when using vascular crushing and modelling
assuming multicompartmental behaviour, it still leaves
another possible error source, namely the dispersion of
the labelled bolus. In a PASL experiment the label is close
to being rectangular at the time of labelling and if the
blood does not experience any resistance, this shape
would persist all the way to the image slices. In reality,
the blood friction with the vessel wall, resistance in
bifurcations and the pulsatile behaviour of the flow make
the flow profile look like something between a parabolic
and ideal plug flow profile [60]. This results in disper-
sion of the bolus while it travels from the labelling slab to
the imaging region, which will lead to underestimation
of the perfusion.
Various attempts to incorporate these effects in the
modelling have been made, and Hrabe et al [61] recently
provided two analytical solutions that might reflect the
The British Journal of Radiology, August 2006
physiology better. While the dispersion might be fairly
homogeneous in healthy volunteers, patients with
vascular diseases can be expected to have completely
different dispersion in affected areas, and the flow
estimation using these advanced models would still be
non-quantitative.
Blood–brain partition coefficient
When a tracer consists of water or is dissolved in
water, it is necessary to know the ratio of water in the
tissue of interest and the feeding blood water in order to
correct the distribution volume and thereby obtain the
correct perfusion values. In the ASL technique, this is
more related to the proton density, which tells us about
how much signal we can expect from different tissues.
This ratio is called the blood–brain partition coefficient
l, and it was first defined by Kety [62] and corrected by
Herscovitch et al [63] to the commonly used values
today: whole brain l 5 0.9 ml g21, grey matter lg 5
0.98 ml g21 and white matter lw 5 0.82 ml g21.
The relevance of this coefficient is controversial; first of
all because the valid use of it requires instantaneous
exchange of the tracer, which, as earlier mentioned, can
be questionable, but also due to the fact that Roberts et al
[64] demonstrated that this coefficient not only varies
between different tissues, but also from region to region.
Nevertheless, a whole brain value of 0.9 ml g21 is
commonly used for the calculations. Notice the unit of
ml g21, which means that the unit of CBF/l in Equation
(5) becomes s21 when CBF is expressed as ml g21 s21.
Blood equilibrium magnetization
The inflowing blood has a proton density different
from the tissue. For absolute quantification, the equili-
brium magnetization of the arterial blood Ma,0 is needed,
i.e. the available longitudinal magnetization from a fully
relaxed blood filled voxel. This could be obtained from a
partial volume-free arterial voxel, but with a standard in-
plane resolution of 3–4 mm this is not possible in
practice. Alternatively, this information can be taken
from the much larger sagittal sinus. However, the T2* of
the deoxygenated venous blood, especially at higher
field strength, is shorter than that of arterial blood. This
results in a underestimated Ma,0. A different approach is
to measure the equilibrium magnetization in a grey or
white matter region and estimate Ma,0 using the blood–
brain partition coefficient. For single inversion time
experiments, this is often done on a voxel by voxel basis
using the control experiment as Mt,0-map after correction
for TR and T1.
Note that Ma,0 is a direct scaling factor of the perfusion
(Equation (6)) and therefore an error in the product
a?Ma,0 will directly change the calculated CBF value.
Functional studies and non-steady state perfusion
Finally, a requirement for using the various models
and formulae for perfusion quantification is that tissue
perfusion is a stationary, linear and time invariant
695

system. This would most often be satisfied in standard
perfusion scans where the subject is resting. However,
ASL has gained more and more popularity in functional
studies where perfusion is monitored during various
stimulation paradigms. Here, a typical experiment
would have a repetition time of 2–3 s, while 4–6 s would
be needed in order to acquire both control and label.
Now, in functional experiments, the haemodynamic
response time is often in the same range, which results
in a violation of the steady state requirement [65, 66]. Not
only will the flow change in between the control and
label experiment, but also in between labelling and
acquisition.
Two main problems arise from this; first, the models
do not assume increasing or decreasing perfusion during
the experiment and calculation errors will occur, and
second, and probably the most important, is that T2* will
change due to the BOLD effect, which will reduce the
signal intensity changes when acquiring at typical echo
times of 15–25 ms.
In the case of a standard block paradigm, where the
rising edges are discarded in the statistical analysis, this
is not a problem. With basic research moving towards
the characterization of the haemodynamic response on
the other hand, the information in the rising and the
falling edges are especially important. Recently, Lu et al
[67] proposed a new subtraction method for reducing the
T2* effect, which is very suitable for this type of
experiments, however, without addressing the problem
of non-steady-state regimen.
Comparison with other imaging modalities
At this point, it should be kept in mind that related
issues such as the ones highlighted in the previous
paragraphs apply to other perfusion modalities like
DSC-MRI, CT-perfusion and positron emission tomogra-
phy (PET) as well. Dynamic susceptibility contrast using
tracers such as gadolinium-DTPA for instance, suffers
from the fact that the relationship between measured
signal and contrast is non-linear and depends on
parameters such as the field strength, shim of the magnet
and the constitution of the vessels [68]. In addition,
correct scaling of the AIF is influenced by partial volume
effects in the voxels from where it is measured.
Altogether, this makes quantification troublesome, and
in the general clinical practice only a relative perfusion
measure is possible. Similar problems exist in CT-
perfusion, mainly related to the extraction of the global
AIF [69]. PET can be said to be a more ‘‘pure’’ method for
measuring perfusion due to the use of free diffusible
tracers and experimental durations that ensure a steady
state. However, a relatively low resolution of typically 6–
10 mm introduces partial volume effects, i.e. a mixture of
grey-matter, white-matter and CSF will be present in
almost all voxels, making direct comparisons to the
abovementioned methods questionable.
Applications
Despite the problems related to the quantification of
perfusion using ASL, this technique has been used in
696
E T Petersen, I Zimine, Y-C L Ho and X Golay
numerous applications, ranging from basic neuroscience
using animal models and human volunteers to clinical
perfusion measurement in pathologies such as stroke
and brain tumours. Furthermore, as long as the same
procedure and parameters are consistently used, repro-
ducible results can be achieved using ASL. Also the fact
that most clinical decisions can be based on relative
differences in perfusion rather than absolute measures
could render this method useful in the daily clinical
practice.
Although most publications on ASL to date have
focused on brain perfusion, other organs like the lungs,
kidneys and the heart have recently gained attention
with improved techniques and hardware available. In
particular, the move toward 3 T high field scanners in the
standard clinical environments seems to push this
method from the research and development stage
towards clinical applications.
The various applications of ASL have recently been
extensively covered elsewhere [4, 35] and only a short
summary of the active fields and most recent applica-
tions will be listed here. Two main categories exist – the
basic science, mainly neuroscience, and the general
clinical use.
Neuroscience
The ability to measure CBF is very important for the
assessment of tissue metabolism and function. The
complete non-invasiveness of ASL, which allows pro-
longed functional studies to be performed on any
volunteer, makes it a preferable choice in many
neuroscience applications. Following a neuronal activa-
tion paradigm, the classical blood oxygen level depen-
dent (BOLD) contrast [70] is a result of T2* changes due
to alterations of CBF and CBV, as well as the cerebral
metabolic rate of oxygen uptake (CMRO2). The ASL
signal, on the other hand, is an absolute measure of CBF
changes, which makes this technique more reproducible
over time, as well as in between subjects [71–73]. In
addition, perfusion functional MRI (fMRI) is believed to
localize regions of activation more accurately [74],
contrary to BOLD, which is affected by the change in
deoxygenated blood in draining venous vessels resulting
in additional signal from ‘‘down stream’’ areas. In
addition to flow information, the BOLD signal can also
be extracted from an ASL [75, 76]. Recent research using
these techniques includes: Lu et al [77], who combined
ASL, BOLD and vascular space occupancy (VASO) and
Obata et al [78], who applied the ‘‘balloon model’’ to
simultaneously acquire BOLD and ASL data. Finally,
Hoge et al [79] combined optical and ASL imaging
methods. All three studies were aimed at developing a
more comprehensive picture of the physiological events
accompanying activation.
For pharmaceutical validation, the suitability of ASL
for prolonged functional studies is advantageous for
studying perfusion-altering drugs. Alternative techni-
ques often require the injection of a tracer, which cannot
be repeated more than a few times and would possibly
interfere with the drug itself. Among investigated
pharmaceutical agents are the vasodilators acetazola-
mide [80, 81] and 2-chloradonosine [82, 83], as well as
The British Journal of Radiology, August 2006
Review article: Non-invasive measurement of perfusion
the vasoconstrictor indomethacin [84, 85] and finally
isoflurane, which is used for anaesthesia [86, 87]. This
field can be expected to expand with broader awareness
of ASL and availability of the sequences used.
Clinical
In the clinics, an important issue is how well ASL
performs when compared with established ‘‘gold-stan-
dard’’ methods. Validation studies have been carried out
in animals [57, 88, 89] using radioactive microspheres
and 14C-iodoantipyrine autoradiography, as well as in
humans [90] comparing ASL and PET. In general, good
correlation is observed for grey matter, whereas white-
matter often shows an underestimation of CBF. This is
mainly attributed the prolonged arrival times and low
perfusion values of white matter, which is in the lower
range of measurable flow values using ASL.
The benefits of ASL particularly suit the requirements
of the paediatric population, where there is a need to
avoid ionizing radiation as procedures are likely to be
repeated for disease monitoring and venous cannulation
can be highly traumatic. In addition, the general higher
blood flow seen in children increases the difference
signal DM and thereby the perfusion SNR [91], making
ASL a very promising tool for paediatric perfusion
studies [92].
Another particularly appropriate clinical arena is that
of cerebrovascular diseases. In fact, stroke is the third
cause of death behind cardiovascular diseases and
cancer. Consequently, a lot of resources are put into the
research of this disease – how to prevent it, the
immediate therapy following stroke and post-stroke
rehabilitation procedures. In recent years, ASL techni-
ques have become an alternative tool to CT and bolus
tracking MRI for studying the mechanisms of stroke and
the underlying processes of brain damage resulting from
ischaemia. Studies of the evolution of acute stroke and
validation of outcome predictors for the identification of
potentially salvageable tissue have been performed in
The British Journal of Radiology, August 2006
animal models [93–96] and patients [97, 98] using both
PASL and CASL. For the assessment of the different
perfusion territories and eventually collateralized flow in
steno-occlusive patients, a range of ASL sequences has
been made capable of labelling individual perfusion
territories of the major feeding vessels one at a time [24,
27, 28, 32, 99–102]. An example of regional perfusion
imaging of a healthy volunteer is shown in Figure 7.
These techniques are very promising in the evaluation of
the successful recruitment of subsidiary blood vessels,
which are believed to play an important role in the
clinical outcome of patients with cerebral artery occlu-
sion. Until recently, diagnostic strategies to evaluate the
collateral circulation could be divided into those that
directly visualized collateral blood vessels, such as
conventional X–ray angiography and indirect methods
that assess tissue perfusion, including acetazolamide-
challenge tests of cerebrovascular reserve. This new class
of ASL techniques makes the combination of both
methods possible for the first time and quantitative
information of the flow territories supplied by each
major brain vessel is obtainable. As such, these methods
could become alternatives to conventional X–ray based
subtraction angiography, which is so far the only
modality giving temporal as well as spatial regional
blood flow information. Using these techniques, Van
Laar et al [103] showed a wide interindividual variability
in the perfusion territories caused mainly by anatomical
variations of the circle of Willis in a population study of
115 healthy ageing volunteers.
ASL has also been used in the evaluation of cancer.
The perfusion change in tumours depends on their
aggressiveness or grade, which makes perfusion infor-
mation important when selecting and evaluating thera-
pies. Repeated ASL perfusion measurements can be used
to monitor the effect of blood flow regulating agents and
other anticancer therapeutics such as anti–angiogenic
agents [104–107].
Finally, ASL has also been applied to organs other than
the brain, although the majority of research and applica-
tions have focused on the latter. Applications range from
Figure 7. Regional perfusion image
(RPI) acquired using QUASAR [124]
on a 23-year-old healthy female
subject.
697

combined lung perfusion using ASL and ventilation
scans using hyperpolarized helium-3 for evaluation of
pulmonary diseases [108–111], to cardiac [112–114] and
renal [6, 115–119] perfusion imaging. However, these
techniques are very challenging, mainly due to non-
rigidity and increased movement of these organs as
compared with the brain.
Conclusion
The evolution of ASL since its invention by Williams et
al [8, 9] in 1992 is remarkable. Numerous sequences have
been developed, each solving problematic issues such as
magnetization transfer effects, insensitivity to transit
times etc. in their own way. Along the way, the
quantitative aspect has been the leading factor and
various models have been proposed for absolute CBF
quantification. However, there are still improvements to
be made and, with the move from traditional brain
perfusion imaging to abdominal organs, there are
certainly more challenges ahead.
Despite the remarkable progress, ASL has still not
overtaken traditional invasive methods in the clinics. One
reason for this is the intrinsically low SNR, making
averaging necessary, and therefore good cooperation from
the subject is essential. This is not always the case with, for
instance, stroke patients or children. Nevertheless, the
great improvements in hardware and the increasing
availability of high field clinical scanners pre-installed
with ASL sequences seem to put more focus on these
techniques. Robust sequences for regional perfusion
imaging and functional studies of either cognitive or
pharmacological nature are also gaining more and more
interest, and the complete non-invasiveness of these
techniques will always be an attractive asset.
Acknowledgments
The authors would like to thank Prof. Alan Jackson for
the opportunity to write this review article. This work
was supported in part by Philips Medical Systems, and
the following grants: # NMRC/0855/2004, NMRC/
CPG/009/2004, NMRC/0919/2004 and NHGA-RPR/
04012.
References
1. Kety SS, Schmidt CF. The nitrous oxide method for the
quantitative determination of cerebral blood flow in man:
theory, procedure and normal values. J Clin Invest
1948;4:476–83.
2. Meier P, Zierler KL. On the theory of the indicator-dilution
method for measurement of blood flow and volume. J Appl
Physiol 1954;6:731–44.
3. Ostergaard L, Weisskoff RM, Chesler DA, Gyldensted C,
Rosen BR. High resolution measurement of cerebral blood
flow using intravascular tracer bolus passages. Part I:
Mathematical approach and statistical analysis. Magn
Reson Med 1996;36:715–25.
4. Barbier EL, Lamalle L, Decorps M. Methodology of brain
perfusion imaging. J Magn Reson Imaging 2001;13:496–520.
5. Grandin CB. Assessment of brain perfusion with MRI:
methodology and application to acute stroke.
Neuroradiology 2003;45:755–66.
698
E T Petersen, I Zimine, Y-C L Ho and X Golay
6. Grenier N, Basseau F, Ries M, Tyndal B, Jones R, Moonen C.
Functional MRI of the kidney. Abdom Imaging
2003;28:164–75.
7. Ostergaard L. Cerebral perfusion imaging by bolus track-
ing. Top Magn Reson Imaging 2004;15:3–9.
8. Williams DS, Detre JA, Leigh JS, Koretsky AP. Magnetic
resonance imaging of perfusion using spin inversion of
arterial water. Proc Natl Acad Sci USA 1992;89:212–6.
9. Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion
imaging. Magn Reson Med 1992;23:37–45.
10. Dixon WT, Du LN, Faul DD, Gado M, Rossnick S. Projection
angiograms of blood labeled by adiabatic fast passage.
Magn Reson Med 1986;3:454–62.
11. Zhang W, Williams DS, Koretsky AP. Measurement of rat
brain perfusion by NMR using spin labeling of arterial
water: in vivo determination of the degree of spin labeling.
Magn Reson Med 1993;29:416–21.
12. Maccotta L, Detre JA, Alsop DC. The efficiency of adiabatic
inversion for perfusion imaging by arterial spin labeling.
NMR Biomed 1997;10:216–21.
13. Utting JF, Thomas DL, Gadian DG, Ordidge RJ. Velocity-
driven adiabatic fast passage for arterial spin labeling:
results from a computer model. Magn Reson Med
2003;49:398–401.
14. Franke C, van Dorsten FA, Olah L, Schwindt W, Hoehn M.
Arterial spin tagging perfusion imaging of rat brain:
dependency on magnetic field strength. Magn Reson
Imaging 2000;18:1109–13.
15. Gach HM, Dai W. Simple model of double adiabatic
inversion (DAI) efficiency. Magn Reson Med 2004;52:941–6.
16. Gach HM, Kam AW, Reid ED, Talagala SL. Quantitative
analysis of adiabatic fast passage for steady laminar and
turbulent flows. Magn Reson Med 2002;47:709–19.
17. Trampel R, Jochimsen TH, Mildner T, Norris DG, Moller
HE. Efficiency of flow-driven adiabatic spin inversion
under realistic experimental conditions: a computer simu-
lation. Magn Reson Med 2004;51:1187–93.
18. Werner R, Norris DG, Alfke K, Mehdorn HM, Jansen O.
Improving the amplitude-modulated control experiment
for multislice continuous arterial spin labeling. Magn Reson
Med 2005;53:1096–102.
19. Marro KI, Hayes CE, Kushmerick MJ. A model of the
inversion process in an arterial inversion experiment. NMR
Biomed 1997;10:324–32.
20. Wolff SD, Balaban RS. Magnetization transfer contrast
(MTC) and tissue water proton relaxation in vivo. Magn
Reson Med 1989;10:135–44.
21. Henkelman RM, Huang X, Xiang QS, Stanisz GJ, Swanson
SD, Bronskill MJ. Quantitative interpretation of magnetiza-
tion transfer. Magn Reson Med 1993;29:759–66.
22. Alsop DC, Detre JA. Multisection cerebral blood flow MR
imaging with continuous arterial spin labeling. Radiology
1998;208:410–6.
23. Talagala SL, Barbier EL, Williams DS, Silva AC, Koretsky
AP. Multi-slice perfusion MRI using continuous arterial
water labeling controlling for MT effects with simultaneous
proximal and distal RF irradiation. Proceedings of the 6th
Annual Meeting of ISMRM, Sydney, 1998.
24. Zhang W, Silva AC, Williams DS, Koretsky AP. NMR
measurement of perfusion using arterial spin labeling
without saturation of macromolecular spins. Magn Reson
Med 1995;33:370–6.
25. Silva AC, Zhang W, Williams DS, Koretsky AP. Multi-slice
MRI of rat brain perfusion during amphetamine stimulation
using arterial spin labeling. Magn Reson Med
1995;33:209–14.
26. McLaughlin AC, Ye FQ, Pekar JJ, Santha AK, Frank JA.
Effect of magnetization transfer on the measurement of
cerebral blood flow using steady-state arterial spin tagging
The British Journal of Radiology, August 2006
Review article: Non-invasive measurement of perfusion
approaches: a theoretical investigation. Magn Reson Med
1997;37:501–10.
27. Zaharchuk G, Ledden PJ, Kwong KK, Reese TG, Rosen BR,
Wald LL. Multislice perfusion and perfusion territory
imaging in humans with separate label and image coils.
Magn Reson Med 1999;41:1093–8.
28. Mildner T, Trampel R, Moller HE, Schafer A, Wiggins CJ,
Norris DG. Functional perfusion imaging using continuous
arterial spin labeling with separate labeling and imaging
coils at 3 T. Magn Reson Med 2003;49:791–5.
29. Floyd TF, Ratcliffe SJ, Wang J, Resch B, Detre JA. Precision
of the CASL-perfusion MRI technique for the measurement
of cerebral blood flow in whole brain and vascular
territories. J Magn Reson Imaging 2003;18:649–55.
30. Talagala SL, Ye FQ, Ledden PJ, Chesnick S. Whole-brain 3D
perfusion MRI at 3.0 T using CASL with a separate labeling
coil. Magn Reson Med 2004;52:131–40.
31. Edelman RR, Siewert B, Darby DG, Thangaraj V, Nobre AC,
Mesulam MM, et al. Qualitative mapping of cerebral blood
flow and functional localization with echo-planar MR
imaging and signal targeting with alternating radio
frequency. Radiology 1994;192:513–20.
32. Golay X, Petersen ET, Hui F. Pulsed star labeling of arterial
regions (PULSAR): a robust regional perfusion technique
for high field imaging. Magn Reson Med 2005;53:15–21.
33. Kwong KK, Chesler DA, Weisskoff RM, Donahue KM,
Davis TL, Ostergaard L, et al. MR perfusion studies with
T1-weighted echo planar imaging. Magn Reson Med
1995;34:878–87.
34. Kim SG. Quantification of relative cerebral blood flow
change by flow-sensitive alternating inversion recovery
(FAIR) technique: application to functional mapping. Magn
Reson Med 1995;34:293–301.
35. Golay X, Hendrikse J, Lim TC. Perfusion imaging using
arterial spin labeling. Top Magn Reson Imaging
2004;15:10–27.
36. Calamante F, Kwong KK, Williams DS, van der GJ, Tyndal
B. Quantification of perfusion in pulsed labeling techni-
ques. Proceedings of the 3rd Annual Meeting of ISMRM,
Nice, France, 870. 1995.
37. Buxton RB, Frank LR, Wong EC, Siewert B, Warach S,
Edelman RR. A general kinetic model for quantitative
perfusion imaging with arterial spin labeling. Magn Reson
Med 1998;40:383–96.
38. Luh WM, Wong EC, Bandettini PA, Hyde JS. QUIPSS II
with thin-slice TI1 periodic saturation: a method for
improving accuracy of quantitative perfusion imaging
using pulsed arterial spin labeling. Magn Reson Med
1999;41:1246–54.
39. Wong EC, Buxton RB, Frank LR. Quantitative imaging of
perfusion using a single subtraction (QUIPSS and QUIPSS
II). Magn Reson Med 1998;39:702–8.
40. Wang J, Alsop DC, Song HK, Maldjian JA, Tang K, Salvucci
AE, et al. Arterial transit time imaging with flow encoding
arterial spin tagging (FEAST). Magn Reson Med
2003;50:599–607.
41. Yang Y, Engelien W, Xu S, Gu H, Silbersweig DA, Stern E.
Transit time, trailing time, and cerebral blood flow during
brain activation: measurement using multislice, pulsed
spin-labeling perfusion imaging. Magn Reson Med
2000;44:680–5.
42. Gonzalez-At JB, Alsop DC, Detre JA. Cerebral perfusion
and arterial transit time changes during task activation
determined with continuous arterial spin labeling. Magn
Reson Med 2000;43:739–46.
43. Gunther M, Bock M, Schad LR. Arterial spin labeling in
combination with a look-locker sampling strategy: inflow
turbo-sampling EPI-FAIR (ITS-FAIR). Magn Reson Med
2001;46:974–84.
The British Journal of Radiology, August 2006
44. Golay X, Stuber M, Pruessmann KP, Meier D, Boesiger P.
Transfer insensitive labeling technique (TILT): application
to multislice functional perfusion imaging. J Magn Reson
Imaging 1999;9:454–61.
45. Hendrikse J, Lu H, van der Grond J, van Zijl PC, Golay X.
Measurements of cerebral perfusion and arterial hemody-
namics during visual stimulation using TURBO-TILT.
Magn Reson Med 2003;50:429–33.
46. Detre JA, Alsop DC, Vives LR, Maccotta L, Teener JW, Raps
EC. Noninvasive MRI evaluation of cerebral blood flow in
cerebrovascular disease. Neurology 1998;50:633–41.
47. Ye FQ, Mattay VS, Jezzard P, Frank JA, Weinberger DR,
McLaughlin AC. Correction for vascular artifacts in cerebral
blood flow values measured by using arterial spin tagging
techniques. Magn Reson Med 1997;37:226–35.
48. Yang Y, Frank JA, Hou L, Ye FQ, McLaughlin AC, Duyn JH.
Multislice imaging of quantitative cerebral perfusion with
pulsed arterial spin labeling. Magn Reson Med
1998;39:825–32.
49. Silver MS, Joseph RI, Hoult DI. Selective spin inversion in
nuclear magnetic resonance and coherent optics through an
exact solution of the Bloch-Riccati equation. Physical
Review A 1985;31:2753–5.
50. Ordidge RJ, Wylezinska M, Hugg JW, Butterworth E,
Franconi F. Frequency offset corrected inversion (FOCI)
pulses for use in localized spectroscopy. Magn Reson Med
1996;36:562–6.
51. Warnking JM, Pike GB. Bandwidth-modulated adiabatic RF
pulses for uniform selective saturation and inversion. Magn
Reson Med 2004;52:1190–9.
52. Pruessmann KP, Golay X, Stuber M, Scheidegger MB,
Boesiger P. RF pulse concatenation for spatially selective
inversion. J Magn Reson 2000;146:58–65.
53. Zhan W, Gu H, Silbersweig DA, Stern E, Yang Y. Inversion
profiles of adiabatic inversion pulses for flowing spins: the
effects on labeling efficiency and labeling accuracy in
perfusion imaging with pulsed arterial spin-labeling.
Magn Reson Imaging 2002;20:487–94.
54. Paulson OB, Hertz MM, Bolwig TG, Lassen NA. Water
filtration and diffusion across the blood brain barrier in
man. Acta Neurol Scand Suppl 1977;64:492–3.
55. Amiry-Moghaddam M, Frydenlund DS, Ottersen OP.
Anchoring of aquaporin-4 in brain: molecular mechanisms
and implications for the physiology and pathophysiology of
water transport. Neuroscience 2004;129:999–1010.
56. Farquhar MG, Palade GE. Junctional complexes in various
epithelia. J Cell Biol 1963;17:375–412.
57. Zhou J, Wilson DA, Ulatowski JA, Traystman RJ, van Zijl
PC. Two-compartment exchange model for perfusion
quantification using arterial spin tagging. J Cereb Blood
Flow Metab 2001;21:440–55.
58. St Lawrence KS, Frank JA, McLaughlin AC. Effect of
restricted water exchange on cerebral blood flow values
calculated with arterial spin tagging: a theoretical investi-
gation. Magn Reson Med 2000;44:440–9.
59. Li KL, Zhu X, Hylton N, Jahng GH, Weiner MW, Schuff N.
Four-phase single-capillary stepwise model for kinetics in
arterial spin labeling MRI. Magn Reson Med 2005;53:511–8.
60. Fung YC. Biomechanics: Circulation. 2nd edition, Springer-
Verlag, New York. 11-21-1996.
61. Hrabe J, Lewis DP. Two analytical solutions for a model of
pulsed arterial spin labeling with randomized blood arrival
times. J Magn Reson 2004;167:49–55.
62. Kety SS. The theory and applications of the exchange of
inert gas at the lungs and tissues. Pharmacol Rev
1951;3:1–41.
63. Herscovitch P, Raichle ME. What is the correct value for the
brain--blood partition coefficient for water? J Cereb Blood
Flow Metab 1985;5:65–9.
699

64. Roberts DA, Rizi R, Lenkinski RE, Leigh JS Jr. Magnetic
resonance imaging of the brain: blood partition coefficient
for water: application to spin-tagging measurement of
perfusion. J Magn Reson Imaging 1996;6:363–6.
65. Friston KJ, Mechelli A, Turner R, Price CJ. Nonlinear
responses in fMRI: the Balloon model, Volterra kernels, and
other hemodynamics. Neuroimage 2000;12:466–77.
66. Friston KJ. Models of brain function in neuroimaging. Ann
Rev Psychol 2005;56:57–87.
67. Lu H, Golay X, Pekar JJ, van Zijl PC. Noise or artifact:
detrimental effects of BOLD signal in arterial spin labeling
fMRI at high field strength. Proceedings of the 13th Annual
Meeting ISMRM, Miami Beach, USA, #35. 2005.
68. Boxerman JL, Hamberg LM, Rosen BR, Weisskoff RM. MR
contrast due to intravascular magnetic susceptibility per-
turbations. Magn Reson Med 1995;34:555–66.
69. Wintermark M, Sesay M, Barbier E, Borbely K, Dillon WP,
Eastwood JD, et al. Comparative overview of brain
perfusion imaging techniques. Stroke 2005;36:2032–3.
70. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic
resonance imaging with contrast dependent on blood
oxygenation. Proc Natl Acad Sci USA 1990;87:9868–72.
71. Aguirre GK, Detre JA, Zarahn E, Alsop DC. Experimental
design and the relative sensitivity of BOLD and perfusion
fMRI. Neuroimage 2002;15:488–500.
72. Detre JA, Wang J. Technical aspects and utility of fMRI
using BOLD and ASL. Clin Neurophysiol 2002;113:621–34.
73. Wang J, Aguirre GK, Kimberg DY, Roc AC, Li L, Detre JA.
Arterial spin labeling perfusion fMRI with very low task
frequency. Magn Reson Med 2003;49:796–802.
74. Duong TQ, Kim DS, Ugurbil K, Kim SG. Localized cerebral
blood flow response at submillimeter columnar resolution.
Proc Natl Acad Sci USA 2001;98:10904–9.
75. Kim SG, Tsekos NV, Ashe J. Multi-slice perfusion-based
functional MRI using the FAIR technique: comparison of
CBF and BOLD effects. NMR Biomed 1997;10:191–6.
76. Kim SG, Ugurbil K. Comparison of blood oxygenation and
cerebral blood flow effects in fMRI: estimation of relative
oxygen consumption change. Magn Reson Med
1997;38:59–65.
77. Lu H, Golay X, Pekar JJ, van Zijl PC. Sustained poststimulus
elevation in cerebral oxygen utilization after vascular
recovery. J Cereb Blood Flow Metab 2004;24:764–70.
78. Obata T, Liu TT, Miller KL, Luh WM, Wong EC, Frank LR,
et al. Discrepancies between BOLD and flow dynamics in
primary and supplementary motor areas: application of the
balloon model to the interpretation of BOLD transients.
Neuroimage 2004;21:144–53.
79. Hoge RD, Franceschini MA, Covolan RJ, Huppert T,
Mandeville JB, Boas DA. Simultaneous recording of task-
induced changes in blood oxygenation, volume, and flow
using diffuse optical imaging and arterial spin-labeling
MRI. Neuroimage 2005;25:701–7.
80. Detre JA, Samuels OB, Alsop DC, Gonzalez-At JB, Kasner
SE, Raps EC. Noninvasive magnetic resonance imaging
evaluation of cerebral blood flow with acetazolamide
challenge in patients with cerebrovascular stenosis. J
Magn Reson Imaging 1999;10:870–5.
81. Yen YF, Field AS, Martin EM, Ari N, Burdette JH, Moody
DM, et al. Test-retest reproducibility of quantitative CBF
measurements using FAIR perfusion MRI and acetazola-
mide challenge. Magn Reson Med 2002;47:921–8.
82. Kochanek PM, Hendrich KS, Robertson CL, Williams DS,
Melick JA, Ho C, et al. Assessment of the effect of 2-
chloroadenosine in normal rat brain using spin-labeled MRI
measurement of perfusion. Magn Reson Med 2001;45:
924–9.
83. Robertson CL, Hendrich KS, Kochanek PM, Jackson EK,
Melick JA, Graham SH, et al. Assessment of 2-chloroade-
nosine treatment after experimental traumatic brain injury
700
E T Petersen, I Zimine, Y-C L Ho and X Golay
in the rat using arterial spin-labeled MRI: a preliminary
report. Acta Neurochir Suppl 2000;76:187–9.
84. St Lawrence KS, Ye FQ, Lewis BK, Weinberger DR, Frank
JA, McLaughlin AC. Effects of indomethacin on cerebral
blood flow at rest and during hypercapnia: an arterial spin
tagging study in humans. J Magn Reson Imaging
2002;15:628–35.
85. St Lawrence KS, Ye FQ, Lewis BK, Frank JA, McLaughlin
AC. Measuring the effects of indomethacin on changes in
cerebral oxidative metabolism and cerebral blood flow
during sensorimotor activation. Magn Reson Med
2003;50:99–106.
86. Liu ZM, Schmidt KF, Sicard KM, Duong TQ. Imaging
oxygen consumption in forepaw somatosensory stimulation
in rats under isoflurane anesthesia. Magn Reson Med
2004;52:277–85.
87. Sicard K, Shen Q, Brevard ME, Sullivan R, Ferris CF, King
JA, et al. Regional cerebral blood flow and BOLD responses
in conscious and anesthetized rats under basal and
hypercapnic conditions: implications for functional MRI
studies. J Cereb Blood Flow Metab 2003;23:472–81.
88. Walsh EG, Minematsu K, Leppo J, Moore SC. Radioactive
microsphere validation of a volume localized continuous
saturation perfusion measurement. Magn Reson Med
1994;31:147–53.
89. Ewing JR, Wei L, Knight RA, Pawa S, Nagaraja TN, Brusca
T, et al. Direct comparison of local cerebral blood flow rates
measured by MRI arterial spin-tagging and quantitative
autoradiography in a rat model of experimental
cerebral ischemia. J Cereb Blood Flow Metab 2003;23:
198–209.
90. Ye FQ, Berman KF, Ellmore T, Esposito G, van Horn JD,
Yang Y, et al. H(2)(15)O PET validation of steady-state
arterial spin tagging cerebral blood flow measurements in
humans. Magn Reson Med 2000;44:450–6.
91. Wang J, Licht DJ, Jahng GH, Liu CS, Rubin JT, Haselgrove J,
et al. Pediatric perfusion imaging using pulsed arterial spin
labeling. J Magn Reson Imaging 2003;18:404–13.
92. Oguz KK, Golay X, Pizzini FB, Freer CA, Winrow N, Ichord
R, et al. Sickle cell disease: continuous arterial spin-labeling
perfusion MR imaging in children. Radiology
2003;227:567–74.
93. Calamante F, Lythgoe MF, Pell GS, Thomas DL, King MD,
Busza AL, et al. Early changes in water diffusion, perfusion,
T1, and T2 during focal cerebral ischemia in the rat studied
at 8.5 T. Magn Reson Med 1999;41:479–85.
94. Lythgoe MF, Thomas DL, Calamante F, Pell GS, King MD,
Busza AL, et al. Acute changes in MRI diffusion, perfusion,
T(1), and T(2) in a rat model of oligemia produced by
partial occlusion of the middle cerebral artery. Magn Reson
Med 2000;44:706–12.
95. Zaharchuk G, Yamada M, Sasamata M, Jenkins BG,
Moskowitz MA, Rosen BR. Is all perfusion-weighted
magnetic resonance imaging for stroke equal? The temporal
evolution of multiple hemodynamic parameters after focal
ischemia in rats correlated with evidence of infarction. J
Cereb Blood Flow Metab 2000;20:1341–51.
96. Pillekamp F, Grune M, Brinker G, Franke C, Uhlenkuken U,
Hoehn M, et al. Magnetic resonance prediction of outcome
after thrombolytic treatment. Magn Reson Imaging
2001;19:143–52.
97. Chalela JA, Alsop DC, Gonzalez-Atavales JB, Maldjian JA,
Kasner SE, Detre JA. Magnetic resonance perfusion imaging
in acute ischemic stroke using continuous arterial spin
labeling. Stroke 2000;31:680–7.
98. Siewert B, Schlaug G, Edelman RR, Warach S. Comparison
of EPISTAR and T2*-weighted gadolinium-enhanced perfu-
sion imaging in patients with acute cerebral ischemia.
Neurology 1997;48:673–9.
The British Journal of Radiology, August 2006
Review article: Non-invasive measurement of perfusion
99. Davies NP, Jezzard P. Selective arterial spin labeling
(SASL): perfusion territory mapping of selected feeding
arteries tagged using two-dimensional radiofrequency
pulses. Magn Reson Med 2003;49:1133–42.
100. Hendrikse J, van der Grond J, Lu H, van Zijl PC, Golay X.
Flow territory mapping of the cerebral arteries with
regional perfusion MRI. Stroke 2004;35:882–7.
101. Trampel R, Mildner T, Goerke U, Schaefer A, Driesel W,
Norris DG. Continuous arterial spin labeling using a local
magnetic field gradient coil. Magn Reson Med 2002;48:543–6.
102. Werner R, Norris DG, Alfke K, Mehdorn HM, Jansen O.
Continuous artery-selective spin labeling (CASSL). Magn
Reson Med 2005;53:1006–12.
103. van Laar PJ, Hendrikse J, Golay X, Lu H, van Osch MJ, van
der GJ. In-vivo flow territory mapping of major brain
feeding arteries. Neuroimage 2006;29:136–44.
104. Brown SL, Ewing JR, Kolozsvary A, Butt S, Cao Y, Kim JH.
Magnetic resonance imaging of perfusion in rat cerebral 9L
tumor after nicotinamide administration. Int J Radiat
Oncol Biol Phys 1999;43:627–33.
105. Gaa J, Warach S, Wen P, Thangaraj V, Wielopolski P,
Edelman RR. Noninvasive perfusion imaging of human
brain tumors with EPISTAR. Eur Radiol 1996;6:518–22.
106. Silva AC, Kim SG, Garwood M. Imaging blood flow in
brain tumors using arterial spin labeling. Magn Reson
Med 2000;44:169–73.
107. Warmuth C, Gunther M, Zimmer C. Quantification of
blood flow in brain tumors: comparison of arterial spin
labeling and dynamic susceptibility-weighted contrast-
enhanced MR imaging. Radiology 2003;228:523–32.
108. Altes TA, Mai VM, Munger TM, Brookeman JR, Hagspiel
KD. Pulmonary embolism: comprehensive evaluation with
MR ventilation and perfusion scanning with hyperpolar-
ized helium-3, arterial spin tagging, and contrast-
enhanced MRA. J Vasc Interv Radiol 2005;16:999–1005.
109. Lipson DA, Roberts DA, Hansen-Flaschen J, Gentile TR,
Jones G, Thompson A, et al. Pulmonary ventilation and
perfusion scanning using hyperpolarized helium-3 MRI
and arterial spin tagging in healthy normal subjects and in
pulmonary embolism and orthotopic lung transplant
patients. Magn Reson Med 2002;47:1073–6.
110. Mai VM, Hagspiel KD, Altes T, Goode AR, Williams MB,
Berr SS. Detection of regional pulmonary perfusion deficit
of the occluded lung using arterial spin labeling in
magnetic resonance imaging. J Magn Reson Imaging
2000;11:97–102.
111. Uematsu H, Levin DL, Hatabu H. Quantification of
pulmonary perfusion with MR imaging: recent advances.
Eur J Radiol 2001;37:155–63.
112. Belle V, Kahler E, Waller C, Rommel E, Voll S, Hiller KH,
et al. In vivo quantitative mapping of cardiac perfusion in
rats using a noninvasive MR spin-labeling method. J Magn
Reson Imaging 1998;8:1240–5.
113. Floyd TF, McGarvey M, Ochroch EA, Cheung AT,
Augoustides JA, Bavaria JE, et al. Perioperative changes
in cerebral blood flow after cardiac surgery: influence of
anemia and aging. Ann Thorac Surg 2003;76:2037–42.
114. Reeder SB, Atalay MK, McVeigh ER, Zerhouni EA, Forder
JR. Quantitative cardiac perfusion: a noninvasive spin-
labeling method that exploits coronary vessel geometry.
Radiology 1996;200:177–84.
115. Berr SS, Hagspiel KD, Mai VM, Keilholz-George S, Knight-
Scott J, Christopher JM, et al. Perfusion of the kidney using
The British Journal of Radiology, August 2006
extraslice spin tagging (EST) magnetic resonance imaging.
J Magn Reson Imaging 1999;10:886–91.
116. De Bazelaire C, Rofsky NM, Duhamel G, Michaelson MD,
George D, Alsop DC. Arterial spin labeling blood flow
magnetic resonance imaging for the characterization of
metastatic renal cell carcinoma(1). Acad Radiol 2005;12:
347–57.
117. Karger N, Biederer J, Lusse S, Grimm J, Steffens J, Heller
M, et al. Quantitation of renal perfusion using arterial spin
labeling with FAIR-UFLARE. Magn Reson Imaging
2000;18:641–7.
118. Prasad PV, Kim D, Kaiser AM, Chavez D, Gladstone S, Li
W, et al. Noninvasive comprehensive characterization of
renal artery stenosis by combination of STAR angiography
and EPISTAR perfusion imaging. Magn Reson Med
1997;38:776–87.
119. Wang JJ, Hendrich KS, Jackson EK, Ildstad ST, Williams
DS, Ho C. Perfusion quantitation in transplanted rat
kidney by MRI with arterial spin labeling. Kidney Int
1998;53:1783–91.
120. Detre JA, Zhang W, Roberts DA, Silva AC, Williams DS,
Grandis DJ, et al. Tissue specific perfusion imaging using
arterial spin labeling. NMR Biomed 1994;7:75–82.
121. Wong EC, Buxton RB, Frank LR. Implementation of
quantitative perfusion imaging techniques for functional
brain mapping using pulsed arterial spin labeling. NMR
Biomed 1997;10:237–49.
122. Jahng GH, Zhu XP, Matson GB, Weiner MW, Schuff N.
Improved perfusion-weighted MRI by a novel double
inversion with proximal labeling of both tagged and
control acquisitions. Magn Reson Med 2003;49:307–14.
123. Chen Q, Siewert B, Bly BM, Warach S, Edelman RR.
STAR–HASTE: perfusion imaging without magnetic sus-
ceptibility artifact. Magn Reson Med 1997;38:404–8.
124. Petersen ET, Lim TC, Golay X. A model-free arterial spin
labeling quantitative approach for perfusion MRI. Magn
Reson Med 2006;55:219–32.
125. Schwarzbauer C, Morrissey SP, Haase A. Quantitative
magnetic resonance imaging of perfusion using magnetic
labeling of water proton spins within the detection slice.
Magn Reson Med 1996;35:540–6.
126. Tanabe JL, Yongbi M, Branch C, Hrabe J, Johnson G,
Helpern JA. MR perfusion imaging in human brain using
the UNFAIR technique. Un-inverted flow-sensitive alter-
nating inversion recovery. J Magn Reson Imaging
1999;9:761–7.
127. Mai VM, Berr SS. MR perfusion imaging of pulmonary
parenchyma using pulsed arterial spin labeling techni-
ques: FAIRER and FAIR. J Magn Reson Imaging
1999;9:483–7.
128. Mai VM, Hagspiel KD, Christopher JM, Do HM, Altes T,
Knight-Scott J, et al. Perfusion imaging of the human lung
using flow-sensitive alternating inversion recovery with
an extra radiofrequency pulse (FAIRER). Magn Reson
Imaging 1999;17:355–61.
129. Zhou J, van Zijl PC. Perfusion imaging using FAIR with a
short predelay. Magn Reson Med 1999;41:1099–107.
130. Zhou J, Mori S, van Zijl PC. FAIR excluding radiation
damping (FAIRER). Magn Reson Med 1998;40:712–9.
131. Schwarzbauer C, Heinke W. BASE imaging: a new spin
labeling technique for measuring absolute perfusion
changes. Magn Reson Med 1998;39:717–22.
701

Book review
Michael Faraday and the electrical century. By I R Morus.
pp. 192, 2004 (Icon Books, Cambridge, UK) £9.99
ISBN 1-84046-540-9
At the beginning of the 20th century the subject of
electricity continued to fascinate many medical men
including the early pioneering radiologists who believed
electricity might have a major therapeutic role. One such
early pioneer was John Mcintyre from Glasgow who was
initially appointed to Glasgow Royal Infirmary in 1885 as
medical electrician before turning his enthusiasm to X-
rays 10 years later on learning of Roentgen’s discovery.
Indeed the precursor of the Radiology Section of the
Royal Society of Medicine was known as the British
Electrotherapeutic Society which was formed in 1902 in
the rooms of the Medical Society of London as a
breakaway group from the Roentgen Society. A journal
called Medical Electrology and Radiology existed and was
the Electrotherapeutic Society’s official publication.
Radiologists in those days were interested in the
therapeutic role of electricity. Indeed in 1907, at the
inauguration of the Electrotherapeutic Section of
the Royal Society of Medicine, Dr Dean Butcher gave
the presidential address entitled ‘‘The Future of
Electricity in Medicine’’. Thus we can see that the subject
of electricity proved to be of fascination to the early
pioneering radiologists soon after the advent of
Roentgen’s discovery.
Michael Faraday of course was greatly interested in
the subject of electricity and became one of the greatest
figures in 19th century science. His experiments trans-
formed the science of electricity and he is credited with,
among other things, inventing the electric motor. He also
investigated the relationship between electricity and
magnetism and showed that currents of electricity could
be produced by a moving magnet, which led to the
eventual development of the electrical power industry.
In this splendid biography of Faraday by Morus an
expert on the scientific study of electricity in the 19th
702
The British Journal of Radiology, 79 (2006), 702
century, Faraday’s seminal contribution to the subject of
electricity is discussed and placed within the context of
the scientific activities of the Victorian age. We are taken
on a tour through his early humble upbringing and given
an insight into the scientific milieu of early Victorian
London in which Farraday worked and would even-
tually climb up to reach the highest echelons of the
scientific establishment of that era. The relationship
between Sir Humphrey Davy and Faraday is covered
and his activities at the Royal Institution are elegantly
described. The biography of Faraday is set against a
backdrop of the scientific activities of London in the
Victorian era. Faraday’s meticulous experiments and
rapid scientific progress are elegantly described. It was
interesting to read of the petty jealousies and quibbles
between the leading scientific personalities of the time
and smile at how little things have changed in the
intervening 150 years regarding scientific rivalries.
Faraday was a scientific genius who rose from humble
beginnings to reach the pinnacles of the scientific
establishment of his day. His research transformed the
subject of electricity and laid the foundations for the
future development of the early electrical industry in
the early 20th century. He influenced a generation of
subsequent scientists including James Clarke Maxwell,
Professor of Physics at the University of Cambridge.
Faraday became a Fellow of the Royal Society and
became a member of several Societies in Europe and
accumulated numerous honours. He was even offered
the presidency of the Royal Society in 1855 but refused.
He died in 1867. This is a splendid short biography of
Faraday beautifully written and a delight to read. It is
written for a wide audience and I am sure many
radiologists and other doctors will find it of particular
interest to learn a little about our scientific heritage.
A K BANERJEE
The British Journal of Radiology, August 2006
The British Journal of Radiology, 79 (2006), 703–705

CASE OF THE MONTH

A young female with galactorrhoea and sudden onset chest pain

1 1 1
C J DAS, MD, DNB, R SHARMA, MD, A SEITH, MD and 2M P BARUAH, DM

Departments of 1Radiodiagnosis and 2Endocrinology, All India Institute of Medical Sciences, New
Delhi-29, India

Received 7 March 2005

Revised 12 July 2005
Accepted 16 August 2005

DOI: 10.1259/bjr/80615044

’ 2006 The British Institute of

Radiology

A 27-year-old married woman presented to the

gynaecology out-patient department with a 3 month
history of galactorrhoea and irregular menstrual cycle.
She also complained of polyuria and polydipsia. No
history of visual problems could be elicited. After initial
evaluation, she was referred to the endocrinology
department for further investigation of her hormonal
status. Clinical examination was essentially within
normal limits except for the galactorrhoea. Her pertinent
laboratory investigations included high serum prolactin
level (150 ng ml21), low levels of gonadotropins (unde-
tectable/ELISA), cortisol (3.8 mg ml21) and thyroid
stimulating hormone (3.8 mg ml21), suggesting panhy-
popituitarism. A high resolution contrast enhanced MRI
(CEMRI) of the sella was performed. MRI revealed a soft
tissue mass in the suprasellar region involving hypotha-
lamus, tuber cinerium and extending into the pituitary
stalk causing its expansion. The mass was hypointense
on T1 weighted images (Figure 1a) and hyperintense
on T2 weighted images (Figure 1b) showing intense Figure 2. Chest radiograph posteroanterior view revealed
reticular opacities and pneumothorax on the left side.

(a) (b) (c)

Figure 1. Contrast enhanced MRI sella showing soft tissue mass in suprasellar region involving hypothalamus, tuber cinerium
extending into the pituitary stalk, which is hypointense on (a) T1 weighted images and hyperintense on (b) T2 weighted images,
showing intense homogeneous enhancement on (c) post-gadolinium images.

The British Journal of Radiology, August 2006 703

 C J Das, R Sharma, A Seith and M P Baruah

(a) (b)

Figure 3. (a,b) Follow-up contrast enhanced MRI after anti-tubercular treatment does not show any change in the size of the
suprasellar mass.

homogeneous enhancement on post-gadolinium images radiotherapy to which she responded. On clinical

(Figure 1c). 2 months later, the patient presented with follow-up the patient is doing well.
one episode of sudden onset dyspnoea and severe non- PLCH is an uncommon disorder of unknown aetiology
productive cough. A posteroanterior (PA) chest radio- that typically affects young adults who are smokers [1].
graph revealed reticular opacities and pneumothorax on Previously called histiocytosis X, this is a group of
the left side (Figure 2). What is the possible diagnosis diseases characterized by proliferation of histiocytes [1].
with this clinical presentation and radiological There are unifocal and multifocal forms of this disease.
appearance? The unifocal disease shows solitary lytic bone lesion.
This form spares the hypothalamic pituitary axis. Adult
PLCH occurs most commonly as a disease isolated to the
Discussion lungs. However, adult PLCH can occur as a part of the
Hand-Schuller-Christian syndrome – a triad of diabetic
A provisional diagnosis of tuberculoma was sug- insipidus, exophthalmos and lytic bone lesions which is
gested, based on the MR appearance and in view of the
high prevalence of this disease in India. The patient was
put on anti-tubercular treatment, but her clinical condi-
tion did not improve even after 12 months and a repeat
CEMRI (Figure 3) at this time did not show any change
in the size of the suprasellar mass. In view of the history
of sudden onset dyspnoea and pneumothorax seen on
the chest radiograph, a high resolution CT (HRCT) scan
of chest was performed which revealed multiple bizar-
rely-shaped cystic lesions in both lungs (Figure 4).
Skeletal survey did not reveal any bony lesions. The
diagnosis was reviewed and, based on the clinical
presentation as well as MR and HRCT appearance, a
final diagnosis of pulmonary Langerhans’ cell histiocy-
tosis (PLCH) with involvement of the hypothalamic-
pituitary axis causing panhypopituitarism and diabetes
insipidus was made. A thoracoscopic lung biopsy was
performed and histopathology confirmed the diagnosis
of PLCH. The patient was treated with hormone
replacement therapy for panhypopituitarism and
Carbamazapine for diabetes insipidus. PLCH was Figure 4. High resolution CT (HRCT) scan of chest shows
treated with a combination of chemotherapy and multiple bizarrely-shaped cystic lesions in both lungs.

704 The British Journal of Radiology, August 2006

Case of the month: A young female with galactorrhoea
commonly seen in children [1]. Taking LCH as a whole,
about one-third of cases are seen in adults and 16–34% of
these have pulmonary involvement [2]. In the past,
PLCH was regarded as a disease with male predomi-
nance, but is now considered to be equally prevalent in
both sexes [3], a change that possibly reflects altered
smoking habits. Diabetes insipidus is the most common
endocrinopathy reported with adult PLCH with the
prevalence being about 5–9% [2]. Other endocrine
abnormalities include growth hormone deficiency,
hypothyroidism and abnormal glucose tolerance [4, 5].
The characteristic radiographic appearance of PLCH
consists of diffuse, symmetric nodular/reticulonodular
opacities predominantly in the mid and upper zones
with relative sparing of the lung bases [3, 6]. For
pulmonary involvement, HRCT is the investigation of
choice [7, 8]. The principal findings are cysts with
irregular margins and nodules, often in combination,
giving a characteristic appearance. These cysts are thin
walled and are usually less than 10 mm in diameter.
Although most of the cysts appear round, they can also
have bizarre shapes, being bilobed, clover-leaf shaped, or
branching in appearance. The nodules are usually solid,
centrilobular in location and measure less than 5 mm in
diameter [7, 8]. The main differential diagnosis of PLCH
on HRCT is lymphangioleiomymatosis (LAM). The most
useful differentiating features are: the distribution of
cysts, which are wide spread in PLCH, decreasing
towards the lower zone and sparing the costophrenic
angles, but evenly diffuse in LAM; the possible presence
of nodules in PLCH and their rarity in LAM; the
tendency for LAM cysts to be uniformly rounded,
whereas some of the cysts in PLCH are irregular [9].
CEMRI is the ideal investigation for delineation of
intracranial involvement. CT and MRI show thickening
and enhancement of the infundibulum and hypoth-
alamus. The posterior pituitary bright spot may be
absent. The lesions are hyperintense on T2 weighted
images [1]. Differential diagnoses include tuberculosis,
sarcoidosis, hypothalamic hamartoma and chiasmatic
glioma. Characteristic bone lesions and a classical clinical
The British Journal of Radiology, August 2006
picture aid in the diagnosis of PLCH. Treatment of PLCH
is problematic because of the unpredictable and variable
outcome. Abstaining from smoking may halt the disease
progression. Radiotherapy and chemotherapy may be
needed for the intraocular involvement. Lung transplan-
tation for terminal respiratory insufficiency has also been
advocated [10].
References
1. Lum C, Kucharczyk W, Montanera WJ, Becker LE. Sella
Turcica and parasellar region. In: Scott Atlas, editor.
Magnetic resonance imaging of the brain and spine, 3rd
edn. Lippincott Williams and Wilkins, 2002:1340.
2. Malpas JS. Langerhans’ cell histiocytosis in adults.
Hematol/Oncol Clin North Am 1998;12:259–68.
3. Friedman PJ, Liebow AA, Sokoloff J. Eosinophilic granu-
loma of lung. Clinical aspects of primary pulmonary
histiocytosis in adult. Medicine 1981;60:385–96.
4. Soliman AT, Alsalmi I, Banna NE, Asfour M. Endocrine
aspects of Langerhans cell histiocytosis. Indian J Pediatr
1996;63:402–6.
5. Rami B, Schneider U, Wandl-Vergesslich K, Frisch H,
Schober E. Primary hypothyroidism, central diabetes
insipidus and growth hormone deficiency in multisystem
Langerhans cell histiocytosis: a case report. Acta Paediatr
1998;87:112–4.
6. Lewis JG. Eosinophilic granuloma and its variants with
special reference to lung involvement. J Med
1964;33:337–59.
7. Grenier P, Valeyre D, Cluzel P, et al. Chronic diffuse
interstitial lung disease: diagnostic value of chest radio-
graphy and high resolution CT. Radiology 1991;179:
123–32.
8. Brauner MW, Grenier P, Mouelhi MM, Mompoint D, Lenoir
S. Pulmonary histiocytosis X: evaluation with high resolu-
tion CT. Radiology 1989;172:255–8.
9. Bonelli FS, Hartman TE, Swensen SJ, et al. Accuracy of high
resolution CT in diagnosing lung diseases. AJR Am J
Roentgenol 1988;170:1507–12.
10. Wilson AG. Pulmonary diseases of unknown origin and
miscellaneous lung disorders in Peter Armstrong’s Imaging
of diseases of the chest, 3rd edn. Mosby, 2000:655–60.
705
BJR
The British Journal
of Radiology

September
2006
Volume 79
Issue 945
September 2006, Volume 79, Issue 945

● Quantitative optical coherence tomographic elastography:

method for assessing arterial mechanical properties

● Low-carbohydrate diet induced reduction of hepatic lipid content

observed with a rapid non-invasive MRI technique

● Translational research in radiotherapy trials

●Oesophageal dysmotility in systemic sclerosis: comparison of

HRCT and scintigraphy

● Relationship between the growth pattern of nasopharyngeal

cancer and the cervical lymph nodes based on MRI findings: can
the cervical radiation field be reduced in patients with
nasopharyngeal cancer?

● Six years experience in intracoronary brachytherapy procedures:

patient doses from fluoroscopy

● What is the optimum breast plan: a study based on the START

trial plans

● Feasibility of automated matching of supine and prone

CT-colonography examinations

● Improved motion compensation in 3D-CT using respiratorycorrelated

segment reconstruction: diagnostic and radiotherapy applications

● A mathematical model for patient skin dose assessment in

cardiac catheterization procedures

● Nurse-led central venous catheter service: Christie experience

 ● Breast movement during normal and deep breathing, respiratory
training and set up errors: implications for external beam partial
breast irradiation

● Book review

● A case of spinal cord compression of unknown cause

The British Journal of Radiology, 79 (2006), 707–711

SHORT COMMUNICATION

Quantitative optical coherence tomographic elastography:

method for assessing arterial mechanical properties
1,2 1 1 1,2
J ROGOWSKA, PhD, N PATEL, BS, S PLUMMER, BS and M E BREZINSKI, MD, PhD

1
Brigham and Women’s Hospital, 75 Francis Street, Boston, MA 02115 and 2Harvard Medical School,
Boston, Massachusetts, USA

ABSTRACT. Optical coherence tomography elastography represents a potentially

attractive new technique for measuring elastic properties of tissues on a micron scale. Received 23 November
In this study, the feasibility of optical coherence tomography (OCT) to study the 2005
Revised 9 March 2006
mechanical properties of phantoms and atherosclerotic arterial samples is reported. Accepted 12 April 2006
The elastic modulus of tissue-mimicking phantoms was measured using OCT and
correlated with mechanical measurements. The results indicate that elastography based DOI: 10.1259/bjr/22522280
on OCT represents an attractive technique for evaluating the mechanical properties of
’ 2006 The British Institute of
tissues.
Radiology

Knowledge of the mechanical properties of vascular
tissue could provide diagnostic information about a
range of vascular diseases, from pulmonary hyperten-
sion to coronary atherosclerosis. Therefore, it is impor-
tant to develop and evaluate new quantitative methods
for measuring the elastic modulus of normal and
abnormal arterial tissue.
The term ultrasound elastography was introduced for
the first time by Ophir and colleagues in 1991 [1]. The
elastography technique was based on applying a
pressure on the examined tissue and in estimating the
induced strain distribution by tracking the tissue motion
[1–3]. An ultrasound elastography is limited by a
resolution between 80 mm and 100 mm, which is likely
insufficient for vascular assessments. A technology that
has shown considerable promise as a method of high
resolution intravascular imaging and elastography is
optical coherence tomography (OCT) [4–9]. OCT is a
micrometre scale imaging technology analogous to
ultrasound, measuring the back-reflection of near-infra-
red light rather than sound. Current OCT systems that
can be used in vivo have a resolution between 10 mm and
20 mm. Recently, several semi-quantitative approaches to
OCT elastography have been demonstrated using
speckle tracking [10–12]. However, in order for the
technique to be used clinically, the quantitative accuracy
of tissue elasticity measurements must be established,
which is the focus of this paper.
In this paper, we first describe and implement an OCT
elastography technique on tissue phantoms, measuring
Address correspondence to: Mark E Brezinski, Brigham and
Women’s Hospital, MRB 106, 75 Francis Street, Boston, MA
02115, USA.
This research is sponsored in part by the National Institutes of
Health, Contracts NIH-RO1-AR44812 (MEB), NIH R01 AR46996
(MEB), NIH R01- HL63953 (MEB), NIH-1-R01-HL55686 (MEB), NIH
R01 EB000419 (MEB) and NIH-1-R29-HL55686 (MEB).
deformation of samples to obtain stress-strain curves that
characterize the linear elastic properties. Second, these
results of OCT elasticity measurements are validated and
quantitated with non-imaging mechanical measurements
of the same phantoms. Finally, this validated OCT
elastography approach was applied to atherosclerotic
arterial samples.
Materials and methods
Experimental elastography system
OCT is analogous to ultrasound, measuring the
intensity of back-reflected infrared light [4, 5].
Ultrashort light pulses or low coherent light is generated
at the sample. The time for the light to be reflected back
or echo delay time is used to measure distances. The
intensity of back-reflection is plotted as a function of
depth. The beam is then scanned across the sample to
produce two- and three-dimensional data sets.
Phantom and aorta images
Phantoms were made by mixing different amounts of
agar, gelatin and water. Using a fixed amount of gelatin
(8 g), phantoms with different hardness were made by
varying the amount of agar: 0.5 g for 1% phantom, 1.0 g
for 2% phantom, 3 g for 3% phantom. Gelatin and agar
were mixed together and dissolved in 50 ml of boiling
water. The 0.25 g of activated charcoal particles were
added as scatterers. After refrigeration, the gelatin
phantoms were cut into 2.5 cm62.5 cm61 cm thick
block shaped samples, covered with a cover slip, and
scanned. The elastic properties of the phantoms were
confirmed by applying weights of 49.81 g, 102.88 g and

The British Journal of Radiology, September 2006 707


133.03 g. The changes in tissue width were measured
both with callipers and OCT imaging. All measurements
were repeated three times for each of the phantoms, thus
producing three sets of data. The OCT measurements
were corrected for refractive index.
The human atherosclerotic aorta samples were
obtained post-mortem and stored at 0 ˚C with 0.1%
sodium azide. The samples were placed under the OCT
system and scanned before and after compression, using
a procedure similar to phantom scanning. The weights
applied to the aorta samples were 2.29 g, 5.60 g and
8.21 g. The OCT scanning was performed in the axial
cross-sections of the samples, covering 2 mm62 mm
areas in the centre, far away from the sample edges. The
OCT image resolution was 400 by 400 pixels, where each
pixel was 0.005 mm60.005 mm.
Calculation of elastic modulus of phantoms
Our goal was to calculate the local values of elastic
modulus in the phantoms. If we assume that the
phantoms are elastic, isotropic and incompressible, and
the stress is applied uniformly in axial direction, then the
Young’s modulus (or elastic modulus) E is defined by
[13]:
s pixels were converted to millimetres and corrected using
E~ ð1Þ
e refraction index. By measuring the strain (Equation (3))
(a)
(c)
Figure 1. A typical optical coherence tomography (OCT) 2% phantom images after applying stress of (b) 0.772 kPa, (c) 1.595 kPa
and (d) 2.062 kPa to original phantom (a). The displacement vectors are shown in red.
708
J Rogowska, N Patel, S Plummer and M E Brezinski
where s is the axial (or normal) stress (defined as the
force F perpendicular to the cross sectional area, divided
by the cross sectional area A):
F
s~ ð2Þ
A
and e is the axial strain (defined as the fractional change
in sample thickness L):
DL
e~ ð3Þ
L
By applying weights, samples were compressed, and
their thicknesses before and after compression were
measured using callipers and OCT displacement vectors.
The image data were processed pixel-by-pixel and the
total axial displacement vectors were calculated using
cross-correlation [10, 11]. For strains up to 5%, the
relationship between force and displacement is usually
linear [14]. Based on Equation (2), the equivalent stress
applied to phantoms was computed by dividing the
force by the surface dimensions of the samples, which
yielded stress values of 0.77 kPa, 1.60 kPa and 2.06 kPa
for three compressions. The optical coherence tomo-
graphic elastography (OCTE) displacement values in
(b)
(d)
The British Journal of Radiology, September 2006
Short communication: Quantitative optical coherence tomographic elastography

Figure 2. Young’s modulus for 1%,

2%, and 3% phantoms estimated
using two techniques: callipers’
measurements and OCT elastogra-
phy. The bars indicate standard
deviation. The differences between
calliper and OCTE measurements
were found to be non-significant
(p50.6, p50.25 and p50.38 for 1%,
2% and 3% phantoms, respectively).

(a) (b)

(c) (d)

Figure 3. Aorta images after applying stress of (b) 0.0355 kPa, (c) 0.0868 kPa and (d) 0.1273 kPa to original image (a). Figure 3a
indicates location of intima (I), media (M) and glass compression plate (G). The displacement vectors are shown in red.

The British Journal of Radiology, September 2006 709


(a)
(c)
Figure 4. Axial displacement maps for compressed aorta images shown on Figure 3.
for three different applied stresses, the stress-strain
relationship was calculated using a linear regression.
The Young’s modulus was estimated from the slope of
the strain-stress linear function. All measurements were
repeated three times for three sets of samples, and then
averaged.
Estimation of the elastic modulus of aorta
Once the stress-strain measurements are obtained on
the phantoms, the slopes of these stress-strain curves are
used to determine the elastic modulus values. Since the
Young’s moduli (E) of the phantoms are known, and
the slope values (S) of the stress-displacement curves for
the phantoms can be measured, the conversion factor C
relating the stress-displacement slope values to the
Young’s modulus could be determined by:
 
1 E1% E2% E3%
C~ z z ð4Þ
3 S1% S2% S3%
where E1%, E2%, and E3% are elastic moduli for 1%, 2%
and 3% phantoms, respectively, and S1%, S2%, and S3% are
slope values from the corresponding stress-displacement
curves.
In order to calculate the estimated elastic modulus of the
aorta (rEAorta), the slope value of the OCTE stress-
displacement curve for the aorta (SAorta) was transformed
710
J Rogowska, N Patel, S Plummer and M E Brezinski
(b)
to Young’s modulus using a conversion factor C [13, 15]:
rEAorta ~C
SAorta ð5Þ
Results
The phantoms underwent compression with weights
and the changes in phantom’s thickness were measured
both with callipers and OCT imaging. The image data
were processed pixel-by-pixel and the total axial dis-
placement vectors were calculated. An example of the
2% phantom is shown on Figure 1.
For all phantoms, stress-strain relationships using
calliper and OCT measurements were calculated and
Young’s modulus was estimated (Figure 2). The differ-
ences between calliper and OCT measurements were
found to be non-significant (paired t-test; p50.6 for 1%
phantom, p50.25 for 2% phantom and p50.38 for 3%
phantom). The relationship between two Young’s mod-
ulus estimators (using callipers and OCTE) was found to
be highly correlated, R250.92.
Aorta samples also underwent compression with
weights (Figure 3). The changes in tissue thickness were
measured using OCT. The image data were processed
pixel-by-pixel and the axial displacement vectors
(Figure 3b–d) and displacement maps (Figure 4) were
The British Journal of Radiology, September 2006
Short communication: Quantitative optical coherence tomographic elastography
calculated. In order to calculate the relative elastic
modulus of the aorta, we used the conversion factor
C512.14, which was derived from stress-displacement
phantom curves. The average estimated Young’s mod-
ulus for the aorta sample shown on Figure 4 was equal to
6.11 kPa.
Discussion
This paper attempts to quantify OCT elastography as a
method for assessing Young’s modulus, and therefore
evaluating tissue mechanical properties. The OCT
elastography system was calibrated with tissue-mimick-
ing phantoms and the elastic modulus of the athero-
sclerotic arterial samples was estimated. A strong
correlation existed between the mechanical measure-
ments and those performed with OCT elastography,
with no significant difference existing between the two
techniques. OCT elastography was then used to estimate
Young’s modulus through the use of the conversion
factor. According to other reports [15], our estimated
elastic modulus of 6.11 kPa for the aorta was within the
range of Young’s modulus values calculated for smooth
muscle (0–100 kPa) and soft tissue (0–200 kPa).
Several obstacles need to be overcome for OCT
elastography application in human arteries in vivo.
First, controlled transmural pressures need to be applied
to make OCT measurements in vivo. This can potentially
be achieved with an intravascular pressure balloon or
saline flushes with measured pressures. Second, it is
assumed that the tissue under investigation is elastic,
uniform and nearly incompressible. Based on work with
ultrasound, the approximation of vascular tissue being
elastic and incompressible is reasonable. However, with
atherosclerotic or severely hypertensive arteries, the
tensile properties will not necessarily be completely
uniform. The importance of this has to be examined with
future studies. Third, for correct conversion of stress-
displacement data to an elastic modulus, it is essential
that the calibration sample and tissue sample have
similar shape [13]. This needs to be taken into account
when calibrating the technique for imaging in vivo
arteries. Sample height and sample compressibility are
non-critical factors for system accuracy. However, the
size of the area where the stress is applied to the sample
is a very critical parameter and needs to be well
controlled to obtain accurate results.
In summary, OCT elastography represents an attrac-
tive technique for evaluating the mechanical properties
of tissue due to its micrometre scale resolution. Other
OCT elastography studies have essentially focused on
qualitative analysis, not on the accuracy of measured
mechanical parameters. The focus of this work was to
The British Journal of Radiology, September 2006
establish quantification of OCT elastography. The study
reported here was also the first attempt to calculate the
elastic modulus of atherosclerotic tissue using OCT.
References
1. Ophir J, Cespedes I, Ponnekanti H, Yanzi Y, Li X.
Elastography: a quantitative method for imaging the elasti-
city of biological tissues. Ultrason Imaging 1991;13:111–34.
2. de Korte CL, Cespedes EI, van der Steen AF, Lancee CT.
Intravascular elasticity imaging using ultrasound.
Ultrasound Med Biol 1997;23:725–46.
3. de Korte CL, van der Steen AF. Intravascular ultrasound
elastography: an overview. Ultrasonics 2002;40:859–65.
4. Brezinski ME, Tearney GJ, Bouma BE, Izatt JA, Hee MR,
Swanson EA, et al. Optical coherence tomography for
optical biopsy: properties and demonstration of vascular
pathology. Circulation 1996;93:1206–13.
5. Boppart SA, Bouma BE, Pitris C, Tearney GJ, Southern JF,
Brezinski ME, et al. Intraoperative assessment of micro-
surgery with three-dimensional optical coherence tomogra-
phy. Radiology 1998;208:81.
6. Jesser CA, Boppart SA, Pitris C, Stamper DL, Nielsen GP,
Brezinski ME, et al. High resolution imaging of transitional
cell carcinoma with optical coherence tomography: feasi-
bility for the evaluation of bladder pathology. Br J Radiol
1999;72:1170–6.
7. Patwari P, Weissman NJ, Boppart SA, Jesser C, Stamper D,
Fujimoto JG, et al. Assessment of coronary plaque with
optical coherence tomography and high frequency ultra-
sound. Am J Cardiol 2000;85:641–4.
8. Podoleanu AGh. Optical coherence tomography. Br J Radiol
2005;78:976–88.
9. Tearney GJ, Brezinski ME, Bouma BE, Boppart SA, Pitris C,
Southern JF, et al. In vivo endoscopic optical biopsy
with optical coherence tomography. Science 1997;276:
2037–9.
10. Rogowska J, Patel NA, Fujimoto JG, Brezinski ME. OCT
elastography of the vascular tissue – importance of cross-
correlation kernel size. OSA Biomedical Topical Meetings,
OSA Technical Digest, (Optical Society of America,
Washington, DC, 2002), pp. PD20.1-20.3.
11. Rogowska J, Patel NA, Fujimoto JG, Brezinski ME.
Quantitative OCT elastography technique for measuring
deformation and strain of the atherosclerotic tissues. Heart
2004;90:556–62.
12. Schmitt JM. OCT elastography: imaging microscopic
deformation and strain in tissue. Opt Express
1998;3:199–211.
13. Erkamp RQ, Wiggins P, Skovoroda AR, Emelianov SY,
O’Donnell M. Measuring the elastic modulus of small tissue
samples. Ultrason Imaging 1998;20:17–28.
14. Chen EJ, Novakofski J, Jenkins WK, O’Brien WD. Young’s
modulus measurements of soft tissues with application to
elasticity imaging. IEEE Trans Ultrason Ferroelectr Freq
Control 1996;43:191–4.
15. Samani A, Bishop J, Luginbuhl C, Plewes DB. Measuring
the elastic modulus of ex vivo small tissue samples. Phys
Med Biol 2003;48:2183–98.
711
 The British Journal of Radiology, 79 (2006), 712–715

SHORT COMMUNICATION

Low-carbohydrate diet induced reduction of hepatic lipid content

observed with a rapid non-invasive MRI technique
1 1 1
K G HOLLINGSWORTH, PhD, M Z ABUBACKER, FRCR, I JOUBERT, 2M E D ALLISON, FRCP and
1
D J LOMAS, FRCR

Departments of 1Radiology and 2Medicine University of Cambridge and Addenbrookes Hospital,

Level 5, Box 219, Addenbrookes Hospital, Cambridge CB2 2QQ, UK

ABSTRACT. Low carbohydrate diets are currently fashionable for inducing weight loss,
but the metabolic effects at organ level are not well understood, especially the effect
on liver fat storage. Such studies require serial hepatic fat measurements, for which
liver biopsy is impractical. In 10 healthy volunteers we demonstrate the use of rapid
(total 2 min acquisition time, 10 min magnet room time), non-invasive, quantitative
MRI to serially measure hepatic fat changes induced by following a low carbohydrate
diet for 10 days. A significant (p,0.01) reduction in hepatic fat after 3 days of dieting
was observed in 5 subjects. All subjects demonstrated significant (p,0.01) reductions in
hepatic fat by day 10. A strong correlation (k50.81) existed between the initial fat Received 6 January 2006
content and the percentage fat content reduction in the first 3 days of the diet. All Revised 3 April 2006
subjects lost weight (average 1.7 kg at day 3 and 3.0 kg at day 10), but this was not Accepted 11 April 2006
correlated with hepatic fat loss after 3 days or 10 days of dieting. The results presented DOI: 10.1259/bjr/23166141
illustrate the potential value of MR hepatic fat quantification in longitudinal studies of
hepatic fat content. ’ 2006 The British Institute of
Radiology

Hepatic steatosis is of particular interest in the western
world owing to the increasing prevalence of the insulin
resistance or ‘‘metabolic’’ syndrome. Hepatic steatosis
and non-alcoholic steatohepatitis (NASH) are now
considered a part of this condition and there is emerging
evidence that, rather than being a consequence of
systemic insulin resistance, they may have a causative
role [1, 2]. As a result, therapeutic or dietary interven-
tions to reduce hepatic steatosis may be more appro-
priate for treating insulin-resistance, rather than by
treating the different aspects separately. It is estimated
that 20–25% of the US population is obese and that
among the obese group there is a high prevalence of non-
alcoholic fatty liver disease (NAFLD). The risk of serious
sequelae such as type II diabetes mellitus or progression
to chronic liver disease means that this is an area
requiring urgent research and assessment [3].
These types of studies require accurate serial quanti-
fication of hepatic steatosis, but currently liver biopsy
(and related biochemical analysis) is the established
method. As an invasive procedure, even with a minimal
related morbidity and mortality, this is difficult to justify
in healthy volunteers and impractical for serial measure-
ments. Additionally the method samples only small
volumes of tissue, which may lead to problems with
regional fat variation and give unrepresentative results.
An alternative approach is to use a validated rapid
imaging method that allows absolute hepatic fat (mainly
Funded by the Fund and Friends of Addenbrookes.
intracellular triglyceride) estimation through fat specific
chemical shift imaging corrected for T2* variation [4, 5].
Previous studies have shown a good correlation between
steatosis assessed by liver biopsy and by MRI methods
[6], and excellent correlation exists between hepatic fat
measured by liver histology, MRI and CT methods [7, 8].
The effect of low carbohydrate diets on weight loss,
insulin resistance and serum triglyceride markers has
recently been the focus of intensive research efforts [9,
10], but to the best of the authors’ knowledge, the effect
of such diets on hepatic fat has never been directly
measured. The aim of this work was to monitor the
hepatic fat response of healthy volunteers during the
induction phase of a low-carbohydrate diet using the MRI
method described above.
Materials and methods
The study was approved by the local ethics committee
and informed consent was obtained from the volunteers
after the procedure was fully explained. Participants
received no monetary incentive. Given the lack of
consensus on the normal hepatic lipid percentage range
measured by MRI we used an arbitrary limit of a pre-diet
hepatic fat estimate of 7.0% or above as the entry
threshold for the study. This threshold was chosen as it
was the median fat percentage observed in our previous
work on healthy volunteers (n525); a similar result has
also been shown with this technique by Fishbein [4]. 10

712 The British Journal of Radiology, September 2006

Short communication: Hepatic lipid content observed with a rapid non-invasive MRI technique

healthy volunteers (3 male, 7 female, age 32–56 years)
were recruited from the community: the volunteers were
not taking regular medication and had no history of
hepatic or biliary disease. Five further volunteers were
excluded from the study after an initial MRI examination
showed their hepatic fat estimate to be below 7.0%.
The participants were asked to follow a low carbohy-
drate diet at home for 10 days, restricting carbohydrate
intake to less than 20 g carbohydrate per day in the form
of green salad or vegetables, but with no other restriction
on total energy intake or food choice. The volunteers
kept a diet diary and abstained from alcohol.
Hepatic fat measurement by MRI was performed at 4
time points; immediately pre-diet, at 3 days and 10 days
on the diet and 7 days after reverting to their normal diet.
The volunteers were examined at the same time of day
on each occasion. The initial body mass index (BMI) was
recorded and the weight of the volunteers was assessed
on each visit to the MRI unit. Weight was measured with
the subjects changed for the MRI examination and with-
out shoes. Adherence to the diet was monitored by the
patients maintaining a food record sheet and by urinary
ketone assessment to ensure that ketosis was initiated
and maintained. Ketones are usually not detectable in
healthy volunteers following a balanced diet.
Examinations were performed on a 1.5 T whole body
MRI (Excite, GEHT, Milwaukee) with an 8-channel body
array. In and out of phase gradient echo scans (matrix
2566128, section (slice) thickness 10 mm, gap 1.5 mm,
repetition time (TR)/echo time (TE)5180/2.2 ms (out of
phase)/4.4 ms (in phase)) were acquired axially at two
different flip angles (20 ˚ and 70 ˚) and a T2* map of the
liver was obtained using a location-matched, multi-
section, multiecho gradient sequence (TR5120 ms, 16
equally spaced echoes, TE152.2 ms, TE254.4 ms). These
required a total of three 20 s breath-holds in addition to
an initial 20 s breath-hold study for checking the
positioning. The T2* data was used to correct the in-
phase and out-of-phase images intensities for T2*
relaxation. The fat percentage was calculated by:
FP~100(Sin
Sout )=2Sin ð1Þ
where FP is the fat percentage, Sin is the intensity from the
in-phase image, Sout is the intensity from the out-of-phase
image (both values corrected for T2* relaxation). The
acquisition of in-phase and out-of-phase images with
different flip angles (and hence T1 weighting) allows us to
distinguish whether fat or water is the majority species
[4, 5]: this resolves the ambiguity that occurs in Equation
(1) for fat percentage greater than 50%.
Four sections centrally placed in the liver were
analysed by a single operator and three circular regions
of interest (ROI) with fixed area (5 cm2) were positioned
over the liver parenchyma on each section, avoiding
large vessels and the gallbladder. The 12 ROIs for each
individual were then averaged to give a result for that
time point. Significant changes between time points were
assessed using a paired t-test (SPSS 12.0). Hepatic fat
percentage was the primary outcome measure, as
assessed from the MR images. Correlation coefficients
(Pearson) were calculated between the initial fat mea-
surement and the percentage reduction in fat after
31 days and 10 days of the diet, and also between the
percentage weight reduction and the percentage reduc-
tion in hepatic fat at days 3 and 10 of the diet.
Results
The diet records indicated that all the volunteers
maintained the diet successfully and this was confirmed
by evidence of ketosis on urine testing. All volunteers
had ‘‘negative’’ ketone readings before the diet, and
these all increased to at least ‘‘moderate’’ by day 3 and
for the duration of the diet. The initial BMI of the
volunteers (range 23–32, median 28) demonstrated that
the volunteers were, at worst, moderately overweight.
Table 1 shows the changes in hepatic fat percentage for
the 10 individuals, at all the time-points in the trial.
Figure 1 shows the relative change (compared with the
pre-diet measurement) in the hepatic fat percentage
graphically. Figure 2 illustrates the time course of results
and error bars for subjects 1 and 2. There was no
significant variation of the measured fat percentage with
the location of the ROI. Five of the subjects experienced a
significant (p,0.01) decrease in hepatic fat in the first
3 days of dieting. Two of these subjects displayed a
further significant change between day 3 and day 10 of
the diet. All 10 subjects experienced a significant

Table 1. Selected BMI, hepatic fat and weight data for the volunteers
Subject Initial BMI Initial hepatic 3 day 10 day 17 day Weight Weight change
(kg m22) fat (%) hepatic fat (%) hepatic fat (%) hepatic fat (%) loss (kg) on (kg) 1 week
completion of after diet
diet (day 10) (day 17)

1 30 12.1 10.0a 9.4a 9.9 4 +0.5

2 32 8.5 7.1a 7.4a 6.7 4 +2.0
3 29 11.8 9.0a 6.7a 6.4 4.5 21.5
4 27 7.0 7.0 5.6a 8.4b 2 21.5
5 23 9.0 7.9a 7.5a 7.7 3 0
6 28 7.2 6.9 6.4a – 3 –
7 26 11.2 9.8 9.0a 9.9 3 +1.0
8 32 8.8 8.2 7.0a 8.6b 3 +0.5
9 26 7.7 7.1 5.0a 6.4b 1 +0.5
10 28 10.2 9.1a 6.1a 5.0b 2 22
a
p,0.01 compared with initial measurement.
b
p,0.01 compared with previous measurement (17 days measurement only).
BMI, body mass index.

The British Journal of Radiology, September 2006 713


Figure 1. Relative change in hepatic fat for all 10 healthy
volunteers compared with hepatic fat percentage at day 0.
Figure 2. Plot of individual hepatic fat measurements for
subjects 1 and 2 with error bars.
(p,0.01) decrease in hepatic fat within 10 days of starting
the diet. It was noted that the volunteers with the highest
initial fat content had the greatest percentage decrease
during the first 3 days of the diet (k50.81, range 0–24%),
although the correlation at day 10 was weaker (k50.42,
range 11–43%). All the subjects lost weight (mean weight
loss: 1.7 kg at day 3 and 3.0 kg at day 10) but this was not
correlated with changes in hepatic fat percentage. 1 week
after stopping the diet, five subjects gained weight, three
lost weight and one maintained the same weight (one
was lost to follow up after cessation of the diet), though
there was no correlation with hepatic fat change in the
week after cessation of the diet.
Discussion
This study evaluates hepatic fat response to a low
carbohydrate diet in 10 healthy volunteers and demon-
strates a significant reduction of hepatic fat levels. To the
best of our knowledge this is the first study to use
quantitative image-based MR to demonstrate a signifi-
cant reduction in hepatic fat during a low carbohydrate
diet intervention. The dominance of fat and protein in
this diet means that a decrease in liver fat reflects
mobilization of hepatic lipid stores as an energy source
and a contributor to the related ketosis. All subjects lost
714
K G Hollingsworth, M Z Abubacker, I Joubert et al
weight, demonstrating that either the diet itself or its
appetite suppressant effect is hypocaloric.
The limitations of this pilot study were the small
number of subjects studied and the lack of biochemical
blood correlates, in particular the evaluation of insulin
resistance by clamping methods. Although all subjects
experienced reduced hepatic fat levels during the study,
there was variation in the percentage of initial fat lost by
day 10 (range 11–43%). Although there was an empirical
correlation between initial fat percentage and the
percentage of hepatic fat lost by day 3, evaluation of
the insulin resistance changes of the subjects may
elucidate the reason for the differing degrees of response,
particularly since there was no correlation between the
percentage hepatic lipid reduction and the reduction in
weight. This would also provide insight into whether a
similar response could be expected from the different
patient groups with hepatic steatosis. A non-dieting
(control) group was not assessed in this study but
previous work on observing liver fat changes in healthy,
non-dieting volunteers has shown weekly changes of no
more than approximately 1% [11]. There was also no
detailed study of the subjects dietary and alcohol habits
for more than 3 days before commencing the diet: such a
standardization step may be important in drawing full
quantitative conclusions from a diet study of short
duration. Future work in larger volunteer and patient
groups will address these limitations.
There have been few comparable studies as the
majority of metabolic studies do not measure hepatic
fat content directly (presumably owing to the ethical
difficulty of performing serial liver biopsies), choosing
rather to measure serum triglycerides. In a study of
overweight subjects fed a high fat (56% of calorie intake)
diet (with carbohydrate) for a period of 2 weeks the liver
accumulated fat (average 35% of the initial fat content)
[12] compared with an isocaloric, low fat (16% of calorie
intake) where there was a reduction in liver fat (average
20% of the initial fat content). A study of the effect of a
hypocaloric low carbohydrate diet followed for 14 days
by patients with type II diabetes found that there was an
improvement in insulin resistance and a decrease in
plasma triglycerides, though liver fat was not directly
measured [10]. Other authors have reported evidence
that low carbohydrate diets can alter body fat composi-
tion [13]. This is of importance given the emerging
literature on the role of hepatic fat in the development of
systemic insulin resistance, leading to type II diabetes
mellitus, hyperlipidaemia, hypertension and increased
artherosclerotic risk. This pilot study suggests that a low
carbohydrate diet may have a role in modifying hepatic
fat and hence insulin resistance.
Alternative non-invasive methods such as ultrasound
and CT have been proposed, but ultrasound is limited by
lack of specificity and CT by the use of ionizing
radiation, which is difficult to justify in healthy indivi-
duals. The method used in this study can be imple-
mented on the majority of currently installed MRI
systems, and is simpler to implement than fat measure-
ment by proton MR spectroscopy, which requires longer
acquisition times and specialist analysis [6, 12]: spectro-
scopy methods cannot easily be used to sample the entire
liver and cannot detect focal fat variation in the liver in a
short examination. One study [14] finds that carbon-13
The British Journal of Radiology, September 2006
Short communication: Hepatic lipid content observed with a rapid non-invasive MRI technique
spectroscopy of hepatic lipids is in excellent agreement
with morphometric analysis of biopsy specimens.
However, this is principally a research technique and is
not available in the vast majority of clinical settings.
This pilot study demonstrates the practical implemen-
tation and utility of MRI fat quantification as a tool in
serial studies of hepatic fat content. In healthy volunteers
the method demonstrated significant hepatic fat reduction
resulting from a low carbohydrate dietary intervention.
Acknowledgments
This study was funded by the Fund and Friends of
Addenbrookes. We thank the staff of the Magnetic
Resonance Imaging and Spectroscopy Unit for support
in running the study
References
1. Samuel VT, Liu ZX, Qu X, Elder BD, Bilz S, Befroy D, et al.
Mechanism of hepatic insulin resistance in non-alcoholic
fatty liver disease. J Biol Chem 2004;279:32345–53.
2. Tiikainen M, Hakkinen AM, Korsheninnikova E, Nyman T,
Makimattila S, Yki-Jarvinen H. Effects of rosiglitazone and
metformin on liver fat content, hepatic insulin resistance,
insulin clearance, and gene expression in adipose tissue in
patients with type 2 diabetes. Diabetes 2004;53:2169–76.
3. Saadeh S, Younossi ZM, Remer EM, Gramlich T, Ong JP,
Hurley M, et al. The utility of radiological imaging in
nonalcoholic fatty liver disease. Gastroenterology
2002;123:745–50.
4. Fishbein MH, Gardner KG, Potter CJ, Schmalbrock P, Smith
MA. Introduction of fast MR imaging in the assessment of
hepatic steatosis. Magn Reson Imag 1997;15:287–93.
The British Journal of Radiology, September 2006
5. Hussain HK, Chenevert TL, Londy FJ, Gulani V, Swanson
SD, McKenna BJ, et al. Hepatic fat fraction: MR imaging for
quantitative measurement and display – early experience.
Radiology 2005;237:1048–55.
6. Thomsen C, Becker U, Winkler K, Christoffersen P, Jensen
M, Henriksen O. Quantification of liver fat using magnetic
resonance spectroscopy. Magn Reson Imag 1994;12:487–95.
7. Longo R, Ricci C, Masutti F, Vidimari R, Croce LS, Bercich
L, et al. Fatty infiltration of the liver: quantification by 1H
localized magnetic resonance spectroscopy and comparison
with computed tomography. Invest Radiol 1993;28:297–302.
8. Joy D, Thava VR, Scott BB. Diagnosis of fatty liver disease:
is biopsy necessary? Eur J Gastro Hepatol 2003;15:539–43.
9. Boden G, Sargrad, K, Homko C, Mozzoli M, Peter Stein T.
Effect of a low-carbohydrate diet on appetite, blood glucose
levels, and insulin resistance in obese patients with Type 2
diabetes. Ann Intern Med 2005;142:403–11.
10. Yancy WS, Olsen, MK, Guyton JR, Bakst RP, Westman EC.
A low-carbohydrate, ketogenic diet versus a low-fat diet to
treat obesity and hyperlipidemia. Ann Int Med
2004;140:769–77.
11. Lomas DJ, Black RT, Pinney J. Hepatic steatosis quantifica-
tion by MRI: serial measurement and normal variation. Proc
Intl Soc Magn Reson Med 2003;11:1424.
12. Westerbacka J, Lammi K, Hakkinen AM, Rissanen A,
Salminen I, Aro A, et al. Dietary fat content modifies liver
fat in overweight nondiabetic subjects. J Clin Endocrin
Metab 2005;90:2804–9.
13. Dansinger ML, Gleason JA, Griffith JL, Selker HP, Schaefer
EJ. Comparison of the Atkins, Ornish, Weight Watchers,
and Zone diets for weight loss and heart disease risk
reduction: a randomized trial. JAMA 2005;293:43–53.
14. Petersen KF, West AB, Reuben A, Rothman DL, Schulman
GI. Non-invasive assessment of hepatic triglyceride content
in humans with carbon-13 nuclear magnetic resonance
spectroscopy. Hepatology 1996;24:114–7.
715

COMMENTARY
Translational research in radiotherapy trials
1
C WEST, BA, PhD and 2S McKEOWN, MA, PhD
1
Academic Department of Radiation Oncology, The University of Manchester, Christie Hospital,
Manchester M20 4BX and 2School of Biomedical Sciences, University of Ulster at Coleraine BT52
1SA, Northern Ireland, UK
In the last 10 years, a range of powerful laboratory
techniques have become available to measure genes at
the DNA, RNA and protein level. This progress in assay
methods is associated with a huge increase in the
amount of data produced and the development of
computational methods for analysing the biological
information. These advances have the potential to
provide the means of assigning molecular signatures
that describe intrinsic differences in how patients
respond to radiotherapy. Currently, the methods are
being validated as tools that provide prognostic or
predictive information on cancer pre-disposition and
treatment outcome on an individual basis. In order to
raise awareness of recent developments in the area, the
British Institute of Radiology’s (BIR’s) Radiation and
Cancer Biology Committee organized a meeting at the
BIR on 10 March 2006 to overview the methods and to
discuss potential applications for advancing translational
research in radiotherapy trials.
The greatly improved opportunities for translational
research in radiotherapy trials in the UK were summar-
ized by Professor Peter Hoskin (Mount Vernon Hospital,
Northwood). In order to improve the design and quality
of trials involving radiotherapy within the UK, trials are
now reviewed by the NCRI Radiotherapy Clinical
Studies Group. The central review encourages multi-
centre involvement to facilitate patient accrual. It also
allows the identification of potential translational
research opportunities at an early stage. There are
currently 44 radiotherapy trials open in the UK with
more planned, giving widening opportunities for trans-
lational questions to be addressed.
High throughput genotyping has the potential to find
common genetic variants reported as single nucleotide
polymorphisms (SNPs) conferring increased radiosensi-
tivity (Dr Alison Dunning, University of Cambridge). In
terms of application in clinical trials, one of the key
advantages of genotyping is the relative ease of sample
collection. A single blood sample taken in EDTA is
required, with some flexibility of the timing of sample
716
The British Journal of Radiology, 79 (2006), 716–718
Received 7 April 2006
Accepted 13 April 2006
DOI: 10.1259/bjr/62971945
’ 2006 The British Institute of
Radiology
collection. An important caveat being that DNA quality
and uniformity matters. Although samples are relatively
stable compared with those collected for RNA or
proteomic analyses, there can be problems with collec-
tions involving DNA extracted from mixed types of
samples, which can yield unusable data. The HapMap
(SNP haplotype map) project has already published its
phase 2 results and most of the .2 million documented
SNPs in the human genome are now tagged and can
therefore be studied. The good news is the technology
involved for analysing SNPs has progressed in recent
years and high-throughput methods incorporating
robotics enable the simultaneous processing of many
thousands of samples within a single day. SNP-tagging
methods, for reducing the number of SNPs that need to
be investigated, are also developing rapidly. Studies can
focus on candidate genes likely to be involved in
determining an individual’s sensitivity to radiation, e.g.
genes involved in the recognition and repair of DNA
damage. Although each gene of interest might contain
several hundred SNPs, the number needed to be studied
directly can be reduced by using tagged SNPs.
‘‘Tagging’’ involves identifying regions of the genome
(haplotype blocks) where SNPs have evolved together
and then finding the SNPs within each block that
provide information on most of the other SNPs.
Nevertheless, even with candidate genes and SNP
tagging, there is an obvious need to collaborate and
pool samples obtained in several trials to enable
exploration of the full human genome. Since there are
many variants, each with probably only a small
individual effect, large samples (.2000 patients) are
required to have any statistical power to detect real
effects due to the variants – the field of association
studies has been plagued with false positive results due
to small study sizes. For radiotherapy, high-throughput
genotyping offers the opportunity to identify a range of
radiosensitivity phenotypes, which in the future should
allow more accurate tailoring of treatment protocols to a
patient’s predicted normal tissue radiosensitivity. This
The British Journal of Radiology, September 2006
Commentary: Translational research in radiotherapy trials
could improve treatment planning for all groups of
patients from the most radiosensitive to radioresistant.
The potential of cDNA and oligonucleotide micro-
arrays was described by Dr Wendy Allen (Queen’s
University Belfast). The techniques are now well
advanced with evaluation of 60 000 genes per sample
possible. The methods are capable of comparing the
expressed genes in tumour samples with patient out-
come following radiotherapy. A disadvantage of the
technology is the requirement for fresh material that is
not always readily obtainable within a multicentre trial
setting. Methods are now becoming available, however,
to extract good quality RNA from wax embedded
material which will increase the applicability of the
approach, since this type of material is easier to store and
much more widely available. Microarrays yield huge
amounts of information, and so the bioinformatic
handling of the data is crucial for identifying potentially
important genes and molecular signatures. Dr Allen
emphasised the need for validation studies to verify
molecular profiles identified in a training set with a
second blinded set of samples. The potential applications
of the technique in identifying predictive/prognostic
gene signatures was also discussed.
Taking the step from expressed gene to translated
protein, Professor Tony Whetton (The University of
Manchester) described the current state of the proteo-
mics revolution. Although the technology is continually
improving, the capabilities are several years behind the
high-throughput DNA and RNA approaches. The
techniques required are technically demanding, combin-
ing highly sophisticated protein separation methods
(two-dimensional liquid chromatography) with mass
spectrometry. This is partly improved by new tagging
techniques and other approaches which allow relative
quantification from several samples as well as protein/
peptide identification. One of the biggest problems,
however, is the dynamic range issue. The separation of
low abundance proteins of interest from high abundance
proteins remains critical, although major progress has
been made in this area with major protein removal
columns.
The preliminary removal of these proteins makes
analysis of low abundance, potentially diagnostic pro-
teins much more likely. As gene expression array and
proteomic data can be linked, there is also potential to
compare transcribed and translated genes in the same
samples. Procedures for blood sampling for proteomic
studies must be carefully standardized, and the use of
proteomic techniques for tissue samples remains proble-
matic and is probably 5 years from being a realistic
proposition. Currently, the development of large-scale
proteomic studies in radiotherapy trials is not realistic.
However, exploration of proteomic techniques in small-
scale studies would be of value.
Dr Tim Helliwell from Liverpool University discussed
the importance of tissue microarrays (TMAs) and their
ability to increase evaluation of multiple samples. With
TMAs, several hundred samples can be examined on a
single slide. There are several advantages of this
approach, although it must be seen as a population
screening tool and should not be used for individual
patient diagnosis. TMAs allow the relative frequency of a
target molecule to be assessed in relation to outcome. It
The British Journal of Radiology, September 2006
can be used for visualizing proteins (using specific
antibodies) or gene amplification/expression (using
fluorescent in situ hybridization). The approach is
limited by a potential sampling error (which can be
minimized by increasing the number of cores per
sample), the ability to select the correct probes and also
probe availability. However, with more probes becoming
available, there is much than can now be done. Again,
with the great increase in information there is a need for
bioinformatic analysis of the results. TMAs have great
potential for obtaining molecular markers associated
with prognosis in past, current and future trails.
Dr Francesca Buffa (Gray Cancer Institute,
Northwood) highlighted the problem of processing the
information that can be produced using the new
techniques. The analysis of a large number of genes/
proteins and a relatively small number of patients poses
several problems regarding variable selection, accuracy
of prediction on future data and interpretation of the
derived model. Analyses must account for multiple
testing and false positives are likely. There are different
approaches for handling the data. Supervised methods
classify according to a pre-selected clinical end-point –
such as locoregional control or disease-specific survival –
or biological function. Unsupervised techniques do not
require a pre-determined outcome or biological knowl-
edge but cluster genes and/or samples that are similar. A
knowledge-based approach can also be used, based on
our understanding of biology to look at the expression of
genes related to a particular phenotype, such as hypoxia.
She discussed the advantages and disadvantages of these
approaches from a statistical point of view and raised
issues that needed to be considered when designing a
trial so that data analysis can be more informative; such
as reproducibility and variability of the data collected
and standardization of laboratory procedures between
centres.
Dr Rob Bristow (Princess Margaret Hospital in
Toronto, Canada) emphasised the need for a focused
tumour type specific approach to the development of
prognostic and predictive biomarker profiles. He
described his work on prostate cancer and the evaluation
of markers related to radiobiologically relevant pheno-
types: hypoxia and activated DNA repair processes. A
considerable amount of work is clearly required in
evaluating markers in different tumours and those useful
in one disease site might have no relevance in another.
The potential was highlighted for using phase I or II trial
data to develop and explore mechanistic hypothesis
applicable to later phase trials, and to establish appro-
priate times for sampling. There is also a need to
evaluate markers not just in clinical trial material but
also in relation to the carcinogenic process of different
tumours. As some markers might prove to be targets for
novel drug development, there is a need to show their
expression in tumours compared with surrounding
normal tissue.
The importance of being aware of any underlying
clinical heterogeneity in biomarker studies was also
highlighted (Mr Priy Silva, The University of
Manchester). Cancers of the head and neck comprise a
heterogeneous group. The multiple sub-sites involved
are associated with differences in radioresponse. A study
in head and neck squamous cell carcinoma from patients
717

who underwent radical radiotherapy highlighted the
potential confounding influence of underlying clinical
heterogeneity. Even within a very clinically homoge-
neous, single sub-site group – oropharyngeal patients
who all received radiotherapy to their primary tumour –
there was variation in outcome probability in relation to
clinical and biomarker data. Differences were shown in
the behaviour of tonsil and base of tongue tumours. For
example, one marker provided highly significant prog-
nostic information in tonsil tumours and another in base
of tongue. Evaluation of potential biomarkers of
response must allow for any underlying clinical hetero-
geneity. Where possible, the powering of studies to
enable meaningful subgroup analyses would be useful.
The quality of samples and reproducibility of data
obtained in both single and multicentre studies is an
important consideration for the design of translational
research in clinical trials. The need to establish detailed
protocols for sample collection particularly in multi-
centre studies was emphasised (Dr Catharine West, The
University of Manchester). A two-centre study was
described involving the collection of head and neck
cancer samples during surgery for RNA profiling. All
samples taken in RNA later yielded high quality RNA,
with significantly greater variation between than within
samples. Using a knowledge-based approach for data
718
C West and S McKeown
analysis the derivation of a radiobiologically relevant
phenotype was described: a hypoxia metagene, which
was shown to yield prognostic information in indepen-
dent datasets.
The meeting provided no obvious quick-fit answers
for what endpoints or approaches to use in the planning
of translational aspects of clinical trials. Indeed all of the
methods have their advantages and disadvantages and,
although they can provide informative data, none can
provide all of the answers at present. What did emerge
from the meeting was a need for a multidisciplinary and
collaborative approach. The development of translational
research protocols will benefit from inclusion, at the
planning stage, of not just the clinical oncologists
involved in running a trial, but also scientists, statisti-
cians, bioinformaticians and pathologists. A strong
theme for all the various approaches described was the
old adage of ‘‘rubbish-in-rubbish-out’’. Prospective
sample collection is clearly important, but must be well
thought out with translational research requirements
discussed at an early stage in the clinical trial develop-
ment process. The current funding opportunities for
clinical trials and translational research should be
exploited by the Clinical Oncology and Radiobiology
community and, towards this goal, communication and
collaboration are essential.
The British Journal of Radiology, September 2006
The British Journal of Radiology, 79 (2006), 719–724

Oesophageal dysmotility in systemic sclerosis: comparison of

HRCT and scintigraphy
1
E H PITREZ, MD, 2M BREDEMEIER, MD, MSc, 2R M XAVIER, MD, PhD, 2K G CAPOBIANCO, MD, MSc,
2
V G RESTELLI, MD, 1M V VIEIRA, MD, PhD, 3D H C LUDWIG, MD, 2J C T BRENOL, MD, PhD,
1
A P A FURTADO, MD, 4L M B FONSECA, MD, PhD and 4B GUTFILEN, MD, PhD

1
Serviço de Radiologia, 2Serviço de Reumatologia and 3Serviço de Medicina Nuclear, Hospital de
Clı́nicas de Porto Alegre, Faculdade de Medicina, Universidade Federal do Rio Grande do Sul, Rua
Ramiro Barcelos 2350, Porto Alegre, RS, Brazil, 90.035-003 and 4Departamento de Radiologia,
Hospital Clementino Fraga Filho, Universidade Federal do Rio de Janeiro, Brigadeiro Trompovsky,
Ilha do Fundão, CEP 21941-590, Rio de Janeiro, RJ, Brazil

ABSTRACT. The aim of this study was to compare oesophageal abnormalities observed
in high-resolution CT with radionuclide transit in patients with systemic sclerosis. 76
patients with systemic sclerosis were evaluated by high-resolution CT and oesophageal
transit scintigraphy. Residual activity >20% (in relation to peak activity) at 15 s after
the beginning of the swallow of the labelled liquid (in supine position) was considered
indicative of oesophageal dysfunction. Supra-aortic and infra-aortic oesophageal
coronal diameters were measured in high-resolution CT. Oesophageal dilatation was
deemed present when the diameters exceeded 10 mm. 19 patients (25%) had supra-
aortic oesophageal dilatation and 48 patients (63.1%) had infra-aortic dilatation. The
prevalence of radionuclide transit delay was 77.6%. All patients (19/19) with supra-
aortic dilatation had oesophageal dysfunction, compared with 70.2% (40/57) of the
patients with no supra-aortic dilatation (p 5 0.004). Oesophageal dysfunction was
present in 97.9% (47/48) of patients with infra-aortic dilatation, compared with 42.9%
(12/28) in patients without it (p , 0.001). Receiver operating characteristic (ROC) curves
have demonstrated that the supra-aortic and infra-aortic diameters had good Received 7 November 2005
discriminatory capacity for oesophageal dysfunction in systemic sclerosis (area under Revised 4 January 2006
the curve, 95% confidence interval: 0.80, 0.70–0.89 and 0.92, 0.86–0.98, respectively). Accepted 30 January 2006
There is a clinically significant association between oesophageal dysmotility and high-
DOI: 10.1259/bjr/17000205
resolution CT findings of oesophageal coronal dilatation. The evaluation of infra-aortic
oesophageal coronal diameter can provide additional useful information about the ’ 2006 The British Institute of
functional and anatomic conditions of the oesophagus in systemic sclerosis. Radiology

Systemic sclerosis (SSc) is a connective tissue disease
characterized by fibrosis in multiple organs, especially in
the skin, lungs and gastrointestinal system [1, 2]. The
oesophagus is the most frequently involved internal
organ [3–5], and atrophy and fibrosis of the smooth
muscle are probably responsible for the impaired
oesophageal motility that is common in this disease [2,
5]. Diminished lower oesophageal sphincter pressure
and associated gastro-oesophageal reflux (GER) occur
frequently, leading to oesophagitis, strictures and ulcera-
tions [4, 6]. Early recognition of the oesophageal
dysfunction is important because patients may be
asymptomatic for a long time despite the presence of
oesophageal motor abnormalities [1, 3, 4].
Manometry has been considered the gold-standard
examination to detect oesophageal dysfunction in early
stages of SSc, but it gives no information on morphology and
has low patient acceptance [1–3]. Radionuclide oesophageal
scintigraphy(RES)standsoutasafrequentlyusedalternative
to manometry, since it is a safe, non-invasive, and sensitive
method that allows quantitative assessment [1, 2, 7–10].
Thorax high-resolution computed tomography
(HRCT) is the preferred radiological examination in the
evaluation of interstitial lung disease in SSc [11].
Oesophageal abnormalities in thorax CT, including
dilatation, air–fluid levels, and food retention have
already been documented in SSc patients [12]. As far as
we are aware, a comparison of the HRCT and RES results
in the evaluation of oesophageal dysfunction has not yet
been performed. Therefore, our objective in this study
was to compare both methods and to test the utility of
HRCT in the diagnosis of oesophageal dysfunction in SSc
using RES as the gold standard.
Material and methods
Patients
76 patients with the diagnosis of SSc (made according
to the clinical evaluation of experienced rheumatologists)
were prospectively studied. The patients met the

The British Journal of Radiology, September 2006 719


American College of Rheumatology criteria for SSc [13]
or the criteria suggested by LeRoy and Medsger for
diagnosis of early forms of SSc [14]. Patients with
overlapping syndromes, active chronic or acute infec-
tions, insulin-dependent diabetes mellitus or with long-
standing diabetes (more than 5 years since diagnosis)
were excluded. Patients with definite diagnosis of SSc
(according to the American College of Rheumatology
criteria) who developed inflammatory myopathy were
not excluded from the analysis. All patients signed
written informed consent before entry in the study. The
study was approved by the Research Ethics Committee
of the Hospital de Clı́nicas de Porto Alegre.
Clinical evaluation
All patients were interviewed and examined by a
rheumatologist (MB). The interview instrument was an
extensive questionnaire directed to the evaluation of
end-organ damage. Duration of disease was defined as
the period of time between the onset of Raynaud’s
phenomenon or skin symptoms (whatever came first)
and the moment of the interview [15].
The sample included 67 women and 9 men, and the
age range varied from 26 years to 78 years (mean
51 years, standard deviation 11 years). The duration of
symptoms ranged from 2–25 years (median 13 years).
Radionuclide oesophageal transit scintigraphy
All subjects were examined after 4 h of fasting. The
study was performed in supine position with liquid
ingestion. The liquid phase was prepared with 6 ml of
water labelled with 1 mCi of 99mtechnetium-phytate for
each swallow. Imaging was performed with a large field-
of-view gamma camera (high sensitivity parallel-hole
collimator, with the camera underneath. The field-of-
view extended from the pharynx to the upper part of the
stomach. A practice run with unlabelled water preceded
the study. The labelled liquid was placed in the mouth
and swallowed on command. The computer acquired a
dynamic study at a rate of 1 s frame21 for 60 s with a
64 6 64 matrix. No dry swallows were allowed. A region
of interest over the whole oesophagus was defined and a
time–activity curve was generated.
The percentage of emptying of the oesophagus was
analysed in 15 s and 60 s after swallow. Residual activity
>20% (in relation to peak activity) at 15 s after the
beginning of the swallow in the supine liquid phase was
considered abnormal and indicative of oesophageal
dysfunction [16].
High-resolution CT
The HRCT was performed with an Elscint Twin
scanner (Elscint, Haifa, Israel) and with a Somatom
Plus 4 (Siemens Medical Systems, Iselin, NJ). HRCT
scans were obtained in supine position during full
inspiration, with the following parameters being used:
1-mm sections at 10-mm intervals, a high-spatial-
frequency algorithm, 512 6 512 matrix, 140 kVp, and
720
E H Pitrez, M Bredemeier, R M Xavier et al
150 mAs (Elscint), 140 kVp and 130 mAs (Siemens). The
scanning time was 1.0 s and the examination was
performed from the apices of the lungs to the lung
bases. Prone scans were obtained whenever subpleural
lesions were observed. HRCT was reviewed with lung
(window width, 2000 H; window level, 2700 H) and
soft-tissue windows (window width, 400 H; window
level, 40 H).
The thoracic oesophagus was divided in supra-aortic
and infra-aortic levels (above and below the superior
limit of aortic arch, respectively). The largest coronal
diameters of the supra-aortic and infra-aortic levels
observed in the HRCT pictures were measured and
recorded, considering the internal limits of inner
oesophageal mucosa (Figure 1). Oesophageal dilatation
was diagnosed if the luminal coronal diameter of the
oesophagus exceeded 10 mm, whatever the content (air,
liquid or solid) [12].
Intraobserver agreement for the oesophageal coronal
diameter measurement in HRCT was tested in 30
examinations. The second reading was performed 12
months apart from the initial evaluation. There was a
good agreement between the two observations for supra-
aortic oesophageal coronal diameter (Fleiss weighted
Kappa 5 0.65) and a moderate agreement for infra-aortic
oesophageal coronal diameter (Fleiss weighted Kappa 5
0.56). Interobserver agreement for oesophageal coronal
diameter measurement in HRCT was tested using
another 30 examinations. The measurements obtained
by another radiologist demonstrated a good agreement
for the supra-aortic (Fleiss weighted Kappa 5 0.64) and
for the infra-aortic oesophageal coronal diameters (Fleiss
weighted Kappa 5 0.60).
All clinical, radiological, and scintigraphic examina-
tions were performed within 6 months. No observer was
informed about the clinical details of the SSc patients or
about the results of the other examinations.
Statistical analysis
Data were analysed using EPI-INFO version 6 and
SPSS for Windows version 11.0. Categorical variables
were presented as numbers and proportions.
Quantitative variables were tested graphically (with
Figure 1. A schematic drawing of how the coronal oeso-
phageal diameter was measured.
The British Journal of Radiology, September 2006
Oesophageal dysmotility in systemic sclerosis
normal probability plots) and statistically (with
Kolmogorov-Smirnov goodness-of-fit test) for the nor-
mality of distribution. Quantitative variable with normal
distribution was presented as mean and standard
deviation (SD). Non-normal quantitative variables were
presented as median and interquartile range (IQR). The
associations between categorical variables were tested
using Fisher’s exact test. Paired comparisons involving
non-normal quantitative variables were performed with
the Wilcoxon signed ranks test. A two-tailed p-value #
0.05 was considered statistically significant.
Receiver operating characteristic (ROC) curves were
used to test the ability of HRCT to differentiate patients
with and without oesophageal radionuclide transit
delay. The area under the ROC curve is a suitable
measure to summarize the discrimination power of a
diagnostic model (representing the accuracy of the
model) and can range from 0.5 (no discrimination) to
1.0 (perfect discrimination). 95% confidence intervals for
the areas under the curves were also calculated.
Sensitivity and specificity values were estimated, along
with 95% confidence intervals (95% CI).
Results
59 out of 76 patients (77.6%) had scintigraphic
evidence of oesophageal dysfunction. The prevalence of
20% or more retention of liquid in supine position at 60 s
(75%) was similar to the evaluation at 15 s.
The thorax HRCT demonstrated ground-glass opaci-
ties or reticular pattern/honeycombing in 43 patients
(56.6% of the sample). Out of these patients, 24 presented
ground-glass opacities and reticular pattern/honey-
combing, 6 presented reticular pattern/honeycombing
without ground-glass opacities, and 13 presented only
ground-glass opacities.
Supra-aortic oesophageal dilatation (Figure 2a) was
present in 19 patients (25.0%). Infra-aortic oesophageal
dilatation (Figure 2b) was detected in 48 patients (63.1%).
Supra-aortic and infra-aortic oesophageal dilatations in
HRCT were significantly associated with oesophageal
Figure 2. A 63-year-old man. High resolution CT demonstrates (a) supra-aortic oesophageal dilatation and (b) infra-aortic
dilatation (black arrows). Note peripheral ground-glass opacities (white arrowheads).
The British Journal of Radiology, September 2006
dysmotility (Tables 1 and 2, respectively). All oesopha-
geal wall calibres were smaller than 3 mm (Figure 3).
As the selection of cut-off values of coronal diameters
is rather arbitrary, we additionally used ROC curves to
test the ability of HRCT to discriminate patients with and
without oesophageal dysfunction on scintigraphy. The
areas under the ROC curves suggest that the oesopha-
geal coronal diameters (particularly the infra-aortic) have
good discriminatory ability for oesophageal dysmotility
(Figure 4). An infra-aortic oesophageal diameter >9 mm
would provide a sensitivity of 83.1% (95% CI: 71.0–91.5)
and a specificity of 94.1% (95% CI: 71.2–99.0) for
oesophageal dysmotility. Supra-aortic oesophageal dia-
meter >4 mm would provide a sensitivity of 64.4% (95%
CI: 50.9–76.4) and a specificity 94.1% (95% CI: 71.2–99.0)
for oesophageal dysmotility.
Discussion
SSc affects the skin in the form of thickening, tightness,
pitting oedema, digital ulcers and telangiectasias, but
also affects visceral organs like the lungs, the kidneys
and the gastrointestinal tract. Although skin changes are
the most remarkable abnormalities, the visceral disease
(especially pulmonary) is responsible for the prognosis
[12].
The gastrointestinal tract is involved in up to 80–90%
of SSc patients [12, 17–19], with the oesophagus being the
internal organ most frequently involved (50–80%) in this
disease [5, 6]. SSc leads to atrophy of smooth muscle, to
impaired peristalsis, and to reduction of the lower
oesophageal sphincter (LES) tone. Reduction in LES tone
leads to gastro-oesophageal reflux and oesophagitis [5–7,
17]. Symptoms include dysphagia, odynophagia, heart-
burn and regurgitation [5, 7, 12, 17], but the absence of
symptoms does not exclude advanced affliction of the
oesophagus [6, 19–21]. The lower two-thirds of the
oesophagus are usually affected, commonly appearing
dilated, with absent peristalsis, and sometimes with
strictures and mucosal changes due to oesophagitis
[4–7, 12].
721
 E H Pitrez, M Bredemeier, R M Xavier et al

Table 1. Association between supra-aortic coronal oesopha- Table 2. Association between infra-aortic coronal oesopha-
geal diameter and oesophageal dysmotility assessed by geal diameter and oesophageal dysmotility assessed by
radionuclide oesophageal scintigraphy* radionuclide oesophageal scintigraphy
Supra-aortic Oesophageal dysmotility p** Infra-aortic Oesophageal dysmotility p**
coronal coronal
oesophageal Yes No oesophageal Yes No
diameter n (%)* n (%)* diameter n (%)* n (%)*

>10 mm 19 (100) 0 (0) 0.004 > 10 mm 47 (97.9) 1 (2.1) ,0.001

,10 mm 40 (70.2) 17 (29.8) , 10 mm 12 (42.9) 16 (57.1)
*The percent values were calculated for the rows. ** Fisher’s *The percent values were calculated for the rows. **Fisher’s
exact test. exact test.

The sclerodermic oesophagus has been traditionally incomplete in individuals with normal manometry after
evaluated by different diagnostic methods, including multiple swallows. Therefore, a definite answer about
manometry [1, 4, 22, 23], endoscopy [5], scintigraphy [1, the ideal protocol to be used in RES has not yet been
3, 4, 7, 9, 10, 16, 24–27], barium-oesophagogram [5, 19, found.
28], cine-oesophagography [23, 29], ultrasound [29] and The lungs are the second most commonly affected
oesophageal pH monitoring [30]. Since the introduction internal organ in scleroderma. The patients often have
of the scintigraphic method in the oesophageal evalua- diffuse interstitial lung disease (fibrosing alveolitis),
tion by Kazem, in 1972 [25], several authors have which is clinically and prognostically relevant. Thorax
demonstrated that the radionuclide oesophageal transit HRCT is the established non-invasive gold-standard
examination is a safe and non-invasive method that technique for the diagnosis of fibrosing alveolitis [11].
could be used as an alternative to oesophageal mano- HRCT abnormalities in scleroderma lung disease are
metry in SSc patients [1, 3, 4, 24, 26]. Oesophageal transit well documented: ground-glass opacification in isolation
scintigraphy has been demonstrated to be sensitive or with reticular patterns are associated with inflamma-
(91–100%) and specific (88%) to detect motor function tory alveolitis and a reticular pattern – with or without
abnormalities diagnosed by manometry [1, 26, 31]. honeycomb changes – is associated with a fibrotic
Although scintigraphy has been considered a good histology [11, 12]. Further than diagnosis, patients with
method for the evaluation of oesophageal motility scleroderma are also evaluated with HRCT to assess the
disorders, there is not a uniformly accepted protocol to activity and progression of pulmonary disease.
guide the execution and the interpretation of the Thorax CT has been previously used to evaluate the
examination. This compromises the applicability and oesophagus in SSc patients. Bhalla et al [12] demon-
reproducibility of the results obtained by different strated oesophageal asymptomatic dilatation (12–
groups of researchers. To illustrate the variability in the 40 mm) in 20/25 patients (80%) with SSc. The authors
methods used, Akesson et al [3] applied a single swallow suggested that oesophageal evaluation by CT could be
protocol with the ingestion of a pineapple puree in used to narrow the differential diagnosis of patients with
sitting position, and oesophageal transit delay was interstitial lung disease. In the present study, infra-aortic
defined as a transit time longer than 300 s. Carette et al
and supra-aortic dilatation (>10 mm) were observed
[26] used a single swallow of 10 ml of labelled water in
in 63.1% and 25% of the cases, respectively. The
supine position, followed by a dry swallow 30 s later.
Davidson et al [24] used a supine colloid ingestion
protocol, defining RES abnormality as the retention of
colloid >5% at 15 s after ingestion. In the present study, in
which oesophageal transit was evaluated after a single
swallow of labelled liquid in supine position, the
prevalence of oesophageal dysfunction was 77.6%. In
previous studies, the reported prevalence of oesophageal
transit abnormalities on scintigraphy varied from 82% to
100% [1, 3, 7, 24, 26] what demonstrates a relative
homogeneity of results despite methodological variations.
In the present report we used an evaluation of
oesophageal transit after a single swallow. Previous
studies have suggested that multiple swallows may be
more appropriate in the evaluation of oesophageal
dysfunction, since the variability observed in single
swallow studies may be reduced [16]. However, as far
as we are concerned, the diagnostic accuracy of single
and multiple swallows (in relation to manometry) has
not been directly compared in previous studies. Ham
et al [32] suggested that the oesophageal transit patterns
observed after multiple swallowing were different from Figure 3. A 55-year-old woman. Infra-aortic slice demon-
those observed after a single swallow. Tolin et al [8] strates patulous thin-wall (white arrow) oesophageal
demonstrated that the oesophageal emptying may be dilatation.

722 The British Journal of Radiology, September 2006

Oesophageal dysmotility in systemic sclerosis
predominance of dilatation in the infra-aortic region
(lower two-thirds) has been largely demonstrated [3, 6, 7,
12] and is related to the greater amount of smooth
muscle fibres in the inferior portions of the oesophagus.
These fibres are affected by extensive fibrosis and
atrophy [7], leading to impaired peristalsis, dilatation,
pouches of air–liquid, and even to solid contents in the
oesophagus (Figure 5).
There are a few studies reporting supra-aortic oeso-
phageal dilatation in SSc patients [21, 33]. Proximal
oesophageal dilatation can be explained examining the
anatomic variability of the distribution of the smooth
musculature in the oesophageal wall. Post-mortem
studies have shown that the level of the transition zone
between striated and smooth muscle is quite variable,
sometimes occurring as proximally as the thoracic inlet
or even in the cervical oesophagus [33]. Therefore,
patients with a higher transition zone would probably
have a greater likelihood of presenting supra-aortic
dilatation.
As far as we are aware, there have been no previous
studies trying to relate the coronal oesophageal dia-
meters with the oesophageal dysfunction in SSc. Our
study demonstrated that oesophageal dilatation (parti-
cularly at the infra-aortic level) is highly associated to
radionuclide transit delay. Additionally, our results also
suggest that the infra-aortic coronal diameter may be
Figure 5. A 61-year-old woman. Dilated infra-aortic oeso-
phagus after overnight fasting demonstrates ingested con-
tents (black arrow).
The British Journal of Radiology, September 2006
Figure 4. Receiver operating char-
acteristic curves of (a) supra-aortic
and (b) infra-aortic coronal dia-
meters for oesophageal dysfunction
detected in scintigraphy.
useful in the diagnosis of oesophageal dysmotility, as
suggested by the ROC curve analysis. Therefore, besides
the routine evaluation of interstitial lung disease on
HRCT, the measurement of coronal oesophageal dia-
meters could provide useful and easily attainable
information about the oesophageal motility status in SSc.
In the present study we had no control group or
patients with other causes of oesophageal dysfunction to
compare with SSc patients. Therefore, our results apply
exclusively to patients with SSc. It would be interest-
ing to compare HRCT and RES in other diseases
characterized by oesophageal dysmotility or structural
abnormalities.
In no patient of this study the oesophageal wall
thickness exceeded the reported normal value of 3 mm.
This has been also described in a previous report [34]
and probably reflects that, although histopathological
oesophageal abnormalities (fibrosis and atrophy) may be
striking, these findings do not correlate with oesophageal
wall thickening, not even in patients with long lasting
disease.
Considering the results presented here, we conclude
that the measurement of oesophageal coronal diameters
in thorax HRCT (directed primarily to the evaluation of
interstitial lung disease) provide useful information
about the functional status of the oesophagus in SSc
patients.
Acknowledgments
We acknowledge Drs Raquel Faccioni, Marcus Franck,
Tatiana Freitas Tourinho, Marcelo Maltchick, Paulo
Sérgio Thys, Patrı́cia Minuzzi da Motta, Max Brenner,
Tamara Mucenic, Adriano Barbiero, Tatiana Karenini
Müller, Carmen Both Schenatto, Charles Lubianca
Kohem, Lilian Scussel-Lonzetti, Claiton Viegas Brenol,
Sandra Helena Machado, and Ilóite Scheibel on their
valuable support. We also thank to Juliana Bredemeier
for the English review of this manuscript.
References
1. Drane WE, Karvelis K, Johnson DA, Curran JJ, Silverman
ED. Progressive systemic sclerosis: radionuclide esophageal
scintigraphy and manometry. Radiology 1986;160:73–6.
2. Kinuya K, Nakajima K, Kinuya S, Michigishi T, Tonami N,
Takehara K. Esophageal hypomotility in systemic sclerosis:
723

close relationship with pulmonary involvement. Ann Nucl
Med 2001;15:97–101.
3. Akesson A, Gustafson T, Wollheim F, Brismar J. Esophageal
dysfunction and radionuclide transit in progressive sys-
temic sclerosis. Scand J Rheumatol 1987;16:291–9.
4. Klein HA, Wald A, Graham TO, Campbell WL, Steen VD.
Comparative studies of esophageal function in systemic
sclerosis. Gastroenterology 1992;102:1551–6.
5. Barros PDS, Barcelos IK, Carvalho ACO. Acometimento do
trato digestivo na esclerose sistêmica. Rev Bras Reumatol
1999;102:81–6.
6. Lock G, Pfeifer M, Straub RH, Zeuner M, Lang B,
Scholmerich J, et al. Association of esophageal dysfunction
and pulmonary function impairment in systemic sclerosis.
Am J Gastroenterol 1998;93:341–5.
7. Kaye SA, Siraj QH, Agnew J, Hilson A, Black CM. Detection
of early asymptomatic esophageal dysfunction in systemic
sclerosis using a new scintigraphic grading method. J
Rheumatol 1996;23:297–301.
8. Tolin RD, Malmud LS, Reilley J, Fisher RS. Esophageal
scintigraphy to quantitate esophageal transit (quantitation
of esophageal transit). Gastroenterology 1979;76:1402–8.
9. Mariani G, Boni G, Barreca M, Bellini M, Fattori B, AlSharif
A, et al. Radionuclide gastroesophageal motor studies. J
Nucl Med 2004;45:1004–28.
10. Fonseca L, Maliska CM, Cruz MG, Castro L, Gutfilen B.
Esophageal reconstruction surgery in oncologic patients:
determination of gastric emptying time. J Exp Clin Cancer
Res 2000;19:137–40.
11. Latsi PI, Wells AU. Evaluation and management of
alveolitis and interstitial lung disease in scleroderma.
Curr Opin Rheumatol 2003;15:748–55.
12. Bhalla M, Silver RM, Shepard JA, McLoud TC. Chest CT in
patients with scleroderma: prevalence of asymptomatic
esophageal dilatation and mediastinal lymphadenopathy.
AJR Am J Roentgenol 1993;161:269–72.
13. Committee SfSCotARADaTC. Preliminary criteria for the
classification of systemic sclerosis (scleroderma). Arthritis
Rheum 1980;23:581–90.
14. LeRoy EC, Black C, Fleischmajer R, Jablonska S, Krieg T,
Medsger TA Jr, et al. Scleroderma (systemic sclerosis):
classification, subsets and pathogenesis. J Rheumatol
1988;15:202–5.
15. Bredemeier M, Xavier RM, Capobianco KG, Restelli VG,
Rohde LE, Pinotti AF, et al. Nailfold capillary microscopy
can suggest pulmonary disease activity in systemic sclero-
sis. J Rheumatol 2004;31:286–94.
16. Tatsch K, Schroettle W, Kirsch CM. Multiple swallow test
for the quantitative and qualitative evaluation of esopha-
geal motility disorders. J Nucl Med 1991;32:1365–70.
17. Rose S, Young MA, Reynolds JC. Gastrointestinal manifes-
tations of scleroderma. Gastroenterol Clin North Am
1998;27:563–94.
18. D’Angelo WA, Fries JF, Masi AT, Shulman LE. Pathologic
observations in systemic sclerosis (scleroderma). A study of
724
E H Pitrez, M Bredemeier, R M Xavier et al
fifty-eight autopsy cases and fifty-eight matched controls.
Am J Med 1969;46:428–40.
19. Poirier TJ, Rankin GB. Gastrointestinal manifestations of
progressive systemic scleroderma based on a review of 364
cases. Am J Gastroenterol 1972;58:30–44.
20. Akesson A, Wollheim FA. Organ manifestations in 100
patients with progressive systemic sclerosis: a comparison
between the CREST syndrome and diffuse scleroderma. Br J
Rheumatol 1989;28:281–6.
21. Turner R, Lipshutz W, Miller W, Rittenberg G, Schumacher
HR, Cohen S. Esophageal dysfunction in collagen disease.
Am J Med Sci 1973;265:191–9.
22. Henry MA, Harbermann MC, Rocha OM. Esophageal
motor disturbances in progressive systemic sclerosis. Dis
Esophagus 1999;12:51–3.
23. Blom-Bulow B, Sundstrom G, Jonson B, Tylen U, Wollheim
FA. Early changes in oesophageal function in progressive
systemic sclerosis: a comparison of manometry and radi-
ology. Clin Physiol 1984;4:147–58.
24. Davidson A, Russell C, Littlejohn GO. Assessment of
esophageal abnormalities in progressive systemic sclerosis
using radionuclide transit. J Rheumatol 1985;12:472–7.
25. Kazem I. A new scintigraphic technique for the study of the
esophagus. Am J Roentgenol Radium Ther Nucl Med
1972;115:681–8.
26. Carette S, Lacourciere Y, Lavoie S, Halle P. Radionuclide
esophageal transit in progressive systemic sclerosis. J
Rheumatol 1985;12:478–81.
27. Tatsch K, Voderholzer WA, Weiss MJ, Schrottle W, Hahn K.
Reappraisal of quantitative esophageal scintigraphy by
optimizing results with ROC analyses. J Nucl Med
1996;37:1799–805.
28. Coggins CA, Levine MS, Kesack CD, Katzka DA. Wide-
mouthed sacculations in the esophagus: a radiographic
finding in scleroderma. AJR Am J Roentgenol
2001;176:953–4.
29. Miller LS, Liu JB, Klenn PJ, Holahan MP, Varga J, Feld RI,
et al. Endoluminal ultrasonography of the distal esophagus
in systemic sclerosis. Gastroenterology 1993;105:31–9.
30. Shoenut JP, Wieler JA, Micflikier AB. The extent and
pattern of gastro-oesophageal reflux in patients with
scleroderma oesophagus: the effect of low-dose omepra-
zole. Aliment Pharmacol Ther 1993;7:509–13.
31. Lally EV, Jimenez SA, Kaplan SR. Progressive systemic
sclerosis: mode of presentation, rapidly progressive disease
course, and mortality based on an analysis of 91 patients.
Semin Arthritis Rheum 1988;18:1–13.
32. Ham HR, Georges B, Froideville JL, Piepsz A. Oesophageal
transit of liquid: effects of single or multiple swallows. Nucl
Med Commun 1985;6:263–7.
33. Meyer GW, Austin RM, Brady CE 3rd, Castell DO. Muscle
anatomy of the human esophagus. J Clin Gastroenterol
1986;8:131–4.
34. Stanley RJ, Lee JKT, Sagel SS. Computed tomography of the
body with MRI correlation. New York, NY: Raven, 1989.
The British Journal of Radiology, September 2006
The British Journal of Radiology, 79 (2006), 725–729

Relationship between the growth pattern of nasopharyngeal

cancer and the cervical lymph nodes based on MRI findings: can
the cervical radiation field be reduced in patients with
nasopharyngeal cancer?
1
N FUWA, MD, PhD, 2Y ARIJI, DDS, PhD, 3T DAIMON, PhD, 4M WAKISAKA, MD, PhD, 4A MATSUMOTO, MD,
1
T KODAIRA, MD, PhD, 1H TACHIBANA, MD, 1T NAKAMUA, MD, PhD and 1Y SATOU, MD, PhD

1
Department of Radiation Oncology, Aichi Cancer Center Hospital, 1-1 Kanokoden, Chikusaku,
Nagoya 464-8681, 2Department of Oral and Maxillofacial Radiology, Aichi-Gakuin University School
of Dentistry, 2-11 Suemori-dori, Chikusaku, Nagoya 464-8651, 3Department of Clinical Research and
Management, Translational Research Information Center, Foundation for Biomedical Research and
Innovation, 1-5-4 Minatojima-nakamachi, Chuo-ku, Kobe 650-0047 and 4Department of Radiology,
Oita Medical University, 1-1 Idaigaoka, Hasama-machi, Oita 879-5593, Japan

ABSTRACT. To identify patients with nasopharyngeal cancer in whom the cervical

radiation field can be reduced, we classified the growth patterns of nasopharyngeal
cancer based on MRI findings into 4 types and performed an evaluation. Based on MRI
findings, we classified the growth patterns of primary cancer in 94 patients with
nasopharyngeal cancer into Type 1 (superficial type), Type 2 (lateral invasive type), Type
3 (upward invasive type), and Type 4 (anterior extension type), and further classified
Type 2, based upon nasopharyngoscopic findings, into Type 2a (unilateral invasive
type) and Type 2b (bilateral invasive type). The cervical lymph node metastasis areas
were evaluated according to these types. Type 2 showed a significantly higher Received 7 March 2005
incidence of cervical lymph node metastasis only on the ipsilateral side than the other Revised 22 March 2006
types (p50.0024). In particular, all patients with Type 2a had cervical lymph node Accepted 27 March 2006
metastasis only on the ipsilateral side (p50.0212). This study suggests that the
DOI: 10.1259/bjr/27870658
distribution of metastasised cervical lymph nodes depends on the pattern of tumour
extent of the primary site. ’ 2006 The British Institute of
Radiology

Cervical lymph node metastasis often occurs in
patients with nasopharyngeal cancer (NPC). It has been
reported that 60–87% of patients demonstrated cervical
lymphadenopathy at the time of diagnosis [1–5].
Previous articles [6–8] have attributed bilateral cervical
lymphadenopathy to the abundance of lymphatic tissues
in the posterior wall of the nasopharynx and the
presence of abundant lymphatic anastomoses crossing
the midline. Therefore, bilateral cervical lymph nodes
have always been included in the target volume on
radical radiotherapy (RT) [6, 7]. As a result, the radiation
fields were wide and complications such as the xerosto-
mia secondary to parotid irradiation have been experi-
enced [9–14].
Recently, MRI has been used to evaluate NPC [15–22].
Multiplanar imaging can show tumour extension in all
planes, and there is improved tumour delineation and
identification of the spread of the tumour to adjacent
areas as a result of the better soft tissue contrast in
comparison with CT.
We previously evaluated the relationship between
tumour growth patterns and lymph node metastasis
based on MRI images in 32 patients with nasopharyngeal
cancer. We showed a high incidence of cervical lymph
node metastasis only on the ipsilateral side in patients
with tumour invasion only to the lateral nasopharynx
(lateral invasive type), and the cervical radiation field
can be reduced in such patients [23].
In this study, to identify patients in whom the cervical
radiation field can be reduced, the relationship between
the growth pattern of nasopharyngeal cancer and lymph
node metastasis on MR images was evaluated in detail in
94 patients, obtained by adding 62 patients to the above
32 patients.
Methods and materials
Patients
Between April 1990 and August 2004, 94 patients (70
males and 24 females) were retrospectively reviewed.
The patient age ranged from 14 years to 80 years (median
51 years). Histology revealed that 17 cases of nasophar-
yngeal cancer were WHO type I, 15 were type II, and 62
were type III. According to the 1997 TNM classification,
the tumour stage was classified as stage I in 3 patients,
stage II A in 1 patient, stage II B in 31 patients, stage III in

The British Journal of Radiology, September 2006 725

 N Fuwa, Y Ariji, T Daimon et al

22 patients, stage IV A in 11 patients, stage IV B in 20

patients and stage IV C in 6 patients.
As a basic treatment method, patients aged # 70 years
were treated with alternating chemoradiotherapy, in
which chemotherapy (CDDP, 5FU) is alternated with
radiotherapy [24], and those aged > 71 years were
treated by radiotherapy alone.

Evaluation items
The evaluation items were the possible relationships
between the growth pattern as well as the size of
nasopharyngeal tumour and cervical lymph node metas-
tasis, between the degree of histological differentiation
and cervical lymph node metastasis, and between the
tumour growth pattern and the degree of histological
differentiation. Figure 1. A 31-year-old woman presenting with Type 1
Differences were analysed by the log-linear model (superficial type) spread. Axial T1 weighted MR image, Gd-
without interaction terms using the contingency table DTPA enhanced, shows no abnormal findings in the
data [25]. nasopharyngeal mucosal space.

MRI techniques
MR studies were performed with a 1.5 T unit (Signa;
General Electric Medical Systems, Milwaukee, WI).
Images were obtained with 5 mm thick contiguous
sections in two or three planes (axial, coronal, sagittal)
depending on the extent of the tumour. All patients
underwent both plain and contrast-enhanced MRI. A
spin-echo (SE) multisection imaging technique was used
in all examinations. A SE 600/25 image and a SE 2000/80
image were considered to be T1 weighted and T2
weighted, respectively.

Classifications of MR image
To determine the pattern of tumour extent at the
primary site and existence of metastasised cervical
lymph nodes, MR images were evaluated by three
experienced radiologists specializing in head and neck
cancers working together as a team. When they did not
reach consensus after an initial reading regarding the
primary tumour and the cervical lymphadenopathy, the
final decision was made by majority rule.
The 94 patients were classified into 4 types with 2
subtypes (Figures 1–4) according to the tumour extent at
the primary site demonstrated on MR images before the
treatment. Type 1 (superficial type) tumours were
limited to the nasopharyngeal mucosal space. Tumour
locations could not be judged on MRI, but were assessed
by nasopharyngoscopy from the appearance of slightly
irregular surface of nasopharyngeal mucosa. Type 2
(lateral invasive type) tumours could be detected on the
lateral pharyngeal wall with involvement of the para-
pharyngeal space laterally. This invasive type was
divided into two subtypes. Type 2a (unilateral invasive
type) tumours were those that remained on either side of
the midline by MRI and did not extend the midline of the
posterior wall of nasopharynx assessed by nasophar-
yngoscopy. Type 2b (bilateral invasive type) tumours
were those that extended a cross the midline by MRI
Figure 2. A 64-year-old man presenting with Type 2a
(unilateral invasive type) spread. Axial T1 weighted MR
image, Gd-DTPA enhanced, shows a right nasopharyngeal
tumour infiltrating the right parapharyngeal space.
and/or nasopharyngoscopy. Type 3 (upward invasive
type) were tumours that mainly invaded cranially
toward the skull base. Type 4 (anterior extension type)
tumours extended anteriorly toward the nasal cavity, but
did not invade the adjacent normal structures. Tumours
that could not be classified into the above 4 types were
considered to be unclassifiable.
Classification of tumour size
All 94 patients were classified into three groups
according to the maximal tumour diameter of axial
imaging: small (S#2 cm), medium (2 cm,M,4 cm) and
large (4 cm #L) tumours. Type 1 tumours that cannot be
measured on MRI images were classified as small.

726 The British Journal of Radiology, September 2006

MR analysis of nasopharyngeal cancer
Figure 3. A 35-year-old man presenting with Type 2b
(bilateral invasive type) spread. Axial T1 weighted MR image,
Gd-DTPA enhanced, shows a left a nasopharyngeal tumour
(arrowheads) invading across the midline to the opposite
side.
Classification of pathology
The 94 patients were divided into two groups
according to the degree of tumour differentiation, those
with lymphoepithelioma, undifferentiated carcinoma
and poorly differentiated squamous cell carcinoma as
Group 1, and those with moderately and well differ-
entiated squamous cell carcinoma as Group 2.
Definition of metastatic lymph nodes
Metastatic cervical lymph nodes were defined as
nodes with a minimal axial diameter of 10 mm in a
(a)
Figure 4. (a) A 60-year-old man presenting with Type 3 (upward invasive type) spread. Axial T1 weighted MR image, Gd-DTPA
enhanced, shows a nasopharyngeal tumour (arrowheads) which is invading toward the left skull. (b) Coronal T1 weighted MR
image, Gd-DTPA enhanced, shows a nasopharyngeal tumour (arrowheads) which is invading mainly toward the left skull base.
The British Journal of Radiology, September 2006
previous study [26]. In this study, in addition to lymph
nodes fulfilling this criterion, the following lymph nodes
were also regarded as metastasis even when the minimal
axis diameter was # 10 mm: necrotic lymph nodes with
a visualized capsule and lymph nodes showing a definite
decrease in size on MR images after treatment.
Results
Of the 94 patients, 19 (20%) had type 1, 12 (13%) had
type 2a, 18 (19%) had type 2b, 14 (15%) had type 3, 29
(31%) had type 4 and 2 (2%) were unclassified. 82 (87%)
out of 94 patients demonstrated unilateral and/or
bilateral cervical lymphadenopathy. 41 (50%) out of 82
patients presented bilateral cervical and/or contralateral
cervical lymphadenopathy.
The relationship between the pattern of tumour extent
at the primary site and the cervical lymph node
metastasis is shown in Table 1. The incidence of
ipsilateral cervical lymph node metastasis was signifi-
cantly higher for Type 2 than for the other types
(p50.0024). In particular, all patients with Type 2a had
cervical lymph node metastasis only on the ipsilateral
side (p50.0148). The incidence of bilateral cervical lymph
node metastases was high for Types 1 (p50.0393) and 4
(p50.0482).
Table 2 shows the size of nasopharyngeal tumour and
areas of cervical lymph node metastasis. No differences
were observed in areas of cervical lymph node metastasis
among these groups, but the incidence of lymph node
metastasis was lower in large tumour groups (p50.0435).
Table 3 shows areas and incidences of cervical lymph
node metastasis according to the degrees of histological
differentiation. No differences were observed in areas of
cervical lymph node metastasis between Group 1
(differentiated type) and Group 2 (poorly differentiated
type), but the incidence of lymph node metastasis was
lower in Group 1 (p50.0008).
(b)
727
 N Fuwa, Y Ariji, T Daimon et al

Table 1. Relationship between the pattern of tumour extent at the primary site and areas of cervical lymph node metastasis
Type 1 Type 2a Type 2b Type 3 Type 4
n519 n512 n518 n514 n529

No lymph node metastasis (N0) 2 1 2 3 3

Ipsilateral lymph node metastasis 5 11 11 5 9
Bilateral and/or contralateral lymph node metastasis 12 0 5 6 (1) 17
Unclassified type (2 cases) was excluded from Table 1.
( ): Number of contralateral neck lymph node metastasis.

Table 2. Relationship between the size of nasopharyngeal Table 3. Relationship between the degrees of histological
tumour and areas of cervical lymph node metastasis differentiation and of cervical lymph node metastasis
Small Middle Large Group 1 n518 Group 2 n576
n519 n541 n534
(S#2 cm) (2 cm, (4 cm#L) No lymph node 7 5
M,4 cm) metastasis
Ipsilateral lymph 8 33
No lymph node 2 3 7(1) node metastasis
metastasis Bilateral and/or 3 38 (1)
Ipsilateral lymph node 5 22 14 contralateral lymph
metastasis node metastasis
Bilateral and/or 12 16 13
contralateral lymph Group 1: differentiated squamous cell carcinoma.
node metastasis Group 2: undifferentiated carcinoma, poorly differentiated
squamous cell carcinoma.
( ): Number of contralateral neck lymph node metastasis. ( ): Number of contralateral neck lymph node metastasis.

Table 4 shows the degree of histological differentiation
and the growth pattern of nasopharyngeal tumour. No
association was observed between Type 1 or 2 and the
degree of histological differentiation. However, patients
with Type 3 (upward invasive type) were frequently
included in Group 1 (p50.0034), and all patients with
Type 4 tumours were included in Group 2 (p50.0383).
Discussion
In the 32 patients in our previous study, tumour
growth patterns were classified into 3 major types (Type
1, superficial type: Type 2, invasive type; and Type 3,
anterior extension type), and Type 2 was further
classified into 3 subtypes (Type 2a, unilateral invasive
type; Type 2b, bilateral invasive type; and Type 2c,
upward invasive type) [23]. In the previous study, the
lymph node metastasis pattern was similar between
Types 2a and 2b, but Type 2c showed a different pattern.
Therefore, in this study, Type 2 was defined as the lateral
invasive type and classified into Type 2a (unilateral
invasive type) and Type 2b (bilateral invasive type),
while Type 2c was separated from Type 2 and changed
to Type 3 (upward invasive type). In the previous study,
Type 2a was differentiated from Type 2b based only on
MRI findings. However, in this study, nasopharyngo-
scopic findings were also used, and Type 2a was defined
as tumour extension not passing the midline of the
nasopharynx.
As Table 1 shows, the incidence of cervical lymph
node metastasis only on the ipsilateral side was
significantly higher for Type 2 than for the other types
(p50.0024). In particular, all patients with Type 2a
showed cervical lymph node metastasis only on the
ipsilateral side, which suggested that the cervical
radiation field can be reduced for this type. Types 1
(superficial type) and 4 (anterior extension type) showed
similar lymph node metastasis patterns and higher
incidences of bilateral cervical lymph node metastasis
than the other types, which suggested that a reduction in
the radiation field is difficult for these types.
No significant association was observed between
tumour size and cervical lymph node metastasis.
However, bilateral cervical lymph node metastasis was
frequently observed in patients with small tumours
(S#2 cm) rather than patients with large tumours
(p50.0272).
Concerning the degree of histological differentiation
and cervical lymph node metastasis, the percentage of
patients with no lymph node metastasis (N0) was
significantly higher in Group 1 (differentiated type) than
in Group 2 (poorly differentiated or undifferentiated
type) (p50.0008). In Group 1, N0 was observed in 7
patients, of whom 3 showed Type 3 on MRI images. As
Table 4 shows, the patients with Type 3 included 7
patients with the differentiated type. Of the 7 patients, 3
had no cervical lymph node metastasis, suggesting that
the incidence of cervical lymph node metastasis is low in
patients with Type 3 and the differentiated type,
compared with those with the other types.
Table 4. Relationship between the pattern of tumour
extent at the primary site and the degrees of histological
differentiation
Type1 Type 2a Type 2b Type 3 Type 4
n519 n512 n518 n514 n529
Group 1 4 2 4 7 0
Group 2 15 10 14 7 29
Group 1: differentiated squamous cell carcinoma.
Group 2: undifferentiated carcinoma, poorly differentiated
squamous cell carcinoma.
Unclassified type (2 cases) was excluded from Table 4.

728 The British Journal of Radiology, September 2006

MR analysis of nasopharyngeal cancer
Figure 5. A 40-year-old woman presenting with Type 4
(anterior extension type) spread. Axial T1 weighted MR
image, Gd-DTPA enhanced, shows a nasopharyngeal tumour
(arrowheads) extending anteriorly without invasion to
adjacent areas.
When the growth pattern of nasopharyngeal cancer
was classified and its association with cervical lymph
node metastasis was evaluated, the incidence of bilateral
cervical lymph node metastasis was high for Types 1 and
4, that of ipsilateral metastasis was high for Type 2, and
intermediate findings were obtained for Type 3. The
incidence of cervical lymph node metastasis was
significantly low in patients with differentiated tumours.
These results suggest that the radiation field and dose
of the neck for nasopharyngeal cancer can be individua-
lized. The radiation field and dose of the neck may be
reduced especially for patients with Type 2a. This
hypothesis will be confirmed by prospective study.
References
1. Lindberg R. Distribution of cervical lymph node metastases
from squamous cell carcinoma of the upper respiratory and
digestive tracts. Cancer 1972;29:1446–9.
2. Mesic JB, Fletcher GH, Goepfert H. Megavoltage irradiation
of epithelial tumors of the nasopharynx. Int J Radiat Oncol
Biol Phys 1981;7:447–53.
3. Bedwinek JM, Perez CA, Keys DJ. Analysis of failures after
definitive irradiation for epidermoid carcinoma of the
nasopharynx. Cancer 1980;45:2725–9.
4. Khoury GG, Paterson ICM. Nasopharyngeal carcinoma: a
review of cases treated by radiotherapy and chemotherapy.
Clin Radiol 1987;38:17–20.
5. Wang CC. Treatment of malignant tumors of the nasophar-
ynx. Otolaryngol Clin North 1980;13:477–81.
6. Fletcher GH, Million RR. Nasopharynx. In: Fletcher GH,
editor. Textbook of radiotherapy, 3rd edn. Philadelphia,
PA: Lea & Febiger, 1980:364–83.
7. Moss WT. The nasopharynx. In: Moss WT, Cox JD, editors.
Radiation oncology, 6th edn. St Louis, MO: The C.V. Mosby
Company, 1989:198–214.
The British Journal of Radiology, September 2006
8. Perez CA. Nasopharynx. In: Perez CA, Brady LW, editors.
Principles and practice of radiation oncology, 2nd edn.
Philadelphia, PA: J.B. Lippincott Company, 1992:617–43.
9. Lee AW, Law SC, Ng SH, Chan DK, Poon YF, Foo W, et al.
Retrospective analysis of nasopharyngeal carcinoma treated
during 1976–1985: late complications following megavol-
tage irradiation. Br J Radiol 1992;65:918–28.
10. Zimmerman RP, Mark RJ, Tran LM, Juillard GF.
Concomitant pilocarpine during head and neck irradiation
is associated with decreased posttreatment xerostomia. Int J
Radiat Oncol Biol Phys 1997;37:571–5.
11. Tokars RP, Griem ML. Carcinoma of the nasopharynx: an
optimization of radiotherapeutic management for tumor
control and spinal cord injury. Int J Radiat Oncol Biol Phys
1979;5:1741–8.
12. Hoppe RT, Goffinet DR, Bagshaw MA. Carcinoma of the
nasopharynx: eighteen years’ experience with megavoltage
radiation therapy. Cancer 1976;37:2605–12.
13. Bedwinek JM, Perez CA, Keys DJ. Analysis of failures after
definitive irradiation for epidermoid carcinoma of the
nasopharynx. Cancer 1980;45:2725–9.
14. Nishioka T, Shirato H, Arimoto T, Kaneko M, Kitahara T,
Oomori K, et al. Reduction of radiation-induced xerostomia
in nasopharyngeal carcinoma using CT simulation with
laser patient marking and three-field irradiation technique.
Int J Radiat Oncol Biol Phys 1997;38:705–12.
15. Teresi LM, Lufkin RB, Hanafee WN. Nasopharynx, oro-
pharynx, and tongue base. In: Stark DD, Bradley WG Jr,
editors. Magnetic resonance imaging, 2nd edn. St Louis,
MO: Mosby-Year Book, 1992:1135–63.
16. Vogl T, Dresel S, Bilaniuk LT, Grevers G, Kang K, Lissner J.
Tumors of the nasopharynx and adjacent areas: MR imaging
with Gd-DTPA. AJR Am J Roentgenol 1990;154:585–92.
17. Hudgins PA, Gussack GS. MR imaging in the management
of extracranial malignant tumors of the head and neck. AJR
Am J Roentgenol 1992;159:161–9.
18. Olmi P, Fallai C, Colagrande S, Giannardi G. Staging and
follow-up of nasopharyngeal carcinoma: magnetic reso-
nance imaging versus computerized tomography. Int J
Radiat Oncol Biol Phys 1995;32:795–800.
19. Naito Y, Honjo I, Nishimura K, Torizuka K. Magnetic
resonance imaging around the eustachian tube. Am J
Otolaryngol 1986;7:402–6.
20. Chong VFH, Fan YF. Detection of recurrent nasopharyngeal
carcinoma: MR imaging versus CT. Radiology
1997;202:463–70.
21. Som PM. Detection of metastasis in cervical lymph nodes:
CT and MR criteria and differential diagnosis. AJR Am J
Roentgenol 1992;158:961–9.
22. Ng SH, Chang TC, Ko SF, Yen PS, Wan YL, Tang LM.
Nasopharyngeal carcinoma: MRI and CT assessment.
Neuroradiology 1997;39:741–6.
23. Wakisaka M, Mori H, Fuwa N, Matsumoto A. MR analysis
of nasopharyngeal carcinoma: correlation of the pattern of
tumor extent at the primary site with the distribution of
metastasized cervical lymph nodes. Preliminary results. Eur
Radiol 2000;10:970–7.
24. Fuwa N, Kano M, Toita T, Shikama N, Kodaira T,
Matsumoto A, et al. Alternating chemoradiotherapy for
nasopharyngeal cancer using cisplatin and 5-fluorouracil - a
preliminary report of Phase II Study. Radiother Oncol
2001;61:257–60.
25. Alan A. Categorial data analysis, 2nd edn. New York, NY:
Wiley, 2000.
26. van den Brekel MW, Stel HV, Castelijns JA, Nauta JJ, van
der Waal I, Valk J, et al. Cervical lymph node metastasis:
assessment of radiologic criteria. Radiology
1990;177:379–84.
729
 The British Journal of Radiology, 79 (2006), 730–733

Six years experience in intracoronary brachytherapy procedures:

patient doses from fluoroscopy
1
C PRIETO, BSc, 1,4E VANO, PhD, 1,4J M FERNÁNDEZ, BSc,
2,4
C GALVAN, MD, PhD,
3
M SABATE, MD, PhD,
4
L GONZALEZ, PhD and 1D MARTINEZ, BSc

1
Medical Physics Service, 2Radiotherapy Service and 3Interventional Cardiology Service, San Carlos
University Hospital, 28040 Madrid and 4Radiology Department, Medicine School, Complutense
University, 28040 Madrid, Spain

ABSTRACT. Typical patient dose levels during intracoronary brachytherapy (ICB)

procedures using beta sources were determined across a sample of 221 treatments.
Dose–area product values, fluoroscopy time and number of frames per procedure, with
median values of 62 Gy cm2, 17.0 min and 1493 images, respectively, resulted in a 20%
to 50% increase in the values measured for percutaneous transluminal coronary
angioplasty procedures in the same medical centre (median values 41 Gy cm2, 14.3 min
and 1078 images). Likely reasons for this increase include the additional complexity of Received 18 August 2005
ICB, the need for recording and reporting every step of the treatment, getting the Revised 17 October 2005
essential parameters for the volume determination of the lesion and therapeutic Accepted 31 October 2005
radiation dose calculation and, finally, the learning curve for this kind of procedure. A
DOI: 10.1259/bjr/75766147
high concentration skin dose distribution during ICB procedures was measured and in
12% of the patients peak skin doses higher than 1.5 Gy were confirmed. 10 patients ’ 2006 The British Institute of
were submitted to clinical follow-up and skin injuries were not identified. Radiology

Restenosis, or re-narrowing of a coronary artery after
dilatation, is the main limitation of percutaneous
transluminal coronary angioplasty (PTCA). Stents have
substantially reduced the incidence of restenosis, but
neo-intimal proliferation within the stent is still a great
problem for a significant percentage of patients.
Intracoronary brachytherapy (ICB) is indicated in
patients with in-stent restenosis [1].
ICB includes delivery of a local radiation dose to the
artery wall (target volume) after stenting. There are
several potential problems in the safety aspects for the
practice of ICB procedures: (a) procedures more complex
than standard PTCA procedures in catheterization
laboratories; (b) introduction of brachytherapy equip-
ment not previously used, resulting in a new challenge in
radiological protection; and (c) coordination among
different specialists (cardiologists, radiotherapists and
medical physicists working in the catheterization labora-
tory). These aspects affect safety of both the patient and
staff involved.
Complex interventional cardiology procedures can
produce deterministic effects (skin injuries) due to the
high radiation doses imparted to some regions of the
patient’s skin [2–4]. PTCA is one of the most frequent
interventional procedures in cardiology, and sometimes
requires long fluoroscopy times and a large number of
cine frames to document the patient’s lesion and the
result of the treatment. The radiation field is usually
‘‘concentrated’’ in one or two specific areas of the skin,
usually between 60 cm2 and 80 cm2 (depending on the
X-ray beam projection, image intensifier field size and
collimation carried out by the cardiologist). Because of
the high restenosis rate, a certain number of patients
require several coronary angiographies and PTCAs,
contributing to the increase in skin irradiation, some-
times over intervals of several months or years [5]. ICB
procedures exhibit a higher concentration of fields than
other interventional cardiology procedures [6], and an
increased dose–area product (DAP), fluoroscopy time
(FT) and number of frames (NF) [7]. The need for
recording and reporting every step of the treatment, the
essential parameters for target volume determination,
and therapeutic dose calculations [8, 9] are important
factors to explain this increment. Therefore, the estima-
tion of the risk of deterministic effects (skin injuries) in
ICB procedures is an aspect of radiation protection that
should be considered as part of the quality assurance
(QA) programme.
In this work, the outcomes in patient radiation
protection aspects from our 6 years experience with a
sample of 221 ICB procedures are shown.
Methods and materials
A total of 221 patients underwent ICB procedures
between November 2000 and May 2005 with beta sources
from a Novoste system (www.novoste.com, Norcross,
Georgia, USA) in 73 cases, and a Guidant system
(www.guidant.com, Indianapolis, USA) in 148 cases.
Intravascular ultrasound (IVUS) was used in a signifi-
cant number of patients to quantify the lesion and to
determine the clinical dosimetry parameters for the
brachytherapy procedures. DAP values, together with
FT and NF were recorded in 202 of the ICB patients as
relevant dosimetric parameters. ICB procedures were

730 The British Journal of Radiology, September 2006

Radiation protection experience in intracoronary brachytherapy
carried out in three dedicated X-ray interventional
cardiology rooms equipped with Philips Integris 3000
and 5000 systems (Philips, Best, The Netherlands; all
with high filtration fluoroscopy pulsed modes and with
the routine use of wedge filters) by one expert interven-
tional cardiologist in collaboration with a radiotherapist
and a medical physicist.
The X-ray systems were under a QA programme,
including periodic constancy checks to evaluate incident
air kerma at the entrance of the image intensifier and
entrance surface air kerma for different thicknesses of
polymethylmethacrylate (PMMA) and copper, following
the protocol proposed by the European DIMOND
consortium [10].
To evaluate the increase in dose and procedure
complexity in ICB, DAP, FT and NF were also recorded
for 1707 PTCAs during the same period and using the
same X-ray systems.
DAP was measured with the built-in calibrated
ionization transmission chambers (PTW, Freiburg,
Germany). Skin dose distributions were measured using
slow film Kodak X-OMAT-V initially, using the proce-
dure previously described [11] and the new EDR2 [12]
introduced in 2002 (Kodak, Rochester, NY). For both
types of film, thermoluminescent dosimetry (TLD) was
used additionally to evaluate some high dose values and
for autocalibration of the films. TLD-100 (LiF:Mg,Ti)
chips and a Harshaw TLD/Bicron/NE-Technology
(BICRON-NE, Solon, OH) reader were used.
Digital recording of all the cine images in DICOM
(Digital Imaging and Communications in Medicine)
format allows, in some complex cases when saturation
of the slow film occurred, a complementary analysis of
the skin dose distribution using the technical parameters
of all the cine series acquired during the procedures [13].
These parameters were retrieved throughout using an
updated version of an ad hoc specific software [14].
DICOM header in the cine series recorded by Philips
systems allows us to identify the X-ray beam projection
(left–right and craniocaudal angulations), image intensi-
fier field size, kilovolts and milliamperes per frame, and
distance from the focus to the image intensifier entrance
during the different cine series.
Methodology for the evaluation of skin dose distribu-
tion and peak skin dose (PSD) has been developed by the
authors [11, 12]. As a part of the clinical follow-up of
patients undergoing ICB procedures, a specific protocol
has been developed in the framework of the DIMOND III
project [15, 16] to detect possible deterministic effects on
the skin of patients whose slow film pattern shows
densities corresponding to doses above 1.5 Gy. Clinical
follow-up was initiated when PSD exceeded 2 Gy, DAP
values exceeded 180 Gy cm2 or whenever it was
recommended by other medical circumstances (e.g.
previous procedures, special skin radiosensitivity). In
these cases, the interventionist should arrange for review
of the patient between 10 days and 14 days after the
procedure. The use of these triggering levels for the
clinical follow-up respond to the likely inability to verify
whether the maximum skin dose exceeded the threshold
for deterministic effects when slow film is clearly
saturated and no TLD chips had been placed in the
region of high dose. The purpose of this review was to
The British Journal of Radiology, September 2006
identify skin effects. A record of the FT, NF and DAP is
part of the documentation of each patient reviewed.
Results
Dosimetric aspects of the X-ray procedure
Table 1 shows a comparison between DAP, FT and NF
in PTCA and ICB procedures in the period 2000–2005.
Previous initial data published for the period 2000–
2001 [6, 7] were similar for PTCA, but greater for ICB.
The reduction of doses during ICB procedures is shown
in Figure 1 for the years 2001–2004. The years 2000 and
2005 have been excluded as few patients were treated in
those years.
Patient skin dose
Skin dose is generally quite widely distributed in
cardiology procedures, but in some procedures a certain
area of the skin could receive a higher dose if the
visualization of the lesion to be treated requires main-
taining a fixed orientation of the X-ray beam. The
introduction of a ‘‘concentration factor’’ has been
proposed in a previous paper [5] to consider this aspect.
The concentration factor has been defined as:
PSD PSD
CF~ ~
ASD DAP
S
where PSD is the ‘‘peak skin dose’’ and ASD is the
‘‘average skin dose’’, obtained as the quotient of DAP and
the total irradiated area S (measured from the slow film).
Figure 2 shows a mosaic of several selected examples
of slow film for PTCAs and for ICB. Note the highest
concentration of the irradiation (more density in some
areas of the films) for the ICB procedures. Conversely,
note the greater number of projections (with different
C-arm angulations) for PTCA procedures.
DAP greater than 180 Gy cm2 was measured in 13% of
all the procedures over the whole study period, although
this threshold was surpassed more frequently during the
introduction of this kind of procedure (27% of the cases
in the period 2000–2001).
For 172 patients (78% of patients treated) skin dose
distribution was measured. In 12% of the patients, peak
skin doses higher than 1.5 Gy were measured (one patient
with 4.6 Gy), but only 10 of these patients were effectively
reviewed, as some patients belonged to other health areas
and some had died. No radiation skin injuries in ICB
procedures were found during follow-up examinations.
Discussion
It is clear that ICB produces a higher DAP, FT and NF,
due to the complexity of the procedure, the need for
recording and reporting every step of the treatment and
essential parameters for volume determination of the
lesion, and therapeutic dose calculation to the target
volume [8, 9]. IVUS is advisable [8] to evaluate in detail
731

Table 1. DAP, fluoro time and number of frames for PTCA and ICB procedures between November 2000 and May 2005
Procedure Sample DAP (Gy cm2) Fluoro time (min)
size
Median Min./Max. 3rd Median
quart.
PTCA 1707 41 10/492 69 14.3
ICB 202 62 11/384 99 17.0
DAP, dose–area product; PTCA, percutaneous transluminal coronary angioplasty; ICB, intracoronary brachytherapy.
the lesion and to obtain the basic data for clinical
dosimetry of the brachytherapy treatment. However, its
use implies an increase in the fluoroscopy time. The
control of the right position of the radioactive source in
the lesion and the removal of the source also require
extra fluoroscopy time and filming series. Sometime,
when ‘‘stepping’’ (irradiation of the lesion in several
steps because of the extension of the lesion plus
appropriate safety margins) is required, the procedure
could become even longer.
The correct position of the radioactive source in the
lesion is one of the basic aspects of the quality of ICB.
Sabate et al [17] state that a recognized limitation of
endovascular beta-radiation therapy is the development
of new stenoses at the edges of the irradiated area. The
term ‘‘geographic miss’’ is used to define cases in which
the radiation source did not fully cover the injured area.
To avoid this problem, fluoroscopy guidance throughout
the procedure and additional cine series to confirm the
correct position of the radioactive source are required.
Images recorded allowed explanation of some unsuc-
cessful treatments as due to geographical miss.
The difficulty in performing IVUS in the vessel to be
treated, the difficulty in positioning the radioactive
source, and the need to use some specific and fixed
projections to correctly document the position of the
source, increase the ‘‘concentration factor’’ of radiation in
some specific skin areas [6]. In these situations, the total
DAP could not always be a good indicator of the level of
risk for deterministic effects (skin injuries), as previously
highlighted [5].
Figure 1. Evolution of median values of dose–area product
(DAP) in intracoronary brachytherapies (ICBs) during the
period 2001–2004. Detail of the distribution (frequency
histogram) of DAP in ICBs.
732
C Prieto, E Vano, J M Fernández et al
Number of frames
Min./Max. 3rd Median Min./Max. 3rd
quart. quart.
3.1/102.5 23.1 1078 143/4887 1536
6.0/70.0 23.3 1493 443/3444 1937
The preliminary European reference level [18] for
PTCA is 94 Gy cm2, which is well above our median
values for both PTCAs and ICBs (Table 1).
The reduction of median DAP for ICB along the period
studied may be due to greater experience and confidence
by the staff involved, the effect of periodical training in
radiation protection highlighting special aspects of ICB
[19], and the use of IVUS in fewer patients (73% of
patients in 2000–2001 compared with 53% in 2002–2005).
Conclusions
ICB leads to an increase in patient dose during the
procedure in comparison with PTCA. The patient dose and
skin dose increase because of the increase in the complexity
of the procedure, the need for recording and reporting
every step of the treatment and essential parameters for
volume determination and the therapeutic dose calcula-
tion. The use of IVUS to quantify the lesion and to
determine the clinical dosimetry parameters for the
brachytherapy procedures may also involve an increase
in patient dose due to the increased fluoroscopy time.
Finally, the learning curve with new techniques implied a
greater increment with respect to conventional PTCA
procedures in the first period of this study.
Continuous training in radiation protection highlight-
ing special aspects of ICB together with greater
experience and ability in these kinds of procedures also
contributed to shortening the difference in patient dose
between ICB and PTCA procedures.
The extra dose associated with ICB procedures and
greater skin dose concentration factor may lead to
further skin injury problems. Notwithstanding, in our
institution, with the X-ray systems submitted to rigorous
quality assurance programmes and with the optimized
technical and clinical protocol, this extra dose has not
been of special concern for patient skin injuries during
the reported period. Nevertheless, in those cases in
which a threshold dose is reached, a specific clinical
follow-up protocol is advisable.
During the study period, satisfactory outcomes and
acceptably low radiation doses to the skin of the patients
were confirmed.
Acknowledgments
This study was partially funded under the European
Commission DIMOND III project (FIGM-CT-2000-
00061), and Coordination Action SENTINEL (FI6R-
012909). Funding was also provided by the National
Program for Scientific Research, Development and
Technological Innovation of the Spanish Department
The British Journal of Radiology, September 2006
Radiation protection experience in intracoronary brachytherapy
Figure 2. Mosaic of several selected examples of slow films
used to measure skin dose distribution for percutaneous
transluminal coronary angioplasties (PTCAs) and for intra-
coronary brachytherapies (ICBs).
for Science and Technology (project BFI2003-09434) and
by the Autonomous Community of Madrid (project GR/
SAL/0272/2004).
References
1. Waksman R, Cheneau E, Ajani AE, White L, Pinnow E,
Torguson R, et al. Intracoronary radiation therapy improves
the clinical and angiographic outcomes of diffuse in-stent
restenostic lesions. Results of the Washington radiation for
instent restenosis trial for long lesions (Long-Wrist) studies.
Circulation 2003;107:1744–9.
2. International Commission on Radiological Protection.
Avoidance of radiation injuries from medical interventional
procedures. ICRP Publication 85. Ann ICRP 2000;30:7.
3. Koenig TR, Mettler FA, Wagner LK. Skin injuries from
fluoroscopically guided procedures: Part 2, Review of 73
cases and recommendations for minimizing dose delivered
to patient. AJR Am J Roentgenol 2001;177:13–20.
4. Vano E, Arranz L, Sastre JM, Moro C, Ledo A, Garate MT,
et al. Dosimetric and radiation protection considerations
based on some cases of patient skin injuries in interven-
tional cardiology. Br J Radiol 1998;71:510–6.
5. Vano E, Goicolea J, Galvan C, Gonzalez L, Meiggs L, Ten JI,
et al. Skin radiation injuries in patients following repeated
coronary angioplasty procedures. Br J Radiol 2001;74:1023–31.
6. Vano E, Prieto C, Fernandez JM, Gonzalez L, Sabate M,
Galvan C. Skin dose and dose–area product values in
patients undergoing intracoronary brachytherapy. Br J
Radiol 2003;76:32–8.
The British Journal of Radiology, September 2006
7. Prieto C, Vano E, Fernandez JM, Sabate M, Gonzalez L,
Aviles P. Staff doses in procedures of intracoronary
brachytherapy using beta sources. IAEA-CN-91/105
Contributed papers. International Conference on
Occupational Radiation Protection: Protecting Workers
against Exposure to Ionizing Radiation. Geneva,
Switzerland 26–30 August 2002.
8. Pötter R, Van Limbergen E, Dries W, Popowski Y, Coen V,
Fellner C, et al. Recommendations of the EVA GEC ESTRO
Working Group: prescribing, recording, and reporting in
endovascular brachytherapy. Quality assurance, equipment,
personnel and education. Radiother Oncol 2001;59:339–60.
9. Nath R, Amols H, Coffey C, Duggan D, Jani S, Li Z, et al.
Intravascular brachytherapy physics: Report of the AAPM
Radiation Therapy Committee Task Group No 60. Med
Phys 1999;26:119–52.
10. Faulkner K. Introduction to constancy check protocols in
fluoroscopic systems. Radiat Prot Dosim 2001;94:65–8.
11. Vano E, Guibelalde E, Fernández JM, González L, Ten JI.
Patient dosimetry in interventional radiology using slow
films. Br J Radiol 1997;70:195–200.
12. Guibelalde E, Vano E, Gonzalez L, Prieto C, Fernández JM,
Ten JI. Practical aspects for the evaluation of skin doses in
interventional cardiology using a new slow film. Br J Radiol
2003;76:332–6.
13. Balter S, Vano E, Gonzalez L. Fluoroscopic patient dosimetry
from DICOM headers. 11th International Congress of the
International Radiation Protection Association. http://www.
irpa11.com. Session 4c1. Madrid. Spain. 23–28 May 2004.
Paper available at: www.irpa11.com/new/pdfs/4c1.pdf
(Accessed 12 August 2005).
14. Vano E, Fernandez JM, Ten JI, Guibelalde E, Gonzalez L,
Pedrosa C. Real-time measurement and audit of radiation
dose to patients undergoing computed radiography.
Radiology 2002;225:283–8.
15. European Commission. DIMOND III project. http://www.
dimond3.org/ (Accessed 6 August 2005).
16. Vano E, Aviles P, Prieto C, Fernandez JM, Guibelalde E,
Galvan C, et al. A dosimetric trial for the clinical follow-up of
potential skin injuries on patients undergoing interventional
cardiology procedures. 11th International Congress of the
International Radiation Protection Association. ID: 1307.
Madrid. Spain. 23–28 May 2004. Paper available at: www.
irpa11.com/new/pdfs/4c11.pdf (Accessed 12 August 2005).
17. Sabaté M, Costa MA, Kozuma K, Kay P, van der Giessen J,
Coen V, et al. Geographic miss: a cause of treatment failure
in radio-oncology applied to intracoronary radiation ther-
apy. Circulation 2000;101:2467–71.
18. Neofotistou V, Vano E, Padovani R, Kotre J, Dowling A,
Toivonen M, et al. Preliminary reference levels in interven-
tional cardiology. Eur Radiol 2003;13:2259–63.
19. Prieto C, Vano E, Fernandez JM, Sabate M, Galvan C,
Meiggs L, et al. Radiation protection training in intra-
coronary brachytherapy. Proceedings of the II International
Conference. Radiation Protection Training: Future
Strategies. 2003. Ed. CIEMA, Madrid, Spain. ISBN: 84-
7834-450-0.
733
 The British Journal of Radiology, 79 (2006), 734–739

What is the optimum breast plan: a study based on the START

trial plans
1
K VENABLES, PhD, MIPEM, 1E A MILES, MPhil, DCR(T), 2E G A AIRD, PhD, FIPEM and 1P J HOSKIN, FCRP, FRCR,
on behalf of the START trial management group

1
Marie Curie Research Wing and 2Medical Physics Department, Mount Vernon Hospital,
Rickmansworth Road, Northwood, Middlesex HA6 2RN, UK

ABSTRACT. Each year thousands of women within the UK are treated with
radiotherapy for breast cancer. The majority of these women are treated using a medial
and lateral tangential field. This study evaluates the plans submitted to the quality
assurance (QA) team of the START trial and investigates some of the differences
between departments. Throughout the START trial, hardcopies of the radiotherapy
dose distribution on the central slice for one in three women were submitted to the QA
team for analysis. The QA team measured physical parameters including breast size and
lung depth as well as noting parameters used for the radiotherapy delivery including
beam energy, field size and wedge angle. Over 1400 plans from 36 centres were
analysed. The mean patient separation was 19.7 cm (SD 2.7 cm) with a mean lung Received 24 October 2005
depth of 1.5 cm (SD 0.7 cm). The modal beam energy was 6 MV and the mean wedge Revised 16 December 2005
angle was 23 ˚. Significant differences in the choice of wedge angle between Accepted 9 January 2006
departments were noted; however, in 90% of cases the resultant plan complied with
DOI: 10.1259/bjr/80814021
the maximum dose gradient of 10% on the central axis specified by the trial protocol.
Less than 3% (37 plans) had dose gradients of greater than 12%. This resulted in a ’ 2006 The British Institute of
mean dose gradient for all patients on the central axis of 5.7% (SD 2.9%). Radiology

Many women with breast cancer are treated with
radiotherapy each year. Guidelines on the production of
isodose distributions have been given by the ICRU [1]. The
START trial is a multicentre UK trial of breast radiotherapy
comparing different fractionation regimens that accrued
4451 patients between January 1999 and December 2002
[2]. There were two randomization options called A and B.
Trial A was a three way randomization to either 50 Gy in
25 fractions, treated daily over 5 weeks, or one of two
schedules treating five times per fortnight to a total dose of
39 Gy or 41.6 Gy in 13 fractions of 3.0 Gy or 3.2 Gy,
respectively. Trial B patients were randomized to 50 Gy in
25 fractions over 5 weeks or 40 Gy in 15 fractions over
3 weeks, treated daily. Guidelines on the production of
isodose distributions were provided in the trial protocol,
which stated that two tangential beams, angled to remove
divergence in the lung, should be used. Departments were
asked to use wedges to achieve a dose gradient of less than
10% across the breast on the central axis slice. In addition,
cobalt-60 was only permitted for patients with a separation
of less than 18 cm.
The trial has had a quality assurance (QA) programme
from the outset. One of the aims of the QA team was to
document the treatment technique in each centre and to
ensure compliance with the protocol.
The treatment of breast cancer with radiotherapy is
evolving and there is an increased awareness of the need
for compensation in three dimensions. Within the
START trial, only two centres used compensation other
than a wedge. For some women, particularly those with
larger breast sizes, this may be necessary to achieve dose
uniformity particularly when doses away from the
central axis are considered. Areas of increased dose
within the breast may be the cause of breast pain and
result in poorer cosmesis [3], while irradiation of large
lung volumes could potentially lead to radiation pneu-
monitis [4]. In current clinical practice the majority of
patients are treated with two opposing tangential fields.
This paper analyses the data for these patients.
Method
Radiation dose distributions for 1 in 3 of the patients,
assigned at the time of randomization, were collected by
the QA team. Hardcopy central axis distributions were
sent to the QA team after the patient had completed
treatment. From these, data on the physical shape of the
patient’s breast, the linac parameters used for treatment
and the resulting isodose distributions were collected.
Data from each plan were checked by a second member
of the QA team. Universal wedges were converted to an
equivalent wedge angle using the formula:
w1 tan (60)
tan h~
w1 zw2
where w1 is the weight of the wedge beam and w2 the
weight of the corresponding open field. The wedge angle
of the universal wedge was assumed to be 60 ˚ for all
departments. During the analysis a further parameter of
patient shape was defined:
max distance from post: field edge to skin
breast shape~
patient separation

734 The British Journal of Radiology, September 2006

START breast plans
Dose gradient was defined as the difference in dose
between a minimum point dose and maximum isodose
encompassing an area of 2 cm2. Unless there was a cold
spot in the centre of the breast, departments were asked
to record the minimum dose at a point 0.5 cm above
the lung boundary on the perpendicular bisector of the
posterior beam edge and at least 1 cm from the 50% field
edge. This point was chosen so as to be in breast tissue
rather than ribs, and to reduce the differences due solely
to the use of different planning systems, which vary in
their ability to account for lack of lateral scatter from the
lung.
The QA team visited each department to observe
simulation, planning and treatment of patients and to
perform measurements on breast shaped phantoms. The
results of the measurements in phantoms have been
previously published [5, 6]. Questions relating to plan-
ning policy were asked of the staff producing the isodose
distributions. Of relevance to the work presented in this
paper were the criteria for their planning distribution,
such as the choice of wedge angle, the resultant posi-
tion and size (area and intensity) of hot spots and den-
sity correction used for lung. In a number of cases, this
changed during the trial as better simulation facilities,
particularly CT scanners, became available.
All analyses were performed using the SPSS program
(SPSS Inc., Chicago, IL). Correlation (Spearman rank
coefficient used if data not normally distributed) was
used to test associations between dose gradient, wedge
angle, breast shape and depth, patient separation, energy
and wedge angle. Differences in dose gradient by centre
were compared using analysis of variance (ANOVA).
Data are presented using boxplots and scatter diagrams.
In boxplots, the horizontal line shows the median
value of the data, the length of the box represents the
interquartile range, and the length of the line indicates
the range excluding outliers and extremes (outliers/
extremes are values more than 1.5/2.5 box-lengths from
the 25th percentile).
Results
1488 plans were collected from 36 departments
(1 centre input patients from 2 sites; these were counted
as separate departments by the QA team since the
planning systems were different on the two sites). Patient
characteristics and treatment techniques are shown in
Tables 1 and 2. The depths of lung incorporated into
the treatment plan varied between departments and is
shown in Figure 1.
A range of energies from Co60 to 10 MV was used
for the treatment of these patients. The majority were
treated using linear accelerators with a nominal energy
of 6 MV. Energy used for treatment is illustrated in
Figure 2. The energy used only correlated weakly with
Table 1. Patient demographics
Minimum Maximum Mean Standard
(cm) (cm) (cm) deviation
Separation 13.4 30.2 19.8 2.7
Breast depth 0.5 13.2 5.2 1.9
Lung depth 0 3.5 1.4 0.6
The British Journal of Radiology, September 2006
Table 2. Patient treatment details
Number of patients
(% of total)
Site Breast 1338 (90 %)
Chest wall 150 (10 %)
Technique Isocentric 1035 (70%)
Fixed FSD 453 (30%)
Use of asymmetry None 1290 (86.7%)
Posterior border 133 (8.9%)
Superior border 65 (4.4%)
FSD, focus to surface distance.
the patient separation (Spearman correlation coefficient
0.21 p,0.001) (Figure 3).
Of the plans received by the QA team, 748 incorpo-
rated a lung correction using either a standard value for
all patients, which varied from 0.2 to 0.33 relative to
water depending on centre, or bulk density corrections
derived from CT data for an individual patient,
which varied from 0.19 to 0.5 (53 plans). Ten patients
were planned using a pixel by pixel correction for
inhomogeneity.
The START trial stated that the recommended dose
gradient across the patient breast be less than 10% on the
central axis and 1390 plans (93.4%) complied with this.
A further 95 plans (6.4%) exhibited dose gradients of
between 10% and 15%. The remaining 3 plans (0.2%) had
a dose gradient of greater than 15%. ICRU guidance
suggests that plans should comply with a dose of 95–
107% prescription dose over the target volume [1]. It was
only possible to assess central sections and of these, 1232
plans (83%) complied with these recommendations. Dose
gradient was found to be weakly correlated to breast
depth (r50.50, p,0.001) (Figure 4) and with patient
separation (r 5 0.57, p,0.001) (Figure 5). This latter
figure shows data separated by energy; as expected, the
dose gradients are less if higher energy machines are
used. The correlation between gradient and separation
also improves if each energy is considered separately,
and this is shown in Table 3.
Figure 1. Box plot illustrating variation in lung depth vs
centre.
735

Figure 2. Energy used for treatment (n51488).
The mean wedge angle used was 22 ˚ for patients after
lumpectomy (range 0–60 ˚) and 33 ˚ (range 0–60 ˚) for
patients post-mastectomy. 690 patients were treated on
linear accelerators with external fixed wedges (although
a small number of these patients may have been treated
using enhanced dynamic wedge (EDW). No significant
difference was found between the medial and lateral
wedge angles. A correlation was found between wedge
angle and patient shape, which was stronger for some
Figure 3. Association between energy used for treatment
and average patient separation. Lines give the range of
patient separations treated at each energy.
736
K Venables, E A Miles, E G A Aird et al
departments than for others. A department with strong
correlation is shown in Figure 6 and the results for all
patients in the trial in Figure 7. For centres where only
fixed wedge angles were available, less correlation is
observed. Wedge angle and patient shape did not
correlate with an increased uniformity of dose gradient
for the same centre. There was a negative correlation
between wedge angle and dose gradient (medial wedge
r520.35, p,0.001, lateral wedge r520.34, p,0.001).
Differences in wedge angles used were also seen between
departments.
Significant differences were seen in the dose gradients
obtained from different centres (ANOVA p,0.001).
Further analysis of these differences revealed that even
when comparing distributions from different depart-
ments with the same planning system and same nominal
energy of accelerator (6 MV), differences were evident,
as shown by Figure 8. Of particular note is the difference
between centres B and E, both of whom put large
numbers of patients into the trial. Centre B did not apply
any lung correction but allowed for the effect of lung by
planning with the apex of the breast hotter than the
medial or lateral edges. Centre E estimated a lung shape
based on the maximum lung depth and applied a lung
correction of 0.25. The dose gradient recorded is not
affected by the incorporation of lung correction. The
planning system was a simple beam library system and
the reduction in dose at the minimum point due to the
loss of lateral scatter when the lung was included would
not be reflected in the calculation. The lower than
average dose gradient for centre B might suggest that
they were not adequately accounting for the effect of the
lung and were using steeper wedge angles than would
have been used on a lung corrected plan. The higher
than average dose gradients recorded by centre E might
suggest that their estimate of lung shape was not
sufficiently accurate and that they were overestimating
the volume of lung in the plan.
The distribution of radiation within the breast was
classified by comparing the dose at the apex of the breast
(taken as 1.5 cm below the skin surface, Figure 9) with the
dose in the most medial and lateral parts of the breast
at least 1 cm from build up or penumbra regions. The
majority of patients (70%) had an even distribution of
radiation between these areas. However, in 11% of plans
the medial and lateral aspects of the breast were hotter
than the apex of the breast by 2–4%. When only those
patients who do not have a lung correction are considered,
the proportion in which the medial and lateral aspects
were hotter than the breast apex was slightly less (8%).
Discussion
There was no correlation between average lung depth
for a centre and average separation for that centre, which
may have been observed if practice in some centres
placed the lateral border more posteriorly than others.
The lung depth included in the treatment field is affected
by the patient’s position; thus it would be necessary for
some clinicians to include more of the lung in order to
comply with the protocol recommendations on the
positioning of the borders. In 1988, Fraass et al recom-
mended the incorporation of lung density correction
The British Journal of Radiology, September 2006
START breast plans
[7]. This is common practice in the UK with just over
half of centres (51%) incorporating a correction and the
majority of others planning to compensate for the effect
of lung.
The weak correlation between energy and patient size
may be related to the availability of different energies. In
departments where high energy machines were available
for any patient, there may be greater tendency for its use
– one department treated 53% of patients using 10 MV
photons. The depth of the 95% isodose increases as
higher energy machines are used. There will therefore be
a decrease in the coverage of the superficial regions of
the planning target volume (PTV) when 10 MV photons
are used compared with 6 MV photons. The depth below
The British Journal of Radiology, September 2006
Figure 4. Relationship between
dose gradient (%) and breast depth.
the skin surface at which the PTV should be drawn is
usually taken to be 0.5 cm, however, compromise in this
region is often preferred to decreased dose at depth. The
exception is post-mastectomy patients for whom the
loss of dose in the superficial tissues, when treated at
high energy, represents a significant percentage of the
planning target volume. Four patients in the trial who
had undergone a mastectomy were treated using 10 MV
photons without the application of bolus.
For the majority of machines in clinical use in the
UK, the wedges are housed within the treatment head.
However, some patients are treated on units with
externally mounted wedges. The scatter from these
external wedges can increase the dose to the patient’s
Figure 5. Relationship between
dose gradient and separation. Data
for 6 MV have been excluded from
the graph to allow the trends for
other energies to be seen.
737
 K Venables, E A Miles, E G A Aird et al

Table 3. Correlations between dose gradient and patient

separation with respect to energy used for treatment
Energy Spearman correlation Number

Co60 0.641 28
4 MV 0.608 50
5 MV 0.709 76
6 MV 0.58 1249
8 MV 0.797 21
10 MV 0.724 66
All patients 0.57 1488

skin and contralateral breast and therefore it has been

recommended by some authors that the wedging be
applied predominantly from the lateral beam [8]. This
practice was not seen in the START data where the
majority of patients had the same wedge used for both
medial and lateral fields. Differences were seen in the
wedge angles used in different departments for the same
‘‘shape’’ of breast. The wedge angle needed will be
affected by the energy of machine used. However, 75% of
departments treated more than 80% of their breast
patients at 6 MV and thus comparisons between depart-
ments are possible. When using fixed wedge angles
there is less scope for optimization of the plan, whilst
maintaining rapid treatment times. Departments using
large wedge angles tend to produce plans in which there
is more dose towards the medial and lateral aspect of the
breast (see Figure 9 for definition of points). In addition,
where lung correction has been omitted, true dose in
these regions may be higher than indicated on the plan,
due to the increased transmission through lung, leading
to a possible increase in the risk of rib fracture.
Care must be taken in comparing distributions
between planning systems due to the subtle differences
in the implementation of the algorithms. However,
differences in average dose gradient were also evident
when comparing departments using the same planning
system and same nominal energy of machine. No further
Figure 7. Correlation between wedge angle and breast
shape for all patients in the trial r520.39, p,0.001. Note the
wide spread in breast shapes for the fixed wedge angles 0 ˚,
15 ˚, 30 ˚, 45 ˚ and 60 ˚.
statistical analysis of this aspect was performed because
of the small numbers in some of the groups, but it reflects
differences in planning policy. Some departments with
conventional rather than CT simulators estimated the
lung shape using either a rule of thumb or a lung
template. Although this will usually give a more real-
istic estimate of the distribution than ignoring the lung
correction completely, it may result in a different
distribution from that which would be produced if more
accurate information was available.
In some instances plans are produced in which no
attempt is made to estimate the lung outline. In these
cases it is not possible to incorporate lung correction to
compensate for the reduced attenuation of the lung
relative to breast tissue. One way to overcome this is to
produce a plan which indicates an artificial dose
gradient across the breast that will not be present in

Figure 6. Correlation between wedge angle and breast

shape (max. distance from post edge to skin/patient separa-
tion) for centre 12. This was a large centre with a wide range Figure 8. Box plot showing difference in dose gradient for
of planning staff, but with rigid guidelines on the final different centres using the same planning system and same
distribution (r520.81 p,0.001, n544). nominal machine energy. Circles represent outliers.

738 The British Journal of Radiology, September 2006

START breast plans
Figure 9. Definitions used for points of the breast.
the patient. For these patients, it is expected that plans
sent to the QA team will aim for the dose at the apex of
the breast on the plan to be hotter than that at the medial
and lateral border. For a small percentage of patients, the
reverse of this situation was seen and the medial and
lateral aspect of the plan was hotter than the apex. In a
few cases, this may be necessary to compensate
adequately for variations in the shape of the breast in
three dimensions, but should not be normal practice if
only the central axis of the breast is considered.
For the majority of patients, only data on the central
slice was sent to the QA team; this will mask the true
dose gradients which will be present in the patient’s
breast as a whole. For some women to conform with
ICRU recommendations of a maximum dose gradient
of 95–107%, intensity-modulated radiotherapy will be
necessary to eliminate hot spots in the superior and/or
inferior aspect of the breast.
Conclusion
The analysis of the plans submitted to the QA team
emphasises the high degree of protocol compliance that
was achieved for this trial. The collection of central axis
distributions by the QA team enabled protocol viola-
tions to be flagged up to centres early in the trial and
thus improved the compliance with the protocol. It
also enabled the QA team to discuss alternative patient
positioning where larger than average lung depths were
seen. The data collected represents a snapshot of clinical
practice in the years the trial was accruing and will
aid the analysis of the main trial once mature survival,
recurrence and late effect data are available.
When producing a dose distribution for a patient, a
number of factors must be considered. Ideal breast plans
have a 95% isodose which covers breast tissue to within
5 mm of the skin surface. No hot spots in excess of 107%
The British Journal of Radiology, September 2006
should be present (lower if only a single slice is assessed)
and the hotspots should be located evenly throughout
the breast tissue. Ideally, patients should have CT scans
to obtain accurate information on the position of the
lung. If this is not possible, the effect of low density lung
tissue should be carefully considered and the distribu-
tion should appear hotter at the apex of the breast. For
some patients, compromise must be reached and the
planner should carefully weigh the increased penetra-
tion available by using higher energy beams against
the decrease in dose to superficial tumours. In particular,
if post-mastectomy radiotherapy is to be given using
photons with energies higher than 6 MV then the use of
bolus should be considered.
Acknowledgments
The START trial management group: Edwin Aird, Jane
Barrett, Peter Barrett-Lee, Judith Bliss, Jackie Brown,
John Dewar, Jane Dobbs, Jo Haviland, Penny Hopwood,
Peter Hoskin, Pat Lawton, Brian Magee, David Morgan,
Roger Owen, Eileen Parkin (RAGE Observer), Joyce
Pritchard (RAGE Observer), Val Speechely, David
Spooner, Mark Sydenham, Karen Venables, Elizabeth
Miles and John Yarnold.
References
1. International Commission on Radiation Units and
Measurements. Prescribing, recording and reporting photon
beam therapy. ICRU Report 50. Bethesda, MD: ICRU, 1993.
2. START trial management group. Standardization of Breast
Radiotherapy (START) Trial. Clin Oncol 1999;11:145–7.
3. Yarnold JR, Donovan EM, Bleakley NJ, Reise SF, Regan J,
Denholm E, et al. Randomised trial of standard 2D radio-
therapy (RT) versus 3D intensity modulated radiotherapy
(IMRT) in patients prescribed breast radiotherapy. Radiother
Oncol 2002;64:S15.
4. Early Breast Cancer Trialists’ Collaborative Group.
Favourable and unfavourable effects on long-term survival
of radiotherapy for early breast cancer: an overview of the
randomised trials. Lancet 2000;355:1757–70.
5. Venables K, Winfield EA, Deighton A, Aird EGA, Hoskin PJ.
The START Trial – measurements in semi-anatomical breast
and chest wall phantoms. Phys Med Biol 2001;46:1937–48.
6. Venables K, Winfield EA, Aird EGA, Hoskin PJ. Three-
dimensional distribution of radiation within the breast: an
intercomparison of departments participating in the START
trial of breast radiotherapy fractionation. Int J Radiat Oncol
Biol Phys 2003;55:271–9.
7. Fraass BA, Lichter AS, McShan DL, Yanke BR, Diaz RF,
Yeakel KS, et al. The influence of lung density corrections on
treatment planning for primary breast cancer. Int J Radiat
Oncol Biol Phys 1988;14:179–90.
8. Ikner CL, Russo R, Podgorsak MB, Proulx GM, Lee RJ.
Comparison of the homogeneity of breast dose distributions
with and without the medial wedge. Med Dosim 1998;23:
89–94.
739
 The British Journal of Radiology, 79 (2006), 740–744

Feasibility of automated matching of supine and prone

CT-colonography examinations
1
A H DE VRIES, MD, 2R TRUYEN, MSc, 2J VAN DER PEIJL, MSc,
1
J FLORIE, MD,
1
RE VAN GELDER, MD,
2
F GERRITSEN, MSc and 1J STOKER, MD

1
Department of Radiology, Academic Medical Center, Amsterdam and 2Philips Medical Systems,
Best, The Netherlands

ABSTRACT. Matching of prone and supine positions in CT colonography may improve

accuracy of polyp detection. The purpose of this study was to investigate the feasibility
of automatic prone-supine matching in CT-colonography using proven polyps as fixed
points of reference. The method is based on similarities in the direction of centre-lines
and allows for compression and extraction of the centre-lines in both positions. To
illustrate the impact of the match error of the new method in practice, the visibility of
the matched polyps in a primary three-dimensional unfolded cube setting was
determined as well. The method was compared with a method that relies on the
normalized distance along the centre-line (NDAC method). The median absolute match
error was 14 mm (range 0–59 mm, average 20 mm) either proximal or distal from the
actual polyp in prone position. In the observer study, 70% (26/37) of the polyps were
Received 12 September
directly visible in prone view. The overall difference in median absolute match error 2005
between both methods was small (2 mm), although half way along the centre-line Revised 9 January 2006
there were polyps with substantial differences in match error (larger with NDAC). We Accepted 18 January 2006
concluded that automated prone-supine matching of CT-colonography studies is
DOI: 10.1259/bjr/55953054
feasible and has a low match error. The difference with the NDAC method was small
and not significant, although half way along the centre-line some differences were ’ 2006 The British Institute of
seen. Radiology

Colorectal cancer is the second leading cause of
cancer-related mortality in the western world. CT-
colonography is considered as a potential screening tool
for colorectal cancer. To improve the accuracy of polyp
detection, patients are both scanned and examined in the
prone and supine position. Combining information from
these scans will assist the reviewer in evaluating colon
segments and differentiating polyps from faeces or folds.
As most faecal material is subject to gravity, combining
both scans may be able to increase specificity of CT-
colonography.
Since reference-points (e.g. hepatic flexure) are often
not fixed, manual verification of findings on supine and
prone positions may be a time-consuming activity. Using
an automated supine-prone matching algorithm may
facilitate this process, and may lead to a more efficient
interpretation of CT-colonography.
The method evaluated in this article is based on
similarities in direction of the centre-line and allows for
compression and extraction of the centre-lines in both
positions.
The first aim of the study was to assess the feasibility
of automated matching of supine and prone CT-
colonography examinations with colonoscopically pro-
ven polyps as fixed points of reference. Therefore the
match error was calculated. The visibility of matched
polyps was determined as well. This was carried out in a
For this research project, a grant from Philips Medical Systems was
received.
three-dimensional setting after a match with this new
method was calculated. Third, the method was com-
pared with a method that relies on the relative location
between start and end of both paths.
Materials and methods
To evaluate the algorithm, CT-colonography exami-
nations with colonoscopically proven polyps (>5 mm)
were used. These examinations formed part of a
comparative study of CT colonography and colonoscopy
in a surveillance population [1]. All polyps could be
identified three dimensionally in both prone and supine
positions. The polyps in both positions were linked,
based on the three criteria of size, location and
morphology. Polyps were excluded in cases where there
could be doubt about the accuracy of these links.
After manual insufflation of colon and rectum and
intravenous administration of either 20 mg butylscopo-
lamine bromide (Buscopan; Boehringer-Ingelheim,
Ingelheim, Germany) or 1 mg glucagon hydrochloride
(Glucagen; Novo Nordisk A/S, Bagsvaerd, Denmark),
patients were scanned with a four-slice CT scanner
(Mx8000; Philips Medical Systems, Best, The
Netherlands) in supine and prone positions. Scan-
parameters were as follows: 120 kV, collimation 4 mm
6 2.5 mm, rotation-time 0.75 s, pitch 1.25, slice-thickness
3.2 mm, reconstruction interval 1.6 mm and reconstruc-
tion filter C. The tube-current varied between 25 mAs

740 The British Journal of Radiology, September 2006

Automated matching of supine and prone CT-colonography

and 70 mAs depending on the waist circumference of the

patient.
Following scanning, centre-lines were calculated semi-
automatically; a ‘‘seed’’ was placed in the proximal colon
and discontinuities of the colon were bridged manually.
For calculation of the centre-line, software was used to
reconstruct the colon three-dimensionally (Easy Vision;
Philips Medical Systems, Best, The Netherlands).
The matching of prone and supine was performed
based on two principles:

1. maximal alignment of centre-line-directions of both

prone and supine positions.
2. limited compression or expansion of the centre-
lines to make the maximal alignment of corre-
sponding segments possible.

In order to reach maximal alignment, the outcome of
the following function needs to be minimum:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 
fi,j ~ xdiri {xdirj z ydiri {ydirj z zdiri {zdirj
Low values of fi,j imply a small difference in centre-line
direction (dir) between sample points i (prone) and j
(supine) in the three dimensions (x, y and z), indicating a
good match. In this step, the points with the same
direction in both positions are therefore matched as well
as possible.
Second, since corresponding directions of both centre-
lines are often not situated in exactly the same part of the
colon, expansion and compression of the paths was
applied. To avoid matching of the hepatic flexure in
supine position to a curvature in the sigmoid colon in
prone position, a penalty-value for expansion and
compression of the paths was applied. This penalty
value was proportional to the expansion or compression
used. Therefore, the more the centre-line was manipu-
lated in order to match parts with a similar direction, the
higher the penalty value was.
The sum of the outcome of the function of maximal
alignment and the penalty value of the expansion/
compression formed the match cost.
With the use of dynamic programming [2], the match
cost was computed for each possible combination of
points on the centre line in both positions. The combina-
tion of points was made such that the total cost (i.e. the
sum of all individual costs) was minimal.
The performance of the algorithm was measured
using the absolute match error (in millimetres). This
match error was calculated by measuring the distance
from the actual polyp location to the calculated
(expected) polyp position, [A] (Figure 1). This was done
along the central axis in prone position after the match
was made.
To illustrate the impact of the match error of the new
method in practice, the visibility of the matched polyps
in a primary three-dimensional unfolded cube setting
was determined as well. This was done by a research
fellow (JF, medical doctor), who had evaluated over 100
primary three-dimensional CT-colonographies (all colo-
noscopically verified). The reviewer was presented a
polyp in supine position that was to be indicated in the
Figure 1. The performance of the algorithm was calculated
by measuring the distance from the actual polyp location in
supine position to the calculated (expected) polyp position in
2  2ffi prone position after the match was made. (A, match error).
adjacent prone position, after the match was calculated
(Figure 2).
The visibility was measured on a four point scale:
I. The corresponding polyp was clearly visible with-
out scrolling along the central axis.
II. As I, but to be certain the reviewer scrolled along
the central axis.
III. The corresponding polyp could not be seen
instantly, but was traceable after scrolling along the
central axis within a margin of 150 mm from the
calculated location.
IV. The corresponding polyp could not be found by
the reviewer within a margin of 150 mm from the
initial spot of the virtual camera.
The match was considered successful if the matched
polyp could be classified under category 1 or 2. This
method was compared with a method based on the
relative location between start and endpoint of this
centre line (Normalized Distance Along the Centre-line,
NDAC) [3]. Here, the beginning of the centre-line, the
caecum, is located at index 0 on the path and the end, the
rectum, is defined as index 1. All other points will have
indices between 0 and 1. The NDAC absolute match
error is then defined as:
Absolute match error5[NDAC polypsupine 2
NDACpolypprone] 6 Length Central Axis prone
The Wilcoxon-test was performed to test for differ-
ences in absolute match error in both methods. The null
hypothesis was that there was no difference between
both methods. With a p-value of less than 0.05, this
hypothesis could be rejected.
Results
32 of the 249 CT-colonography studies from the former
study included one or more polyps >5 mm that were
visible in both supine and prone examinations. These 32

The British Journal of Radiology, September 2006 741

 A H de Vries, R Truyen, J van der Peijl et al

Figure 2. User interface panel shows both the supine and prone dataset in an unfolded three-dimensional rendering, in a two-
dimensional axial rendering and in a volume rendered overview of the colon. On the left the supine position was displayed, on
the right the prone position. Interactive navigation along the pre-calculated centre-line was possible. This panel shows a
calculated match. In this example a match error of 27 mm was measured and the corresponding polyp was immediately visible
(category I). This match error must not be confused with the difference in ‘‘distance along path’’. The matching was evaluated
with the use of a colonoscopically proven polyp (arrow) (Prototype based on EasyVision; Philips Medical Systems, Best, The
Netherlands).

examinations included 53 polyps (5 mm or larger) visible
in both positions. One polyp in two patients was
excluded because of invisibility on the three dimensional
display. These polyps in the rectum were both hidden
behind the balloon of the inserted catheter and could
only be seen in a two-dimensional read.
One patient was excluded because 12 of the 14 polyps
were situated in the rectum and sigmoid. Here there
could be doubt about the correctness of the exact linking
of the polyps seen in colonoscopy and colonography
since many morphologically less specific polyps were
seen. In total, 16 polyps in three patients were excluded.
Of the remaining 37 polyps, 26 (70%) were sessile, 8
(22%) were pedunculated and 3 (8%) were flat. 24 polyps
were 5–9 mm, 11 polyps were 10–14 mm and 2 polyps
were larger than 15 mm. These were both carcinomas
(5%). Of the remaining polyps 14 (38%) were adenomas,
10 (27%) were non-adenomas and of the remaining
polyps histology was not obtained.
Two polyps were situated in the descending colon and
four polyps in the rectum. In the remaining four
segments, 7 to 9 polyps were situated. Four of the 29
patients had undergone a hemicolectomy.
The median absolute match error was 14 mm (range 0–
59 mm, average 20 mm) either proximal or distal from
the actual polyp in prone position. In the observer study,
70% (26/37) of the polyps was directly visible in prone
view. Of these directly visible polyps, 20 polyps (54%)
were seen instantly without flying though the colon
(category I), the remaining 6 polyps (16%) were also
instantly visible, but the observer chose to move the
virtual camera over a small distance to verify the polyp
(category II). The remaining 11 polyps (30%) could not be
seen instantly but all were found after a flight through
the colon within 150 mm of the matching position
(category III).
When the NDAC method was applied to our CT-
colonography examinations, this resulted in a median

742 The British Journal of Radiology, September 2006

Automated matching of supine and prone CT-colonography
Figure 3. (a) The absolute match error of the method based on the maximal alignment of the centre-line plotted against the
relative location of a polyp along the centre-line in supine position. The 0 on the x-axis corresponds to caecum and 1 to rectum.
(b) The absolute match error of the NDAC method plotted against the relative location of a polyp along the centre-line in supine
position. The 0 on the x-axis corresponds to caecum and 1 to rectum. The three circled arrows indicate polyps from three
different patients with a relatively large absolute match error.
absolute match error of 16 mm (range 0.5–105 mm,
average 26 mm), slightly larger than the other method.
There was no significant difference (p50.502) in match
error between both methods.
Half way along the centre-line three polyps were
visible, with a substantial difference in match error in
favour of the method of maximal alignment of centre-
lines (Figure 3). These polyps were all situated in the
transverse colon. None of these three patients had
undergone a hemicolectomy.
Discussion
This study shows that the matching of prone and
supine CT-colonography examinations is feasible; the
median absolute match error was 14 mm and 70% of all
polyps were visible after the match was made.
The overall difference in median absolute match error
between both methods was small (2 mm), although half
way along the centre-line there were polyps with sub-
stantial differences in match error (larger with NDAC). All
these polyps were situated in the transverse colon.
The transverse colon is the largest intraperitoneally
situated colonic segment [4] that can move relatively
freely in the abdominal cavity. We assume that this is the
most mobile part of the colon, based on our observations
of the three-dimensional overviews in both positions.
The fact that polyps with a substantial difference in
match error are situated in this segment might suggest
that the method that relies on maximal alignment of
centre-lines has an advantage in this mobile region.
The methods of matching prone and supine positions
were evaluated based on the assumption that the
included polyps were (immobile) true polyps and,
second, that the actual link between these polyps in
both positions was correct. In case there was doubt about
the status of the polyp and linking, the polyp was
excluded. With these exclusion criteria a very robust
reference standard was created. Consequently, using this
The British Journal of Radiology, September 2006
reference standard we were able to determine most
precisely the matching quality of the methods.
Matching is important in determining whether a
suspected lesion is a polyp or faecal material. A rule of
thumb is that a suspected lesion with an unchanged
position in prone and supine is most likely a polyp,
while a change of position is related to stool. An accurate
matching algorithm is therefore helpful to reduce the
number of false positive findings in an efficient way.
Pitfalls in diagnosis can be caused by adherent stool
mimicking a polyp, in case faecal tagging is absent or
inadequate. On the other hand, pedunculated polyps
(large stalk) or polyps situated in mobile segments may
move (or seem to move) when both positions are
compared. Therefore, when assessing whether a suspi-
cious finding is a polyp, it is important not only to take
the location into account, but also homogeneity and
morphology of the lesion.
Incorporated in a CAD algorithm, this matching tool
could also be used to reduce the number of false positive
findings. Näppi et al [5] used a region-growing scheme
with distance calculations to divide the colonic lumen
into overlapping segments that match in the supine and
prone data sets. Polyp candidates detected by means of a
CAD scheme were eliminated if they could be seen in
only one of the two corresponding segments.
A limiting factor for generalizing the visibility of the
matched polyps is the use of a 3D unfolded cube method
for reviewing the data. Using this method, 70% of the
polyps could be seen instantly when the match was
made. This review method may have had a positive
influence on the percentage of polyps that were seen
instantly, since the unfolded cube display method
creates a 360 ˚ view without major distortion of the image.
Although evaluated here in a 3D-setting, the method of
matching prone and supine datasets can be applied in a 2D-
setting as well. This can be done on condition that a centre-
line is calculated through both corresponding colons.
Theoretically, the matching quality of the method of
maximal alignment is not altered by absence of a part of
743

the beginning or ending of a centre-line in one position
(e.g. caecum or rectum). In other words, the absence of
baseline adjustment, a condition in other clinically
evaluated methods [3, 6, 7], will not have a large
influence on the match error in this new method. This
baseline adjustment may not be possible in patients with
an inadequately distended or insufficiently cleansed
caecum or rectum in one position. Although this
condition was not present in any of the data sets used,
this proposition was confirmed in a test case. Here we
removed 50 cm of the beginning of the centre-line
without large consequences for the match error in the
remaining colon segments.
The design of this feasibility study, with selected
polyps as fixed reference points, precludes an assess-
ment of the clinical value of the matching tool. Further
study should assess matching quality and time efficiency
of the matching tool in a series of unselected CT-
colonography examinations.
From this study we conclude that automated prone-
supine matching of CT-colonography studies is feasible
and the difference in median absolute match error from
the NDAC method was small and not significant, although
half way along the centre-line (transverse colon) more
pronounced differences were seen in some cases.
744
A H de Vries, R Truyen, J van der Peijl et al
References
1. van Gelder RE, Nio CY, Florie J, Bartelsman JF, Snel P, De
Jager SW, et al. Computed tomographic colonography
compared with colonoscopy in patients at increased risk
for colorectal cancer. Gastroenterology 2004;127:41–8.
2. Rouchka E. Dynamic programming. Stockholm
Bioinformatics Center 2004 June 1. Available from: http://
www.sbc.su.se/,per/molbioinfo2001/dynprog/dynamic.
html [Accessed 23 March 2006].
3. Iordanescu G, Summers RM. Automated centerline for
computed tomography colonography. Acad Radiol
2003;10:1291–301.
4. Saunders BP, Fukumoto M, Halligan S, Jobling C, Moussa
ME, Bartram CI, et al. Why is colonoscopy more difficult in
women? Gastrointest Endosc 1996;43(2 Pt 1):124–6.
5. Nappi J, Okamura A, Frimmel H, Dachman A, Yoshida H.
Region-based supine-prone correspondence for the reduc-
tion of false-positive CAD polyp candidates in CT colono-
graphy. Acad Radiol 2005;12:695–707.
6. Acar B, Napel S, Paik DS, Li P, Yee J, Beaulieu CF.
Registration of supine and prone CT colonography data:
Method and evaluation. Radiology 2001;221:332.
7. Li P, Napel S, Acar B, Paik DS, Jeffrey RB, Beaulieu CF.
Registration of central paths and colonic polyps between
supine and prone scans in computed tomography colono-
graphy: pilot study. Med Phys 2004;31:2912–23.
The British Journal of Radiology, September 2006
The British Journal of Radiology, 79 (2006), 745–755

Improved motion compensation in 3D-CT using respiratory-

correlated segment reconstruction: diagnostic and radiotherapy
applications
S MORI, PhD, M ENDO, PhD, R KOHNO, PhD and S MINOHARA, PhD

National Institute of Radiological Sciences, Inage-ku, Chiba 263-8555, Japan

ABSTRACT. Conventional respiratory-gated CT and four-dimensional CT (4DCT) are

disadvantaged by their low temporal resolution, which results in the inclusion of
anatomic motion-induced artefacts. These represent a significant source of error both
in radiotherapy treatment planning for the thorax and upper abdomen and in
diagnostic procedures. In particular, temporal resolution and image quality are vitally
important to accurate diagnosis and the minimization of planning target volume
margin due to respiratory motion. To improve both temporal resolution and signal-to-
noise ratio (SNR), we developed a respiratory-correlated segment reconstruction
method (RS) and adapted it to the Feldkamp-Davis-Kress algorithm (FDK) with a 256
multidetector row CT (256MDCT). The 256MDCT scans approximately 100 mm in the
craniocaudal direction with a 0.5 mm slice thickness in one rotation. Data acquisition
for the RS-FDK relies on the assistance of a respiratory sensing system operating in cine
scan mode (continuous axial scan with the table stationary). We evaluated the RS-FDK
for volume accuracy and image noise in a phantom study with the 256MDCT and
compared results with those for a full scan (FS-FDK), which is usually employed in
conventional 4DCT and in half scan (HS-FDK). Results showed that the RS-FDK gave a
more accurate volume than the others and had the same SNR as the FS-FDK. In a Received 16 August 2005
subsequent animal study, we demonstrated a practical sorting process for projection Revised 6 January 2006
data which was unaffected by variations in respiratory period, and found that the RS- Accepted 24 January 2006
FDK gave the clearest visualization among the three algorithms of the margins of the
DOI: 10.1259/bjr/63249054
liver and pulmonary vessels. In summary, the RS-FDK algorithm provides multi-phase
images with higher temporal resolution and better SNR. This method should prove ’ 2006 The British Institute of
useful when combined with new radiotherapeutic and diagnostic techniques. Radiology

Continuing rapid progress in computer hardware and
software has led to better radiation therapy planning and
dramatic improvements in delivery. New types of
conformal planning and delivery technology, of which
intensity-modulated radiation therapy (IMRT) is a promi-
nent example [1, 2], have the potential to achieve a much
higher degree of target conformity and normal tissue
sparing than existing treatment techniques. Since higher
target conformity generally requires more accurate defini-
tion of the target, the new radiation therapy requires a
reduction in the target volume error arising from the
respiratory motion of organs such as the lung and liver.
Moreover, the human respiratory cycle is not strictly
regular, but varies in amplitude and period from one cycle
to the next [3–5], and greater time is spent in exhalation
than in inhalation [6, 7]. These complexities hamper the
accuracy of radiotherapy, for example in the determina-
tion of peak-to-peak amplitude of external chest motion.
Voluntary or imposed breath-hold techniques have been
proposed to reduce or eliminate these effects of breathing
during both imaging and radiotherapy treatment [8–11],
but these prolong treatment and in any case cannot be
tolerated by many patients. Respiratory motion during CT
acquisition may produce artefacts that resemble disease
symptoms [12], and these remain a significant source of
error in radiotherapy treatment planning for the thorax
and upper abdomen [12–14].
Many investigators have introduced automatic respira-
tory-gated CT [15, 16] or four-dimensional CT (4DCT) [17,
18] acquisition techniques which provide wide craniocau-
dal (CC) coverage beyond the CT detector width without
image gaps during free breathing, and have adapted them
to radiation therapy [19–21]. In the respiratory-gated CT
technique, CT images are taken by an axial scan gated to
respiratory signals from a respiratory motion detection
system. In contrast, the 4DCT technique obtains CT images
by cine scan and sorts the CT images obtained in the same
respiratory phase. The temporal resolution, therefore, is
determined by the gantry rotation time. Due to their
relatively low temporal resolution, however, these techni-
ques do not remove motion artefacts completely.
Segment reconstruction correlated to a physiological
signal was first developed using electrocardiograms
(ECG), and ECG-correlated reconstruction has been
widely used to delineate coronary arteries with intrave-
nous injection of contrast agent [22–26]. To our knowledge,
however, few papers have appeared on respiratory-
correlated segment reconstruction. Koenig et al [27]
introduced data acquisition processing on several
half-turns in order to reduce the dose delivered per

The British Journal of Radiology, September 2006 745


rotation with the same signal-to-noise ratio (SNR). Sonke
et al [28] reported the use of respiratory-correlated cone
beam CT integrated with a linear accelerator, but their
rotation speed of approximately 4 min is considerably
longer than that with our 256 multidetector row CT
(256MDCT), rendering their experience not directly
applicable to the present study.
To improve temporal resolution and image quality, and
thereby minimize planning target volume (PTV) margin
due to respiratory motion, we developed a respiratory-
correlated segment reconstruction method (RS) and
adapted it to the Feldkamp-Davis-Kress algorithm (FDK)
[29] with a 256MDCT [30]. The RS-FDK algorithm
provides multiphase images with higher temporal resolu-
tion and a better SNR than the conventional respiratory-
gated CT and 4DCT techniques, and should therefore
prove useful when combined with new radiotherapy
techniques such as four-dimensional (4D) radiation
therapy [31–34]. Here, with a view to clinical utilization,
we evaluated the use of RS-FDK with the 256MDCT in
phantom and animal studies.
Materials and methods
Prototype 256 multidetector row CT-scanner
(256MDCT)
The 256MDCT was developed at the National Institute
of Radiological Sciences (NIRS) [30, 35]. A wide-area
two-dimensional (2D) detector was designed on the basis
of existing CT technology [36] and mounted on the
gantry frame of an advanced MDCT. The number of
elements was 912 (transverse)6256 (craniocaudal), each
of approximately 0.5 mm60.5 mm at the centre of
rotation. Rotation time was 1.0 s. Owing to disk storage
limitations, continuous scan time was limited to 15 s for
acquisition at 256 mm60.5 mm, 30 s at
128 mm61.0 mm, and 60 s at 64 mm62.0 mm. The
scanner could scan approximately 100 mm in the
craniocaudal direction in one rotation. Data sampling
rate was 900 views/s, and dynamic range of the A–D
converter was 16 bits. The detector element consisted of a
scintillator and photodiode, the former being the same as
that used in the MDCT (Toshiba Aquilion; Toshiba
Medical Systems, Japan). A FDK algorithm was used for
reconstruction. Reconstruction of volume data of
51265126128 voxels with a high-speed image processor
in a field-programmable gate array (FPGA)-based archi-
tecture took less than 1 s.
External respiratory signal tracking system and
signal processing
Generally, two kinds of external respiratory signal
tracking system are used. These have tagging points with
artificial markers that reflect or emit light, termed
‘‘passive’’ and ‘‘active’’. The latter type is routinely used
for gated irradiation and CT acquisition during heavy ion
radiotherapy at our institution [20]. A number of problems
with these markers have been reported, such as a residual
error between the marker position and surface point of
interest, patient setup, and a reduction in resolution in
746
S Mori, M Endo, R Kohno and S Minohara
orthogonal directions [37–39]. Here, however, the passive
marker was used to obtain respiratory phase alone, and
not absolute distance of motion.
The respiratory sensing system consisted of a work-
station (Dell, Roundrock, TX) equipped with a real-time
digital video analyser, in-house gating software and user
interface within the PV-WAVE programming package
(Visual Numerics, San Ramon, CA), and a charged-
coupled-device (CCD) camera (XC-EI50; Sony Corp.,
Tokyo, Japan) with an attached infrared illuminator
(OTR, Tokyo, Japan) (Figure 1a). SNR of the camera signal
was 60 dB and vibration tolerance was 10 G (20–200 Hz).
Since the reflected marker was captured around the centre
of view, distortion was negligible. The respiratory sensing
system was affixed to the patient table in this study. To
perform an RS scan using conventional CT, movement of
the couch to the adjacent position is necessary to obtain the
next respiratory cycle. Because this process is repeated until
the entire scan range is completed, the respiratory sensing
system may induce significant mechanical disturbances
that would be amplified by zoom optics. In contrast, couch
movement is not necessary for RS-FDK using the
256MDCT, and the vibration induced by rotation of the
gantry does not significantly disturb the couch. Mechanical
disturbance with this equipment is therefore negligible.
The cycle of respiratory phase was monitored with a
CCD camera focused on a reflective marker seal on the
patient’s chest or abdominal area. Video signals from the
camera were transferred to the workstation and software
routines were run to determine respiratory function from
the detected marker position. Tracking of the marker was
facilitated by observation of the seal at very high contrast.
A spline curve was fitted to the respiratory function to
reduce noise in the detected signal and to correlate the
sampling rate of 30 frame/s for respiratory function with
the 900 frames/s for the projection data [40].
Reconstruction algorithms
Three types of reconstruction algorithm based on the
FDK were used, namely full scan (FS-FDK), half scan
(HS-FDK) and respiratory-correlated segment FDK (RS-
FDK). The usual reconstruction algorithm, FS-FDK, uses
a uniform weight over the 2p of the projections, and is
usually employed in conventional respiratory-gated CT.
To increase temporal resolution, we employed the HS-
FDK, which uses only p plus the fan angle of the
projections with a Parker weighting function applied
prior to the filtered back projection operation [41].
Results showed an effective scan time of 500 ms (central
ray) when a 1.0 s rotation mode was used. Details of the
HS-FDK have been described elsewhere [42].
With regard to the RS-FDK, its essential concept is to
sort projection data in the same respiratory phase, rather
than reconstructing CT images as in the conventional
respiratory-gated CT technique. Data acquisition for
anatomical sites subject to breathing motion such as the
lung and abdomen relies on the assistance of the
respiratory sensing system to reduce the impact of
respiratory motion. The 256MDCT uses a cine scan mode
(continuous axial scanning with the table stationary) to
acquire all respiratory phase projection data (Figure 1a).
The respiratory phase is determined from the respiratory
The British Journal of Radiology, September 2006
Improved motion compensation in 3D-CT

Figure 1. Sorting process in the

respiratory-correlated segment
reconstruction method (RS). (a) A
respiratory signal is obtained from
the infrared video camera.
(b) Projection data with projection
angle are obtained from the
256MDCT. (c) Relationship between
projection angle and moving object
position. (d) Projection data for the
same respiratory phase are sorted to
obtain four data sections (sections
A–D). (e) The four projection-
quadrant sections – each section
corresponds to p/2 of projections.
(f) RS-FDK images are obtained
after generating cone-beam back
projection (FDK).

signal collected by the respiratory sensing system (see data PRS(b) as follows:
below) during the cine scan (Figure 1b,c).
8
Projection angle shift per respiratory period, PS, > 2p
>
> p(b) 0ƒbƒ
between rotation time Trot and respiratory period Tres is >
> Ns
>
>
defined as: > :::::::
>
<
PRS (b)~ p(bz2np) 2np ƒbƒ 2(nz1)p ð3Þ
1 >
> Ns Ns
PS~ mod (Tres ,Trot ) ð1Þ >
> :::::::
Trot >
>
>
>
>
: p(bz2(Ns {1)p) 2(Ns {1)p ƒbƒ2p
N s
where PS is normalized by Trot, and is given by a fraction
of rotation. The number of sections, Ns, is then obtained as:
where p(b) is the projection datum obtained from the
8 cine scan at projection angle b, and n is an integer
> 1 PS~0
>
>
{1  between 0 and Ns–1.
>
< PS 0vPSv1=2
The 2p projection data set thereby acquired is used for
Ns ~ 2 PS~1=2 ð2Þ
>
> FDK reconstruction. For RS-FDK, the total scan time TS
>h
> i
: ð1{PSÞ{1 1=2vPSv1 and temporal resolution TR are given as follows:

TR~Trot =Ns ð4Þ

[ ] denotes floor function, which yields the greatest
integer lower or equal. Projection data for the same
respiratory phase are sorted to obtain RS-FDK projection TS~Tres :Ns ð5Þ

The British Journal of Radiology, September 2006 747


We noted that it was necessary to desynchronize the
respiratory period from the gantry rotation time, because if
the respiratory period is a harmonic (PS50) or subharmo-
nic (PS50.5) of this time, the restricted temporal resolution
Trot or Trot/2 is obtained, respectively.
In this study, the data set was divided into four
projection-quadrant sections (Ns54), with each section
therefore corresponding to p/2 of the projections (sections
A–D) (Figure 1d), covering 2p (Figure 1e). Since temporal
resolution was increased in proportion to the number of
sections (Equation 5), temporal resolution of the RS-FDK
was four and two times as high as those of the FS-FDK and
HS-FDK, respectively. However, total scan time was also
increased in proportion to the number of sections
(Equation 5). To avoid high patient doses, total scan time
was conservatively limited to 17 s, less than that proposed
by the Institutional Review Board (IRB) of the National
Institute of Radiological Sciences (NIRS).
Multiphase images were obtained by shifting the start
projection angle for FS-FDK and HS-FDK. Both algo-
rithms provide volumetric multiphase images over the
total acquisition period. In contrast, RS-FDK provided
multiphase images by sorting projection data for each
respiratory phase, and thereby produced images for only
one respiratory cycle.
Image noise ratios for FS-FDK, HS-FDK and RS-FDK
are given in Appendix 1.
Phantom study
Volume accuracy results for RS-FDK were compared
with those for FS-FDK and HS-FDK following evaluation
with a moving phantom designed for the 256MDCT
(Figure 2). The phantom contained a 30 mm diameter
acrylic ball as the target volume placed on a moving table
connected to a mechanically driven motor with speed
adjustment. To simulate human respiratory motion,
sinusoidal movement was employed in an oblique
direction by setting the moving phantom at 45 ˚ to the
CC direction. Motion distance and time period of the
moving phantom were 40 mm and 3.74 s, respectively.
The phantom position was denoted using a motion
function f(t) as follows:
f (t)~20:sin(2p:t=3:74½s
)½mm
ð6Þ
 pffiffiffi pffiffiffi
ðLR(t),AP(t),CC(t)Þ~ f (t)= 2,0,f (t)= 2 ð7Þ
748
S Mori, M Endo, R Kohno and S Minohara
where t denotes time [s], and (LR(t), AP(t), CC(t)) are
relative coordinates of the ball’s position where LR, AP
and CC denote the left–right, anterior–posterior and
craniocaudal directions, respectively. A slow-motion
phase was defined as the ball positioned at f(t1)5¡20
[mm] and having a displacement in 1 s of 6.8 mm. A
fast-motion phase was defined as the ball positioned at
f(t2)50 [mm] and having a displacement in 1 s of
30.0 mm.
The ball volume was obtained by applying a threshold
to images, where the actual volume of the ball was
14 137 mm3 (54/36p6153). Volume error in the moving
phantom was assessed as the percent ratio (%) of volume
from images of the moving phantom to the actual
volume.
Scan conditions were 120 kV, 150 mA using a 0.5 mm
256 row detector and 15 s acquisition time. Reconstruction
parameters were a voxel size of 0.35 mm6
0.35 mm60.35 mm, matrix size of 51265126256 and
0.35 mm reconstruction increment. The convolution kernel
was a standard body kernel (FC10).
Animal study
Image quality of RS-FDK was evaluated in an animal
study in comparison with those for FS-FDK and HS-
FDK. Four domestic pigs were used to simulate a human
model. The animals were aged 21–23 weeks, weighed 20–
25 kg and had a diameter and circumference of 130 mm
and 590 mm, respectively. Although diameter and
circumference were smaller than those of adult humans,
pigs were selected owing to their ease of handling, and
well-developed interlobular septa and anatomic struc-
tures that are similar to those of the human lung [43]. All
animal procedures were approved by the IRB of the
NIRS. The pigs were given an intramuscular injection of
a mixture of 10 mg kg21 of ketamine hydrochloride
(Sankyo Yell, Tokyo, Japan) and 7 mg kg21 of xylazine
2% (Bayer, Tokyo, Japan) and sedated and ventilated
with a respiratory pump with isoflurane 2–2.5%. The
breathing cycle was adjusted to 4.2–4.3 s. A video camera
was focused on a reflective marker seal set on the
interseptum to track respiratory motion and obtain the
respiratory signal.
Scan conditions were 120 kV, 200 mA using a 1.0 mm
128 row detector and 17 s acquisition time, and cine scan
mode. Reconstruction parameters were a voxel size of
Figure 2. Photographs of the mov-
ing phantom. (a) The phantom was
set 45 ˚ to the caudocranial direction
and connected to a mechanically
driven motor with speed adjust-
ment. (b) The phantom contained a
30 mm diameter acrylic ball as the
target volume placed on a moving
table.
The British Journal of Radiology, September 2006
Improved motion compensation in 3D-CT

Figure 3. Motion functions for oblique motion (a). CT images of the 30 mm diameter acrylic ball in oblique motion in (b) the
slow- and (c) fast-motion phases. Images in each of the vertical frames from top to bottom were reconstructed with FS-FDK, HS-
FDK and RS-FDK. The left–right, anterior–posterior and caudocranial directions are denoted as LR, AP and CC, respectively.
Window level is –373 HU and window width is 1190 HU for all images.

0.47 mm60.47 mm60.47 mm, matrix size of 5126
5126216 and 0.47 mm reconstruction increment. The
convolution kernel was FC10. Effective dose was
estimated as 38.6 mSv (52.27 mSv s21617 s).
Results
Phantom study
Motion functions (a) and reconstructed images for (b)
slow- and (c) fast-motion phases are shown in Figure 3.
With regard to the slow-motion phase, FS-FDK images
were severely degraded and distorted while those with
HS-FDK did not seem spherical. In contrast, the RS-FDK
images visualized the ball as spherical and were of better
quality than the HS-FDK images due to their better
temporal resolution. Furthermore, image quality with
RS-FDK was the same for three different cross-sections.
For the fast-motion phase, although the reconstructed
images showed greater degradation than those of the
slow-motion phase for all algorithms, image quality with
RS-FDK was nevertheless better than with the others. In
particular, the RS-FDK images show the ball as round in
both coronal and sagittal sections. The magnitude of
artefacts differed between the transverse and long-
itudinal sections.
Volume using the threshold of 50% of CT number in
the static image was 14 123 mm3, which nearly equals the
actual value of 14 137 mm3. The volume ratios decreased
as threshold CT number increased. The threshold was
fixed at 50% of the difference between the CT number of
the ball and the background CT number, because this
gives the approximate actual volume of the ball.
Figure 4 shows the volume percentage of the ball
obtained from images of oblique motion. The volume
percentage decreased from unity for all cases, with that
by RS-FDK larger than with the other two, and that by

The British Journal of Radiology, September 2006 749


Figure 4. Volume percentage obtained from the axial image
where the ball motion is in the oblique direction.
FS-FDK being the smallest. For the fast phase, the
volume ratio by RS-FDK was 5% larger than that for
HS-FDK and more than 40% larger than that by FS-FDK.
Animal study
The phase of respiratory signals for all animals as
sensed by the respiratory sensing system is shown in
Figures 5 and 6, and respiratory periods for all animals
are summarized in Table 1. Animal respiration was
controlled with almost complete regularity by the
ventilator, with only a few irregular periodic motions
due to voluntary breathing. Breathing motion at peak
inhalation was slightly faster than that at peak exhala-
tion. This is similar to the behaviour seen for human
respiratory signals, which are not sinusoidal; dwell time
at peak exhalation is often longer than is shown by a
sinusoidal signal.
Figure 7 shows projection data of the pig in the AP
direction at peak exhalation, mid inhalation and peak
750
S Mori, M Endo, R Kohno and S Minohara
inhalation phases. The length of respiratory motion was
approximately 17 mm in the CC direction.
Coronal and sagittal images for each respiratory phase
(pig 1) are shown in Figures 8 and 9, respectively. These
images captured each respiratory phase exactly. The
margins of the liver and the pulmonary vessels were
sharper in the RS-FDK than in the FS-FDK and HS-FDK
images. FS-FDK gave the worst image quality due to its
low temporal resolution. For mid-exhalation and mid-
inhalation, geometrical distortion of the pulmonary
vessels and interlobular septa became visible in FS-
FDK and HS-FDK images, and the latter exhibited a
streak artefact on the ribs in spite of their good temporal
resolution.
These results indicate that the RS-FDK algorithm
provides the best image quality for all respiratory phases
due to its good temporal resolution.
Discussion
In the present study, we developed the RS-FDK
algorithm to allow the precise capture of anatomical
structures under free breathing conditions, and com-
pared results with those for two other reconstruction
algorithms, FS-FDK and HS-FDK, in experiments using a
moving phantom and in animal studies. Results showed
that RS-FDK gives better temporal resolution than HS-
FDK and FS-FDK and the equivalent SNR as FS-FDK
(Appendix 1).
Although conventional respiratory-gated CT and
4DCT techniques provide wide CC coverage beyond
the detector width for the same respiratory phase, the
inclusion of motion artefact is not completely mitigated
due to the relatively low temporal resolution. This
disadvantage means that the actual position of a moving
object in radiotherapy treatment planning CT is uncer-
tain, as is the volume of a moving object relative to its
actual volume. These uncertainties result in planning
errors. Furthermore, respiratory motion during irradia-
tion can cause the radiation beam to miss part of the
target itself. To avoid these errors, PTV is defined with a
margin of several centimetres added to the clinical target
volume (CTV); but this in turn carries the risk that an
excessively wide margin will unnecessarily irradiate
normal tissue. It is therefore necessary to minimize the
Figure 5. Respiratory signal
obtained from the external respira-
tory signal tracking system (pig 1).
The solid line shows raw respiratory
data and the broken one shows a
spline fitting to the raw data.
The British Journal of Radiology, September 2006
Improved motion compensation in 3D-CT

Figure 6. Respiratory signal for four pigs; (a) pig 1, (b) pig 2, (c) pig 3 and (d) pig 4. Four respiratory cycles are overlapped to
show reproducibility of each cycle.

PTV margin. As seen in the phantom study, the analysis of the trade-off between temporal resolution and
reconstructed images did not visualize the actual shape image quality (e.g. artefact and image noise). In contrast,
due to distortion or motion artefacts. These errors may RS provides multiphase images with better temporal
result in the delivery of an insufficient dose to the target resolution than HS and the same SNR as FS, and may
volume or an excess dose to normal tissue. supply useful information for 4D radiation therapy
HS and RS allow the precise imaging of a periodically planning. With the multiphase images we may make a
moving object. The shorter acquisition time with HS, dynamic treatment plan that enables irradiation with
however, may lead to a lower SNR than that provided by continually adjusted beam shape to match the respira-
a conventional acquisition time [44], necessitating careful tory phase. If such a dynamic treatment could be
realised, it brings greater accuracy in radiation therapy
than respiratory-gated radiation therapy, which irradi-
Table 1. Respiratory cycle for four pigs ates at the most stable point, such as peak exhalation [45].
The limitation of RS is its higher patient dose. This
Animal Respiratory Average SD
cycle (s)
occurs because the CT scans the same position con-
tinuously, and because the effective dose increases in
1 2 3 4 proportion to acquisition time. Acquisition time should
1 4.24 4.45 4.35 4.30 4.33 0.09 be minimized even for patients receiving radiation
2 4.20 4.24 4.40 4.20 4.26 0.09 therapy. Because the respiratory period at small PS
3 4.21 4.32 4.46 4.38 4.34 0.10
values such as 0.1 requires a longer scan time, rotation
4 4.13 4.23 4.33 4.28 4.24 0.09
time should be adjusted so that total scan time is

The British Journal of Radiology, September 2006 751


Figure 7. Images of projection data in the anterior–posterior direction. (a) Peak exhalation, (b) mid inhalation and (c) peak
inhalation. Window level is –373 HU and window width is 1190 HU for all images.
decreased. If this is difficult, an alternative is to scan in
one or two respiratory cycles, and thereafter to scan in
appropriate respiratory phases only. Devices providing
such functionality are essential to avoiding excessive
patient dose. In the present study, scan time was
minimized by adjusting respiratory period to 3.74 s for
the phantom study and to 4.2–4.3 s for the animal study.
Many authors have reported the limitations of the
abdomen as an external surrogate for the respiratory
phase, namely that longer CT scanning time under free
breathing and the placement of the marker result in
inconsistency between the position of the marker and
752
S Mori, M Endo, R Kohno and S Minohara
Figure 8. Coronal images obtained
at four respiratory phases (peak
inhalation, mid exhalation, peak
exhalation and mid inhalation).
Reconstruction increment is 0.4 mm
and slice thickness is 0.4 mm. (a) FS-
FDK, (b) HS-FDK and (c) RS-FDK.
Window level is –373 HU and win-
dow width is 1190 HU for all images.
Figure 9. Sagittal images obtained
at four respiratory phases (peak
inhalation, mid exhalation, peak
exhalation and mid inhalation).
Reconstruction increment is 0.4 mm
and slice thickness is 0.4 mm. (a) FS-
FDK, (b) HS-FDK and (c) RS-FDK.
Window level is –373 HU and win-
dow width is 1190 HU for all images.
internal anatomy motion [16, 18, 46]. Lujan et al [6] have
reported that the motion of the diaphragm due to
respiration is predominantly in the CC direction and is
periodic but asymmetric, with more time spent at the
end of expiration than at the end of inspiration. This may
lead to erroneous prediction of the dose delivered to the
patient, and when examination time is prolonged, may
result in the degradation of RS-FDK image quality.
However, examination time in the present study was
only 17 s, and the pigs were sedated and ventilated with
a respiratory pump with an almost constant respiratory
cycle. The relationship between the respiratory signal
The British Journal of Radiology, September 2006
Improved motion compensation in 3D-CT
and target motion is therefore better correlated using the
amplitude and cycle of the respiratory cycle. Since the
breathing characteristics of patients are not always as
reproducible as those of sedated pigs, the respiratory
control discussed below is necessary when the RS-FDK is
used in free-breathing patients.
Moreover, it should be remembered that a tumour
located in a lower lobe of the lung has a larger range of
tumour motion than one in an upper lobe due to
proximity to the diaphragm [47]. This may also impact
RS-FDK in clinical use, which requires correction of the
correlation between respiration and target motion, such
as by control of the patient’s breathing by operator
guidance [8–10, 16, 48, 49] or by means of an occlusion
valve [11, 50, 51]. It will be necessary to control for this
time dependency of respiratory asymmetry through
suitable correction methods.
With regard to the projection data sorting process,
adaptation of spline curve fitting may induce over-
lapping or a lack of projection data at adjacent quadrant
sections (A–D in Figure 1). However, the spline curve of
respiratory function showed good regulation (Table 1).
In this study, respiratory period for all animals was
controlled with a ventilator to between 4.2 s to 4.3 s.
Generally, if the period shifts between respirations,
errors in the sorting process may lead to incorrect
classification of the projection data into the wrong
projection data section, and thereby result in image
artefacts in the resultant RS-FDK. The same problem is
seen in cardiac imaging with ECG gating [52, 53]. In this
case, compensation of the projection data must be done
by extending or shortening the range of the quadrant in
both quadrant sections. Here, however, thanks to the use
of the ventilator, the respiratory period was controlled
and these errors were minimized, resulting in acceptable
image quality.
We have carried out clinical trials at NIRS using
carbon-ion beams [54]. Gated irradiation is a useful
method and several findings have been reported [20, 21].
Depth dose distributions of charged particles exhibit a
strong Bragg peak at the very end of the maximum
distance that a charged particle travels in a tissue (range),
beyond which the dose very rapidly falls to zero [55].
Charged particle beam radiotherapy may therefore be
valuable when the organs at risk are closely proximal to
the target volume. However, the presence of a range-
shortening tissue inhomogeneity such as the soft tissue
moving into the lung would not only reduce the target
dose from the value in the stretched Bragg peak to
essentially zero, but the dose would also put nearby
organs at risk. Range is therefore an important factor in
charged particle radiation therapy, even more so than in
photon beam therapy. The use of RS-FDK in charged
particle therapy as well as 4D radiation therapy should
allow a significant increase in dose distribution accuracy
and provide sufficient information to allow the mini-
mization of the PTV margin in radiation therapy
planning. We are now investigating these points.
References
1. Sternick ES. The theory and practice of intensity modulated
radiation therapy. Madison, WI: Advanced Medical
Publishing, 1997.
The British Journal of Radiology, September 2006
2. Webb S. Intensity-modulated radiation therapy. Bristol and
Philadelphia: IOP Publishing Ltd, 2000.
3. Liang P, Pandit J, Robbins P. Non-stationarity of breath-by-
breath ventilation and approaches to modeling the phe-
nomenon. In: Semple S, Adams L, Whipp B, editors.
Modeling and control of ventilation. New York, NY:
Plenum, 1995:117–21.
4. Hugelin A, Vibert J. Is the respiratory rhythm multistable in
man respiratory control. In: Swanson G, Grodins F, RI H,
editors. A modeling perspective. New York, NY: Plenum,
1989:353–9.
5. Benchetrit G. Breathing pattern in humans: diversity and
individuality. Respir Physiol 2000;122:123–9.
6. Lujan AE, Balter JM, Ten Haken RK. A method for
incorporating organ motion due to breathing into 3D dose
calculations in the liver: sensitivity to variations in motion.
Med Phys 2003;30:2643–9.
7. Murphy MJ. Tracking moving organs in real time. Semin
Radiat Oncol 2004;14:91–100.
8. Hanley J, Debois MM, Mah D, Mageras GS, Raben A,
Rosenzweig K, et al. Deep inspiration breathhold technique
for lung tumors: the potential value of target immobiliza-
tion and reduced lung density in dose escalation. Int J
Radiat Oncol Biol Phys 1999;45:603–11.
9. Mah D, Hanley J, Rosenzweig KE, Yorke E, Braban L, Ling
CC, et al. Technical aspects of the deep inspiration breath-
hold technique in the treatment of thoracic cancer. Int J
Radiat Oncol Biol Phys 2000;48:1175–85.
10. Rosenzweig KE, Hanley J, Mah D, Mageras G, Hunt M,
Toner S, et al. The deep inspiration breath-hold technique in
the treatment of inoperable non-small-cell lung cancer. Int J
Radiat Oncol Biol Phys 2000;48:81–7.
11. Dawson LA, Brock KK, Kazanjian S, Fitch D, McGinn CJ,
Lawrence TS, et al. The reproducibility of organ position
using active breathing control (ABC) during liver radio-
therapy. Int J Radiat Oncol Biol Phys 2001;51:1410–21.
12. Tarver RD, Conces DJ, Godwin JD. Motion artifacts on CT
simulate bronchiectasis. AJR Am J Roentogenol
1988;151:1117–9.
13. Mayo JR, Muller NL, Henkelman RM. The double-fissure
sign: a motion artifact on thin-section CT scans. Radiology
1987;165:580–1.
14. Shimizu S, Shirato H, Kagei K, Nishioka T, Bo X, Dosaka-
Akita H, et al. Impact of respiratory movement on the
computed tomographic images of small lung tumors in
three-dimensional (3D) radiotherapy. Int J Radiat Oncol
Biol Phys 2000;46:1127–33.
15. Kini VR, Keall PJ, Vedam SS, Arthur DW, Kavanagh BD,
Cardinale RM, et al. Preliminary results from a study of a
respiratory motion tracking system: underestimation of
target volume with conventional CT simulation. Int J Radiat
Oncol Biol Phys 2000;48(Suppl. 1):164.
16. Mageras GS, Yorke E, Rosenzweig K, Braban L, Keatley E,
Ford E, et al. Fluoroscopic evaluation of diaphragmatic
motion reduction with a respiratory gated radiotherapy
system. J Appl Clin Med Phys 2001;2:191–200.
17. Low DA, Nystrom M, Kalinin E, Parikh P, Dempsey JF,
Bradley JD, et al. A method for the reconstruction of four-
dimensional synchronized CT scans acquired during free
breathing. Med Phys 2003;30:1254–63.
18. Vedam SS, Kini VR, Keall PJ, Ramakrishnan V, Mostafavi
H, Mohan R. Quantifying the predictability of diaphragm
motion during respiration with a noninvasive external
marker. Med Phys 2003;30:505–13.
19. Kubo HD, Len PM, Minohara S, Mostafavi H. Breathing-
synchronized radiotherapy program at the University of
California Davis Cancer Center. Med Phys 2000;27:346–53.
20. Minohara S, Kanai T, Endo M, Noda K, Kanazawa M.
Respiratory gated irradiation system for heavy-ion radio-
therapy. Int J Radiat Oncol Biol Phys 2000;47:1097–103.
753

21. Ohara K, Okumura T, Akisada M, Inada T, Mori T, Yokota
H, et al. Irradiation synchronized with respiration gate. Int J
Radiat Oncol Biol Phys 1989;17:853–7.
22. Grass M, Manzke R, Nielsen T, Koken P, Proksa R,
Natanzon M, et al. Helical cardiac cone beam reconstruction
using retrospective ECG gating. Phys Med Biol
2003;48:3069–84.
23. Kachelriess M, Kalender WA. Electrocardiogram-correlated
image reconstruction from subsecond spiral computed
tomography scans of the heart. Med Phys 1998;25:2417–31.
24. Kachelriess M, Knaup M, Kalender WA. Extended parallel
backprojection for standard three-dimensional and phase-
correlated four-dimensional axial and spiral cone-beam CT
with arbitrary pitch, arbitrary cone-angle, and 100% dose
usage. Med Phys 2004;31:1623–41.
25. Kachelriess M, Ulzheimer S, Kalender WA. ECG-correlated
image reconstruction from subsecond multi-slice spiral CT
scans of the heart. Med Phys 2000;27:1881–902.
26. Taguchi K, Anno H. High temporal resolution for multislice
helical computed tomography. Med Phys 2000;27:861–72.
27. Koenig A, Grangeat P, Bonnet S, Hugonnard P. Dynamic
reconstruction for radiotherapy planning. In: CARS 2002:
Computer Assisted Radiology and Surgery; 2002; Paris, 20–
26 June 2002: Berlin: Springer; 2002.
28. Sonke J, Zijp L, Remeijer P, Herk M. Respiratory correlated
cone beam CT. Med Phys 2005;32:1176–86.
29. Feldkamp LA, Davis LC, Kress JW. Practical cone-beam
algorithm. J Opt Soc Am 1984;1:612–9.
30. Endo M, Mori S, Tsunoo T, Kandatsu S, Tanada S, Aradate
H, et al. Development and performance evaluation of the
first model of 4D CT-scanner. IEEE Trans Nucl Sci
2003;50:1667–71.
31. Keall PJ, Joshi S, Vedam SS, Siebers JV, Kini VR, Mohan R.
Four-dimensional radiotherapy planning for DMLC-based
respiratory motion tracking. Med Phys 2005;32:942–51.
32. Keall PJ, Kini VR, Vedam SS, Mohan R. Motion adaptive x-
ray therapy: a feasibility study. Phys Med Biol 2001;46:1–10.
33. Ozhasoglu C, Murphy MJ. Issues in respiratory motion
compensation during external-beam radiotherapy. Int J
Radiat Oncol Biol Phys 2002;52:1389–99.
34. Vedam SS, Keall PJ, Docef A, Todor DA, Kini VR, Mohan R.
Predicting respiratory motion for four-dimensional radio-
therapy. Med Phys 2004;31:2274–83.
35. Mori S, Endo M, Tsunoo T, Kandatsu S, Tanada S, Aradatte
H, et al. Physical performance evaluation of a 256-slice CT-
scanner for four-dimensional imaging. Med Phys
2004;31:1348–56.
36. Saito Y, Aradate H, Igarashi K, Ide H. Large area 2-
dimensional detector for real-time 3-dimensional CT (4D
CT). Proc SPIE 2001;4320:775–82.
37. Moore C, Lilley F, Sauret V, Lalor M, Burton D. Opto-
electronic sensing of body surface topology changes during
radiotherapy for rectal cancer. Int J Radiat Oncol Biol Phys
2003;56:248–58.
38. Baroni G, Ferrigno G, Orecchia R, Pedotti A. Real-time
three-dimensional motion analysis for patient positioning
verification. Radiother Oncol 2000;54:21–7.
39. Baroni G, Troia A, Troia A, Orecchia R, Pedotti A. [Opto-
electronic techniques and 3D body surface reconstruction
for the control of patient positioning in the radiotherapy of
breast cancer]. Radiol Med (Torino) 2001;102:168–77.
40. Zeng R, Fessler JA, Balter JM. Respiratory motion estima-
tion from slowly rotating x-ray projections: theory and
simulation. Med Phys 2005;32:984–91.
41. Parker DL. Optimal short scan convolution reconstruction
for fanbeam CT. Med Phys 1982;9:254–7.
42. Taguchi K. Temporal resolution and the evaluation of
candidate algorithms for four-dimensional CT. Med Phys
2003;30:640–50.
754
S Mori, M Endo, R Kohno and S Minohara
43. Murata K, Herman PG, Khan A, Todo G, Pipman Y, Luber JM.
Intralobular distribution of oleic acid-induced pulmonary
edema in the pig. Evaluation by high-resolution CT. Invest
Radiol 1989;24:647–53.
44. Gould RG. Ultrafast computed tomography in cardiac
imaging: principles and practice. Futura: Mount Krisco; 1992.
45. Sontag M. Respiratory gated radiotherapy. Med Phys
1999:1112.
46. Vedam SS, Keall PJ, Kini VR, Mohan R. Determining
parameters for respiration-gated radiotherapy. Med Phys
2001;28:2139–46.
47. Hoisak JD, Sixel KE, Tirona R, Cheung PC, Pignol JP.
Correlation of lung tumor motion with external surrogate
indicators of respiration. Int J Radiat Oncol Biol Phys
2004;60:1298–306.
48. Balter JM, Lam KL, McGinn CJ, Lawrence TS, Ten Haken
RK. Improvement of CT-based treatment-planning models
of abdominal targets using static exhale imaging. Int J
Radiat Oncol Biol Phys 1998;41:939–43.
49. Kim DJ, Murray BR, Halperin R, Roa WH. Held-breath self-
gating technique for radiotherapy of non-small-cell lung
cancer: a feasibility study. Int J Radiat Oncol Biol Phys
2001;49:43–9.
50. Stromberg JS, Sharpe MB, Kim LH, Kini VR, Jaffray DA,
Martinez AA, et al. Active breathing control (ABC) for
Hodgkin’s disease: reduction in normal tissue irradiation
with deep inspiration and implications for treatment. Int J
Radiat Oncol Biol Phys 2000;48:797–806.
51. Wong JW, Sharpe MB, Jaffray DA, Kini VR, Robertson JM,
Stromberg JS, et al. The use of active breathing control
(ABC) to reduce margin for breathing motion. Int J Radiat
Oncol Biol Phys 1999;44:911–9.
52. Dhanantwari AC, Stergiopoulos S, Iakovidis I. Correcting
organ motion artifacts in x-ray CT medical imaging systems
by adaptive processing. I. Theory. Med Phys 2001;28:1562–76.
53. Dhanantwari AC, Stergiopoulos S, Zamboglou N, Baltas D,
Vogt HG, Karangelis G. Correcting organ motion artifacts
in x-ray CT systems based on tracking of motion phase by
the spatial overlap correlator. II. Experimental study. Med
Phys 2001;28:1577–96.
54. Kanai T, Endo M, Minohara S, Kato H, Tsujii H. Estimating
uncertainties of the geometrical range of particle radio-
therapy during respiration. Int J Radiat Oncol Biol Phys
1999;44:201–10.
55. Bragg WH, Kleeman R. On the ionisation curves of radium.
Phil Mag 1904;8:726–38.
Appendix 1
The image noise ratios for FS-FDK, HS-FDK and RS-
FDK were calculated as follows. If sP is the noise
magnitude (the standard deviation) of the measured
projection data, noise magnitude of FS-FDK sFS, HS-FDK
sHS and RS-FDK sRS can be calculated by noise
propagation analysis by integrating the squared weights
over the range of 2p:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð 2p ð cm ffi
1 2 2
sFS ~ ( ) sp dbdc
0 {cm 2
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð ð
2p cm
sHS ~ v2HS (c,b)sp 2 dbdc~1:36sFS
0 {cm
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð 2p ð cm ffi
1 2 2
sRS ~ ( ) sp dbdc~sFS
0 {cm 2
The British Journal of Radiology, September 2006
Improved motion compensation in 3D-CT

times larger than that for FS-FDK, whereas RS-FDK gives

8
> p b the same magnitude of image noise as FS-FDK.
>
> sin2 ( ) b [ ½0:2cm {2p

>
> 4 cm {c
<
vHS (c,b)~ 1 b [ ½2cm {2p,p{2c

>
>
>
> p pz2c {b
> 2
: sin (
m
) b [ ½p{2c,pz2cm

4 cm zc

where vHS(c, b) is the Parker weighting, cm is the fan

angle, b denotes the range of the projection angle and c
denotes the ray-sum angle within a projection. RS also
uses a uniform weight over the 2p of projections, as for
FS. Figure A1 shows 3D images of weighting functions of Figure A1. 3D images of the weighting functions of (a) FS-
FS-FDK, HS-FDK and RS-FDK. From the above analysis, FDK and RS-FDK, and (b) HS-FDK. b denotes the projection
the magnitude of the image noise for HS-FDK is 1.36 angle and c denotes the ray-sum angle within a projection.

The British Journal of Radiology, September 2006 755

 The British Journal of Radiology, 79 (2006), 756–761

A mathematical model for patient skin dose assessment in

cardiac catheterization procedures
R E MORRELL, PhD, MIPEM and A T ROGERS, MSc, MIPEM

Medical Physics and Clinical Engineering, City Hospital Campus, Nottingham University Hospitals
NHS Trust, Hucknall Road, Nottingham NG5 1PB, UK

ABSTRACT. A mathematical model has been developed for the assessment of patient
skin doses from cardiac catheterization procedures. This uses exposure and projection
data stored in the DICOM image files. Since these contain only information about the Received 22 April 2005
acquisition runs, a correction is needed to estimate and include the contribution from Revised 17 May 2006
fluoroscopy. Maximum skin doses calculated by the model were found to correlate well Accepted 20 June 2006
with those measured on Kodak EDR2 film. Three methods for including the
DOI: 10.1259/bjr/57359387
contribution from fluoroscopy were investigated, and all successfully identified
patients receiving skin doses in excess of 1 Gy. It is hoped to automate this tool for ’ 2006 The British Institute of
routine assessment of skin doses in our cardiac catheterization laboratories. Radiology

Cardiac catheterization procedures can result in high
radiation doses to the patient’s skin, sufficient to cause
deterministic effects [1–6]. In order to comply with the
recommendations of the International Commission on
Radiological Protection [7], a robust method for routine
assessment of patient skin dose is needed.
Dose–area product (DAP) alone is not an adequate
indicator of skin dose, since in many cases there is a poor
correlation between the two quantities [8–12]. Slow radio-
graphic film can be used to measure the dose distribution
across the patient’s skin. However, its useful range is
limited by its saturation point. Kodak EDR2 film (Eastman
Kodak Company, Rochester, NY) has previously been
used for skin dosimetry [12–14], and is the least sensitive of
the slow films designed for portal imaging and quality
control in radiotherapy. However, it saturates at around
1 Gy to 1.5 Gy, depending on the processing conditions
applied [13–15]. There is now a growing range of
‘‘Gafchromic’’ films (International Specialty Products,
Wayne, NJ), which saturate at higher radiation doses and
do not require processing, but at present these are much
more expensive than those in the Kodak range. Smaller
detectors such as thermoluminescent dosemeters and
scintillation detectors cover only a small area, so are liable
to miss the region of maximum dose.
In modern cardiac catheterization laboratories, infor-
mation about the exposure parameters is stored in the
DICOM file for each image series. The DICOM standard
[16] specifies fields for data such as the number of frames
in the series, angulation of the X-ray imaging unit,
detector position and field size, imaging mode, beam
energy and tube current. If these fields are populated,
they allow the position and magnitude of the radiation
dose to the patient’s skin to be estimated, for each image
series. Since fluoroscopic images are not usually stored,
this detailed dose information is only available for the
acquisition runs.
The purpose of this study was to develop a mathema-
tical model to calculate the skin dose distribution across
the patient’s back, using the exposure and projection data
stored in the image files. Maximum skin doses predicted
by the model were compared with film dosimetry
measurements for coronary angiography (CA) and percu-
taneous transluminal coronary angioplasty (PTCA) proce-
dures. Three methods for including an estimated
contribution from fluoroscopy were investigated.
Method
The dose model was developed in Matlab version 7
(The Mathworks, Inc., Natick, MA). It was designed to
extract and use the data from DICOM image files from
an Integris H5000F C-arm imaging unit (Philips Medical
Systems, Best, The Netherlands).
The dose model
The exposure information is first extracted from all of
the image files in the patient examination folder, and
written to a structure array. An example of these data is
shown in Table 1. Each row relates to one file, i.e. one
image series. The columns contain the series or run
number (Run), number of frames in that series (Frames),
imaging protocol (Protocol), peak beam kilovoltage
(kVp), tube current (mA), pulse width (ms), primary
and secondary angles describing the orientation of the
imaging unit in degrees (Ang1 and Ang2), source to
image distance in millimetres (SID) and detector field
size in millimetres (II).
‘‘Ang1’’ relates to left–right rotation of the gantry, and
is positive for left anterior oblique and left lateral views,
when the detector is towards the patient’s left-hand side.
‘‘Ang2’’ describes craniocaudal rotation, and is positive
when the detector is closer to the patient’s head.
Additional data required by the model are:
N centreheight – the height of the centre of rotation of
the C-arm from the floor (107 cm).

756 The British Journal of Radiology, September 2006

A mathematical model for patient skin dose assessment

Table 1. Example of exposure data extracted from DICOM image files

Run Frames Protocol kVp mA ms Ang1 Ang2 SID II

1 62 12.5 FPS Coronary 81 835 7 38 22 987 170

2 52 12.5 FPS Coronary 83 870 7 42 –28 1031 170
3 54 12.5 FPS Coronary 78 720 7 –27 –20 1019 170
4 53 12.5 FPS Coronary 74 710 7 –27 27 1019 170
5 53 12.5 FPS Coronary 76 758 7 –1 –28 1034 170
6 48 12.5 FPS Coronary 71 625 7 –1 24 983 170
7 61 12.5 FPS Coronary 83 860 7 40 0 976 170
8 44 12.5 FPS Coronary 73 650 7 –30 0 962 170
9 8 12.5 FPS Coronary 74 720 7 –28 –20 1037 170
10 29 12.5 FPS Coronary 81 889 7 –28 –20 1037 170
kVp, peak beam kilovoltage (kV); mA, tube current (mA); ms, pulse width (ms); Ang1, left-right angulation of imaging unit;
Ang2, craniocaudal angulation of imaging unit; SID, focal spot to detector distance (mm); II, field size at detector face (mm).

N radius – the distance from the focal spot to the centre The distance (ffd, in centimetres) from the focal spot to
of the C-arm (77.5 cm). each point in the plane is then calculated using
N couchheight – the height of the couch above the floor. Pythagoras’ law. The dose at each point on the film, in
This can be adjusted during the examination and is the absence of beam collimation, would be:
not recorded by the imaging unit. A typical value
(87 cm) was estimated from the recorded SIDs and
measured field sizes for a selection of patient films kVp 2 |mA|ms|frames
dose~ESDRphantom | ð5Þ
from a previous dose survey [12]. ffd2
N ESDRphantom – the dose rate (ESDR) at the entrance
surface of a 20 cm polymethyl methacrylate (PMMA) where kVp2, mA, ms and frames are taken from the
phantom, normalized for kVp, mA, pulse time (ms), DICOM file for that particular run.
frame rate and focus to PMMA distance in centi- Beam limitation is calculated assuming square radia-
metres (fsd), as shown in Equation (1). This has a tion fields, and ignoring any secondary collimation
value of 5.9761025 mGycm2 (kVp2 mA ms)21 per applied by the user. The angle (h) between the beam
frame. central axis and its collimated outer edge is calculated
from the field size and SID.
ESDR|fsd2 ‘‘SIDcoll’’ is the radius of a sphere centred at the focal
ESDRphantom ~ ð1Þ
kVp 2 |mA|ms|frame rate spot that passes through the four corners of the
collimated field, at the detector face. It is calculated from
The entrance dose rate to the phantom was measured field size and SID, using Pythagoras’ law. The positions
using a Radcal 9010 series dosemeter, with a 60 cm3 of the field corners at the detector face are calculated in
ionization chamber (Radcal, Monrovia, CA), as shown in Cartesian coordinates, using Equations (6) to (8).
Figure 1. The standard clinical acquisition mode (12.5
FPS Coronary) was used. This mode does not employ
any copper filtration.
Since the imaging unit stores no information about
couch position, the model assumes that the couch
remains at a fixed height throughout the procedure,
and that no panning is used.
The dose distribution is calculated at 1 mm intervals in
a plane at the height of the couch top, that extends from
230 cm to +30 cm in the x (cross-couch) direction and
220 cm to +20 cm in the y (parallel to couch axis)
direction. Positive x is defined towards the right-hand
side of the patient, and positive y towards the patient’s
head, so that the dose map appears as though viewing
the patient from behind.
For each acquisition run, the position of the focal spot
(xspot, yspot, zspot) is calculated in Cartesian coordinates,
using Equations (2) to (4).

xspot ~radius| sinðAng1Þ| cosðAng2Þ ð2Þ

yspot ~{radius| sinðAng2Þ ð3Þ

Figure 1. Set-up for determining the entrance dose rate to
phantom. PMMA, polymethyl methacrylate; fsd, as used in
zspot ~centreheight{radius| cosðAng1Þ| cosðAng2Þ ð4Þ Equation (1).

The British Journal of Radiology, September 2006 757


ycoll ~yspot zSIDcoll | sinðAng2+hÞ
xcoll ~xspot {SIDcoll | sinðAng1+hÞ| cosðAng2+hÞ ð7Þ
zcoll ~zspot zSIDcoll | cosðAng1+hÞ| cosðAng2+hÞ ð8Þ
They are then translated from the detector face onto the
plane of the couch top, by means of scaling.

measured maximum doses were computed. This subset
xcoll {xspot
included 14 CA and 20 PTCA procedures.
xcouch ~xspot z
| couchheight{zspot
zcoll {zspot

Results
ycoll {yspot

ycouch ~yspot z
| couchheight{zspot ð10Þ
zcoll {zspot
To apply the collimation, the values of the dose array
are set to zero at all locations outside the trapezium
formed by these four points.
The completed dose maps from each run are summed
to give a total dose array, which is then displayed as a
filled contour map.
The model described above considers only the acquisi-
tion run data, which is stored in the image files. In clinical
practice, some examinations involve extended fluoroscopy
times, and where fluoroscopy is performed primarily in
one projection this can raise the maximum skin dose
considerably. Three options for estimating and including
the contribution from fluoroscopy were investigated.
1. Dose–area product. The DAP was calculated from the
exposure parameters for each acquisition run, and
summed to give acquisition DAP for the whole proce-
dure. The final dose map was multiplied by the ratio of
displayed total DAP to calculated acquisition DAP.
2. Fluoroscopy time. The entrance dose rate to the 20 cm
PMMA phantom in the standard fluoroscopy mode
(Low Fluoro), at the 23 cm field size, with an SID of
100 cm and fsd of 67 cm has been measured as
approximately 40 mGy min21. The fluoroscopy time
(in minutes) was divided by the number of runs,
multiplied by this dose rate, and added to the dose
array for each run, before applying beam collimation.
3. Concentration factor. An attempt was made to relate
maximum skin dose to DAP for individual proce-
dures, by means of a ‘‘concentration factor’’. This was
simply the ratio of the maximum calculated skin dose,
to the sum of doses at all points in the dose map. Peak
skin dose was predicted for each patient using the
product of DAP and concentration factor, together
with a constant that forced the regression line
between predicted and measured doses to have a
gradient of one.
Comparison of calculated and measured doses
A dose survey has previously been carried out in our
cardiac catheterization laboratory, using Kodak EDR2
film to measure the skin dose distribution across the
758
R E Morrell and A T Rogers
ð6Þ patient’s back [12]. The current study employed the dose
model to calculate a dose map for each of the patients
included in the survey. The maximum doses determined
by the two methods were compared.
The calculated skin dose maps were first compared
visually with the films. Patients were excluded from the
study if the region of maximum dose predicted by the
model was outside the area of the film.
For the subset of films that showed no saturation, the
Pearson correlation coefficients between calculated and
ð9Þ
Figure 2 shows an example of the data output from the
skin dose model. A visual comparison with the corre-
sponding dosimetry film (Figure 3) shows that the
region of maximum dose has been correctly identified.
The fields visible on the calculated dose map can be
broadly matched with some of those appearing on the
film. However, fields arising from purely fluoroscopic
exposures are seen only on the film, since the model has
no projection data for fluoroscopy.
Figure 4 shows the calculated doses for procedures
resulting in film saturation, for each version of the dose
model. The film saturates at about 1 Gy, so saturation
indicates a skin dose of 1 Gy or more [15].
The most basic form of the dose model, which includes
only acquisition data, generally predicted doses of at
least 800 mGy for these procedures. However, in one
case, a dose of only 332 mGy was computed. Although
this patient had a maximum skin dose of at least 1 Gy
and may therefore be at risk of deterministic effects, the
model would not have identified her as having received
a dose that may be of concern.
On applying any of the correction methods to include
the contribution from fluoroscopy, the calculated dose to
this patient increased to more than 1 Gy. When corrected
by fluoroscopy time or concentration factor, the model
successfully identified all procedures that resulted in
film saturation. When adjusted by the ratio of displayed
Figure 2. Example output from dose model.
The British Journal of Radiology, September 2006
A mathematical model for patient skin dose assessment

Figure 5. Calculated versus measured doses, using only

acquisition data.

Figure 3. Dosimetry film for the patient whose calculated

skin dose map is shown in Figure 2.

to calculated DAP, the model predicted one of these

patients to have a maximum skin dose of 928 mGy, and
all others to have maximum doses of more than 1 Gy.
Figures 5–8 show calculated versus measured doses
for all the procedures that did not result in film
saturation, i.e. for which measured doses were less than
1 Gy. The error bars show the expected uncertainty in
film dosimetry measurements [12]. In each case, a trend
line has been fitted, that passes through the origin. The Figure 6. Calculated versus measured doses, using dose–
equation of the trend line and the square of the Pearson area product (DAP) to estimate the contribution from
correlation coefficient are shown on each figure. fluoroscopy.
When using only the acquisition data, the gradient of
the trend line was 0.735 (Figure 5), suggesting that the line to 1 (Figure 8). This method gave the strongest
model underestimates doses by 26% on average. There correlation between calculated and measured doses
was a strong correlation between calculated and mea- (R250.735).
sured doses (R250.672).
On using DAP or fluoroscopy time to estimate the
contribution from fluoroscopy the gradient of the trend Discussion
line was closer to 1 (Figures 6 and 7). The strength of the
correlation also increased (R250.708 and R250.716, For doses of up to 1 Gy, maximum skin doses
respectively). computed using the model correlate well with those
The product of concentration factor and DAP was measured using Kodak EDR2 film. Unlike film dosime-
multiplied by 181 000, to force the gradient of the trend try, the model has no limit on the magnitude of the doses
that can be evaluated.
Using only the acquisition data stored in the DICOM
files can lead to large errors in dosimetry for examinations

Figure 4. Calculated doses for procedures that resulted in

film saturation, using each version of the dose model. Fluoro
Correction 1 uses DAP, 2 uses fluoroscopy time, 3 uses
concentration factor to estimate the contribution from Figure 7. Calculated versus measured doses, using fluoro-
fluoroscopy. scopy time to estimate the contribution from fluoroscopy.

The British Journal of Radiology, September 2006 759


Figure 8. Calculated versus measured doses, using concen-
tration factor to estimate the contribution from fluoroscopy.
involving long fluoroscopy times and few projections.
Three options for including an estimated contribution
from fluoroscopy were explored. Each improved the
accuracy of the model for procedures involving a large
proportion of fluoroscopy, and predicted doses of at least
928 mGy for all procedures resulting in film saturation.
The method using displayed DAP together with a
‘‘concentration factor’’ gave the strongest correlation
between calculated and measured doses. Using this
method, 95% of calculated doses were within
¡ 270 mGy of measured doses, for films showing no
saturation.
Each method for including the contribution from
fluoroscopy requires additional information (either
DAP or fluoroscopy time) that is not stored in the image
files and must therefore be obtained from another source.
At present, displayed DAP and fluoroscopy time for all
procedures are manually recorded in a book, and in an
Oracle database (Oracle Corporation, Redwood Shores,
CA). In the near future, these may be automatically
stored in a new catheterization laboratory information
system.
It is hoped that clinical application of the dose model
can eventually be completely automated – to extract the
relevant data from the image files and obtain the
fluoroscopy time from the information system or
database as each examination is completed, to calculate
the skin dose distribution and alert a member of staff if
the maximum dose exceeds a certain threshold. This
would enable staff to follow up at-risk patients by
examining their skin and warning them of potential
effects before they leave the hospital.
The accuracy of the model is limited by a number of
unknown variables, about which assumptions have had
to be made. No exposure factors are stored for fluoro-
scopy, as has been previously discussed, necessitating an
estimation of the contribution to total dose from the
fluoroscopic parts of the procedure. The actual couch
position for each procedure is not known, so it is
necessary to assume a certain couch height, and to
ignore any horizontal movement. No information about
secondary collimation or the use of the wedge filter is
available, so these must be assumed not to have been
used.
Improved accuracy is dependent on manufacturers
building in the facilities to make this information
available. Whilst den Boer et al [17] and Chugh et al
[18] have worked with equipment manufacturers to
760
R E Morrell and A T Rogers
develop real-time skin dose monitoring software that
utilizes much of this information, none is currently
available for purchase.
As with film dosimetry, the model considers only
those radiation beams that pass through the plane of the
couch, and ignores any contributions to skin dose from
lateral views. A potential improvement would be to use a
three-dimensional model of the patient to estimate skin
doses over the whole surface of the thorax. However, this
would require couch positioning data, to achieve any
degree of accuracy.
Conclusion
A dose model has been developed to calculate the skin
dose distribution across the patient’s back from cardiac
catheterization procedures. This utilizes the exposure
and projection data stored in the DICOM image files, as
well as DAP or fluoroscopy time. Maximum doses
calculated by the model correlated well with those
measured using Kodak EDR2 film. After applying a
correction to include the estimated contribution from
fluoroscopy, the model successfully identified patients
receiving skin doses in excess of 1 Gy. It is hoped to
automate the dose model for use as a routine dosimetry
tool in the cardiac catheterization laboratory.
The accuracy of the model is limited by several
unknown variables that are not recorded by the imaging
system. Improvement of dosimetric accuracy is depen-
dent upon manufacturers developing methods for stor-
ing and accessing this information.
Acknowledgments
We are grateful to Dr Nick Gibson for his assistance
with the Matlab code, and to Prof. Alan Perkins for
helpful discussions.
References
1. Kawakami T, Saito R, Miyazaki S. Chronic radiodermatitis
following repeated percutaneous transluminal coronary
angioplasty. Br J Dermatol 1999;141:150–3.
2. D’Incan M, Roger H, le Boudec MCF, Souteyrand P.
Radiodermatitis following cardiac catheterization. Arch
Dermatol 1997;133:242–3.
3. Dehen L, Vilmer C, Humiliere C, Corcos T, Pentousis D,
Ollivaud L, et al. Chronic radiodermatitis following cardiac
catheterisation: a report of two cases and a brief review of
the literature. Heart 1999;81:308–12.
4. Shope TB. Radiation-induced skin injuries from fluoro-
scopy. Radiographics 1996;16:1195–9.
5. Vanõ E, Goicolea J, Galvan C, Gonzalez L, Meiggs L, Ten JI,
et al. Skin radiation injuries in patients following repeated
coronary angioplasty procedures. Br J Radiol 2001;74:
1023–31.
6. Vanõ E, Arranz L, Sastre JM, Moro C, Ledo A, Garate MT,
et al. Dosimetric and radiation protection considerations
based on some cases of patient skin injuries in interven-
tional cardiology. Br J Radiol 1998;71:510–6.
7. International Commission on Radiological Protection.
Report 85: Avoidance of radiation injuries from medical
interventional procedures. Ann ICRP 2000;30:45–7.
8. van de Putte S, Verhaegen F, Taeymans Y, Thierens H.
Correlation of patient skin doses in cardiac interventional
The British Journal of Radiology, September 2006
A mathematical model for patient skin dose assessment
radiology with dose-area product. Br J Radiol 2000;73:
504–13.
9. Vanõ E, Gonzalez L, Ten JI, Fernandez JM, Guibelalde E,
Macaya C. Skin dose and dose-area product values for
interventional cardiology procedures. Br J Radiol
2001;74:48–55.
10. Waite JC, Fitzgerald M. An assessment of methods for
monitoring entrance surface dose in fluoroscopically
guided interventional procedures. Radiat Prot Dosim
2001;94:89–92.
11. Delichas MG, Psarrakos K, Giannoglou G, Molyvda-
Athanasopoulou E, Hatziioannou K, Papanastassiou E.
Skin doses to patients undergoing coronary angiography
in a Greek hospital. Radiat Prot Dosim 2005;113:
449–52.
12. Morrell RE, Rogers AT. Kodak EDR2 film for patient skin
dose assessment in cardiac catheterization procedures. Br J
Radiol 2006;79:603–7.
13. Guibelalde E, Vanõ E, Gonzalez L, Prieto C, Fernandez JM,
Ten JI. Practical aspects for the evaluation of skin doses in
The British Journal of Radiology, September 2006
interventional cardiology using a new slow film. Br J Radiol
2003;76:332–6.
14. Vanõ E, Prieto C, Fernandez JM, Gonzalez L, Sabate M,
Galvan C. Skin dose and dose-area product values in
patients undergoing intracoronary brachytherapy. Br J
Radiol 2003;76:32–8.
15. Morrell RE, Rogers A. Calibration of Kodak EDR2 film for
patient skin dose assessment in cardiac catheterization
procedures. Phys Med Biol 2004;49:5559–70.
16. National Electrical Manufacturers Association. Digital
Imaging and Communications in Medicine (DICOM). Part
6: Data Dictionary (PS 3.6-2006). Rosslyn, Virginia: National
Electrical Manufacturers Association, 2006.
17. den Boer A, de Feijter PJ, Serruys PW, Roelandt JRTC. Real-
time quantification and display of skin radiation during
coronary angiography. Circulation 2001;104:1779–84.
18. Chugh K, Dinu P, Bednarek DR, Wobschall D, Rudin S,
Hoffmann K, et al. A computer-graphic display for real-
time operator feedback during interventional x-ray proce-
dures. Proc Soc Photo-Opt Instrum Eng 2004;5367:464–73.
761
 The British Journal of Radiology, 79 (2006), 762–765

Nurse-led central venous catheter service: Christie experience

K GOPAL, MRCP, FRCR, L FITZSIMMONS, RGN and J A L LAWRANCE, MRCP, FRCR

Christie Hospital, Wilmslow Road, Manchester M20 4BX, UK

ABSTRACT. The aim of this study was to evaluate the success and complication rate of
Nurse-led subclavian central venous catheter (CVC) insertion using the landmark
technique. A prospective study was performed on all subclavian CVC insertion between
13/01/03 to 01/07/03. Data recorded included indication for catheter insertion, type of
catheter, complications during the procedure and patient satisfaction. A total of 348
subclavian cuffed tunnelled catheters were inserted over the study period. The age
group ranged from 31 years to 84 years with a mean age of 53 years. This included 192
women and 156 men. The majority of CVC insertions were through the right subclavian Received 11 July 2005
(79%) and were single lumen (76%). In total, complications were encountered in 48 Revised 21 December 2005
patients (14%). These included misplaced tip in 29 (8%), arterial puncture in 16 (4%), Accepted 8 February 2006
pneumothorax in 3 (1%) and the procedure failed in 3 (1%). Of these multiple
DOI: 10.1259/bjr/79794134
complications were seen in 3 (1%). No interventions were required for the
pneumothoraces or for the arterial punctures. In conclusion, nurse-led subclavian CVC ’ 2006 The British Institute of
placements using the landmark technique are both safe and effective. Radiology

Venous access is a critical issue in the care and
management of patients with a wide variety of malig-
nancies. The insertion of tunnelled central venous
catheters (CVC) for patients requiring long-term venous
access is now common. In the NHS, approximately
200 000 CVC are inserted in adult patients per year [1].
This is partly due to increased usage in cancer patients.
CVC are inserted by surgeons, anaesthetists, interven-
tional radiologists, medical oncologists and nurses. The
British Committee of Standards in Haematology (BCSH)
guidelines [2] state that insertion should be performed by
experienced operators, regardless of speciality. In our
centre, these lines are inserted by a team of experienced
nurses. Nurses undergo a supervised training pro-
gramme for 5 weeks (discussed below) in which they
not only observe but perform at least 50 CVC insertions.
They are then formally assessed to determine whether
they may perform the line insertions independently.
Various studies have been published supporting the
extension of nurse’s roles in this area. One of the main
advantages of nurse-led CVC insertion is that the
procedure can be done at the bedside. Others include
reduced waiting time for patients, stable workforce and
successful outcome due to increased frequency of
practice [3, 4]. Although central venous access devices
are clearly advantageous with respect to delivery of
therapy, their placement and maintenance is not without
potential complications. McBride and colleagues [5]
demonstrated that there is a steep operator learning
curve and the complication rate improves notably with
experience. In recent years, The National Institute of
Clinical Excellence (NICE) has recommended the use of
ultrasound guided placement of CVC [6]. However, this
has been received with a mixed response form various
specialities. At our institution, a large oncology centre in
the UK, central venous access is performed by the
Address correspondence to: Dr J A L Lawrance.
landmark technique led by a team of nurse specialists.
We undertook this study to evaluate the success and
complication rate of nurse-led CVC insertion via the
subclavian route.
Materials and methods
We perform approximately 950 nurse-led CVCs per
year. These include both CVC and peripherally inserted
central catheters (PICC). A prospective study was
performed on all CVC insertion between 13/01/03 and
01/07/03. The following information was recorded,
including: diagnosis and indication for catheter inser-
tion, type and position of the catheter, complications
during the procedure and patient satisfaction.
Insertion technique
Pre-procedure investigations include full blood count
and coagulation profile. Local anaesthesia (5–10 ml 2%
lidocaine) is used to anaesthetise the skin. Prophylactic
intravenous antibiotics are not routinely administered
[7]. Catheters are inserted blindly using the landmark
technique. The right subclavian vein is used unless
contraindicated, e.g. due to thrombosis, infection, left
pulmonary compromise or right sided mastectomy and
radiotherapy. The catheters are secured with ethilon
sutures at the insertion site. The intended catheter tip
position is the distal superior vena cava (SVC), SVC/
right atrial junction or right atrium [8]. Following the
insertion, a chest radiograph is performed to check
catheter position and identify potential complications
such as pneumothorax. As most of the patients received
continuous infusion of chemotherapy as out-patients,
they are taught to look carefully for signs of catheter
infection, blockage or accidental removal. Written

762 The British Journal of Radiology, September 2006

Nurse lead venous catheter service

instructions are provided for reference. Patients were
then asked to fill out an audit questionnaire regarding
the procedure, which included patient satisfaction,
waiting time, etc.
Results
348 CVC were inserted over this 6 month period.
Similar numbers of PICC lines were also inserted, but
these were excluded from the study group. Indications
for catheter insertion include carcinomas such as breast,
gastrointestinal, leukaemia, lymphoma, ovarian and
multiple myeloma (Table 1).
The age group ranged from 31 years to 84 years, with a
mean age of 53 years. This included 192 women and 156
men. The majority of the patients were of normal build
except for 16% who were either obese or emaciated,
which was calculated based on their body mass index
(Table 2). The main indication for CVC insertion in
patients was for chemotherapy. Others include bone
marrow transplant, total parental nutrition, difficult
venous access and combination of the above (Table 3).
The right subclavian route was the preferred route of
insertion unless contraindicated (Table 4). All insertions
Table 1. Subset of disease groups
Subset Frequency (%)
Gastrointestinal 190 (54%)
Breast 38 (11%)
Leukaemia 24 (7%)
Lymphoma 24 (7%)
Ovarian 12 (3%)
Multiple myeloma 8 (3%)
Other 52(15%)
were done using the blind technique under local
anaesthesia. Sedation was not routinely administered
and in our study only 78 patients (22.4%) received
sedation. Two patients had to have the procedure under
general anaesthesia. Single, double or triple lumen
catheters were used. Although multi-lumen catheters
facilitate concurrent administration of different medica-
tions and fluids, we preferred the usage of single lumen
catheters to reduce the rate of infection (Table 5). Nurse-
led CVC insertions are usually performed at the bedside
in the Day Case Unit (Table 6). The procedures are
usually performed as day cases. CVC catheters were
inserted with a single or double needle pass in the
majority of patients (Table 7). Insertion related complica-
tions were noted in a total of 48 patients (14%). There
were three pneumothoraces (1%), and 16 arterial
punctures (4%). None of the former required chest
drains and none of the latter caused significant haemo-
dynamic problems, such as hypotension or tachycardia.
In three patients there was failure of catheter insertion
and an alternative approach, i.e. internal jugular vein was
used. Chest radiographs confirmed tip malposition in 29
patients (8%). In cases where the tip is malpositioned,
catheters are repositioned under fluoroscopy guidance
by the nurse specialists themselves. This can usually be
achieved by a technique developed by one of the authors
(JALL) consisting of a combination of a forced saline
injection and hyperventilation, the combination of which
will usually reposition a contralateral line tip to the SVC.
In cases where this fails, the tip of the CVC line is
withdrawn and manipulated to lie within the distal
SVC/right atrium fluoroscopically. In three patients (1%)
multiple complications were seen (Tables 8 and 9). The
latter patients were hospitalized, but all complications
resolved conservatively. 327 patients (94%) tolerated the
procedure very well. Minor problems, e.g. discomfort

Table 5. Type of line inserted

Table 2. Patients’ body habitus
Lumen
Body habitus Frequency (%)
Single 264 (76%)
Normal 287 (82.5%) Double 71 (20%)
Obese 37 (10.6%) Triple 7 (2%)
Emaciated 21 (6%) Not known 6 (2%)
Not known 3

Table 6. Place of insertion

Table 3. Indication for central venous catheter (CVC)
insertion Location Frequency (%)

Indication Frequency (%) Bed side 342 (98%)

Radiology department 2 (0.5%)
Chemotherapy 312 (90%) Not known 6 (1.5%)
Bone marrow transplant 18 (5%)
Difficult venous access 5 (1.5%)
Total parental nutrition 7 (2%)
Other 6 (1.5%) Table 7. Number of attempts to cannulate the subclavian
vein
No. of passes Frequency (%)
Table 4. Site of insertion 1 168 (48%)
Site Frequency (%)
2 67 (19%)
3 27 (8%)
Right subclavian 274 (79%) 4 6 (2%)
Left subclavian 71 (20%) Several 45 (13%)
Not known 3 (1%) Not known 35 (10%)

The British Journal of Radiology, September 2006 763


Table 8. Summary of insertion related complications
Summary Frequency (%)
No. of patients 348 (100%)
No complications 299 (86%)
Complications 48 (14%)
Table 9. Subdivision of complications
Complication Frequency (%)
Misplaced tip 29 (8%)
Arterial puncture 16 (4%)
Pneumothorax 3 (1%)
Failed procedure 3 (1%)
Multiple complications 3 (1%)
and pains were reported in 6%. Long term follow-up of
these catheters, in relation to thrombosis, was not
assessed in this study as this has been done in a previous
publication from our institution [8].
Discussion
Initially implemented by Broviac et al [9] in 1973 and
subsequently modified by Hickman et al [10] in 1979,
indwelling central venous access devices have revolu-
tionized the care and treatment of the oncology patient.
Various studies have been published supporting the
extension of the nurse’s role in this area. Indeed, BCSH
guidelines state that insertion should be performed by
experienced operators regardless of speciality [2]. In our
institution, all nurses undergo intensive training for
5 weeks. This is divided into Phase 1 and Phase 2, which
last for 1 week and 4 weeks, respectively. During Phase
1, the nurses not only observe the insertion of CVC but
also become familiar with all types of CVC access
devices, their uses and the advantages and disadvan-
tages associated with the various types of catheters. Later
in Phase 2 they undertake at least 50 CVC insertions
under direct supervision and with the assistance of a
mentor. Concluding the training period, a formal
documented assessment of competency is undertaken
by an independent assessor and, once satisfactory,
nurses are then allowed to perform independently.
Nurse placements of CVC at the patient’s bedside mean
that insertions can take place at the optimal time in a
patient’s management as waiting lists for theatres or X-
ray suites are no longer barriers to insertion, which helps
to reduce the patients waiting time for CVC insertion.
The nurse specialists are a stable workforce and the
frequency of practice correlates with successful outcome
[3, 4, 11]. In addition, the numbers of patients encoun-
tered are too great to be accommodated in the X-ray
department with the limited resources available. Nurses
are also being trained to use fluoroscopy, which they use
to reposition CVC lines if blind insertions lead to
misplaced catheter tips. In our centre, CVC are inserted
by a team of experienced nurses under blind percuta-
neous placement. Image guidance is not used for venous
puncture. One of the authors (JALL) has trained the team
in the use of fluoroscopy and the team has free access to
the interventional suite when uncertain about CVC
764
K Gopal, L Fitzsimmons and J A L Lawrance
placement or in patients in whom they suspect potential
problems might arise. In rare cases, patients are referred
for radiologically inserted internal jugular lines, using
ultrasound and fluoroscopic guidance.
The right subclavian vein is punctured and the
operator advances the catheter blindly to the lower part
of the SVC/right atrium. Fluoroscopy is usually only
used to reposition misplaced catheter tips. We have used
the subclavian route historically and our CVC service
was started before NICE guidelines and before the
widespread use of ultrasound guidance for central line
insertion. Due to the high success and acceptable
complication rates over the years, this service has been
continued. Given the high number of line insertions we
perform and our limited resources, we have continued
with this service. Some advantages of bedside CVC
insertion are the fact that they can be inserted on the
ward and patients may be less anxious of the ward
environment compared with operating theatres or X-ray
suites, as well as having cost implications. The main
disadvantage is risk of catheter tip misplacement. As the
use of tunnelled CVC is increasing, it is important to
recognize that their use is not without complications. The
experience of the operator is an important factor in
keeping complications to a minimum. In our study, the
total incidence of insertion related complications was
14%, which compares favourably with other reports [12,
13]. The majority of these complications was related to
catheter tip malposition (8%) with data from other
studies ranging between 12% and 32% [13, 14]. Early
complications such as pneumothorax (1%) and arterial
puncture (4%) are also similar to previous reports in the
literature [15, 16]. It is said that the risk increases with
blind punctures where the operator is solely reliant on
anatomical landmarks for the venipuncture. However, in
our case the risks were well within the limitations and all
the pneumothoraces encountered were small and none
required chest drain insertion. In our study, 1% of
procedures failed and had to be abandoned [17] with
multiple complications seen in a similar number [18],
which are similar in number to those reported earlier.
Sedation is not routinely administered for this proce-
dure and when used we perform this within the remit of
The Safe Sedation Policy prescribed by the trust. The
hospital policy on sedation is widely published in our
department and intranet. As part of the training, the
nurses undergo a formal training on sedation and, in
addition, all nurse practitioners have a Post Graduate
Diploma assessment on their ability to use drugs and
side effects. All the procedures are performed under
indirect supervision of a doctor in the outpatient or
radiology department. To reduce the risk of infection, we
prefer the insertion of single lumen catheters (76%) as it
is apparent that the greater the number of catheter
lumens, the greater the potential for sepsis [19]. The total
life span of the catheters depends on need. For example,
in haematology patients they can remain in situ for
12 months from diagnosis to post-transplantation and in
solid tumour patients receiving chemotherapy, approxi-
mately 4 months. With respect to immediate complica-
tions involved in the placement of CVC, our results are
comparable with those described previously by several
investigators. Although the cost for nurse-led CVC was
not within the remit of the study, we believe that the cost
The British Journal of Radiology, September 2006
Nurse lead venous catheter service
of the procedure as performed by nurses, including the
complications, should be less than radiologically inserted
lines. Given the European Working Time Directive and
the move to roll redesign, this is an excellent example of
the successful deployment of nurses to perform what
was previously the doctors’ role.
In 2002, NICE recommended the use of ultrasound for
the insertion of CVC into the internal jugular vein in
adults and children [6]. This has been received with a
mixed response across various specialities. NICE had
largely concentrated on the complication of inadvertent
arterial puncture and had made the above recommenda-
tion. Ultrasound does reduce the risk, but it is usually
small [20]. Two recent studies make clear that ultrasound
guidance offered incomplete protection against arterial
injury [21] and did not improve the result of right IJV
cannulation, compared with a meticulous landmark
based technique [22]. NICE had considered only one
randomized control study that analysed the effect of
ultrasound and landmark technique in the placement of
subclavian CVC catheters. The operators in both groups
were relatively inexperienced with failure rates as high
as 55% in the landmark technique, which is far higher
that reported in trials (9–19%) [6]. Muhm concluded that
ultrasound guidance improved the number of attempts
per cannulation and successful first attempts for cathe-
terization of the internal jugular vein, but not the
subclavian approach [23]. In addition, the Department
of Health has suggested that the subclavian route was
associated with less infection [24]. Furthermore, none of
the trials quoted by NICE involved nurses [6]. NICE
admits that the landmark method is safe in experienced
hands and operators should maintain their ability to use
the landmark method. The Royal College of
Anaesthetists agrees and advises that utilization of the
landmark method is still an acceptable alternative
whether ultrasound is available or not [25].
Conclusion
The use of CVC is now well established in the
treatment of cancer patients. With its increasing demand,
manpower and facilities are its main limitation. Our
study has indicated that complication rates are well
within acceptable limits using a landmark technique for
the subclavian route. Operator experience, irrespective of
the speciality, is the key to limiting the complication rate
and regular audits are necessary to prove their efficiency.
Nurse-led CVC placements using the landmark techni-
que are both safe and effective.
References
1. Elliot TSJ, Faroqui MH, Armstrong RF, Hanson GC.
Guidelines for good practise in central venous catheterisa-
tion. J Hosp Infect 1994;28:163–76.
2. BCSH Working party. BCSH guidelines on the insertion
and management of central venous lines. Br J Haematol
1997;98:1041–7.
3. Fitzsimmons CL, Gilleece MH, Ranson MR, Wardley A,
Morris C, Scarffe H. Central venous cathter placement:
extending the role of nurses. J R Coll Physicians Lond
1997;31:533–4.
The British Journal of Radiology, September 2006
4. Hamilton H, O’Byrne M, Nicholai L. Central lines inserted
by clinical nurse specialists. Nurs Times 1995;91:38–9.
5. McBride KD, Fisher R, Warnock N, Winfield DA, Reed
MW, Gaines PA. A comparative analysis of radiological and
surgical placement of central venous catheters. Cardiovasc
Intervent Radiol 1997;20:17–22.
6. Guidance of the use of ultrasound locating devices for
placing central venous catheters. National Institute for
Clinical Excellence Technology appraisal guidance No 49,
September 2002. National Institute for Clinical Excellence.
7. Ranson MR, Oppenheim BA, Jackson A, Kamthan AG,
Scarffe JH. Double-blind placebo controlled study of
vancomycin prophylaxis for central venous catheter inser-
tion in cancer patients. J Hosp Infect 1990;15:95–102.
8. Cadman A, Lawrance JA, Fitzsimmons L, et al. To clot or
not to clot? That is the question in central venous catheters.
Clin Radiol 2004;59:349–55.
9. Broviac JW, Cole JJ, Scribner BH. A silicone rubber atrial
catheter for prolonged parental alimentation. Surg Gynecol
Obstet 1973;136:602–6.
10. Hickman RO, Buckner CD, Clift RA, et al. A modified right
atrial catheter for access to the venous system in marrow
transplant recipients. Surg Gynecol Obstet 1979;148:871–5.
11. Hamilton HC. Selecting the correct intravenous device:
nursing assessment. Br J Nurs 2000;9:968–78.
12. Ray S, Stacey R, Imrie M, Filshie J. A review of 560 Hickman
catheter insertions. Anaesthesia 1996;51:981–5.
13. DeJong PCM, Meyenfeldt MR, Rouflart M, Wesdorp RIC,
Soeters PB. Complications of central venous catheterization
of the subclavian vein: the influence of a parenteral
nutrition team. Acta Anaesth Scand 1985;Suppl. 81:48–52..
14. Dewey JC, Robert WH. Aberrant locations and complica-
tions in initial placement of subclavian vein catheters. Arch
Surg 1984;119:293–5.
15. Gray RR. Radiological placement of indwelling central
venous lines for dialysis, TPN and chemotherapy. J Interv
Radiol 1991;6:133–44.
16. McBride KB, Warnock N, Fisher R, Reed NW, Winfield DA,
Gaines PA. A comparative review of surgically and
radiologically placed Hickman catheters. Cardiovasc
Interv Radiol 1994;17:S122.
17. Rosen M, Latto P, Ng S. Percutaneous central venous
catheterisation. London: W.B. Saunders, 1992.
18. Mansfield PF, Hohn DC, Fornage BD, Gregurich MA, Ota
DM. Complications and failures of subclavian-vein cathe-
terization. N Engl J Med 1994;331:1735–8.
19. Eastridge BJ, Lefor AT. Complications of indwelling venous
access devices in cancer patients. J Clin Oncol
1995;13:233–8.
20. Chalmers N. NICE should reconsider its guidance on the
use of ultrasound for placing central venous catheters, (18
December 2002) Rapid response for Muhm. BMJ
2002;325:1373–4.
21. Augoustides JG, Diaz D, Weiner J, et al. Current practise of
internal jugular venous cannulation in a University anaes-
thesia department. J Cardiothor Vasc Anesth
2002;16:567–71.
22. Hayashi H, Amano M. Does ultrasound imaging before
puncture facilitate internal jugular vein cannulation? J
Cardiothor Vasc Anesth 2002;16:572–5.
23. Muhm M. Ultrasound guided central venous access. BMJ
2002;325:1373–4.
24. Pratt RJ, Pellowe CM, Loveday HP, et al. The epic project:
developing national evidence-based guidelines for prevent-
ing healthcare associated infections. Phase1: Guidelines for
preventing hospital acquired infections. J Hosp Infect
2001;47(Suppl):S1–82.
25. Ball DR. Evidence based central venous access. Rapid
response for Muhm. BMJ 2002;325:1373–4.
765
 The British Journal of Radiology, 79 (2006), 766–773

Breast movement during normal and deep breathing, respiratory

training and set up errors: implications for external beam partial
breast irradiation
1 1 1
S CHOPRA, MD, DNB, K A DINSHAW, DMRT, FRCR, R KAMBLE, MSc and 2R SARIN, MD, FRCR

1
Radiation Oncology, Tata Memorial Hospital, Mumbai, Maharashtra and 2ACTREC, Kharghar,
NaviMumbai, Maharashtra, India

ABSTRACT. This study was designed to evaluate interfraction and intrafraction breast
movement and to study the effect of respiratory training on respiratory indices. Five
patients were immobilized in supine position in a vacuum bag and three-dimensional
set up errors, respiratory movement of the breast during normal and deep breathing,
tidal volume and breath hold time were recorded. All patients underwent respiratory
training and all the respiratory indices were re-evaluated at the end of training.
Cumulative maximum movement error (CMME) was calculated by adding directional
maximum set up error and maximum post training movement during normal
breathing. The mean set up deviation was 1.3 mm (SD ¡ 0.5 mm), 1.3 mm (SD ¡
0.3 mm) and 4.4 mm (SD ¡ 2.6 mm) in the mediolateral, superoinferior and
anteroposterior dimensions. Pre-training mean of the maximum marker movement
during normal breathing was 1.07 mm, 1.94 mm and 1.86 mm in the mediolateral,
superoinferior and anteroposterior dimensions. During deep breathing these values
were 2 mm, 5.5 mm and 4.8 mm. While respiratory training had negligible effect on
breast movement during normal breathing, it resulted in a modest reduction during Received 11 January 2006
deep breathing (p50.2). The mean CMME recorded for these patients was 3.4 mm, Revised 25 March 2006
4.5 mm and 7.1 mm in the mediolateral, superoinferior and anteroposterior dimension. Accepted 27 March 2006
Respiratory training also resulted in an increase in breath hold time from a mean of 31 s
DOI: 10.1259/bjr/98024704
to 44 s (p50.04) and tidal volume from a mean of 560 cm3 to 1160 cm3 (p50.04). With
patients immobilized in the vacuum bag the CMMEs are relatively less. Individualized ’ 2006 The British Institute of
directional margins may aid in reduction of planning target volume (PTV). Radiology

Accelerated partial breast irradiation (APBI) is pre-
sently the focus of research in radiotherapy for early
breast cancer (EBC) [1]. In addition to interstitial
brachytherapy, three-dimensional conformal radiation
therapy (3DCRT) and intensity-modulated radiation
therapy (IMRT) are currently being evaluated for APBI
in this group of patients [1–5]. For successful implemen-
tation of these high precision techniques both intrafrac-
tion and interfraction movements of the breast have to be
considered. Efforts have been made to decrease respira-
tory movement by the use of active breathing control
(ABC), respiration correlated cone beam CT (RC-CBCT),
real time tumour tracking and helical tomotherapy while
implementing 3DCRT for APBI [6–10].
While efforts are being made to incorporate image
guidance for precise localization of planning target
volume (PTV) for APBI, there is a paucity of data
regarding cumulative positional uncertainties due to
intrafraction respiratory movement and interfraction-re-
positioning errors with different positioning devices in
patients with breast cancer. This study was designed
with the aim:
Address correspondence to: Prof. Rajiv Sarin, Director, ACTREC,
Tata Memorial Centre, Kharghar, NaviMumbai, Maharashtra, India;
E-mail: rsarin@actrec.res.in.
(a) To ascertain intrafraction respiratory movement of
the breast.
(b) To evaluate interfraction movement of breast clinical
target volume (CTV) by studying the re-positioning
set up errors with individualized vacuum cast
immobilization.
(c) To evaluate the possible efficacy of respiratory
training in reducing breast movement for the purpose
of external beam APBI with 3DCRT and IMRT.
(d) To devise guidelines for CTV to PTV margins based
on cumulative maximum movement error (CMME)
combining intrafraction or interfraction maximum
respiratory motion during normal breathing and
maximum re-positioning set up errors.
Methods and materials
Five patients with EBC who had undergone breast con-
servative surgery at our institution and were potentially
eligible for APBI as per standard eligibility criteria for
APBI [1] were included in this study. Patients with
restricted arm movements after surgery and pre-
existing respiratory problems were excluded. Written
informed consent was obtained from all the patients.
While these women participated in this investigational

766 The British Journal of Radiology, September 2006

Breast movement and implications for partial breast irradiation

study for respiratory movements and set up errors

during positioning for APBI, they were treated with the
standard radiotherapy protocol of our department. The
standard treatment included 5 weeks of daily whole
breast radiotherapy with bitangential portals with
6 MV photons on an inclined breast board followed
by an en face electron boost. For the purpose of this
study, the women underwent a separate process of
evaluation of set up errors for which they were
immobilized in an individualized body cast (vacuum
bag). Baseline movements during normal and deep
breathing, breath hold time and tidal volume were
evaluated for all the women. Breath hold time and tidal
volume were measured on a daily basis during the
course of respiratory training. After a short course of 8–
10 days of respiratory training, breast movement
during normal and deep breathing was re-evaluated.

Immobilization and surface markers

Patients were immobilized in individualized body
casts made of a polyurethane bag filled with tiny
polystyrene pellets, which sets according to the body
shape upon application of vacuum. Patients were
positioned in the body cast in supine position with arms
above the head without using the breast board. The
lumpectomy cavity was localized and outlined on the
breast surface using the information from pre-operative
clinical description, mammography films, intraoperative
findings and lumpectomy scar. With an additional
margin of 1.5 cm in the x and y dimensions the CTV
was delineated on the patient’s surface. Lead markers
(2 mm diameter) were placed at the centre of each of the
four borders of the outlined CTV and at the centre of the
CTV (Figure 1). For the purpose of the study it was
assumed that the movement of these five markers on the
anterior surface of the CTV would represent the move-
ment of the entire CTV.
Interfraction movement: re-positioning set up
errors
After carefully positioning a patient in the customized
body cast on the simulator (Ximavision 2.1H; Varian
Medical Systems, Palo Alto, CA), a field size of 10 cm 6
10 cm was centred on the CTV at a target to skin distance
of 95 cm. Using an isocentric technique, orthogonal
anterior and lateral images were acquired. Patients were
instructed to hold their breath after deep inspiration
during image acquisition. This served as a baseline
image against which set up reproducibility was evalu-
ated. After the baseline image acquisition patients were
then taken off the couch and re-positioned four times at
different time intervals on the same day and orthogonal
images were repeated. For each patient, four pairs of
orthogonal films were compared against the baseline
orthogonal pair of films. On each of these films, the
perpendicular distance between the centre of each
spherical marker and the nearest delineator wire or
the central crosswire of the simulator was measured
using an inbuilt measuring scale with an accuracy of
0.1 mm. The differences between the measurements in
Figure 1. Localization of clinical target volume (CTV) on the
breast surface with lead markers.
the baseline and subsequent films were recorded for each
of these markers in the mediolateral, superoinferior and
anteroposterior dimensions. Internal surgical titanium
clips demarcating the tumour bed were present in one
patient. However, the visualization of these titanium
clips was poor in orthogonal films and not suitable for
evaluation of tumour bed movement.
Systematic error (S), defined as the variation between
the planned position and average position on re-
positioning, was calculated as the average value of the
mean deviation of each patient. Random errors (s)
defined as fraction to fraction variations around the
mean deviation were calculated as the average of the
standard deviation around the mean [11].
Intrafraction movement: respiratory movement of
breast
Baseline evaluation of respiratory movements was
carried out in normal as well as deep breathing, using
the cine-acquisition mode of XimavisionH. Orthogonal
cine-fluoroscopic images were acquired at the rate of two
images per second with the gantry at 0 ˚ and 90 ˚ (or 270 ˚).
At the end of image acquisition, cine recordings were
available for each patient and the orthogonal image sets
could be used for evaluation of movement in mediolat-
eral, superoinferior and anteroposterior dimensions. A
total of 80 images (40 orthogonal images in normal
breathing and 40 orthogonal images in deep
breathing) were available per patient. For the purpose
of measurement or tracking of movements of breast
markers during respiration, the projection of the field
delineator and the central cross wires were considered as
a stationary structure against which the position of
individual markers was measured in serial cine images

The British Journal of Radiology, September 2006 767

 S Chopra, K A Dinshaw, R Kamble and R Sarin

Table 1. Set up variation and cumulative set up errors in mediolateral, superoinferior and anteroposterior dimensions with the
use of individualized body cast
Sr. No: Mediolateral (mm) Superoinferior (mm) Anteroposterior (mm)
Mean¡SD (Range) Mean¡SD (Range) Mean¡SD (Range)

Patient 1 0.5¡0.6 (–0.3 to +0.9) 1.5¡1.3 (+0.3 to +3.3) 4.6¡1.2 (–2.7 to –5.2)
Patient 2 2.0¡1.7 (–2.7 to +0.2) 1.4¡0.5 (–1.9 to +1.7) 1.8¡1.7 (–0.5 to +5.0)
Patient 3 1.3¡1.4 (–0.6 to –3.9) 1.0¡0.8 (–0.3 to +1.9) 6.0¡0.8 (–5 to –7.2)
Patient 4 1.5¡0.9 (0 to –2.1) 1.9¡1.7 (–0.4 to +4.2) 7.8¡1.9 (0 to –9.2)
Patient 5 1.3¡0.8 (–0.5 to +2.3) 1.0¡0.7 (–0.2 to –1.9) 1.8¡0.5 (–0.8 to –2.7)
Mean 1.3¡0.5 1.3¡0.3 4.4¡2.6

acquired during the respiratory cycle. On each of these
images, the perpendicular distance between the marker
and the nearest delineator wire or central crosswire of
the simulator was measured as for set-up errors.
Differences between the measurements in the baseline
and subsequent images were recorded for each of these
markers for ascertaining the breast movement in the
mediolateral, superoinferior and anteroposterior direc-
tion during a few respiratory cycles.
Respiratory training
After baseline assessment of respiratory movement, all
patients were given a short course of respiratory training
for a period of 8–10 days. Patients were trained by an
occupational therapist (RK). Deep breathing exercises,
including inspiratory and expiratory manoeuvres and
forced abdominal expiration technique, were taught to
all the patients. A spirometer was used on a daily basis
for training. Patients carried out all these exercises for
15–20 min once daily under the supervision of the
occupational therapist and repeated them at least twice
daily without the instructor. Breath holding time and
tidal volume were recorded before starting respiratory
training. While patients received respiratory training
daily breath hold time and tidal volume were noted.
Detailed cine-fluoroscopic evaluation of breast move-
ment as described above was done before starting the
respiratory training and repeated once after completion
of respiratory training. Any reduction in the movement
of the skin markers during respiration, increase in the
tidal volume or breath holding time was used as an
end point to evaluate the efficacy of short course
respiratory training. The Wilcoxon Sign rank test was
used to evaluate the statistical significance of observed
differences.
Observations and results
Re-positioning errors using the customized vacuum
bag were evaluated in the mediolateral, superoinferior
and anteroposterior dimensions for all patients. Cranial,
anterior and right sided deviations were recorded in the
positive direction whereas caudal, posterior and left
sided deviations were recorded in the negative direction.
The details of the set up errors on re-positioning are
shown in Table 1. The mean deviations for all markers
from their baseline positions were 1.3 mm (SD ¡
0.5 mm) in the mediolateral; 1.3 mm (SD ¡ 0.3 mm) in
the superoinferior and 4.4 mm (SD ¡ 2.6 mm) in the
anteroposterior dimension. Rotational errors were not
calculated.
Breast movement, as judged by five surface markers,
was minimal during normal breathing with pre-training
mean marker movement of 1.07 mm (0.6–1.4 mm);
1.94 mm (0.7–4.5 mm) and 1.86 mm (0.6–1.9 mm) in the
mediolateral, superoinferior and anteroposterior direc-
tions, respectively, during normal breathing. The breast
movements during normal breathing were not affected
by respiratory training (Table 2).
As expected, the marker movement was more pro-
nounced during deep breathing with mean values of

Table 2. Maximum movement during normal breathing in anteroposterior, mediolateral and superoinferior dimension before
and after respiratory training
Mediolateral (mm) Superoinferior (mm) Anteroposterior (mm)

Pre-training Post-training Pre-training Post-training Pre-training Post-training

Patient 1 1.4 1.4 4.5 4.5 1.9 1.7

Patient 2 NA 1.6 0.8 1.6 0.8 0.9
Patient 3 1.6 0.4 2.3 1.2 4.9 1.6
Patient 4 0.6 0.6 1.4 1.4 1.1 0.7
Patient 5 0.7 1.1 0.7 0.6 0.6 1.2
Mean of max. 1.07 (0.6–1.4) 1.02 (0.4–1.4) 1.94 (0.7–4.5) 1.86 (0.6–4.5) 1.86 (0.6–4.9) 1.09 (0.7–1.7)
movement in
five patients
Wilcoxon Signed 0.6 0.5 0.3
Rank sum test
(p-value)
NA, not available.

768 The British Journal of Radiology, September 2006

Breast movement and implications for partial breast irradiation

Table 3. Maximum movement of breast in anteroposterior, mediolateral and superoinferior dimension during deep breathing
before and after respiratory training
Mediolateral (mm) Superoinferior (mm) Anteroposterior (mm)

Pre-training Post-training Pre-training Post-training Pre-training Post-training

Patient 1 1.7 2.0 8.2 6.6 9 6

Patient 2 NA 2.2 5.7 4.1 3.8 1.3
Patient 3 3.4 2.8 5.0 5.4 5.5 3.8
Patient 4 2.0 0.9 7.4 4.2 1.7 2.5
Patient 5 1.0 1.0 1.4 1.2 4.1 5.0
Mean of max. 2.0 (1.0–3.4) 1.7 (0.9–2.8) 5.5 (1.4–8.2) 4.3 (01.2–6.6) 4.8 (1.7–9) 3.7 (1.3–6)
Movement in
five patients
Wilcoxon Signed 0.2 0.1 0.2
Rank sum test
(p-value)
NA, not available.

2 mm (1–3.4 mm), 5.5 mm (1.4–8.2 mm) and 4.8 mm Discussion

(1.7–9 mm) in the mediolateral, superoinferior and
anteroposterior directions, respectively (Table 3).
Respiratory training produced a non-significant reduc-
tion of movement of marker in mediolateral, supero-
inferior and anteroposterior directions during deep
breathing.
After training the breath hold time increased from a
pre-training mean of 31 s to a mean of 44 s (p50.04). The
tidal volume increased in all the patients from an
average of 560 cm3 in the pre-training phase to
1160 cm3 in the post-training phase (p50.04) (Table 4,
Figures 2 and 3).
CTV to PTV margins based on cumulative maximum
movement error
For determining the CTV to PTV margin for each
study patient, we derived the CMME by combining the
maximum intrafraction movement as judged by cine-
fluoroscopic assessment of breast respiratory motion
during normal breathing and maximum interfraction re-
positioning set up errors obtained in these test conditions
(Table 5). The CMME did not exceed 10 mm for any of
these patients in any of the axes. The mean of the CMME
recorded for these 5 patients was 3.4 mm (range 2.3–
4.3 mm), 4.5 mm (range 2.5–7.8 mm) and 7.1 mm (range
3.9–9.9 mm) in the mediolateral, superoinferior and
anteroposterior dimensions, respectively.
While several studies have previously evaluated set up
errors in women undergoing daily breast radiotherapy
[12–17], this is probably the first study in which both the
interfraction (re-positioning set up errors) and intrafrac-
tion (respiratory) components of variability in CTV
localization have been ascertained individually for each
study patient. Our findings confirm that re-positioning
set up errors and respiratory movements both contribute
to the total uncertainty. The maximum interfraction and
intrafraction movement readings have been combined to
derive the cumulative maximum movement error for
each patient. In our opinion, this represents a simple and
perhaps more accurate and safe method of deriving the
CTV to PTV margin for individual patients.
In this study instead of relying on the surgical clips,
which are usually placed on the posterior surface of the
lumpectomy cavity, we have evaluated respiratory
movements by skin surface markers representing the
anterior surface of the CTV. We feel that it is reasonable
to assume that the movement of the breast surface
represents movement of the CTV, except in patients with
large pendulous breasts where surface markers may not
be so reliable for representing the underlying CTV. As
we observed in one patient, titanium surgical clips are
not visualized well in all projections that are required for
assessing set up errors and respiratory movements,
especially in the true lateral projection. Moreover, it
has been reported that the surgical clips may be

Table 4. Pre- and post-respiratory training breath hold time and tidal volume
Patient no. Breath hold time in seconds Tidal volume in ml

Pre-training Post-training Pre-training Post-training

Patient 1 25 40 400 800

Patient 2 25 40 800 1600
Patient 3 45 46 800 1200
Patient 4 23 46 400 1200
Patient 5 39 50 400 1000
Mean 31.4 44.5 560 1160
Median 25 46 400 1200
Wilcoxon Signed Rank Test (p-value) 0.04 0.04

The British Journal of Radiology, September 2006 769


displaced by up to 3 mm (range 0–11 mm) [18] especially
if they are not anchored to the muscle. Other authors
have described using bony or soft tissue reference points
for measuring set up errors or respiratory movements
[15–17]. In our opinion, none of the soft tissue or bony
reference points in this region are so well defined that
they can be reliably and accurately localized in all
projections in serial imaging.
Interfraction movement: re-positioning set up
errors
Previous studies have measured re-positioning set up
errors using bony or soft tissue reference points [12–17]
during a 5–6 week course of daily radiotherapy. In
contrast, we have ascertained the set up errors by re-
positioning the patients five times during the same day
and by using skin markers. While our approach of re-
positioning assessments over 1 day may not reflect time
trends in set up variation over 5 weeks [19–21] it has
certain other advantages. Most importantly, it removes
the confounding variability in the placement of markers
or problems in precisely localizing the same reference
point in the thoracic bony cage or soft tissues in serial
orthogonal films. For abbreviated treatment like APBI
which is delivered over 5 days, time trends if any, are
likely to be much less than that observed over a
treatment course of 35 days [21]. Assessment over
1 day allows the use of individualized decision regarding
CTV to PTV margins, the benefits of which are discussed
later.
770
S Chopra, K A Dinshaw, R Kamble and R Sarin
Figure 2. Box whisker plot showing
variation of breath hold time during
respiratory training over 1 week (BH
1–75 breath hold time day 1–7).
y-axis: breath hold time in seconds;
x-axis: breath hold time over 1 week
of respiratory training; N5 number
of patients.
The set up errors with this individualized body cast
immobilization were quite small under these test condi-
tions, except in the anteroposterior direction (Table 1).
However, we have not compared the re-positioning set
up errors without body cast. Similar findings have been
reported by other groups using various immobilization
devices [12–15]. Nalder [12] et al evaluated set up
reproducibility for breast radiotherapy using a breast
board with and without a Vacfix device (PAR Scientific
A/S, Denmark). They noted reduction in set up errors in
the superoinferior direction using the Vacfix device.
Giraud et al [22] have demonstrated that a personalized
body cast leads to a significant reduction in lung apex
movements along with reduction in lateral displacement
of the thoracic cage.
Vicini et al [4] have used vacuum bags for immobiliza-
tion for APBI and 5 mm margin has been incorporated
for re-positioning errors.
Intrafraction movement: respiratory movement of
breast
Ideally the entire treatment period should be con-
tinuously monitored and orthogonal images should be
acquired simultaneously for assessing intrafraction
motions. However, for reasons of logistics and in order
to minimize radiation exposure to the opposite breast
with lateral beams, we used cine-fluoroscopic assess-
ment of breast motion for only 10 s. Since the beam on
time for each radiation field of APBI using 3DCRT is
usually in the range of 20–30 s, this 10 s assessment may
Figure 3. Box whisker plot showing
the variation in tidal volume during
respiratory training over 1 week.
(Tidal 1–7 5 tidal volume from day
1–7). y-axis: tidal volume in milli-
litres; x-axis: tidal volume over
1 week of respiratory training; N5
number of patients.
The British Journal of Radiology, September 2006
Breast movement and implications for partial breast irradiation
Table 5. Measured cumulative maximum movement error (CMME in mm) combining maximum respiratory motion in normal
breathing after respiratory training and maximum set up errors
Patient Mediolateral CMME (mm)
Patient 1 2.3 (1.4 + 0.9)
Patient 2 4.3 (1.6 + 2.7)
Patient 3 4.3 (0.4 + 3.9)
Patient 4 2.7 (0.6 + 2.1)
Patient 5 3.4 (1.1 + 2.3)
Mean CMME 3.4
**Use of pre-respiratory training measurements of respiratory movement would have made the combined maximum movement
over 10 mm (12.1 mm and 10.3 mm).
therefore be a reasonable representation of intrafraction
movement.
The findings of our study confirm that respiratory
movements during normal breathing are negligible
(,2 mm) and may be ignored for conventional whole
breast treatment, whereas deep breathing could cause
significant intrafraction error. However, respiratory
motion even during normal breathing merits special
consideration when planning external beam APBI using
3DCRT or IMRT and even for tangential whole breast
treatment if using IMRT. With IMRT, not only the CTV is
moving with respiration but the lack of synchrony
between respiratory motion and the dynamic sliding
multileaf collimator could result in perturbation in
radiation dose delivery [23].
Hence for IMRT, respiratory movements are a matter
of concern even for whole breast irradiation. Other
authors have evaluated movements of the breast in
various dimensions with varying respiratory patterns
ranging from no breathing to deep breathing and
reported 2–16 mm, 2–6 mm and 2–16 mm movement in
the mediolateral, superoinferior and anteroposterior
dimensions, respectively, and documented that the best
reproducibility was achieved when patients were asked
to hold their breath [23, 24]. In our study we recorded
baseline average movement of less than 2 mm in
mediolateral, superoinferior and anteroposterior dimen-
sions during normal breathing. At the William Beaumont
Hospital, ABC was used in the initial few patients
undergoing external beam APBI [3, 6]. However, ABC is
not used currently for APBI and a generic margin of
10 mm is being incorporated (5 mm for set up errors and
5 mm for breathing movements) for CTV to PTV
generation. Formenti et al [2, 5] have treated patients
with APBI in prone position with the aim of reducing
breathing movements. However, up to 1 cm allowance
has to be incorporated for enhanced set up errors in this
position. Improved accuracy, both in re-positioning and
minimizing target motion, has recently been reported
with the incorporation of video based surface registra-
tion, optoelectronic surface registration and RC-CBCT
[8–10]. This may further aid in reducing CTV to PTV
expansion, but limited access to these devices may limit
its wider applicability.
Although there was only minimal impact of the short
course of respiratory training in reducing breast move-
ments, the tidal volume and breath hold time increased
appreciably in all the patients. The post-training mean
breath hold time was 44 s as compared with 12–16 s in
untrained patients as reported by other studies [25]. This
paradox of increased tidal volume with similar or
The British Journal of Radiology, September 2006
Superoinferior CMME (mm) Anteroposterior CMME (mm)
7.8 (4.5 + 3.3) 7.2 (1.7 + 5.2)
3.5 (1.6 + 1.9) 5.9 (0.9 + 5.0)
3.1 (1.2 + 1.9) **8.8 (1.6 + 7.2)
5.6 (1.4 + 4.2) **9.9 (0.7 + 9.2)
2.5 (1.6 + 1.9) 3.9 (1.2 + 2.7)
4.5 7.1
reduced thoracic movement may be explained by an
increase in the diaphragmatic component of breathing.
However, this was not specifically measured in the
present study. Increasing the diaphragmatic component
of respiration and increased breath hold time may have
two implications in thoracic radiotherapy. First, respira-
tory training imparted in this fashion has the possibility
of exacerbating lung movements, especially of the lower
lobe. Hence, the impact of respiratory exercises, which
are sometimes imparted to patients with lung cancer in
order to improve their pulmonary functions, should be
studied with special reference to their impact on
respiratory lung movement during shallow or normal
breathing. Movements during deep breathing are irrele-
vant since it is unlikely that any clinical context would
require delivering radiotherapy during continuous deep
breathing. Second, such short course respiratory training
could help deep inspiration breath hold (DIBH) assisted
treatment of thoracic or abdominal tumours by increas-
ing the breath hold time, thereby shortening the
treatment delivery time and possibly improving accu-
racy in treatment delivery.
CTV to PTV margins
In the literature, there are several recommendations
for deriving the CTV to PTV margins [11, 26]. Using a
mathematical notation and assuming that the distribu-
tion of set up errors is gaussian, Stroom et al have
recommended that a safety margin of 2.0S + 0.7s should
be incorporated where S and s are as previously defined
[11]. The criteria for this recommendation are based on
an assumption that 99% of the CTV would be covered by
95% of the prescribed dose. While this recommendation
has been favoured by other investigators [27] we feel that
there are two caveats with this recommendation. First,
set up errors may not always have a normal or gaussian
distribution as seen in our data as well as in other studies
[28, 29]. Second, such values of systematic and random
errors and the CTV to PTV margins based on these
formulae are derived for a particular study population
and then recommended to be used for generating the
CTV to PTV margin for individual patients. The present
study as well as other investigators [30] have shown that
there could be significant patient to patient variation
both in the magnitude and vector of the mean or
maximum set up error. For highly conformal radiation
plans that are delivered in very abbreviated schedules,
such as APBI over 5 days or highly hypofractionated
schedules such as extracranial body frame stereotaxy
771

[31, 32] use of generic values of CTV to PTV margins may
result in under coverage or an inappropriately large
margin for a substantial part of the treatment. For such
highly abbreviated or hypofractionated schedules, the
maximum re-positioning set up error measured over
repeated re-positioning on a single day along with the
measured respiratory movements can be incorporated in
the initial individualized CTV to PTV generation algo-
rithm. This would ensure the use of minimum yet safe
CTV to PTV margins and in patients where the margins
thus obtained are unacceptably large, the process of
immobilization may be repeated or changed or the
patient may be given special instruction. Finally, the
appropriateness of these margins could be further
validated with online portal imaging and if necessary
online or offline correction [7].
Conclusions
Interfraction re-positioning set up errors and intrafrac-
tion respiratory movement of the breast contribute to
total uncertainty in irradiating the intended CTV.
Moreover, there are significant interindividual differ-
ences in the maximum values of interfraction and
intrafraction movement errors. With the use of indivi-
dualized vacuum bag fixation and individually ascer-
tained values for set up errors and respiratory
movements on a single day it is possible to use tight
CTV to PTV margins for external beam APBI delivered
during normal breathing. It is unlikely that more
advanced techniques such as ABC, real time tumour
tracking and image guided radiotherapy would result in
‘‘clinically significant’’ reduction in the CTV to PTV
margin for APBI over and above what is achieved by the
technique described here. Although not of much use for
APBI, by increasing the breath hold time, a short course
of respiratory training may be useful for patients
undergoing DIBH for intrathoracic or abdominal
tumours.
References
1. Sarin R. Partial breast treatment for early breast cancer:
emergence of a new paradigm. Nat Clin Pract Oncol
2005;2:40–7.
2. Formenti SC, Rosenstein B, Skinner KA, Jozsef G. TI stage
breast cancer: adjuvant hypofractionated conformal radia-
tion therapy to tumor bed in selected postmenopausal
breast cancer patients-pilot feasibility study. Radiology
2002;222:171–8.
3. Baglan KL, Sharpe MB, Jaffray D, Frazier RC, Fayad J,
Kestin LL, et al. Accelerated partial breast irradiation using
3D conformal radiation therapy (3DCRT). Int J Radiat Oncol
Biol Phys 2003;55:302–11.
4. Vicini FA, Remouchamps V, Wallace M, Sharpe M, Fayad J,
Tyburski L, et al. Ongoing clinical experience utilizing 3D
conformal external beam radiotherapy to deliver partial
breast irradiation in patients with early stage breast cancer
treated with breast conserving therapy. Int J Radiat Oncol
Biol Phys 2003;57:1247–53.
5. Formenti SC, Truong MT, Goldberg JD, Mukhi V,
Rosenstein B, Roses D, et al. Prone accelerated partial
breast irradiation (p-apbi) after breast conserving surgery:
preliminary clinical results and dose volume histogram
772
S Chopra, K A Dinshaw, R Kamble and R Sarin
(DVH) analysis. Int J Radiat Oncol Biol Phys
2004;60:493–504.
6. Wong JW, Sharpe MB, Jaffray DA, Kini VR, Robertson JM,
Stromberg JS, et al. The use of active breathing control
(ABC) to reduce margin for breathing motion. Int J Radiat
Oncol Biol Phys 1999;44:911–9.
7. Kim L, Vicini F, Yan D, Vargas C, Martinez A, Wong J.
Reduction of PTV margin for accelerated partial breast
irradiation using on-line detection of surgical clips. Int J
Radiat Oncol Biol Phys 2004;60:S336.
8. Shirato H, Shimizu S, Kitamura K, Nishioka T, Kagei K,
Hashimoto S, et al. Four dimensional treatment planning
and fluoroscopic real time tumor tracking: radiotherapy
for moving tumor. Int J Radiat Oncol Biol Phys 2000;48:
435–42.
9. Riboldi M, Baroni G, Orecchia R, Pedotti A. Enhanced
surface registration techniques for patient positioning
control in breast cancer radiotherapy. Technol Cancer Res
Treat 2004;3:51–8.
10. Hui SK, Das RK, Kapatoes J, Oliviera G, Becker S, Odau H,
et al. Helical tomotherapy as a means of delivering
accelerated partial breast irradiation. Technol Cancer Res
Treat 2004;3:639–46.
11. Stroom JC, de Boer HC, Huizenga H, Visser AG. Inclusion
of geometrical uncertainties in radiotherapy treatment
planning by means of coverage possibility. Int J Radiat
Oncol Biol Phys 1999;43:905–19.
12. Nalder CA, Bidmead AM, Mubata C D, Tait D, Beardmore
C. Influence of vac-fix immobilization device on the
accuracy of patient positioning during routine breast
radiotherapy. Br J Radiol 2001;74:249–54.
13. Pradier O, Schmidberger H, Weiss E, Bouscayrol H, Daban
A, Hess CF. Accuracy of alignment in breast irradiation: a
retrospective analysis of clinical practice. Br J Radiol
1999;72:685–90.
14. Bohmer D, Feyer P, Harder C, Korner M, Sternemann M,
Dinges S, et al. Verification of set up deviations in patients
with breast cancer using portal imaging in clinical practice.
Strahlenther Onkol 1998;174 S2:36–9.
15. Lirette A, Pouliot J, Aubin M, Larochelle M. Efficacy of
portal imaging in tangential breast irradiation: a prospec-
tive study. Radiother Oncol 1995;37:241–5.
16. Herman MG. Clinical use of electronic portal imaging.
Semin Radiat Oncol 2005;15:157–67.
17. Hurkmans CW, Remeijer P, Lebesque JV, Mijnheer BJ. Set
up verification using portal imaging: review of current
clinical practice. Radiother Oncol 2001;58:105–20.
18. Weed D, Yan D, Martinez A, Vicini FA, Wilkinson TJ, Wong
J. The validity of surgical clips as a radiographic surrogate
for the lumpectomy cavity in image guided accelerated
partial breast irradiation. Int J Radiat Oncol Biol Phys
2004;60:484–92.
19. Ghizelan MJ, Jaffray DA, Siewerdsen JH, Van Herk M,
Shetty A, Sharpe MB, et al. Prostate gland motion assessed
with cine magnetic resonance imaging. Int J Radiat Oncol
Biol Phys 2005;62:406–17.
20. Langen KM, Pouliot J, Anezinos RTT, Aubin M, Gottschalk
AR, Hsu IC, et al. Evaluation of ultrasound based prostate
localisation for image-guided radiotherapy. Int J Radiat
Oncol Biol Phys 2003;57:635–44.
21. Trog D, Garbe S, Lutterbey G, Barwig P, Boldt I, Stolz A, et
al. Volumetric changes of the breast during radiotherapy. Is
replanning necessary for the electron boost? Strahlenther
Onkol 2005;181:255–9.
22. Giraud P, De Rycke Y, Dubray B, Helfre S, Voican D, Guo L,
et al. Conformal radiotherapy (CRT) planning for lung
cancer: analysis of intrathoracic organ motion during
extreme phases of breathing. Int J Radiat Onc Biol Phys
2001;51:1081–92.
The British Journal of Radiology, September 2006
Breast movement and implications for partial breast irradiation
23. George R, Keall PJ, Kini VR, Vedam SS, Siebers JV, Wu Q, et
al. Quantifying the effect of intrafraction motion during
breast IMRT planning and dose delivery. Med Phys
2003;30:552–62.
24. Vedam SS, Kini VR, Keall PJ, Ramakrishnan V, Mostafavi
H, Mohan R, et al. Quantifying the predictability of
diaphragm motion during respiration with a non-invasive
external marker. Med Phys 2003;30:505–13.
25. Hanley J, Debois MM, Mah D, Mageras GS, Raben A,
Rozenweig K, et al. Deep inspiration breath-hold technique
for lung tumors: the potential value of target immobiliza-
tion and reduced lung density in dose escalation. Int J
Radiat Oncol Biol Phys 1999;45:603–11.
26. Herk MV, Remeuer P, Rasch C, Lebesque JV. The
probability of correct target dosage: dose-population
histograms for deriving treatment margins in radio-
therapy. Int J Radiat Oncol Biol Phys 2000;47:
1121–35.
27. Alheit H, Domfeld S, Dawel M, Alheit M, Henzel B, Steckler
K, et al. Patient position reproducibility in fractionated
stereotactically guided conformal radiotherapy using the
The British Journal of Radiology, September 2006
BrainLab mask system. Strahlenther Onkol 2001;177:
264–8.
28. Lujan AE, Ten haken RK, Larsen EW, Balter JM.
Quantization of set up uncertainties in 3D dose calculations.
Med Phys 1999;26:2397–402.
29. Huizenga H, Levendag PC, De Porre, Visser AG. Accuracy
in radiation field alignment in head and neck cancer: a
prospective study. Radiother Oncol 1988;11:181–7.
30. Erridge SC, Seppenwolde Y, Muller SH, Van Herk M, De
Jaeger K, Belderbos JS, et al. Portal imaging to assess set up
errors, tumor motion and tumor shrinkage during con-
formal radiotherapy of non small cell lung cancer.
Radiother Oncol 2003;66:75–85.
31. Zimmerman FB, Geinitz H, Schill S, Grosu A,
Schratzenstaller U, Molls M, et al. Stereotactic hypo
fractionated radiation therapy for stage I non-small cell
lung cancer. Lung Cancer 2005;48:107–14.
32. Balter JM, Brock KK, Lam KL, Tatro D, Dawson LA,
Mcshan DL, et al. Evaluating the influence of set up
uncertainties on treatment planning for focal liver tumors.
Int J Radiat Oncol Biol Phys 2005;63:610–4.
773

Book review
Computed tomography, 2nd edn. By WA Kalender. pp. 306,
2005 (Publicis, Erlangen, Germany) J49.90
ISBN 3-89578-216-5
This is an excellent well-written and well-illustrated
book, which deals with the scientific and technical
aspects of computed tomography. CT scanning has
developed rapidly in the last few years and now makes
up half of the collective dose delivered to patients from
diagnostic radiology (47% in Germany in 2003). There is
a need for medical physicists, radiologists and radio-
graphers to have a thorough understanding of the
technique and the information it can provide in order
that its full potential is realised. This book by an author
who has played an important role in the development of
spiral CT addresses this need. The book explains the
fundamentals and application of CT with the main focus
on physics and technology, and deals with problems
relevant to the user of CT equipment. It is directed at a
multidisciplinary readership and the concepts are
explained clearly for all staff involved in use of the
technique. This is the second edition of the textbook,
which has been revised to include important new
developments in detector technology and applications.
The first chapter covers fundamentals, starting from
first principles. Those working with the topic can skim
quickly through the chapter before delving into the more
meaty content of the later chapters. Others will find it a
useful introduction. Chapter 2 contains technical con-
cepts, including scanner design, collimation, filtration,
detector systems, can modes and finishing off with a
section on important aspects of modern equipment, such
as the use of cone-beam geometry in multislice scanners
and combined CT/PET scanners. Chapter 3 deals with
spiral CT and chapter 4 image quality, including
considerations of requirements for 64 slice scanners.
774
The British Journal of Radiology, 79 (2006), 774
Chapter 5 is devoted to radiation protection with
explanations of the standard dose quantities. Questions
regarding the possibility of dose reduction are discussed
in a separate section. The chapter promotes the thesis
that CT when applied properly should not be regarded
as a high dose technique. Two and three dimensional
approached to display of information are explained in
chapter 6 with consideration of interactive approached
for assessment of the large volume of data generated.
Techniques are illustrated with a CD-ROM issued with
the book. Example images can be loaded and viewed
interactively, so that readers can undertake exercises in
image viewing and adjustment themselves. This pro-
vides an invaluable teaching aid for those who do not
have easy access to CT data themselves. Moreover, the
author indicates in the preface to the new edition that
readers may use the figures in their own presentations
and teaching material. Chapter 7 deals with clinical
applications, such as CT angiography, interventional CT
and cardiac imaging, while in chapter 8 the author
provides his view on future developments and includes
some wishes about how perception and use might
change in the future. The full mathematical treatment
of image reconstruction is left until chapter 9 for those
with the mathematical background and the need to
understand the image reconstruction methods in more
depth, so that others are not put off by encountering the
more complex formulae in the earlier chapters.
This is an excellent text covering every aspect of the
physics and technology of CT, which is suitable for a
wide range of staff involved in use and applications of
CT scanners. It would make a useful addition to any
Radiology or Medical Physics Department library.
C J MARTIN
The British Journal of Radiology, September 2006
The British Journal of Radiology, 79 (2006), 775–777

CASE OF THE MONTH

A case of spinal cord compression of unknown cause

D BUTTERISS, FRCR and C SOH, FRCR

Department of Neuroradiology, Regional Neurosciences Centre, Newcastle General Hospital,

Westgate Road, Newcastle-upon-Tyne NE4 6BE, UK

Received 21 June 2005

Revised 27 July 2005
Accepted 16 August 2005

DOI: 10.1259/bjr/29157329

’ 2006 The British Institute of

Radiology

Case report sensation below the shoulders and perianal/saddle

A 65-year-old woman presented to the accident and
emergency department complaining of a 3-month history
of falls, reduced power and numbness in her arms and
legs, and urinary incontinence. Her status had deterio-
rated over the previous 4 weeks and she was bed-bound
on admission. There was no significant past medical
history. Examination revealed reduced power in all
limbs, increased tone in the right upper limb, altered
Address correspondence to: Dr C Soh.
paraesthesia. A presumptive diagnosis of spinal cord
compression in the mid-cervical spine was made and the
patient was referred for spinal MRI. Selected axial and
sagittal images of the cervical (Figure 1) and lumbar
spine (Figure 2) are shown.
What do the images show?
Is there evidence of myelopathy?
What is your differential diagnosis?
CT of the cervical spine was subsequently performed,
prior to surgery (Figure 3).
Does this alter or narrow your differential diagnosis?

(a) (b) (c)

Figure 1. Sagittal (a) T1 weighted, (b) T2 weighted MR images through the cervical spine. (c) Representative axial T1 weighted
image through the cervical spine.

The British Journal of Radiology, September 2006 775


(a)
Figure 2. Sagittal (a) T1 and (b) T2 weighted MR images through the lumbar spine. (c) Representative axial T2 weighted image
through the lumbar spine.
MRI demonstrates low/intermediate signal masses on
T1 and T2 imaging, involving the C3–C6 vertebral bodies,
with destruction of C4 and C5 extension into the
paravertebral soft tissues. At C4–C5 there is extradural
compression of the cervical cord, with high intramedul-
lary cord signal on T2 weighted imaging consistent with
myelopathy. CT of this level shows a relatively high
attenuation mass containing fragmentary calcification
that is destroying the cervical vertebral bodies and
extending into the surrounding soft tissues and into the
spinal canal with thecal compression.
MRI of the lumbar spine shows involvement of the L3–
S1 vertebral bodies, posterior elements and facet joints
with small extradural masses, but no evidence of neural
compression. Further imaging demonstrated further
lesions in the sacroiliac joints, both hips and the pubic
symphysis.
Neurosurgical C3–C6 partial anterior vertebrectomies
with iliac crest grafting and plate fixation revealed pale,
cheesy deposits, shown to be gouty tophi on histologi-
cal investigation. Urate-lowering medical therapy was
instituted.
Discussion
Gout is a relatively common metabolic condition
characterized by deposition of monosodium urate crystals
in joints and soft tissues. In Europe the incidence is
approximately 0.2–0.5% and the male to female ratio is
approximately 5:1. The acute presentation is with a
monoarthritis, in 25% of cases affecting the first tarsome-
tatarsal joint (podagra). Chronicity results in the formation
of tophi within the soft tissues around joints and in the
pinnae of the ears due to deposits of monosodium urate
and associated inflammatory cells. This is rare before at
776
D Butteriss and S Soh
(b) (c)
least 10 years of acute attacks. The peripheral joints are
most commonly affected, but occasionally spinal involve-
ment with gouty tophi occurs [1].
The presentation of spinal tophaceous gout is non-
specific, with local pain, radicular symptoms and
paraparesis or tetraparesis occurring due to neural
compression. There is usually, but not invariably, a
history of prior acute attacks of gout, or of hyperur-
icaemia [2].
MRI appearances are variable. Tophi appear as
homogeneous intermediate/low intensity masses on T1
weighted imaging, but have a variable T2 weighted
appearance ranging from homogeneous low through
heterogeneous signal to homogeneous high signal. Tophi
may contain small areas of signal drop-out that have
been shown to represent calcification on CT imaging.
Enhancement post-gadolinium administration is also
variable. Rim enhancement is the most common finding,
but both complete lack of enhancement and homo-
geneous enhancement have both been described [3].
Appearances may mimic spinal infection, malignancy or
haemorrhage [4].
Symptomatic spinal gout usually involves the epidural
space, with variable involvement of the intradural
component, ligamentum flavum, facet joints, posterior
elements, vertebral bodies and paraspinal soft tissues [3].
Acute management in cases with neural compression
usually requires surgical decompression, where the
tophus appears as a chalky, cheesy or fibrous mass.
Often the diagnosis is made on histological examination
of the surgical sample. Local pain may respond to anti-
inflammatory medication or colchicine. Long-term allo-
purinol therapy may result in reduction in size or even
disappearance of tophi.
In conclusion, spinal tophaceous gout is a rare cause of
spinal neural compression, but should be considered when
The British Journal of Radiology, September 2006
Case of the month: Unknown cause of spinal cord compression
(a)
Figure 3. Axial CT images of the cervical spine in (a) bone and (b) soft tissue windows.
MRI demonstrates intermediate/low signal on T1
weighted deposits with low signal foci on all sequences,
especially if there is involvement of the posterior elements.
References
1. Huskisson EC, Drury PL. Rheumatology and bone disease.
In: Kumar P, Clark M, editors. Clinical medicine. London,
UK: Bailliere Tindall, 1994:409–11.
The British Journal of Radiology, September 2006
(b)
2. Paquette S, Lach B, Guiot B. Lumbar radiculopathy
secondary to gouty tophi in the filum terminale in a patient
without systemic gout: case report. Neurosurgery
2000;46:986–8.
3. Hsu C-Y, Shih T T-F, Huang K-M, Chen P-Q, Sheu J-J,
Li Y-W. Tophaceous gout of the spine: MR imaging features.
Clin Radiol 2002;57:919–25.
4. Barrett K, Miller ML, Wilson JT. Tophaceous gout of the
spine mimicking epidural infection: case report and review
of the literature. Neurosurgery 2001;48:1170–2.
777
BJR
The British Journal
of Radiology

October
2006
Volume 79
Issue 946
October 2006, Volume 79, Issue 946

● Virtual pulmonary arterioscopy in pulmonary embolic disease

● Breast radiotherapy in women with pectus excavatum (funnel

chest): is the lateral decubitus technique an answer? A dosimetric
study

● Hypoxia in head and neck cancer

● Is routine chest radiography a useful test in the follow up of all

adult patients with soft tissue sarcoma?

● Multiple hyperechoic testicular lesions are a common finding on

ultrasound in Cowden disease and represent lipomatosis of the
testis

● Duplex ultrasound of the superior mesenteric artery in chronic

pancreatitis

● Accuracy and precision of an external-marker tracking-system for

radiotherapy treatments

● Normalized data for the estimation of fetal radiation dose from

radiotherapy of the breast

● Dosimetric and treatment planning considerations for

radiotherapy of the chest wall

● Patient and staff radiation doses from early radiological

examinations (189921902)

● Ultrasound spectrum in intraductal papillary neoplasms of breast

The British Journal of Radiology, 79 (2006), 779–784

SHORT COMMUNICATION

Virtual pulmonary arterioscopy in pulmonary embolic disease

C HOSKINS, BSc, FRCR and M CARPENTER, BSc

Department of Diagnostic Imaging, Mayday University Hospital, Croydon CR7 7YE, UK

ABSTRACT. 16 slice multidetector CT provides virtual endoscopic views of the inside of

arteries, or any other hollow structures. This is performed non-invasively using post-
processing of three-dimensional isotropic image data sets, acquired during standard CT
examinations. These virtual endoscopic views are simultaneously correlated with the
standard multiplanar reconstructions, with the ability to navigate a virtual camera
through the hollow structure under study. Normal and abnormal volume rendered
images of the pulmonary arteries are presented in correlation with the multiplanar Received 6 January 2006
reformats. The abnormal images show the volume rendered appearances of acute and Revised 12 April 2006
chronic pulmonary embolic disease. It is also postulated that this technique has a Accepted 19 May 2006
problem solving role in the differential diagnosis of chronic mural emboli from
DOI: 10.1259/bjr/40749658
extravascular structures such as adjacent lymph nodes or bronchiolar impaction. This
technique may also have a role in medical education, providing clinicians and medical ’ 2006 The British Institute of
students with interactive three-dimensional representations of disease processes. Radiology

The establishment of CT pulmonary angiography as
the primary imaging technique for assessment of
pulmonary embolic disease is widely recommended in
many centres, as it combines the ability to directly
visualize thrombus and evaluate the mediastinum and
the lung parenchyma within one examination [1, 2].
Production of near isotropic data sets with 16+ slice
multidetector CT has enabled the introduction and or
refinement of numerous image processing techniques,
avoiding the inherent distortion and artefacts associated
with non-isotropic data [3].
Volume rendering (VRT) is one such technique that
produces a 3D coloured image with depth perception.
Unlike the earlier method of shaded surface display
(SSD), VRT evaluates all the voxel intensities from the
standard isotropic data set normally acquired for a
routine examination. The VRT software assigns differing
colours and degrees of transparency to CT number
ranges within the volume and is thus able to display
overlapping structures. It is also possible to position a
virtual ‘‘endoscopic’’ camera within the arterial lumen
guided by the multiplanar reconstruction (MPR) dis-
plays. This technique of ‘‘fly through’’ or perspective
volume rendering (pVRT) allows visualization of the
inner wall and any intraluminal contents of the arteries.
Development of VRT relied heavily on the advance-
ment of computing power [4], but is now universally
applied to virtual colonoscopy, virtual bronchoscopy and
virtual angioscopy/interior vessel analysis [5]. Identified
applications for virtual angioscopy include ‘‘fly
through’’ coronary angiography [6] and qualitative
assessment of carotid artery stenosis [7]. The prerequi-
sites for good angioscopic VRT imaging are the
reconstruction of thin overlapping slices and the exis-
tence of a strong interface between the Hounsfield Units
of the structure of interest and its surroundings. For
virtual angioscopy this interface is achieved with good
contrast opacification of the blood vessels.
In this paper we show the pVRT appearance of the
pulmonary arteries, both in normal cases and in cases of
acute and chronic thromboembolic disease. These are
correlated with the MPR reformats. We postulate that
this post-processing non-invasive technique has the
potential to be used as a problem solving tool in the
differential diagnosis of chronic thromboembolic disease
versus periarterial lymph nodes and bronchiolar impac-
tion. We also briefly discuss a possible role in medical
education and the visualization of disease processes.
Methods
Records were compiled of cases where a thoracic CT or
pulmonary CT demonstrated acute or chronic pulmon-
ary emboli, periarterial lymph nodes or a normal
pulmonary arterial tree. These cases were protected on
a workstation to enable future study.
CT scanning
CT pulmonary angiograms were performed using a
Siemens Somatom Sensation 16 (Siemens AG Medical
Systems, Germany). 100 ml of intravenous contrast was
administered at a rate of 3 ml s21 using an automated
injector pump. Timing for the scan was influenced by the
use of bolus triggering software (Siemens CARE Bolus)
with a threshold of 100 HU and a region of interest
positioned over the main pulmonary artery. Once
triggered, a volume scan was undertaken following a
6 s delay through the whole chest caudocranially, using

The British Journal of Radiology, October 2006 779


a slice collimation of 16 mm 6 0.75 mm. Following raw
data reconstruction (slice width of 1.0 mm, reconstruc-
tion interval of 0.7 mm) the resultant overlapping axial
images were transferred to a workstation.
Data processing
Case images were reviewed using the interactive axial,
sagittal and coronal MPR display with any additional
reformatting (angulation, curved MPR, magnification)
undertaken when required. Examinations were first pre-
selected for adequate contrast enhancement of the
pulmonary arteries, which was judged subjectively.
MPR images of normal arteries, arteries with adjacent
lymphatic tissue and arteries with features of acute or
chronic emboli were created and saved. Chronic embolic
features included mural thrombus and web formation.
From the selected images, pVRT or ‘‘fly through’’ views
were acquired utilizing the standard post-processing
software supplied with the scanner. The standard isotropic
data set is manipulated to display 3D imagery by assigning
different colour and degrees of transparency to different
voxel Hounsfield Units within the volume. Whilst this
rendering can be altered interactively by the user with
regards to colour, opacity and brightness, our study
employed the default setting. Further manipulation of
the pVRT images is performed, locating areas of interest
within the arteries by moving the virtual camera display
on the MPR images. Subsequent views of the arterial
lumen and emboli were obtained and saved.
Results
Figure 1 shows an axial reconstruction from a CT
pulmonary angiogram with a diagrammatic representa-
tion of the virtual camera observing a normal main
pulmonary arterial bifircation; together with a pVRT
(a)
Figure 1. Axial CT pulmonary angiography of normal main pulmonary bifurcation with volume rendered (pVRT) image.
780
C Hoskins and M Carpenter
image of this area using different virtual camera
positions within the arterial lumen.
Figure 2 shows a sagittal multiplanar reformat demon-
strating an acute embolus appearing as a filling defect in
the pulmonary arterial branch. The pVRT image clearly
shows the occluding clot and an adjacent patent vessel.
Figure 3 shows a chronic embolus appearing as a
flattened eccentric defect in contrast filling at an obtuse
angle with the vessel wall on the right side. There is a
web in the left interlobar pulmonary artery. The pVRT
image demonstrates the intraluminal web.
Figure 4 shows chronic mural embolus in the left
pulmonary artery on both coronal and sagittal MPRs,
with the sagittal MPR demonstrating the position of the
virtual camera. The pVRT image shows the intra-arterial
nature of the embolus.
Figure 5 shows both axial and coronal views of a follow
up scan in a patient with long term perivascular lymphatic
tissue unchanged from previous scans. PVRT imaging
confirms that this perivascular tissue is not intra-arterial.
Discussion
With the advent of 16 (and more) slice CT and
subsequent acquisitions of near perfect isotropic volume
data, there has been an explosion in the use of post-
processing techniques. These include maximum intensity
projections (MIP), SSD, VRT, pVRT and MPR reconstruc-
tions which have benefited from a remarkable increase in
resolution as well as software improvements, which have
allowed them to be used easily and quickly with the
minimum of training. Post-processing with CT pulmonary
angiography includes the standard multiplanar reformats
as well as curved or other reformats individually tailored
to the arterial branch under study. MIP algorithms have
been most commonly used for the assessment of vascular
structures, extracting the highest attenuating voxels along
a specified direction and adding this to the projected
(b)
The British Journal of Radiology, October 2006
Short communication: Virtual arterioscopy in pulmonary emboli

(a) (b)

Figure 2. Sagittal multiplanar reconstruction of acute embolus with corresponding pVRT image.

The British Journal of Radiology, October 2006 781

 C Hoskins and M Carpenter

(a) (b)

(c)

Figure 3. Multiplanar reconstruction and perspective volume rendering (pVRT) images of chronic emboli and webs.

782 The British Journal of Radiology, October 2006

Short communication: Virtual arterioscopy in pulmonary emboli
(a)
(c)
Figure 4. Multiplanar reconstruction and perspective volume rendering (pVRT) images of chronic embolus.
image, providing a display demonstrating the blood
vessels well but suppressing the background tissue.
Resultant images are rotatable, allowing elimination of
overlap. However, unlike VRT and pVRT there is no sense
of depth as the MIP image is essentially a collapsed 3D
structure imposed onto a 2D surface without perspective
[4]. Furthermore, as high attenuation suppresses low
attenuation, this reconstruction technique can obscure
low attenuation thrombus. Chronic embolic features such
as webs and mural thrombus are also difficult to see even
with relatively thin slice MIP imaging [3].
Characteristic CT features of chronic pulmonary
thromboembolic disease include webs, bands, intimal
irregularities and eccentric flattened defects at an obtuse
angle with the vessel wall. Thrombi can be simulated by
periarterial lymph nodes [8].
The British Journal of Radiology, October 2006
(b)
(d)
While to our knowledge this volume rendering
imaging has not been studied in relation to pulmonary
embolic disease, we feel it may prove useful in problem
solving cases where the differential lies between intra-
vascular but perimural pathology, such as chronic
embolus, and extravascular pathology, such as bron-
chiolar impaction or periarterial lymphatic tissue, situ-
ated adjacent to the arterial wall. It may also prove useful
in the display of the intravascular findings of chronic PE
such as mural thrombus and webs, which are often
extremely subtle [3].
Volume rendering can play a role in medical education
with the improved visual presentation of disease
processes. Several CT image reformatting approaches
such as sagittal/coronal, oblique, curves and variable
thickness viewing helps to orientate referring clinicians
783

(a)
(c)
Figure 5. Multiplanar reconstruction of perivascular lymphatic tissue with confirmatory perspective volume rendering (pVRT)
image of arterial lumen.
and radiologists to particular anatomical structures and
pathology, allowing selective display and enhancement
of relevant findings [3, 5]. However, these methods still
require the user to think in sections whereas volume
rendered imagery provides realistic three-dimensional
pictures of intraluminal disease processes, which if
studied on an interactive CT workstation can provide
excellent learning opportunities to health care students.
References
1. Schoepf UJ, Goldhaber SZ, Costello P. Spiral computed
tomography for acute pulmonary embolism. Circulation
2004;109:2160–7.
2. Reidel M. Diagnosing pulmonary embolism. Postgrad Med J
2004;80:309–19.
784
C Hoskins and M Carpenter
(b)
(d)
3. Gruden JF, Tigges S, Baron MG, Pearlman H. MDCT
pulmonary angiography: image processing tools. Semin
Roentgenol 2005;40:48–63.
4. Edelman RR, Rabin DN. 3D imaging technologies. J Imaging
Technology Management. www.ImagingEconomics.com, 2001.
5. Cody DD. AAPM/RSNA physics tutorial for residents: topics
in CT. Imaging and Therapeutic Technology 2002;22:1255–68.
6. Van Ooijen PM, Oudkerk M, Van-Geuns RJ, Rensing BJ, De-
Feyter PJ. Coronary artery fly through using electron beam
computed tomography. Circulation 2000;102:E6–10.
7. Do-Yeon Kim, Jong-Wong P. Visualisation of the internal
carotid artery using MRA images. Magn Reson Imaging
2005;23:27–33.
8. Han D, Kyung SL, Franquet T, Muller L, Sung Kim T,
Hojoong K, et al. Thrombotic and nonthrombotic pulmonary
arterial embolism: spectrum of imaging findings.
RadioGraphics 2003;23:1521–39.
The British Journal of Radiology, October 2006
The British Journal of Radiology, 79 (2006), 785–790

SHORT COMMUNICATION

Breast radiotherapy in women with pectus excavatum (funnel

chest): is the lateral decubitus technique an answer? A dosimetric
study
M A BOLLET, MD, F CAMPANA, MD, Y M KIROVA, MD, R DENDALE, MD, M-G SALIOU, MD,
J-C ROSENWALD, PhD and A FOURQUET, MD

Department of Radiation Oncology, Institut Curie, Paris, France

ABSTRACT. Breast radiotherapy is a technical challenge in women with pectus

excavatum. We aim to assess isocentric lateral decubitus (ILD) technique as a means to
irradiate breasts for patients with pectus excavatum. Four women presenting with left-
sided breast cancers and found to have pectus excavatum were offered breast-
conserving treatments. Post-operative breast radiotherapy was indicated (50 Gy) in two
patients, with an additional boost to the tumour bed (16 Gy). Both ILD and supine
techniques were simulated. We report the dosimetric comparison of these techniques
and the acute skin toxicity of ILD radiotherapy. ILD permitted the same breast dose-
homogeneity as the supine technique while decreasing breast thickness by 4.5–6.8 cm.
The width of lung and/or heart receiving . 20 Gy ranged between 2.1 cm and 4.3 cm Received 6 January 2006
with the supine technique and between 0.5 cm and 1.1 cm with ILD. The estimated Revised 3 May 2006
percentage of ipsilateral lung receiving . 20 Gy ranged from 21% to 34% with the Accepted 30 May 2006
supine technique and from 0% to 5% with ILD. Acute skin toxicity was scored 1 for all
DOI: 10.1259/bjr/23839243
patients at completion of ILD radiotherapy. ILD is an effective breast radiotherapy
technique for patients with pectus excavatum that preserves the underlying heart and ’ 2006 The British Institute of
lung from unnecessary toxicity. Radiology

Breast conserving treatment is considered as standard
management for early-stage breast cancer when techni-
cally feasible. Post-operative radiotherapy to the breast is
an important part of the treatment and has been shown
to significantly diminish the rate of local relapse, the rate
of breast cancer mortality and the risk of overall
mortality [1]. However, the irradiation of normal tissues,
especially of the underlying lung and left ventricle, can
lead to lethal complications explaining the increase in
non-cancer related deaths. The risk of late toxicities such
as radiation-induced pneumonitis or coronary morbidity
has been related to the volume of, respectively, under-
lying lung and left ventricle receiving a dose higher than
20–25 Gy [2–4].
The standard breast radiation technique uses tangen-
tial fields delivered to the breast in a supine position [5].
In the case of women presenting with pectus excavatum,
a congenital deformation of the anterior chest wall
leading to a deep depression of the sternum, also
described as funnel chest, this supine technique exposes
too much of the underlying lung and/or left ventricle in
the treatment fields. To try and comply with the two
contradictory breast radiotherapy requirements of breast
dose homogeneity and lung/heart preservation presents
Address correspondence to: Marc A Bollet, MD, Institut Curie,
Radiotherapy Department, 26, rue d’Ulm, 75005 Paris, France.
E-mail: marc.bollet@curie.net.
a particularly difficult technical challenge that, up to
now, has been dealt with by using sophisticated three-
dimensional, intensity-modulated radiotherapy (IMRT)
or proton therapy techniques [6–8]. In this study we
present a convenient solution: the isocentric lateral
decubitus technique (ILD). Breast irradiation in the
lateral decubitus set-up was originally developed for
women with large, pendulous breasts and has been used
to treat thousands of patients [9]. The need to comply
with modern requirements has prompted our depart-
ment to adapt the technique to an isocentric set-up as
described by Campana et al [10]. To ensure a good
reproducibility of position, the alignment of the patient is
verified by using laser projections marked on the
patient’s skin. The position of the epoxy breast-support
is also marked on the patient’s skin. Day-to-day
reproducibility is verified using the distance from the
top of the treated breast to the corner of the epoxy
breast-support. Portal imaging controls are regularly
performed.
Methods and materials
Four women presenting with cancer of the left breast
and found to have pectus excavatum were offered
breast-conserving treatments at the Institute Curie from
October 2002 to September 2004. Surgery consisted of

The British Journal of Radiology, October 2006 785


lumpectomy with either axillary lymph node clearance
or sentinel node excision. Breast irradiation alone was
indicated (50 Gy at ICRU point in 25 fractions over
5 weeks) with, in the cases of patients A and B who were
younger than 50 years old at the time of diagnosis, an
additional boost to the tumour bed (16 Gy at ICRU point
in 8 fractions over 1.5 weeks). The breast ICRU point was
located at the central slice at the mid-bridge for tangents
[11, 12].
Both ILD and supine breast radiation techniques were
evaluated at the time of simulation in order to decide on
the optimal treatment using dose distribution as a
criterion.
For practical reasons we decided to focus our interest
on the whole-breast irradiation, common to the four
patients.
Full descriptions of the ILD technique and the supine
technique can be obtained in the articles of Campana et
al and Perez et al, respectively [5, 10]. They both comply
with the recommendations of the EORTC quality
assurance in conservative treatment of early breast
cancer [13]. A schematic presentation of the ILD
technique is reproduced in Figure 1. For illustrative
purposes Figure 2 shows patient A in both the supine
(Figure 2a) and the ILD (Figure 2b) positions. In short,
the ILD technique consists of treating a patient in a
lateral decubitus position with the treated breast resting
on an individually chosen breast support made of a thin
layer of carbon fibre and the contralateral breast brought
upwards, out of the field, by elastic straps.
For both techniques we used a simulator – CT (Varian
Ximascan) to produce localization films of the fields and
at least three CT slices, i.e. one at the central beam axis and
one in each mid-field (z50; z5+25% of the field size;
z5–25% of the field size). The CT images were
transferred to the ISIS 3D treatment planning system.
After having checked the position of the beams, dose
distributions were calculated, without heterogeneity
correction, using either 60-cobalt (source–axis distance
(SAD) 80 cm) or 4 MV X-ray (SAD 100 cm). Whenever
needed, standard wedges were added in order to
786
M A Bollet, F Campana, Y M Kirova et al
obtain the best possible dose distributions for both
techniques.
Measures on CT slices were done with the ISIS 3D
measurement tool. The breast thickness was defined as
the distance between the entry and the exit points of the
fields at the central axis. The width of lung and/or left
ventricle that received a dose equal to or higher than
20 Gy was measured on all CT-scan slices, on an axis
perpendicular to that of the lateral field (Figure 3a).
Figure 3 shows an example of two CT-scan slices from
the same patient (patient B) in both the supine technique
(Figure 3a) and the lateral decubitus technique
(Figure 3b). The central lung distance (CLD) was
measured in centimetres as the perpendicular distance
from the field edge to the posterior part of the anterior
chest wall at the centre of the field (Figure 4) [14]. The
maximal heart distance (MHD) was measured in centi-
metres as the maximal width of heart in the tangent
fields (Figure 4) [15]. The percentage of the ipsilateral
lung volume and of heart receiving 20 Gy or more,
known respectively as IPV20 and HV20, were estimated
using Kong’s formulae [15]:
IPV20 (%)~9:8|CLD (cm) 4:9
HV20 (%)~6:1|MHD (cm) 1:3
The acute skin toxicity of the breast alone radiotherapy
was assessed on the last day of treatment (50 Gy for patients
C and D and 66 Gy for patients A and B) and scored
according to the Radiation Therapy Oncology Group
(RTOG) acute radiation morbidity scoring criteria [16].
Results
Dosimetry study
Both techniques fulfilled the recommendations of the
ICRU 62 report with a dose heterogeneity in the planning
Figure 1. Schematic presentation of
the isocentric lateral decubitus (ILD)
technique for radiotherapy of the
left breast.
The British Journal of Radiology, October 2006
Short communication: ILD breast irradiation – pectus excavatum
(a)
Figure 2. Pictures of the same patient with pectus excavatum at simulation in both (a) the supine and (b) the lateral decubitus
set-ups.
target volume within a range corresponding to 95–107%
of the prescription dose. In the case of patient D, the dose
coverage of the breast with the supine technique was
optimal when only 48 Gy were prescribed at the ICRU
point in 24 fractions.
Breast thickness ranged between 3.6 cm and 6.4 cm
with the ILD technique and between 9.4 cm and 12 cm
with the supine technique.
The internal tangent fields’ simulation films of both
techniques are shown in Figure 4. An overview of the
results of the dosimetry study is reported in Table 1. The
central lung distance ranged from 2.6 cm to 4 cm with
the supine technique and from 0 cm to 1 cm with the ILD
technique. The estimation of ILV20 ranged from 21% to
34% with the supine technique and from 0% to 5% with
the ILD technique. The maximal heart distance ranged
between 0 cm and 1.7 cm and the estimation of the HV20
from 0% to 9% with the supine technique. They were
both zero with the ILD technique.
The maximal width of lung and/or left ventricle that
received 20 Gy or more ranged between 2.1 cm and
4.3 cm with the supine technique and between 0.5 cm
and 1.1 cm with the ILD technique.
Acute toxicity
No patient needed to have her radiotherapy suspended.
Acute skin toxicity at the completion of treatment was
mild and scored 1 (RTOG) for all four patients.
The British Journal of Radiology, October 2006
(b)
Discussion
Breast conserving treatments for breast cancers can
sometimes be denied to women with pectus excavatum,
where radiotherapy would involve too high a risk of
pulmonary or cardiac toxicity.
As previously discussed, the ILD is a simple, repro-
ducible breast irradiation technique that can easily be
implemented in radiotherapy departments [10]. More
than 500 patients have already been treated using this
technique in our centre. Most often, this technique was
decided on in the case of patients with large pendulous
breasts who needed breast alone irradiation. In this
study, we address the possibility of using this technique
for women with pectus excavatum who need breast-alone
irradiation. We performed dosimetric studies on four
women and report here the acute toxicity of this
technique. Because the set-up device we used was too
cumbersome to fit into our dosimetric CT-scan, we could
not perform whole lung and heart scanning and thus
dose–volume histogram comparisons were not available.
We could nevertheless perform a number of CT slices that
accurately reflected the dosimetry of both the supine and
the ILD breast radiotherapy techniques. In addition, we
used the two-dimensional information available from the
simulation field to estimate the percentage of the
ipsilateral lung and heart receiving 20 Gy or more [15].
The formula we used was generated by Kong et al from
data of 40 patients simulated for breast radiotherapy (22
left side) in a supine position with a prescribed dose of
787

(a)
Figure 3. CT-scan slices at the level of the central axis (z50) of the same patient with pectus excavatum taken in (a) the supine
and in (b) the lateral decubitus set-ups. Dosimetry of whole-breast radiotherapy was made for a prescribed dose of 50 Gy at mid-
thickness using 4 MV photons.
46.8 Gy at the central slice at the mid-bridge for tangents.
This formula is to be handled with caution when it comes
to estimating the lung and heart volumetric doses in our
patients. First, because we prescribed 50 Gy, i.e. 7% more
than in Kong et al series. Second, because this formula
was not meant to be applied to ILD. We assumed that, if
anything, the results using this formula would under-
estimate the difference between the supine and the ILD
techniques in terms of volumes of lung and heart
receiving 20 Gy or more as the craniocaudal lengths of
either lung or heart within the simulation field were
diminished in the lateral position (data not shown).
788
M A Bollet, F Campana, Y M Kirova et al
(b)
Pectus excavatum is a rare condition and therefore this
study suffers from a small number of patients making a
formal statistical comparison of the two techniques
inappropriate. We can nevertheless observe that, as
expected, the change of position from the supine to the
lateral decubitus position, with the breast spread out on an
individually chosen dedicated breast support, made the
thickness of breast smaller by 41–62% and thus also made
the dose in depth more homogeneous. The breast took the
shape of a parallelogram making it easier to obtain a
homogeneous dose with the use of a single wedge
(Figure 3b). The most important benefit is that the breast
Figure 4. Simulation films of the
internal fields of the four patients
presenting pectus excavatum both
with the isocentric lateral decubitus
(ILD) and the standard supine (Sup)
radiotherapy techniques. Each gra-
duation represents a centimetre.
Central lung distance (CLD) is the
perpendicular distance from the
tangential field edge to the poster-
ior part of the anterior chest wall at
the centre of the field. Maximal
heart distance (MHD) is the maximal
width of heart in the fields.
The British Journal of Radiology, October 2006
Short communication: ILD breast irradiation – pectus excavatum

Table 1. Dosimetric study comparing the isocentric lateral decubitus (ILD) and the supine breast radiotherapy techniques in four
women presenting with pectus excavatum. Dosimetries of whole-breast radiotherapy were made for a prescribed dose of 50 Gy
at mid-thickness using either 60-Cobalt gammas or 4 MV photons
Patient Breast thickness Central lung 20 Gy ipsilateral Maximal heart 20 Gy heart Maximal dose to
(cm) distance (cm) lung volume (%) distance (cm) volume (%) breast (Gy)

Supine ILD Supine ILD Supine ILD Supine ILD Supine ILD Supine ILD

A 10.9 6.4 2.6 1.0 21 5 0.0 0.0 0 0 52 52

B 12.0 5.2 4.0 0.0 34 0 1.7 0.0 9 0 54 53
C 9.4 3.6 3.2 0.0 26 0 0.0 0.0 0 0 53 52
D* 11.4 5.6 2.7 0.4 22 0 0.8 0.0 4 0 54 52
*For the dosimetry study of this patient, we chose to reduce the prescribed dose to 48 Gy in 24 fractions with the supine
technique in order to keep the maximum dose below 54 Gy (this patient has however been treated with the ILD technique to
a total dose of 50 Gy in 25 fractions).

tissue is located at a greater distance from the chest wall, Acknowledgments

making it easier to encompass the whole breast within the
radiation field without taking in any of the heart and less
than 1 cm of the underlying lung. The resultant estimated
percentage of ipsilateral heart and lung are thus reduced to
a maximum of 0% and 5%, respectively. In contrast to this,
the standard supine position had central lung depths of at
least 2.5 cm resulting in the estimated volume of lung
receiving 20 Gy always being greater than 20%. The
protocols in use at the Institut Curie define the maximum
depth of lung and/or left ventricle receiving greater than
20 Gy to be 2 cm, thus the standard supine technique
would be unacceptable.
IMRT has been proposed as a means to decrease the
volume of the ipsilateral lung receiving a radiation dose
that is above the tolerance threshold [6, 7]. Fogliata et al
have suggested that three-field conformal radiotherapy
technique permitted a reduction of the ipsilateral lung
volume receiving more than 20 Gy from 24% with two-
field tangential treatment to less than 20%. This reduction
was the same whether using IMRT or not [8]. The
drawback of IMRT and the 3D conformal technique
described by Fogliata et al, contrary to the ILD technique,
is that it increases the volume of surrounding normal
tissues (the heart, spinal cord, and contralateral breast and
lung) receiving low-dose irradiation compared with the
supine technique [6]. The other advantage the ILD
technique has over a three-dimensional radiotherapy
approach is its cost-effectiveness, both financially and
time-wise. However, we have to bear in mind that the ILD
technique is not designed to treat the breast in the case of a
treatment with radiotherapy to the lymph node bearing
areas; in such cases, IMRT with or without active breathing
control, or even proton therapy [8], could be considered.
The acute skin toxicity of the ILD technique was good,
with no need for a suspension of treatment and only a
grade I RTOG skin toxicity in all of our four patients of
whom two were treated with a boost of up to 66 Gy.
Conclusions
The ILD technique is an easy, cost-effective means to
offer breast preservation to patients with pectus excava-
tum, without exposing the underlying heart and, more
particularly, ipsilateral lung to unnecessary toxicity.
We thank J Y Kristner for designing the treatment set-
up device, Chantal Dauphinot and Yann Brunet for their
help in the dosimetry study and all the members of the
Groupe Sein at the Institut Curie for their participation in
the care of the patients.
References
1. Clarke M, Collins R, Darby S, et al. Effects of radiotherapy
and of differences in the extent of surgery for early breast
cancer on local recurrence and 15-year survival: an over-
view of the randomised trials. Lancet 2005;366:2087–106.
2. Lind PA, Pagnanelli R, Marks LB, et al. Myocardial
perfusion changes in patients irradiated for left-sided breast
cancer and correlation with coronary artery distribution. Int
J Radiat Oncol Biol Phys 2003;55:914–20.
3. Graham MV, Purdy JA, Emami B, et al. Clinical dose-
volume histogram analysis for pneumonitis after 3D
treatment for non-small cell lung cancer (NSCLC). Int J
Radiat Oncol Biol Phys 1999;45:323–9.
4. Yu X, Prosnitz RR, Zhou S, et al. Symptomatic cardiac
events following radiation therapy for left-sided breast
cancer: possible association with radiation therapy-induced
changes in regional perfusion. Clin Breast Cancer
2003;4:193–7.
5. Perez CA, Garcia DM, Kuske RR. Breast: Stage Tis, T1, and
T2 tumors. In: Company JL, editor. Principles and practice
of radiation oncology. Second Edition. Philadelphia: JB
Lippincott Company, 1994:877–947.
6. Teh BS, Lu HH, Sobremonte S, et al. The potential use of
intensity modulated radiotherapy (IMRT) in women with
pectus excavatum desiring breast-conserving therapy.
Breast J 2001;7:233–9.
7. Thilmann C, Zabel A, Kuhn S, et al. Inversely planned
intensity modulated radiotherapy for irradiation of a
woman with breast cancer and funnel chest. Strahlenther
Onkol 2002;178:637–43.
8. Fogliata A, Bolsi A, Cozzi L. Critical appraisal of treatment
techniques based on conventional photon beams, intensity
modulated photon beams and proton beams for therapy of
intact breast. Radiother Oncol 2002;62:137–45.
9. Fourquet A, Campana F, Rosenwald JC, et al. Breast
irradiation in the lateral decubitus position: technique of
the Institut Curie. Radiother Oncol 1991;22:261–5.
10. Campana F, Kirova YM, Rosenwald JC, et al. Breast
radiotherapy in the lateral decubitus position: a technique
to prevent lung and heart irradiation. Int J Radiat Oncol Biol
Phys 2005;61:1348–54.

The British Journal of Radiology, October 2006 789


11. ICRU. Report 62: Prescribing, recording, and reporting
photon beam therapy (supplement to ICRU Report 50).
Bethesda, MD: International Commission on Radiation
Units and Measurements; 1999.
12. ICRU. Report 50: Prescribing, recording, and reporting
photon beam therapy. Washington DC: International
Commission on Radiation Units and Measurements; 1993.
13. Bartelink H, Garavaglia G, Johansson KA, et al. Quality
assurance in conservative treatment of early breast cancer.
Report on a consensus meeting of the EORTC Radiotherapy
and Breast Cancer Cooperative Groups and the EUSOMA
(European Society of Mastology). Radiother Oncol
1991;22:323–6.
790
M A Bollet, F Campana, Y M Kirova et al
14. Bornstein BA, Cheng CW, Rhodes LM, et al. Can simulation
measurements be used to predict the irradiated lung
volume in the tangential fields in patients treated for breast
cancer? Int J Radiat Oncol Biol Phys 1990;18:181–7.
15. Kong FM, Klein EE, Bradley JD, et al. The impact of central
lung distance, maximal heart distance, and radiation
technique on the volumetric dose of the lung and heart
for intact breast radiation. Int J Radiat Oncol Biol Phys
2002;54:963–71.
16. Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation
Therapy Oncology Group (RTOG) and the European
Organization for Research and Treatment of Cancer
(EORTC). Int J Radiat Oncol Biol Phys 1995;31:1341–6.
The British Journal of Radiology, October 2006
The British Journal of Radiology, 79 (2006), 791–798

REVIEW ARTICLE

Hypoxia in head and neck cancer

1
A Y ISA, MBChB, MRCS, M Phil, 2T H WARD, PhD,
3
C M L WEST, BA, PhD,
4
N J SLEVIN, MBChB, FRCP, FRCR.BW
and 5J J HOMER, MD, FRCS (ORL-HNS)

1
Department of Surgery, Christie Hospital, Manchester, 2Drug Development Group, Paterson
Institute for Cancer Research, Manchester, 3Academic Department of Radiation Oncology,
University of Manchester, Manchester, 4Department of Clinical Oncology, Christie Hospital,
Manchester and 5Departments of Surgery, Christie Hospital and University Department of
ORL-HNS, Manchester Royal Infirmary, Manchester, UK

ABSTRACT. A high level of hypoxia in solid tumours is an adverse prognostic factor for
the poor outcome of cancer patients following treatment. This review describes the Received 21 July 2005
status of research into finding a practical method for measuring hypoxia and treating Revised 19 April 2006
Accepted 22 May 2006
hypoxic tumours. The application of such methodology would enable the selection of
head and neck cancer treatment based on an individual’s tumour oxygenation status. DOI: 10.1259/bjr/17904358
This individualization would include the selection not only of surgery or radiotherapy,
but also of novel hypoxia-modification strategies. ’ 2006 The British Institute of
Radiology

Hypoxia, the inadequate supply of blood-borne
oxygen, has been linked with a poor tumour response
to radiotherapy [1–4]. This resistance relates, in part, to
the radiobiological effects of hypoxia. However, hypoxia
is also a marker of an aggressive tumour phenotype and
is associated with a poor outcome following surgery [5,
6]. Furthermore, hypoxia is implicated in the resistance
of tumours to some chemotherapeutic agents [7]. This
review discusses the biology of tumour hypoxia,
methods for measuring tumour hypoxia, possible clinical
applications of hypoxia measurements and potential
hypoxia modification approaches for patients with head
and neck cancer.
The biology of tumour hypoxia
Tumour cells require a host vasculature for their
supply of nutrients and oxygen, but oxygen cannot
diffuse further than around 150 mm through tissues. As
tumour growth outstrips its vasculature, the cells
become hypoxic [8]. This hypoxia, termed chronic or
diffusion-limited hypoxia, occurs adjacent to areas of
necrosis. Tumour hypoxia also arises from the transient
or intermittent occlusion of tumour blood vessels, called
acute or perfusion-limited hypoxia. Although the
mechanisms behind the development of acute hypoxia
are unknown, it may arise from the blocking of vessels
This work was supported by the Wolfson Foundation, the Christie
Hospital Endowment Fund and the National Translational Cancer
Research Network.
Address correspondence to: Mr J J Homer, University Department
of Otolaryngology-Head and Neck Surgery, Manchester Royal
Infirmary, Oxford Road, Manchester M13 9DL, UK.
by circulating blood and tumour cells, blood vessel
collapse from high tumour interstitial pressure, and/or
the interruption of tumour blood flow. All these factors
could occur because of the chaotic, disorganized and
fragile nature of tumour compared with normal tissue
vasculature [9, 10].
Thomlinson and Gray first proposed the existence of
hypoxia in human tumours in 1955 [11] and confirmed
the radioresistance of hypoxic mammalian cells, which
was described in 1936 by Mottram [12]. Thomlinson and
Gray suggested that the presence of quiescent but viable
hypoxic cells, capable of re-oxygenation during fractio-
nated radiotherapy, would limit the success of treatment.
This suggestion spurred research into finding methods
for measuring tumour hypoxia and for overcoming
hypoxic cells in patients undergoing radiotherapy. It is
only in the last decade, however, that hypoxia has
emerged as a key factor in driving malignant progres-
sion. Hypoxia is involved in the transcriptional regula-
tion of a number of genes. Hypoxia-inducible factor-1a
(HIF-1a) and HIF-2a are transcription factors that
mediate cell responses to hypoxia. Cells express HIF-1a
continuously, but the protein degrades rapidly under
normoxic conditions. In hypoxia, the protein is stabilized
and induces the transcription of a number of genes
including those involved in angiogenesis, glycolysis, pH
control and oxygen delivery. HIF-1a transcribed genes
include the key angiogenic growth factor, vascular
endothelial growth factor (VEGF) [13], as well as glucose
transporter 1 (Glut-1) [14] and the pH regulator, carbonic
anhydrase IX (CAIX) [15]. Thus, cellular response to
hypoxia strives to prevent cell death. HIF-1a up-
regulation occurs within 2 min of hypoxia, accumulating
rapidly over 30 min to peak after another 30 min [16].

The British Journal of Radiology, October 2006 791


Clinical importance of hypoxia
The biological effect of radiation depends on the
degree of oxygenation, and hypoxic cells are approxi-
mately three-fold more resistant than well-oxygenated
cells. This oxygen effect is due to the interaction between
oxygen and the free radicals produced when radiation is
absorbed in tissues. Radiation absorbed in tissues
produces highly reactive, short-lived free radicals, which
produce double strand breaks in DNA leading to cell
death. The oxygen increases the damage produced by
radiation by increasing the lifetime of the free radicals.
Due to the short life-span of the free radicals, oxygen
needs to be present at the time of irradiation to be
effective [17]. Studies showed poorer locoregional con-
trol and survival in patients with hypoxic compared with
oxygenated head and neck squamous cell cancers treated
with radiation [1–4].
There is evidence that hypoxia limits the effectiveness of
not only radiotherapy but also surgery. In studies measur-
ing tumour oxygenation using polarographic oxygen
microelectrodes, tumours treated with primary surgery
with or without radiation had a poor outcome [5, 6]. Hockel
et al found the median partial pressure of oxygen (pO2) to
be the strongest independent predictor of overall and
disease-free survival in patients with locally-advanced
uterine cervical cancer irrespective of treatment modality.
Similar findings reported with high-grade non-metastatic
soft tissue sarcomas [6] indicate the involvement of factors
other than hypoxia-mediated radioresistance. There is
evidence that hypoxia enhances genetic instability in
tumour cells and selects for tumour cell populations with
diminished apoptotic potential, and increased aggressive-
ness and metastatic potential [17–19].
Hypoxia has also been implicated in the resistance of
tumours to some chemotherapeutic agents, including
those used in the treatment of head and neck cancers [7].
Hypoxic cells are resistant to 5-fluorouracil, doxorubicin,
bleomycin and cisplatin [20–23]. Although the mecha-
nisms behind this drug resistance are not understood,
reduced cellular proliferation, low pH and hypoxia-
induced alterations in gene expression may play a role.
Measuring hypoxia
Various methods are under development for measuring
tumour hypoxia in cancer patients. There is a need for a
method that is practical, fast and reliable, i.e. suitable for
routine clinical application. Although intratumour varia-
bility in oxygenation is a potential confounding factor
affecting many of the measurement strategies, tumour-to-
tumour variability is greater [17]. The ability to differ-
entiate between necrotic and hypoxic areas is also
important, but not possible with all techniques. Whilst
the ultimate result of hypoxia is necrosis (and therefore
necrotic cells might be a measure of the process of
hypoxia), it is the measurement of hypoxic cells that
remain viable that are of the greatest clinical importance.
Oxygen microelectrodes
The current resurgence of interest in tumour hypoxia
stems from data obtained during the 1990s using
792
A Y Isa, T H Ward, C M L West et al
computerized Eppendorf oxygen microelectrodes.
Advantages of Eppendorf over older style electrodes
were the use of fine needles to minimize tissue trauma,
an automatic stepper motor to enable rapid movement
through tissue and to avoid tissue compression artefacts
and computerization to enable the quick collection of
multiple measurements. The electrodes are used with or
without image guidance, depending on the location and
size of the tumour. Le et al showed that median pO2
readings from advanced head and neck cancer nodal
metastases were significantly lower than normal sub-
cutaneous tissue (14.6 mmHg vs 51.2 mmHg; p,0.001)
and in each patient, median tumour metastases pO2 was
consistently lower than that of normal subcutaneous
tissue. 40% of the study group had median pO2
measurements of less than 10 mmHg [24]. Nordsmark
et al [3] found in their study of pre-treatment oxygena-
tion in neck metastasis of advanced squamous cell
carcinoma of the head and neck that the fraction of
pO2,2.5 mmHg was significant as a continuous variable
of local failure following radiotherapy, but median pO2
was not. Another study of head and neck squamous cell
carcinoma (HNSCC) treated with chemoradiotherapy/
radiotherapy showed that the hypoxic subvolume
defined as the percentage of tumour multiplied by
percentage of pO2-values below 5 mmHg were signifi-
cant in multivariate analysis for poorer overall
survival where no correlation was found with median
pO2 [18]. The approach, however, is limited to tumours
that are accessible. Also, oxygen electrodes cannot
differentiate viable hypoxic regions from necrotic areas
[25, 26].
Comet assay
The comet assay uses individual tumour cells dis-
sociated from needle biopsies. It is a sensitive method of
measuring DNA strand breaks in tumour cells after
single doses of 3.5–10 Gy. The approach is based on the
fact that ionizing radiation produces around three-fold
more damage in well-oxygenated compared with
hypoxic cells (i.e. it measures radiobiologically hypoxic
cells) [27, 28]. The comet assay was compared with the
oxygen microelectrode system in a group of patients
with heterogeneous advanced tumours (including non-
HNSCC). Aquino-Parsons et al found a correlation
between hypoxic fraction measured by comet assay
and percentage of pO2 values ,5 mmHg measured by
the oxygen microelectrode (r250.46, p,0.001). In their
study, tumours defined as hypoxic with a median
pO2,10 mmHg were found to have .20% radiobio-
logically hypoxic cells as measured by the comet assay
[28]. Another study compared the two methods in node
positive Stage III–IV HNSCC. The majority had Stage IV
disease and were randomized to cisplatin-based induc-
tion chemotherapy followed by concurrent chemora-
diotherapy with or without tirapazamine. A weak but
statistically significant negative correlation between
oxygen microelectrode median pO2 and 1 min median
tail moment as a measure of DNA damage post-
irradiation from the comet assay was found (r25 0.08,
p50.05). There was a significant correlation between
response and comet median tail moment (p50.001);
The British Journal of Radiology, October 2006
Review article: Hypoxia in head and neck cancer
however, no correlation was found between response
and median pO2 readings [24]. The lack of correlation
between median pO2 and response is not unexpected as
Nordsmark et al had described this in an earlier study
[3]. The use of median pO2 instead of percentage
pO2,5 mmHg may also explain the negative correlation
between the oxygen microelectrode and the comet assay.
The advantage of the comet assay lies in its ability to
measure radiobiologically hypoxic cells only and not
necrosis. Radiation doses in excess of 3.5 Gy are required
to produce sufficient DNA damage to distinguish a
hypoxic subpopulation, thus limiting its application to
patients receiving larger fractions or doses.
Contamination of the fine needle aspirate sample by
circulating white blood cells (which are unirradiated)
could falsely reduce the median tail moment making the
tumour seem more hypoxic [28, 29].
Using nitroimidazoles as hypoxic markers
Nitroimidazoles are nitro aromatic compounds that
bind to hypoxic cells. Nitroimidazoles diffuse easily into
hypoxic cells due to their high solubility and low
metabolism [27]. A non-therapeutic dose of a nitroimi-
dazole is administered, systemically prior to biopsy or
resection of a tumour, and hypoxic cells in histological
sections can be identified using immunohistochemistry
(IHC), flow cytometry and immunofluorescence.
Nitroimidazoles used as hypoxic markers include mis-
onidazole, pimonidazole, etanidazole, EF5 and nitroimi-
dazole-theophylline [17, 30]. An advantage of the
approach is its applicability to tumours inaccessible to
oxygen electrodes. However, a biopsy might not be
representative of the heterogeneity of hypoxia within a
whole tumour. Although no relationship was found
between pimonidazole binding and oxygen electrode
data in carcinoma of the cervix [31], a study in head and
neck cancer showed that high pimonidazole binding did
predict adverse treatment outcome. Locoregional
tumour control was statistically significantly lower for
patients who had hypoxic tumours with 2-year control
rates of 48% vs 87% for tumours with high and low
pimonidazole binding levels, respectively [32]. This
finding raises the possibility that measurements of
pimonidazole binding in head and neck cancer might
be useful as a selection tool for hypoxia-modifying
treatments. However, pimonidazole staining in highly
differentiated or keratanized tumour tissue has been
described. The question as to whether these areas are
hypoxic has been raised. This is especially pertinent in
head and neck cancers that have a considerable amount
of keratinization as pimonidazole staining of keratinized
but non-hypoxic areas would lead to an overestimation
of the level of hypoxia. Janssen et al found in 25% of their
head and neck cancer series up to 30% of staining was in
well-differentiated areas [33]. Further investigation is
required.
Measuring endogenous markers of hypoxia
Hypoxia-inducible proteins are under investigation
as potential endogenous markers of hypoxia. The
The British Journal of Radiology, October 2006
expression of HIF and the proteins it regulates tran-
scriptionally can be measured using IHC [34]. The
advantage of the approach for routine clinical use is
that, as for nitroimidazole probes, only a biopsy is
required. Disadvantages are that a single biopsy might
not be representative of a whole tumour and that the
proteins investigated can be upregulated by factors other
than hypoxia, i.e. they are not hypoxia specific. Potential
endogenous markers of hypoxia investigated include
HIF-1a, HIF-2a, Glut-1 and CAIX. Strong HIF-1a expres-
sion was associated with a significantly poorer outcome
following radiotherapy in oropharyngeal tumours
expressing HIF-1a [35]. High HIF-1a and HIF-2a expres-
sion was associated with poor outcome in patients with
advanced squamous cell carcinoma of the head and neck
treated with concurrent carboplatin chemoradiotherapy
[36]. Complete response was lower in tumours expres-
sing HIF-1a and/or HIF-2a 49% vs 86% (p50.004) and
high expression of HIF-1a and HIF-2a was significantly
associated with poor local relapse-free survival (p50.003
and p50.003, respectively) and overall survival (p50.05
and p50.001, respectively) in univariate analysis. HIF-2a
over expression was also an adverse prognostic factor for
locoregional control (p50.002) and overall survival
(p50.0004) in head and neck cancers treated with
accelerated radiotherapy [37]. However, the expression
of HIF-1a in surgically treated head and neck squamous
cell carcinoma was associated with improved disease-
free survival (p50.016) and overall survival (p50.027)
where no difference in outcome was seen with HIF-2a
expression [34]. A more recent study of T1/2 squamous
cell carcinomas of the floor of mouth confirmed the
positive prognostic effect of HIF-1a expression in
surgically treated patients. In this series of 85 patients,
HIF-1a expression was associated with a significantly
improved 5-year survival rate (p,0.01) and a signifi-
cantly increased disease-free period (p50.01) [38]. It
remains to be established whether the importance of the
expression a hypoxia-inducible marker that is upregu-
lated by factors other than hypoxia will vary with disease
stage, but clearly this is an area that needs further study.
Glut-1 is a ubiquitously expressed facilitative glucose
transporter that is over-expressed in a number of
tumours [39]. Although Glut-1 is upregulated by a
variety of agents and conditions, it has potential as an
intrinsic marker of hypoxia because of dual control of
expression in hypoxic conditions via HIF-1 and reduced
oxidative phosphorylation [14]. Its expression is hetero-
geneous in cell cytoplasm and membrane, and over-
expression has been shown in tumours including those
of the breast [40], lung [41], thyroid [42], cervix [39],
hypopharynx [43] and oral cavity [44]. Glut-1 staining
tends to localize in necrotic and perinecrotic areas of
tumours [39]. Over expression of Glut-1 was associated
with a poor prognosis in oral squamous cell carcinoma
[44, 45] and hypopharyngeal carcinoma [43]. For exam-
ple, survival times were 138 months and 60 months for
patients with surgically treated oral squamous cell
carcinoma expressing low versus high levels of Glut-1,
respectively [44]. Recent experiments on human tumour-
derived xenografts revealed that the chemosensitivity
of certain alkylating agents may be influenced by Glut-1
expression [46]. However, there is only a very weak
[39] or no [47] relationship between tumour Glut-1
793

expression and oxygen variables (pO2, hypoxic fractions
2.5 and 5). The latter, along with the lack of prognostic
significance of Glut-1 expression in multivariate analysis,
has raised questions on the suitability of Glut-1 as an
endogenous hypoxia marker [47].
CAIX is a transmembrane glycoprotein that is induced
by hypoxia via the HIF-1 pathway and may be a useful
endogenous marker of hypoxia [15]. Expression of CAIX
in hypoxic tumours is localized to perinecrotic regions
and is thought to confer a survival advantage by
maintaining intracellular pH [15]. CAIX expression
correlated with tumour oxygenation status measured
using oxygen electrodes in cervical squamous cell
carcinomas and is associated with poor outcome follow-
ing radiotherapy [48]. CAIX expression is associated
with poor complete response rate (40% vs 70%, p50.02)
to chemoradiotherapy in advanced squamous head and
neck cancer [49]. Interestingly, the co-expression of HIF-
1a and CAIX were associated with a poorer progression-
free survival (p50.04) in chemoradiated locally advanced
nasopharyngeal cancer than the expression of either
marker alone [50]. The combined expression of CAIX and
Glut-1 in HNSCC treated radiotherapy with or without
chemotherapy was also associated with poorer local
control (p50.02) and disease-free survival (p50.04)
where expression of either marker alone was not
significantly correlated with outcome [51].
More recently, interest has grown in osteopontin – an
extracellular matrix protein involved with tumour cell
invasion, migration, angiogenesis and tumour growth
[52, 53]. Osteopontin is secreted in plasma making it an
attractive, easy method of assessing hypoxia. Plasma
osteopontin levels were inversely correlated with pO2
(p50.003, r5 –0.42) in HNSCC and osteopontin was an
independent predictor for freedom from relapse and
survival on multivariate analysis [54]. The prognostic
significance of osteopontin in HNSCC treated by radio-
therapy was supported by Overgaard et al [55].
Moreover, the prognosis of patients with high levels of
osteopontin was improved by addition of hypoxia
modifier nimorazole. Locoregional failure (RR 0.19
[95% CI 0.08–0.44]) and disease specific mortality (RR
0.25, 95% CI 0.11–0.59) was more frequent in patients
with high concentration of osteopontin assigned placebo
than those assigned nimorazole [55].
Hypoxic regions within solid tumours will be not only
oxygen deficient, but also acidic and nutrient depleted.
Recent studies indicating that glucose concentration [56]
and pH [57] affect the expression of endogenous markers
of hypoxia raise concern over the generalized use of
these proteins or genes as hypoxic markers. In vitro
experiments using FADU human pharyngeal carcinoma
and HT1080 human fibrosarcoma cells showed a lack of
hypoxic HIF-1a accumulation in glucose-depleted con-
ditions. Serum-depleted conditions also caused
decreased hypoxic HIF-1a accumulation in FADU cells
despite the presence of glucose [56]. Even though
Vordermark et al [56] showed no effect of pH on HIF-
1a expression, Sorensen et al [57] found that the
expression of CA9, GLUT1 and Osteopontin (OPN)
genes were inhibited in hypoxia when extracellular pH
was reduced to 6.3. As CA9 and GLUT1 are HIF-1
regulated genes it is likely that HIF-1 will be affected by
pH as well. A possible explanation for the findings of
794
A Y Isa, T H Ward, C M L West et al
Vordermark et al may be that the lowest pH level of 6.7
was not acidic enough to show any inhibitory effect. The
time course and the levels to which the genes were
upregulated were different, as was the oxygen concen-
tration at which upregulation was maximal, e.g. after
24 h exposure CA9 upregulation was maximal at 1%
oxygen whereas GLUT1 and OPN were maximal at
0.01% and 0% oxygen, respectively [57]. More work is
required to reveal any other factors that may effect these
proteins or genes that have until now been loosely
termed endogenous markers of hypoxia. Perhaps the
term ‘‘markers of poor prognosis’’ may be more apt. It
may be that instead of utilizing a single protein, tumours
should be tested for multiple markers to create a
hypoxia-associated molecular profile.
Non-invasive imaging
2-Fluoro-2-deoxy-D-glucose (FDG) positron emission
tomography (PET) utilizes the concept that most
tumours exhibit accelerated glycolysis allowing 18F-
labelled FDG to be trapped in tissues with a higher
metabolic rate than normal tissues [58]. There are several
studies with contradictory findings with regards the
correlation of Glut-1 expression and 18F-FDG uptake in
tumours [59–62]. Most studies compared standardized
uptake ratios or values from pre-operative 18F-FDG PET
scans and compared them with immunohistochemical
Glut-1 expression from a section of the tumour or a
biopsy specimen. Interestingly, a recent article based on
a rat tumour model showed a good correlation between
18
F-FDG uptake and Glut-1 expression (r50.829,
p,0.001) at a micro-regional level [60]. Even so, the
ability of 18F-FDG to detect tumour hypoxia is question-
able [63, 64]. The imaging of radiolabelled nitroimida-
zoles by single-photon emission computer tomography
(SPECT) and PET as a non-invasive measure of hypoxia
is also being studied, amongst them 18F-fluoromisonida-
zole (18FMISO) [63], 18F-fluoroerythronitroimidazole
(18F-FETNIM) [65], 18F-fluoroetanidazole (18F-FETA)
[66], and 18F-EF3 [67]. Metal labelled hypoxia markers
are also currently being developed, e.g. Cu ATSM [68].
Preliminary clinical testing of some of these markers has
been promising with small series indicating a prediction
of radiotherapy outcome [65, 69]. Advantages of the PET
based approach include the ability to assess hypoxia of
the whole tumour using a non-invasive technique that
evaluates viable regions within the tumour. Deep lesions
that are otherwise inaccessible to the microelectrodes can
be studied and sequential scans can indicate response
to treatment. A disadvantage is the limited spatial
resolution.
Cross sectional imaging methods are also being
studied, in particular MRI. The methods are not hypoxia
specific, but the wide availability of MRI is an advantage
of the approach. Dynamic contrast-enhanced MRI has
been the most widely studied method. The method
yields parameters that reflect, amongst other parameters,
tumour perfusion and vascularity. There is some
evidence that the approach can yield data that reflect
tumour oxygenation [70], and predict radiotherapy
outcome in patients with carcinoma of the cervix and
advanced head and neck cancer [71, 72].
The British Journal of Radiology, October 2006
Review article: Hypoxia in head and neck cancer
Hypoxia modification strategies
There are a number of approaches under investigation
to improve the oxygenation status of tumours. Oxygen
delivery to a tumour can be raised by increasing the
oxygen content of inspired air or raising haemoglobin
levels. There are also hypoxia specific cytotoxins such as
bioreductive drugs. For tumours treated with radiation,
hypoxic cell radiosensitizers and radiotherapy modula-
tion strategies are being developed.
Increasing oxygen delivery
One of the earliest hypoxia modification strategies
studied was hyperbaric oxygen. Although breathing
oxygen during radiotherapy was used to some success
in the past, delivery was complex and patient tolerance
poor. Carbogen (2–5% carbon dioxide and oxygen) was
suggested as an alternative to pure oxygen because of it
causing vasodilation, an increase in respiratory drive and
a right shift in the oxygen-haemoglobin dissociation.
Patients breathe carbogen through a face mask during
their radiotherapy. Accelerated radiotherapy combined
with carbogen and nicotinamide (ARCON) is being
investigated in patients with head and neck cancer. The
accelerated radiotherapy targets rapidly proliferating
tumour cells, the carbogen maximizing arterial pO2 and
nicotinamide is a vasodilator that aims to minimize
diffusion-limited hypoxia. The radiosensitizing effect of
ARCON can also be expected in normal tissue hence the
need for reduced doses to the spinal cord and laryngeal
cartilage [73, 74] and the importance of its selective use
based perhaps on hypoxia measurement. So far, results
with ARCON in clinical trials in squamous cell carci-
noma of the head and neck have been mixed [74, 75].
How much additional benefit with ARCON is due to
hypoxia reduction as opposed to reduction in tumour
clonogenic repopulation by accelerated fractionated
radiotherapy has not been ascertained.
Raising haemoglobin levels
Anaemia is associated with a poor outcome following
chemoradiotherapy and radiotherapy [18, 76, 77]. The
detrimental effect of anaemia on radiotherapy outcome
has led to the routine use of blood transfusions prior to
radiotherapy in patients with low haemoglobin levels.
An alternative approach under investigation is the use of
erythropoietin. Erythropoietin increases red cell produc-
tion and theoretically reduces tumour hypoxia. In a
retrospective study of patients with oropharyngeal
cancer treated with chemoradiotherapy prior to surgical
resection, Glaser et al found that a low pre-operative
haemoglobin level was an independent adverse prog-
nostic factor for locoregional control and survival [76].
This was reversed with the use of recombinant human
erythropoietin during chemoradiotherapy. However, a
recent double-blind, placebo-controlled randomized trial
using erythropoietin in head and neck cancer patients
undergoing radiotherapy did not show any improve-
ment in treatment outcome and actually suggested
poorer locoregional progression-free survival [78].
The British Journal of Radiology, October 2006
Further carefully controlled trails are required to study
this more closely.
Hypoxia specific cytotoxins
Bioreductive drugs are reduced under hypoxia or by
reducing enzymes to cytotoxic metabolites [79]. Quinone
(e.g. mitomycin C, porfiromycin), nitroimidazole (e.g.
RSU-1069) and N-oxide (e.g. tirapazamine) compounds
can be bioreduced to cytotoxic species. Although
mitomycin C was used to treat head and neck cancer
patients before it was known to be a bioreductive agent,
it is not hypoxia selective and there is little differential
cell killing between oxygenated and hypoxic cells. Of the
new agents developed for their preferential toxicity
toward hypoxic cells, tirapazamine is probably of most
interest. Tirapazamine exhibits selective cytotoxicity for
hypoxic cells. It is reduced in hypoxia to a highly reactive
radical, which is capable of causing single and double
strand DNA breaks. In normoxia this radical is back
oxidized to the inert parent compound. This cycle is
thought to confer hypoxic selectivity [80]. Tirapazamine
can be given with other cytotoxic agents or with
radiotherapy. A phase I trial of fractionated radiotherapy
with concomitant tirapazamine and cisplatin in
advanced oropharyngeal cancer, found dose limiting
toxicity in the form of febrile neutropenia. This was
overcome by omitting tirapazamine in weeks 5 and 6.
Measurement of hypoxia by 18F-misonidazole PET
scanning revealed a decrease in tumour hypoxia follow-
ing treatment [81]. A recent randomized phase II trial
comparing tirapazamine, cisplatin and radiation versus
fluorouracil, cisplatin and radiation in 120 stage III and
IV HNSCC patients found a trend in favour of the
tirapazamine arm for both locoregional control (84% vs
66%) and failure-free survival (55% vs 44%). This became
statistically significant for locoregional control following
adjustment for known prognostic factors [82].
Gene therapy strategies are also under investigation to
target hypoxic cells in solid tumours. These include
using anaerobic bacteria, e.g. Clostridium Spp as a vector
to deliver genetic material preferentially to hypoxic
tissue at sufficient levels to cause a therapeutic effect
[79]. HIF-1a downregulation by intratumoural delivery
of antisense HIF-1a plasmid showed promising results
[83].
Hypoxic cell radiosensitizers
As discussed above, nitroimidazoles can diffuse into
hypoxic cells due to their high solubility and low
metabolism. The radiosensitizing effects of these com-
pounds are related to their electron affinity, mimicking
the oxygen effect, increasing DNA damage and restoring
radiosensitivity. The use of first generation nitroimida-
zoles (e.g. misonidazole) was limited due to intolerable
side effects including irreversible peripheral neuropathy.
However, nimorazole, a 5-nitroimidazole derivative, has
fewer side effects [84]. A phase III trial in Denmark
showed significantly increased locoregional control in
supraglottic and pharyngeal tumours treated with
nimorazole and conventional radiotherapy [77].
795

Improving the delivery of radiotherapy
Radiotherapy can be modulated to deliver more
radiation to hypoxic tumours. Intensity-modulated-
radiation therapy (IMRT) and three-dimensional con-
formal radiation therapy (3D-CRT) [85] enables small
irregular volumes to be ‘‘painted’’ with radiotherapy.
With better methods of detecting hypoxic microregions
in the future, possibly with PET and SPECT, it may be
possible to select hypoxic regions within tumour for
treatment with higher radiation doses using these
techniques.
The future: individualizing treatment according to
hypoxic status
The choice of treatment for patients with head and
neck cancer is often a matter of clinician and centre
preference. There is an unmet need for biological
parameters to individualize treatments. The various
ways of measuring hypoxia provide potentially promis-
ing ways of predicting response to radiotherapy. This
may allow clinicians to choose primary surgery or
(chemo)radiotherapy on a scientific, individualized
basis, rather than by protocol and acumen. This
individualization of treatment has potential to improve
patient management in two ways. First, a decrease in
radiation failures would reduce the need for salvage
surgery, which is associated with greater surgical
morbidity than primary surgery [86]. Second, there
would be a more appropriate selection of patients for
organ preserving chemoradiotherapy. Furthermore,
there are number of specific hypoxia modification
strategies that can be applied to patients on an
individualized basis.
Progress in the development of a tool for measuring
tumour hypoxia should enable the future informed
clinical choice between primary (chemo)radiotherapy,
(chemo)radiotherapy with hypoxia modification strate-
gies or radical surgery for locally advanced tumours.
References
1. Brizel DM, Dodge RK, Clough RW, Dewhirst MW.
Oxygenation of head and neck cancer: changes during
radiotherapy and impact on treatment outcome. Radiother
Oncol 1999;53:113–7.
2. Gatenby RA, Kessler HB, Rosenblum JS, Coia LR,
Moldofsky PJ, Hartz WH, et al. Oxygen distribution in
squamous cell carcinoma metastases and its relationship to
outcome of radiation therapy. Int J Radiat Oncol Biol Phys
1988;14:831–8.
3. Nordsmark M, Overgaard M, Overgaard J. Pretreatment
oxygenation predicts radiation response in advanced
squamous cell carcinoma of the head and neck. Radiother
Oncol 1996;41:31–9.
4. Nordsmark M, Bentzen SM, Rudat V, Brizel D, Lartigau E,
Stadler P, et al. Prognostic value of tumor oxygenation in
397 head and neck tumors after primary radiation therapy.
An international multi-center study. Radiother Oncol
2005;77:18–24.
5. Hockel M, Schlenger K, Aral B, Mitze M, Schaffer U, Vaupel
P. Association between tumor hypoxia and malignant
progression in advanced cancer of the uterine cervix.
Cancer Res 1996;56:4509–15.
796
A Y Isa, T H Ward, C M L West et al
6. Brizel DM, Scully SP, Harrelson JM, Layfield LJ, Bean JM,
Prosnitz LR, et al. Tumor oxygenation predicts for the
likelihood of distant metastases in human soft tissue
sarcoma. Cancer Res 1996;56:941–3.
7. Teicher BA. Hypoxia and drug resistance. Cancer
Metastasis Rev 1994;13:139–68.
8. Horsman MR. Measurement of tumor oxygenation. Int J
Radiat Oncol Biol Phys 1998;42:701–4.
9. Vaupel P. Oxygen transport in tumors: characteristics and
clinical implications. Adv Exp Med Biol 1996;388:341–51.
10. Giatromanolaki A, Harris AL. Tumour hypoxia, hypoxia
signaling pathways and hypoxia inducible factor expres-
sion in human cancer. Anticancer Res 2001;21(6B):4317–24.
11. Thomlinson RH, Gray LH. The histological structure of
some human lung cancers and the possible implications for
radiotherapy. Br J Cancer 1955;9:539–49.
12. Mottram J. A factor of importance in the radio sensitivity of
tumours. Br J Radiol 1936;9:606–64.
13. Richard DE, Berra E, Pouyssegur J. Angiogenesis: how a
tumor adapts to hypoxia. Biochem Biophys Res Commun
1999;266:718–22.
14. Behrooz A, Ismail-Beigi F. Stimulation of glucose transport
by hypoxia: signals and mechanisms. News Physiol Sci
1999;14:105–10.
15. Beasley NJ, Wykoff CC, Watson PH, Leek R, Turley H,
Gatter K, et al. Carbonic anhydrase IX, an endogenous
hypoxia marker, expression in head and neck squamous
cell carcinoma and its relationship to hypoxia, necrosis, and
microvessel density. Cancer Res 2001;61:5262–7.
16. Jewell UR, Kvietikova I, Scheid A, Bauer C, Wenger RH,
Gassmann M. Induction of HIF-1alpha in response to
hypoxia is instantaneous. Faseb J 2001;15:1312–4.
17. Hockel M, Vaupel P. Tumor hypoxia: definitions and
current clinical, biologic, and molecular aspects. J Natl
Cancer Inst 2001;93:266–76.
18. Stadler P, Becker A, Feldmann HJ, Hansgen G, Dunst J,
Wurschmidt F, et al. Influence of the hypoxic subvolume on
the survival of patients with head and neck cancer. Int J
Radiat Oncol Biol Phys 1999;44:749–54.
19. Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ,
Lowe SW, et al. Hypoxia-mediated selection of cells with
diminished apoptotic potential in solid tumours. Nature
1996;379:88–91.
20. Grau C, Overgaard J. Effect of etoposide, carmustine,
vincristine, 5-fluorouracil, or methotrexate on radiobio-
logically oxic and hypoxic cells in a C3H mouse mammary
carcinoma in situ. Cancer Chemother Pharmacol
1992;30:277–80.
21. Liang BC. Effects of hypoxia on drug resistance phenotype
and genotype in human glioma cell lines. J Neurooncol
1996;29:149–55.
22. Koch S, Mayer F, Honecker F, Schittenhelm M, Bokemeyer
C. Efficacy of cytotoxic agents used in the treatment of
testicular germ cell tumours under normoxic and hypoxic
conditions in vitro. Br J Cancer 2003;89:2133–9.
23. Sakata K, Kwok TT, Murphy BJ, Laderoute KR, Gordon GR,
Sutherland RM. Hypoxia-induced drug resistance: compar-
ison to P-glycoprotein-associated drug resistance. Br J
Cancer 1991;64:809–14.
24. Le QT, Kovacs MS, Dorie MJ, Koong A, Terris DJ, Pinto HA,
et al. Comparison of the comet assay and the oxygen
microelectrode for measuring tumor oxygenation in head-
and-neck cancer patients. Int J Radiat Oncol Biol Phys
2003;56:375–83.
25. Airley RE, Loncaster J, Raleigh JA, Harris AL, Davidson SE,
Hunter RD, et al. GLUT-1 and CAIX as intrinsic markers of
hypoxia in carcinoma of the cervix: relationship to
pimonidazole binding. Int J Cancer 2003;104:85–91.
26. Raleigh JA, Dewhirst MW, Thrall DE. Measuring tumor
hypoxia. Semin Radiat Oncol 1996;6:37–45.
The British Journal of Radiology, October 2006
Review article: Hypoxia in head and neck cancer
27. Stone HB, Brown JM, Phillips TL, Sutherland RM. Oxygen
in human tumors: correlations between methods of mea-
surement and response to therapy. Summary of a workshop
held November 19-20, 1992, at the National Cancer
Institute, Bethesda, Maryland. Radiat Res 1993;136:422–34.
28. Aquino-Parsons C, Luo C, Vikse CM, Olive PL. Comparison
between the comet assay and the oxygen microelectrode for
measurement of tumor hypoxia. Radiother Oncol
1999;51:179–85.
29. Olive PL, Durand RE, Jackson SM, Le Riche JC, Luo C,
Ma R, et al. The comet assay in clinical practice. Acta Oncol
1999;38:839–44.
30. Evans SM, Hahn SM, Magarelli DP, Koch CJ. Hypoxic
heterogeneity in human tumors: EF5 binding, vasculature,
necrosis, and proliferation. Am J Clin Oncol 2001;24:467–72.
31. Nordsmark M, Loncaster J, Aquino-Parsons C, Chou SC,
Ladekarl M, Havsteen H, et al. Measurements of hypoxia
using pimonidazole and polarographic oxygen-sensitive
electrodes in human cervix carcinomas. Radiother Oncol
2003;67:35–44.
32. Kaanders JH, Wijffels KI, Marres HA, Ljungkvist AS, Pop
LA, van den Hoogen FJ, et al. Pimonidazole binding and
tumor vascularity predict for treatment outcome in head
and neck cancer. Cancer Res 2002;62:7066–74.
33. Janssen HL, Hoebers FJ, Sprong D, Goethals L, Williams KJ,
Stratford IJ, et al. Differentiation-associated staining with
anti-pimonidazole antibodies in head and neck tumors.
Radiother Oncol 2004;70:91–7.
34. Beasley NJ, Leek R, Alam M, Turley H, Cox GJ, Gatter K,
et al. Hypoxia-inducible factors HIF-1alpha and HIF-2alpha
in head and neck cancer: relationship to tumor biology and
treatment outcome in surgically resected patients. Cancer
Res 2002;62:2493–7.
35. Aebersold DM, Burri P, Beer KT, Laissue J, Djonov V,
Greiner RH, et al. Expression of hypoxia-inducible factor-
1alpha: a novel predictive and prognostic parameter in the
radiotherapy of oropharyngeal cancer. Cancer Res
2001;61:2911–6.
36. Koukourakis MI, Giatromanolaki A, Sivridis E, Simopoulos
C, Turley H, Talks K, et al. Hypoxia-inducible factor
(HIF1A and HIF2A), angiogenesis, and chemoradiotherapy
outcome of squamous cell head-and-neck cancer. Int J
Radiat Oncol Biol Phys 2002;53:1192–202.
37. Koukourakis MI, Bentzen SM, Giatromanolaki A, Wilson
GD, Daley FM, Saunders MI, et al. Endogenous markers of
two separate hypoxia response pathways (hypoxia induci-
ble factor 2 alpha and carbonic anhydrase 9) are associated
with radiotherapy failure in head and neck cancer patients
recruited in the CHART randomized trial. J Clin Oncol
2006;24:727–35.
38. Fillies T, Werkmeister R, van Diest P, Brandt B, Joos U,
Buerger H. HIF1-alpha overexpression indicates a good
prognosis in early stage squamous cell carcinomas of the
oral floor. BMC Cancer 2005;5:84.
39. Airley R, Loncaster J, Davidson S, Bromley M, Roberts S,
Patterson A, et al. Glucose transporter glut-1 expression
correlates with tumor hypoxia and predicts metastasis-free
survival in advanced carcinoma of the cervix. Clin Cancer
Res 2001;7:928–34.
40. Brown RS, Goodman TM, Zasadny KR, Greenson JK, Wahl
RL. Expression of hexokinase II and Glut-1 in untreated
human breast cancer. Nucl Med Biol 2002;29:443–53.
41. Minami K, Saito Y, Imamura H, Okamura A. Prognostic
significance of p53, Ki-67, VEGF and Glut-1 in resected
stage I adenocarcinoma of the lung. Lung Cancer
2002;38:51–7.
42. Schonberger J, Ruschoff J, Grimm D, Marienhagen J,
Rummele P, Meyringer R, et al. Glucose transporter 1 gene
expression is related to thyroid neoplasms with an
The British Journal of Radiology, October 2006
unfavorable prognosis: an immunohistochemical study.
Thyroid 2002;12:747–54.
43. Mineta H, Miura K, Takebayashi S, Misawa K, Araki K,
Misawa Y, et al. Prognostic value of glucose transporter 1
expression in patients with hypopharyngeal carcinoma.
Anticancer Res 2002;22(6B):3489–94.
44. Kunkel M, Reichert TE, Benz P, Lehr HA, Jeong JH, Wieand
S, et al. Overexpression of Glut-1 and increased glucose
metabolism in tumors are associated with a poor prognosis
in patients with oral squamous cell carcinoma. Cancer
2003;97:1015–24.
45. Oliver RJ, Woodwards RT, Sloan P, Thakker NS, Stratford
IJ, Airley RE. Prognostic value of facilitative glucose
transporter Glut-1 in oral squamous cell carcinomas treated
by surgical resection; results of EORTC Translational
Research Fund studies. Eur J Cancer 2004;40:503–7.
46. Airley RE, Phillips RM, Evans AE, Double J, Burger AM,
Feibig HH, et al. Hypoxia-regulated glucose transporter
Glut-1 may influence chemosensitivity to some alkylating
agents: results of EORTC (First Translational Award) study
of the relevance of tumour hypoxia to the outcome of
chemotherapy in human tumour-derived xenografts. Int J
Oncol 2005;26:1477–84.
47. Mayer A, Hockel M, Wree A, Vaupel P. Microregional
expression of glucose transporter-1 and oxygenation status:
lack of correlation in locally advanced cervical cancers. Clin
Cancer Res 2005;11:2768–73.
48. Loncaster JA, Harris AL, Davidson SE, Logue JP, Hunter
RD, Wycoff CC, et al. Carbonic anhydrase (CA IX)
expression, a potential new intrinsic marker of hypoxia:
correlations with tumor oxygen measurements and prog-
nosis in locally advanced carcinoma of the cervix. Cancer
Res 2001;61:6394–9.
49. Koukourakis MI, Giatromanolaki A, Sivridis E, Simopoulos
K, Pastorek J, Wykoff CC, et al. Hypoxia-regulated carbonic
anhydrase-9 (CA9) relates to poor vascularization and
resistance of squamous cell head and neck cancer to
chemoradiotherapy. Clin Cancer Res 2001;7:3399–403.
50. Hui EP, Chan AT, Pezzella F, Turley H, To KF, Poon TC,
et al. Coexpression of hypoxia-inducible factors 1alpha and
2alpha, carbonic anhydrase IX, and vascular endothelial
growth factor in nasopharyngeal carcinoma and relation-
ship to survival. Clin Cancer Res 2002;8:2595–604.
51. De Schutter H, Landuyt W, Verbeken E, Goethals L,
Hermans R, Nuyts S. The prognostic value of the hypoxia
markers CA IX and GLUT 1 and the cytokines VEGF and IL
6 in head and neck squamous cell carcinoma treated by
radiotherapy +/- chemotherapy. BMC Cancer 2005;5:42.
52. Hirama M, Takahashi F, Takahashi K, Akutagawa S,
Shimizu K, Soma S, et al. Osteopontin overproduced by
tumor cells acts as a potent angiogenic factor contributing
to tumor growth. Cancer Lett 2003;198:107–17.
53. Philip S, Bulbule A, Kundu GC. Osteopontin stimulates
tumor growth and activation of promatrix metalloprotei-
nase-2 through nuclear factor-kappa B-mediated induction
of membrane type 1 matrix metalloproteinase in murine
melanoma cells. J Biol Chem 2001;276:44926–35.
54. Le QT, Sutphin PD, Raychaudhuri S, Yu SC, Terris DJ, Lin
HS, et al. Identification of osteopontin as a prognostic
plasma marker for head and neck squamous cell carcino-
mas. Clin Cancer Res 2003;9:59–67.
55. Overgaard J, Eriksen JG, Nordsmark M, Alsner J, Horsman
MR. Plasma osteopontin, hypoxia, and response to the
hypoxia sensitiser nimorazole in radiotherapy of head and
neck cancer: results from the DAHANCA 5 randomised
double-blind placebo-controlled trial. Lancet Oncol
2005;6:757–64.
56. Vordermark D, Kraft P, Katzer A, Bolling T, Willner J,
Flentje M. Glucose requirement for hypoxic accumulation
797

of hypoxia-inducible factor-1alpha (HIF-1alpha). Cancer
Lett 2005;230:122–33.
57. Sorensen BS, Hao J, Overgaard J, Vorum H, Honore B,
Alsner J, et al. Influence of oxygen concentration and pH on
expression of hypoxia induced genes. Radiother Oncol
2005;76:187–93.
58. Zimmer LA, Branstetter BF, Nayak JV, Johnson JT. Current
use of 18F-fluorodeoxyglucose positron emission tomogra-
phy and combined positron emission tomography and
computed tomography in squamous cell carcinoma of the
head and neck. Laryngoscope 2005;115:2029–34.
59. Marom EM, Aloia TA, Moore MB, Hara M, Herndon JE
2nd, Harpole DH Jr, et al. Correlation of FDG-PET imaging
with Glut-1 and Glut-3 expression in early-stage non-small
cell lung cancer. Lung Cancer 2001;33:99–107.
60. Zhao S, Kuge Y, Mochizuki T, Takahashi T, Nakada K, Sato
M, et al. Biologic correlates of intratumoral heterogeneity in
18F-FDG distribution with regional expression of glucose
transporters and hexokinase-II in experimental tumor. J
Nucl Med 2005;46:675–82.
61. Brown RS, Leung JY, Fisher SJ, Frey KA, Ethier SP, Wahl
RL. Intratumoral distribution of tritiated-FDG in breast
carcinoma: correlation between Glut-1 expression and FDG
uptake. J Nucl Med 1996;37:1042–7.
62. Tian M, Zhang H, Nakasone Y, Mogi K, Endo K. Expression
of Glut-1 and Glut-3 in untreated oral squamous cell
carcinoma compared with FDG accumulation in a PET
study. Eur J Nucl Med Mol Imaging 2004;31:5–12.
63. Rajendran JG, Mankoff DA, O’Sullivan F, Peterson LM,
Schwartz DL, Conrad EU, et al. Hypoxia and glucose
metabolism in malignant tumors: evaluation by
[18F]fluoromisonidazole and [18F]fluorodeoxyglucose posi-
tron emission tomography imaging. Clin Cancer Res
2004;10:2245–52.
64. Bentzen L, Keiding S, Horsman MR, Falborg L, Hansen SB,
Overgaard J. Feasibility of detecting hypoxia in experi-
mental mouse tumours with 18F-fluorinated tracers and
positron emission tomography--a study evaluating
[18F]Fluoro-2-deoxy-D-glucose. Acta Oncol 2000;39:629–37.
65. Lehtio K, Eskola O, Viljanen T, Oikonen V, Gronroos T,
Sillanmaki L, et al. Imaging perfusion and hypoxia with
PET to predict radiotherapy response in head-and-neck
cancer. Int J Radiat Oncol Biol Phys 2004;59:971–82.
66. Barthel H, Wilson H, Collingridge DR, Brown G,
Osman S, Luthra SK, et al. In vivo evaluation of
[18F]fluoroetanidazole as a new marker for imaging tumour
hypoxia with positron emission tomography. Br J Cancer
2004;90:2232–42.
67. Mahy P, De Bast M, Leveque PH, Gillart J, Labar D,
Marchand J, et al. Preclinical validation of the hypoxia
tracer 2-(2-nitroimidazol-1-yl)- N-(3,3,3-[(18)F]trifluor-
opropyl)acetamide, [(18)F]EF3. Eur J Nucl Med Mol
Imaging 2004;31:1263–72.
68. Dehdashti F, Grigsby PW, Mintun MA, Lewis JS, Siegel BA,
Welch MJ. Assessing tumor hypoxia in cervical cancer by
positron emission tomography with 60Cu-ATSM: relation-
ship to therapeutic response-a preliminary report. Int J
Radiat Oncol Biol Phys 2003;55:1233–8.
69. Eschmann SM, Paulsen F, Reimold M, Dittmann H, Welz S,
Reischl G, et al. Prognostic impact of hypoxia imaging with
18F-misonidazole PET in non-small cell lung cancer and
head and neck cancer before radiotherapy. J Nucl Med
2005;46:253–60.
70. Cooper RA, Carrington BM, Loncaster JA, Todd SM,
Davidson SE, Logue JP, et al. Tumour oxygenation levels
correlate with dynamic contrast-enhanced magnetic reso-
nance imaging parameters in carcinoma of the cervix.
Radiother Oncol 2000;57:53–9.
71. Loncaster JA, Carrington BM, Sykes JR, Jones AP, Todd SM,
Cooper R, et al. Prediction of radiotherapy outcome using
798
A Y Isa, T H Ward, C M L West et al
dynamic contrast enhanced MRI of carcinoma of the cervix.
Int J Radiat Oncol Biol Phys 2002;54:759–67.
72. Hoskin PJ, Saunders MI, Goodchild K, Powell ME, Taylor
NJ, Baddeley H. Dynamic contrast enhanced magnetic
resonance scanning as a predictor of response to accelerated
radiotherapy for advanced head and neck cancer. Br J
Radiol 1999;72:1093–8.
73. Kaanders JH, Bussink J, van der Kogel AJ. ARCON: a novel
biology-based approach in radiotherapy. Lancet Oncol
2002;3:728–37.
74. Kaanders JH, Pop LA, Marres HA, Bruaset I, van den
Hoogen FJ, Merkx MA, et al. ARCON: experience in 215
patients with advanced head-and-neck cancer. Int J Radiat
Oncol Biol Phys 2002;52:769–78.
75. Bernier J, Denekamp J, Rojas A, Minatel E, Horiot J, Hamers
H, et al. ARCON: accelerated radiotherapy with carbogen
and nicotinamide in head and neck squamous cell
carcinomas. The experience of the Co-operative Group of
Radiotherapy of the European Organization for Research
and Treatment of Cancer (EORTC). Radiother Oncol
2000;55:111–9.
76. Glaser CM, Millesi W, Kornek GV, Lang S, Schull B,
Watzinger F, et al. Impact of hemoglobin level and use of
recombinant erythropoietin on efficacy of preoperative
chemoradiation therapy for squamous cell carcinoma of
the oral cavity and oropharynx. Int J Radiat Oncol Biol Phys
2001;50:705–15.
77. Overgaard J, Hansen HS, Overgaard M, Bastholt L,
Berthelsen A, Specht L, et al. A randomized double-blind
phase III study of nimorazole as a hypoxic radiosensitizer
of primary radiotherapy in supraglottic larynx and pharynx
carcinoma. Results of the Danish Head and Neck Cancer
Study (DAHANCA) Protocol 5-85. Radiother Oncol
1998;46:135–46.
78. Henke M, Laszig R, Rube C, Schafer U, Haase KD, Schilcher
B, et al. Erythropoietin to treat head and neck cancer
patients with anaemia undergoing radiotherapy: rando-
mised, double-blind, placebo-controlled trial. Lancet
2003;362:1255–60.
79. Wouters BG, Weppler SA, Koritzinsky M, Landuyt W,
Nuyts S, Theys J, et al. Hypoxia as a target for combined
modality treatments. Eur J Cancer 2002;38:240–57.
80. Gandara DR, Lara PN Jr, Goldberg Z, Le QT, Mack PC, Lau
DH, et al. Tirapazamine: prototype for a novel class of
therapeutic agents targeting tumor hypoxia. Semin Oncol
2002;29(1 Suppl. 4):102–9.
81. Rischin D, Peters L, Hicks R, Hughes P, Fisher R, Hart R, et
al. Phase I trial of concurrent tirapazamine, cisplatin, and
radiotherapy in patients with advanced head and neck
cancer. J Clin Oncol 2001;19:535–42.
82. Rischin D, Peters L, Fisher R, Macann A, Denham J, Poulsen
M, et al. Tirapazamine, cisplatin, and radiation versus
fluorouracil, cisplatin, and radiation in patients with locally
advanced head and neck cancer: a randomized phase II trial
of the Trans-Tasman Radiation Oncology Group (TROG
98.02). J Clin Oncol 2005;23:79–87.
83. Sun X, Kanwar JR, Leung E, Lehnert K, Wang D, Krissansen
GW. Gene transfer of antisense hypoxia inducible factor-1
alpha enhances the therapeutic efficacy of cancer immuno-
therapy. Gene Ther 2001;8:638–45.
84. Timothy AR, Overgaard J, Overgaard M. A phase I clinical
study of Nimorazole as a hypoxic radiosensitizer. Int J
Radiat Oncol Biol Phys 1984;10:1765–8.
85. Chapman JD, Schneider RF, Urbain JL, Hanks GE. Single-
photon emission computed tomography and positron-
emission tomography assays for tissue oxygenation.
Semin Radiat Oncol 2001;11:47–57.
86. Makitie AA, Irish J, Gullane PJ. Pharyngocutaneous fistula.
Curr Opin Otolaryngol Head Neck Surg 2003;11:78–84.
The British Journal of Radiology, October 2006
The British Journal of Radiology, 79 (2006), 799–800

Is routine chest radiography a useful test in the follow up of all

adult patients with soft tissue sarcoma?
1
H K LORD, MRCP, 2D M SALTER, MD, FRCPath, FRCPE,
1
R H MACDOUGALL, FRCS, FRCR, FRCPE and
1
G R KERR, MSc

1
Department of Clinical Oncology, Edinburgh Cancer Centre, Crewe Road, Edinburgh EH4 2XU and
2
Department of Pathology, Royal Infirmary of Edinburgh, 51 Little France Crescent, Edinburgh
EH16 4SA, UK

ABSTRACT. Following treatment for localized soft tissue sarcoma the risk of relapse is
either locally or in the lungs. In Edinburgh patients are reviewed every 6 months with a
chest X-ray (CXR). The radiation exposure over a 10 year follow up remains small, but it
is unclear if all patients, irrespective of the initial grade of their primary tumour,
require this. To determine the pick up rate of lung metastases by routine CXR over a
10 year period and to review the primary histology. Adult patients on routine follow up
between 1994 and 2004 were identified and the notes of those with lung metastases
reviewed. Data was collected on their initial histology, and date and method of
diagnosis of lung metastases. 179 patients were under follow up. 24 (13%) developed
lung metastases. For 2, notes were not found. 6 (27%) had metastases diagnosed by
routine CXR, 9 (41%) had metastases diagnosed by non routine CXR and 7 (32%) had
metastases diagnosed by CT. On review of histology none were grade 1, 4 (18%) were Received 9 December 2005
grade 2 and 18 (82%) were grade 3. 155 patients received. 6 monthly CXR for 10 years Revised 16 February 2006
with no detection of lung metastases. Lung metastases occurred in a minority of Accepted 10 March 2006
patients (13%) and most (82%) occurred in patients with grade 3 tumours. No patients
DOI: 10.1259/bjr/69175634
with grade 1 tumours developed lung metastases. Thus routine CXR may be
appropriate on grade 3 tumours, but not on lower grade tumours where other risk ’ 2006 The British Institute of
factors are absent. Radiology

Soft tissue sarcoma in adulthood is relatively rare.
Following treatment for localized disease, the risk of
relapse is either locally or in the lungs. Prognostic
indicators for recurrence are well documented [1–6] and
include age at diagnosis, tumour depth, tumour size,
histological type, histological grade, positive surgical
margins and tumour site. All patients in our centre are
followed up for 10 years regardless of initial pathology,
with 6 monthly chest radiographs (CXR). This is justified
because surgical intervention may be curative if lung
metastases are diagnosed early. The effective radiation
dose from a single CXR is quoted as between 0.02 mSv and
0.1 mSv, which is the equivalent to between 2.4 days and
10 days of background radiation. Thus whilst the overall
radiation dose over 10 years from 6 monthly CXR remains
small, it was unclear if CXR was a useful screening tool in
this setting, and if it was indicated in all patients regardless
on initial pathology. This retrospective audit was therefore
performed to answer these questions.
Methods and materials
Adult patients newly diagnosed with a localized
primary soft tissue sarcoma and all patients on routine
follow up between 1994 and 2004 were identified on the
departmental database and the notes of those with lung
metastases were reviewed.
Data were collected on the initial grade of tumour and
the method of diagnosis of lung metastases. The
pathology was reviewed where possible and graded
according to Federation Nationale des Centre de Lutte
Contre le Cancer (FNCLCC) classification.
Results
A total of 179 patients were under follow up during
this 10 year period. 24 (13%) developed lung metastases,
and 22 sets of notes were retrieved. Two sets of notes
were unavailable. Histology was reviewed for 21
patients and known from documentation for 1 patient.
None of the 22 cases with pulmonary metastases had
grade 1 tumours. 4 cases (18%) had grade 2 tumours and
18 (82%) had grade 3 tumours at initial diagnosis
(Table 1). Six (27%) had their lung metastases diagnosed
by routine screening CXR, 9 (42%) had their lung
metastases diagnosed on a non-routine CXR, prompted
by symptoms, and 7 (32%) had their metastases
diagnosed by a CT scan, performed as part of re-staging
for local recurrence or performed to investigate a decline
in general health (Table 2). 155 patients received 6
monthly CXR for 10 years and remained relapse free.
Total patient years at risk were 512.24, equating to 1
patient developing lung metastases for every 21 years of
follow up.

The British Journal of Radiology, October 2006 799


Table 1. Number of patients developing lung metastases
according to grade of primary
Tumour grade No. developing lung metastases
Grade 1 0 Non routine CXR
Grade 2 4 CT
Grade 3 18
Discussion
The small size of this cohort prevents meaningful
statistical analysis, but the trend shown is in keeping
with previously published data. Whilst this audit has not
identified the metastatic potential for grade 1 tumours,
there are data that this potential exists. In a paper by Le
Doussal [1] the 5 year risk of metastases for grade 1 and 2
tumours together was 20%, although it is not clear how
many were grade 1.
In a paper by Le [2] looking at prognostic factors in
head and neck sarcomas, which included 65 patients, the
5 year cause specific survival was 100% for grade 1
tumours. This is in concordance with a series by Willers
et al [7] demonstrating similar outcomes for grade 1
tumours.
However, the interplay of prognostic factors beyond
tumour grade alone is relevant. Mandard et al [8]
performed a multivariate analysis of 109 patients with
soft tissue sarcoma. Four prognostic groups were
identified depending on the presence or absence of
necrosis, invasive tumour, adequate surgical margins
and size of tumour less than or greater than 5 cm. Three
patients with FNCLCC grade 1 tumours fell into a poor
prognostic due to tumour size greater than 5 cm and/or
inadequate surgical margins and 2 of these developed
metastatic disease from which they died. A multivariate
analysis performed by Heise [9] also demonstrated
tumour site to be relevant, with retroperitoneal and
mediastinal disease associated with poorer outcomes
compared with disease in the extremities or head and
neck. Thus it is clear that the prognosis of an individual
patient is dependent on factors other than tumour grade.
The usefulness of routine CXR to diagnose metastatic
disease in soft tissue sarcoma has been less well
documented. Kane [10] has summarized the data
demonstrating a lack of cost effectiveness for the use of
CT, and highlights how regular clinic visits and
thorough clinical examination detect the majority of
recurrent disease. Early pulmonary disease is frequently
asymptomatic however and, if detected, can still be
cured by surgical intervention. The low cumulative
dose of radiation received from 6 monthly CXR makes
this a safe, simple and appropriate first tool [11]. The
optimal follow up for patients with soft tissue sarcoma
remains unknown [12], but in this series routine CXR
detected 27% of the pulmonary metastatic cases and
thus should be recommended. Whilst it may be omitted
in patients with grade 1 tumours, other prognostic
800
H K Lord, D M Salter, R H M MacDougall and G Kerr
Table 2. Method of diagnosis of lung metastases
Method of diagnosis Number diagnosed
Routine CXR 6
9
7
CXR, chest radiograph.
factors should also be borne in mind, and the decision
based on the overall prognostic grouping, rather than
grade alone.
References
1. Le Doussal V, Coindre JM, Leroux A, Hacene K, Terrier P,
Bui NB, et al. Prognostic factors in patients with localised
primary malignant fibrous histiocytoma: a multicenter
study of 216 patients with multivariate analysis. Cancer
1996;77:1823–30.
2. Le QT, Fu KK, Kroll S, Fitts L, Massullo V, Ferrell L, et al.
Prognostic factors in adult soft tissue sarcomas of the head
and neck. Int J Radiat Oncol Biol Phys 1997;37:975–84.
3. Coindre JM, Terrier P, Guillou L, Le Doussal V, Collin F,
Ranchere D, et al. Predictive value of grade for metastases
development in the main histologic types of adult soft
tissue sarcomas: a study of 1240 patients from the French
Federation of Cancer Centers Sarcoma Group. Cancer
2001;91:1914–26.
4. Vraa S, Keller J, Nielson OS, Sneppen O, Jurik AJ, Jensen
OM. Prognostic factors in soft tissue sarcomas: the Aarhus
experience. Eur J Cancer 1998;34:1876–82.
5. Collin CF, Friedrich C, Godbold J, Hajdu S, Brennan MF.
Prognostic factors for local recurrence and survival in
patients with localised extremity soft tissue sarcoma. Semin
Surg Oncol 1988;4:30–7.
6. Pezzi ME, Rawling MS, Esgro JJ, Pollock RE, Rohmsdahl
MM. Prognostic factors in 227 patients with malignant
fibrous histiocytoma. Cancer 1992;69:2098–103.
7. Willers H, Hug EB, Spiro IJ, Efird JT, Rosenberg AE, Wang
CC. Adult soft tissue sarcomas of the head and neck treated
by radiation and surgery or radiation alone: patterns of
failure and prognostic factors. Int J Radiat Oncol Biol Phys
1995;33:585–93.
8. Mandard AM, Petiot JF, Marnay J, Mandard JC, Chasle J, de
Ranieri E, et al. Prognostic factors in soft tissue sarcoma; a
multivariate analysis of 109 cases. Cancer 1989;63:1437–51.
9. Heise HW, Myers MH, Russel WO, Suit HD, Enzinger FM,
Edmonson JH, et al. Recurrence free survival time for
surgically treated soft tissue sarcoma patients. Cancer
1986;57:172–7.
10. Kane JM. Surveillance strategies for patients following
surgical resection of soft tissue sarcomas. Curr Opin Oncol
2004;16:238–321.
11. Chang AE, Schaner EG, Conkle DM, Flye MW, Doppman
JL, Rosenberg SA. Evaluation of computed tomography in
the detection of pulmonary metastases: a prospective study.
Cancer 1979;43:913–6.
12. Sakata K, Johnson FE, Beitler AL, Kraybill WG, Virgo KS.
Extremity soft tissue sarcoma patient follow up; tumour
grade and size affect surveillance strategies after potentially
curative surgery. Int J Oncol 2003;22:1335–43.
The British Journal of Radiology, October 2006
The British Journal of Radiology, 79 (2006), 801–803

Multiple hyperechoic testicular lesions are a common finding on

ultrasound in Cowden disease and represent lipomatosis of the
testis
1
J WOODHOUSE, MRCS and 2M M FERGUSON, FDSRCPS

1
Department of Urology, Christchurch Hospital, Christchurch and 2Department of Stomatology,
University of Otago, Dunedin, New Zealand

ABSTRACT. Cowden disease (CD) is a genetic disease associated with multiple

hamartomas and malignant neoplasms. During investigations for possible subnormal
fertility, a series of eight males with CD underwent ultrasound scanning of their testes.
Our findings detail the seven adult patients that were found to have multiple and
bilateral testicular lesions on ultrasound. These lesions in Cowden’s patients represent a
newly described testicular pathology – lipomatosis of the testis. Here we detail the Received 5 October 2005
radiological findings. Ultrasound findings showed multiple (estimated 40+) discreet Revised 22 March 2006
lesions randomly scattered thoughout the parenchyma of each affected testis. The Accepted 30 March 2006
lesions themselves had heterogeneous echo texture, but all being hyperechoic and with
DOI: 10.1259/bjr/50628431
a variable size from 1 mm to 6 mm. One patient underwent MR examination of the
lesions in which a high T1 signal was seen, but no useful characteristic features were ’ 2006 The British Institute of
identifiable. Radiology

Cowden disease (CD; also known as Multiple Methods

Hamartoma Syndrome), is marked by increased cellular
proliferation of ectodermal, mesodermal and endoder-
mal tissues [1, 2]. These neoplasms can be benign or
malignant. In particular, cancer of the breast, thyroid,
endometrium and skin has been associated with CD.
Diagnosis is based on operational criteria laid out by the
‘‘International Cowdens syndrome consortium’’ and
supported by genetic testing. Mucocutaneous lesions
comprise the predominant feature and are a principal
way of suspecting the diagnosis. Phenotypic expression
is variable and thus some cases probably remains
undiagnosed. Despite it being a dominantly inherit-
able condition, the prevalence of the disease remains
relatively low with more recent estimates at 1
per 250 000 [3].
A singular lesion within the parenchyma of a testis is
not an uncommon finding with ultrasound scanning, but
when found they are regarded as highly suspicious of
malignancy. Multiple lesions of the testes are infrequent
and usually represent microlithiasis. Multiple testicular
lesions in the context of CD has been previously reported
[4], but only in the context of a single patient. No
evidence was provided to establish the histopathology
in this case. Our findings are suggestive that these
lesions are a very frequent occurrence in CD. This
ultrasound finding is now known to represent a new
histopathological entity called lipomatosis of the testis
[5]. Here we detail the radiological aspects of this new
condition.
Address correspondence to: Professor Martin M Ferguson,
Department of Stomatology & Health Sciences, University of
Otago, P.O. Box 647 Dunedin, New Zealand.
All male patients with CD known to the authors were
involved in this study, which was initially directed as a
study into fertility issues and CD and subject to local
ethics approval. The study was prospective and there
were no exclusions made. All of the eight males had
clinical features of CD and the diagnosis was confirmed
by identification of the mutation on chromosome 10q.23.
Ages ranged from 16 years to 58 years (mean 5 38 years).
The patients were interviewed and examined with
attention initially on issues surrounding possible low
fertility. Structural assessment of the testes in all eight
was performed with ultrasound. Scanning was per-
formed with a range of machines in four separate
institutions from both the private and public sectors.
Scanners included Philips HDI 5000 using high fre-
quency linear probes. Exact scanner models were unable
to be determined for all scans performed. All sonogra-
phers held the diploma of Australian Society of
Ultrasound in Medicine and Biology. All scans were
reviewed retrospectively by consultant radiologists at
Christchurch Public Hospital. In some cases, some scans
were assessed in real time by those consultants.
Following identification of anatomical testicular lesions,
all patients had serum tumour marker assays performed
(alpha fetoprotein, beta human chorionic gonadotropin,
and lactate dehydrogenase) and were offered interval
ultrasound. Two patients had had previous testicular
ultrasound for comparison. Overall, five of the eight had
repeated scans with intervals ranging from 5 months to
89 months for comparison. One patient underwent MRI
(General Electric Sigma Horizon LX 1.5 Tesla, Wisconsin,
USA) with T1, T1 fat saturated, T1 fat saturated post-
gadolinium enhancement (repetition time (TR) 600 ms,

The British Journal of Radiology, October 2006 801


echo time (TE) 14 ms, field of view (FOV) 10 cm, slice
thickness 3 mm, gap 0.5 mm, phase 224, frequency 384,
electronic train length 3; 3 inch surface coil) and T2
sequences (TR 3500 ms, TE 96 ms, FOV 10 cm, slice
thickness 3 mm, gap 0.5 mm, phase 192, frequency 256,
electronic train length 12; 3 inch surface coil), attempting
to further characterize these lesions. Four of eight
subjects elected to undergo open testicular biopsy under
general anaesthetic due to lack of any definitive
pathological correlation.
Results
Of the cohort, all were normal on urogenital physical
examinations with none of the lesions described being
palpable. One patient previously had a testicular lump
observed by ultrasound and discharged as benign, and
this was still palpable and non tender. Another patient
reported generalized low grade testicular tenderness. All
were of normal sexual development.
All patients had at least one ultrasound and five had
repeat interval scans. Ultrasound showed that all had
testes of normal size and contour, but the seven adults
had multiple (estimates of 40+ per testes) hyperecho-
genic foci of approximately 1–6 mm size within the
parenchyma of both testicles (Figure 1). Most lesions
were small and non-shadowing. Only in one testis were
acoustic shadows seen (Figure 2). The lesions were
randomly scattered throughout the testis in all cases.
The lesions did not demonstrate increased vascularity on
Doppler imaging. There were no other focal parenchy-
mal abnormalities seen. Five of seven had interval repeat
ultrasound for comparison (range 5–89 months). Within
the limits of ultrasound analysis, all lesions were
reported as remaining stable. However, owing to the
number of lesions in each testis it was not possible to
categorically determine this.
Figure 1. Longitudinal ultrasound view of the right testis
demonstrates well defined multiple non-shadowing small
round lesions that are hyperechoic. The lesions are of various
sizes and are non-confluent. No hypoechoic regions or
calcifications are seen.
802
J Woodhouse and M M Ferguson
Figure 2. Longitudinal ultrasound of testis in one case
demonstrated that some of the lesions displayed acoustic
shadowing. The lesions here are similarly hyperechoic. No
hypoechoic regions or calcifications are seen.
Two patients had undergone previous testicular
ultrasound performed for other reasons available for
retrospective comparisons. One patient, 3 years pre-
viously, had four serial scans for a suspicious scrotal
lump. These had shown a calcified nodule on the surface
of the testicle, but with poorly defined regions of
echogenicity within the parenchyma. The original inter-
pretation was that these regions had represented micro-
lithiasis. Films were reviewed in light of our findings and
the opinion was that these lesions were not microlithia-
sis. Another patient had had two scans carried out
7 years previously. Identical small echogenic non-
shadowing foci were present at that point. With both
patients, as initial scans had been for other reasons, it
was impossible to comment with certainty as to the
stability of the lesions; although it did seem that these
lesions had not progressed significantly.
In an effort to further characterize the lesions non-
invasively, one patient underwent MRI of the scrotum.
MRI showed multiple (greater than 15) tiny focal high-
signal areas within both testes, probably corresponding
to the lesions seen on the ultrasound images, although
these were very poorly visualized in all sequences used.
Lesions were best seen on the T1 series (Figure 3).
Insufficient detail was gained to be diagnostic.
Discussion
CD is a rare disease and as such our cohort of eight
patients represents a relatively large population of CD
patients. The identification of the above described
multiple bilateral testicular lesions on ultrasound and
MR was initially documented by Lindsay [4] in a single
CD patient, and has not been documented outside the
context of CD. In this single case, the lesions were
assumed to be hamartoma based on the lesions having a
fat component on MRI, but no histopathological proof
was offered for this diagnosis. Our series of CD
patients were all impalpable incidentalomas found while
The British Journal of Radiology, October 2006
Ultrasound of testicular lesions in Cowden disease can represent lipomatosis
Figure 3. Sagittal T1 weighted spin echo sequence (repeti-
tion time (TR) 600 ms, echo time (TE) 14 ms, 3 mm slice
thickness, field of view (FOV) 10 cm) of the left testis. The
scan demonstrates tiny hyperintense areas (white arrows)
within the testis parenchyma consistent with fat.
investigating subfertility, probably explaining why they
have not been extensively documented previously. Our
group has established that these lesions represent
lipomatosis of the testis in the four patients who
underwent biopsy. Biopsy microscopic findings showed
‘‘interstitial lipomatosis consisting of nests of adipocytes
within the testicular interstitium. No intratubular calci-
fication or intratubular germ cell neoplasia was seen’’.
There appeared to be no detrimental effect on sperma-
togenesis, nor on testicular biochemical function [5].
The diagnosis of malignancy was considered before
diagnostic biopsy was performed. Serial scanning using
ultrasound could not absolutely qualify the stability of
the lesions as lesions were too numerous to enable the
identification of individual lesions for comparative
assessment on repeated scans. MRI unfortunately proved
inadequate to be useful diagnostically. Hence biopsy was
a diagnostic option that was offered to all these patients
and testicular biopsies were performed on some of these
patients (but not on all, once it was established that the
lesions seen on ultrasound were not tumours). The
appearances seen in our series represent benign testi-
cular lipomatosis (hamartomatous growths). Ultrasound
findings of primary testicular malignancies tend to be
hypoechoic and homogeneous (seminoma) or are well
defined with cystic elements and calcifications (tera-
toma). Lymphoma, leukaemia, and metastatic disease is
uncommon and has variable appearance on ultrasound
from focal hypoechoic lesions to a diffusely altered
echogenicity [6, 7] – not similar to the findings of
testicular lipomatosis seen in our cohort, nor the lesions
described by Lindsay [4].
Benign testicular lesions are uncommon and are
difficult to distinguish from malignancy on ultrasound.
Because of this, they often result in orchidectomy. The
high signal on MR T1 imaging supported a fatty
component to the lesions and does lend itself to a
diagnosis of hamartomas (as suggested by Lindsay) or
lipomas as is the diagnosis. Testicular lipomas are
The British Journal of Radiology, October 2006
reported, although they are very rare and have only
been described only as singular lesions [8, 9].
Microlithiasis is the most commonly documented
testicular pathology occurring as a multiple entity, and
consists of multiple randomly scattered punctuate
lesions in the testicular parenchyma (with variable
acoustic shadowing) of approximately 1–3 mm in size
representing calcified deposits in the lumen of semini-
ferous tubules. As they represent calcifications, they
have a very high echogenicity and are thus distinctly
different to the lesions we find in the Cowden’s testes.
Testicular lipomatosis is a novel entity. We have
detailed how the ultrasound appearances differ distinc-
tively from other pathologies. These lesions are not
described outside the context of CD and thus are
apparently pathognomic for CD and could be used as a
major diagnostic criterion for CD in adult males. A
finding on ultrasound of these lesions should thus alert
the radiologist to the diagnosis of CD. In our series, 7 of 8
of males with CD had these lesions. Phenotypic expres-
sion in CD often is not manifest until aged 20–30 years
[10] and it remains to be seen if the eighth male –
currently with no testicular lesions and aged 16 years –
will eventually develop these lesions. The long term
behaviour of these lesions is unknown to date.
Acknowledgments
The Authors would like to acknowledge the kind help
of Hamish Fraser, Andrew Lang, Diane Leighton and
Tim Buckingham for their help and advice regarding this
work. We would also like to acknowledge the staff of
Christchurch Public Hospital Radiology department and
the Southern Cross Radiology department where the
majority of scans were performed.
References
1. Lloyd KM, Dennis M. Cowden’s disease. A possible new
symptom complex with multiple system involvement. Ann
Intern Med 1963;58:136–42.
2. Weary PE, Gorlin RJ, Gentry WC Jr, Comer JE, Greer KE.
Multiple hamartoma syndrome (Cowden’s disease). Arch
Dermatol 1972;106:682–90.
3. Nelen MR, et al. Novel PTEN mutations in patients with
Cowden disease: absence of clear genotype-phenotype
correlations. Eur J Hum Genet 1999;7:267–73.
4. Lindsay C, Boardman L, Farrell M. Testicular hamartomas
in Cowden disease. J Clin Ultrasound 2003;31:481–3.
5. Woodhouse JB, Delahunt B, English SF, Fraser HH,
Ferguson MM. Testicular lipomatosis in Cowden’s syn-
drome. Mod Pathol 2005;18:1151–6.
6. Hamm B. Differential diagnosis of scrotal masses by
ultrasound. Eur Radiol 1997;7:668–79.
7. Howlett DC, Marchbank ND, Sallomi DF. Pictorial review.
Ultrasound of the testis. Clin Radiol 2000;55:595–601.
8. Harper M, Arya M, Peters JL, Buckingham S, Freeman A,
O’Donoghue EP. Intratesticular lipoma. Scand J Urol
Nephrol 2002;36:223–4.
9. Honore LH. Fatty metaplasia in a postpubertal undes-
cended testis: a case report. J Urol 1979;122:841–2.
10. Nelen MR, et al. Localization of the gene for Cowden
disease to chromosome 10q22–23. Nat Genet 1996;13:114–6.
803
 The British Journal of Radiology, 79 (2006), 804–807

Duplex ultrasound of the superior mesenteric artery in chronic

pancreatitis
1 1 3
M HORNUM, MD, S LARSEN, MD, PhD, O OLSEN, MD, PhD and 2J F PEDERSEN, MD, PhD

Departments of 1Gastroenterology and 2Radiology, Glostrup Hospital, Copenhagen and

3
Department of Surgical Gastroenterology, Amager Hospital, University of Copenhagen,
Copenhagen, Denmark

ABSTRACT. Blood flow in the superior mesenteric artery (SMA) increases after a meal
due to a vasoactive effect of the decomposed food. In exocrine pancreatic insufficiency,
the digestion of food is compromised. We used duplex ultrasound to test the
hypothesis that blood flow in the SMA after a meal increases less in patients with
pancreatic insufficiency than in control persons. We studied 16 patients with chronic
pancreatitis, eight of them with exocrine insufficiency, and eight healthy volunteers.
The resistive index (RI) in the SMA was determined before and after a liquid meal. The
RI reflects the downstream circulatory resistance, giving a precise description of
mesenteric hyperaemia. Both groups of patients with chronic pancreatitis unexpectedly
had lower fasting RI than controls, 0.818 and 0.815 vs 0.851, p50.028 and p50.0030,
respectively. Postprandialy there was significantly less decrease in RI (less increase in
flow) in patients with exocrine insufficiency than in controls, 0.055 vs 0.099, p50.0047.
There was a significant trend for a less pronounced postprandial decrease in RI with Received 11 January 2006
more impaired pancreatic function (p50.0036). Our study thus demonstrates a reduced Revised 5 April 2006
postprandial increase in SMA flow in patients with exocrine pancreatic insufficiency, Accepted 24 April 2006
and suggests an increased fasting SMA flow in chronic pancreatitis. Further studies are
DOI: 10.1259/bjr/41916828
needed to evaluate the possible role of the test-meal-induced shift in RI in the SMA and
of a lower-than-normal fasting RI in the diagnosis and monitoring of chronic ’ 2006 The British Institute of
pancreatitis. Radiology

At ultrasound scanning combined with the pulsed
Doppler technique we can non-invasively monitor
alterations in blood flow velocity. Applied to the super-
ior mesenteric artery (SMA), the method confirms that
blood flow in this vessel increases in response to a meal
[1–3]. The vasoactive components seem to be influenced
by the digestive products of the diet, so that the effect of
a meal on splanchnic blood flow may depend on the
intraluminal digestion [1]. Patients with end stage
chronic pancreatitis are characterized by maldigestion
due to exocrine pancreatic insufficiency, but it is not
known whether the reduced intraluminal food digestion
is mirrored in the mesenteric blood flow. It was our
hypothesis that these patients would show less post-
prandial increase in mesenteric blood flow than control
persons, so that the postprandial shift in Doppler
velocity pattern might be used in the diagnosis and
monitoring of chronic pancreatitis.
In this study we examine the velocity pattern in the
SMA before and after a liquid test meal in patients with
chronic pancreatitis with and without preserved exocrine
function, and in healthy controls.
Methods and patients
We studied 16 patients with chronic pancreatitis, eight
with minimal or no exocrine pancreatic function and
eight with reduced exocrine function, and eight
healthy volunteers (Table 1). The aetiology of chronic
pancreatitis was alcohol in 14 patients and unknown in
two. All patients had moderate or marked imaging
findings according to the Cambridge classification [4]
plus reduced exocrine pancreatic function judged from
the intraduodenal meal-stimulated lipase concentration
(Lundh test), or steatorrhoea. The eight patients with
exocrine insufficiency were characterized by a Lundh
test result below 10% of the normal lower limit [5], or
excretion of more than 7 g of fat per day in faeces [6]
(plus morphological changes). Four of them had diabetes
but no signs of peripheral vascular disease, and seven
had enzyme substitution. The patients with residual
pancreatic exocrine function had no steatorrhoea and
meal stimulated intraduodenal lipase concentration far
over 10% of lower normal limit (plus morphological
changes) and had no other known gastrointestinal
disorder, and no diabetes or peripheral vascular disease.
The controls had no known gastrointestinal disease and
were not taking any kind of medication.
All participants gave written informed consent to take
part in the study, which was approved by the Regional
Research Ethics Committee (KA 0129).
The ultrasound examinations were performed with a
Philips HDI 5000 Sono CT unit (Philips Medical Systems,
Bothell, WA). The SMA was examined in its long axis in
the sagittal plane. The sampling cursor was placed
within the diameter of the vessel, 2–3 cm distal to its
origin, and the angle between the ultrasound beam and

804 The British Journal of Radiology, October 2006

Mesenteric blood flow and pancreatic insufficiency

Table 1. Patients and control persons
Chronic pancreatitis Normal control
persons
Exocrine No exocrine
insufficiency insufficiency
No. of subjects 8 8 8
Females/males 2/6 4/4 3/5
Median age, 55 (40–65) 51 (36–73) 44 (31–69)
years (range)
Median body 21 (17–25) 24 (19–28) 23 (19–27)
mass index
kg m2 (range)
the SMA was kept smaller than 60 ˚. The resistive index
(RI) was calculated in accordance with the formula:
peak-systolic velocity minus end-diastolic velocity
divided by the peak-systolic velocity [7]. Each RI value
is the average of three measurements.
To monitor gastric emptying the antrum was localized
on a sagittal image in front of the superior mesenteric
vein [8]. The image was frozen and the antral area was
measured by means of the built-in calliper system. After
the meal the antral area was determined at 5–10 min
intervals until the area approached baseline, and the time
from the meal until the area had decreased to 150% of
baseline was used as surrogate expression of gastric
emptying time [8].
Patients and healthy volunteers were examined in
random order. Any pancreatic enzyme substitution was
discontinued for 72 h before the study. The examination
commenced after an overnight fast, with the subjects
resting in the supine position for 45 min. RI in the SMA
was recorded at least twice during the last 30 min before
the meal and the baseline antral area was determined.
The subjects then ingested a test meal consisting of 74.8 g
NAN 1 (Nestlé Danmark A/S), containing 5.9% fat, 2.3%
protein, and 11.6% carbohydrate, in 300 ml of water,
with a total energy load of 1598 kJ, and SMA flow
characteristics were recorded five times at 15 min
intervals. Antral area was monitored as described. In
most subjects, gastric emptying was not complete after
80 min. In these situations antral area was monitored for
an additional 20 min.
For the data analysis, we calculated the difference
between baseline RI and the mean of the five post-meal
RI determinations in each participant. This difference
thus expresses the integrated RI response to the test
meal. Results were analysed for statistical significance
using the Mann-Whitney two-sample rank sum test for
unpaired data and Wilcoxon test for paired data. The
Jonckheere-Terpstra test for ordered alternatives [9] was
used to test the hypothesis that the postprandial shift in
RI would be smaller with poorer exocrine function.
Results are expressed as medians and ranges, and p,0.05
was considered statistically significant.
Results
The fasting RI was significantly lower in patients with
exocrine insufficiency and in patients with reduced
exocrine function, median values being 0.82 in both
groups as compared with 0.85 in healthy volunteers,
p50.028 and p50.0030, respectively (Table 2, Figure 1).
After the test meal, the RI decreased less in patients
without and with residual exocrine function, 0.055 and
0.072, respectively, compared with controls, 0.099
(Table 2, Figures 2 and 3). The difference between
patients with exocrine insufficiency and controls was
statistically significant, p50.0047, indicating less post-
prandial increase in mesenteric blood flow in exocrine
insufficiency, whereas there was no significant difference
between patients with residual exocrine function and the
control group. There was a significant trend for less
postprandial decrease in RI (less decrease in downstream
circulatory resistance) from controls through patients
with chronic pancreatitis without clinical insufficiency to
patients with manifest pancreatic insufficiency
(p50.0036, Figure 3).
There was no difference in the mean antral emptying
time between the three groups (Table 2).
Discussion
We used RI to characterize downstream circulatory
resistance in the SMA. It is calculated from two Doppler
shift frequencies measured at the same image, and is
independent of the angle of insonation. The RI is shown
to increase linearly with the peripheral resistance at a
constant pressure [7, 10]. It therefore makes sense that RI
in the SMA has been reported significantly lower in
active than in inactive ulcerative colitis [11] and Crohn’s
disease [12].
Median fasting RI in our control persons was 0.85
which is in agreement with results from other studies
[11, 12]. Unexpectedly, both patients with chronic
pancreatitis and patients with pancreatic insufficiency

Table 2. Median resistive index (RI) in the superior mesenteric artery before and after standard meal and antral emptying time
after meal in eight patients with chronic pancreatitis and exocrine insufficiency, eight patients with chronic pancreatitis and
preserved exocrine function and in eight healthy control persons
Chronic pancreatitis Normal control persons

Exocrine insufficiency No exocrine insufficiency

Fasting RI, median (range) 0.818 (0.753–0.850) 0.815 (0.757–0.838) 0.851 (0.821–0.893)
Mean RI 0–80 min after meal, 0.753 (0.691–0.869) 0.712 (0.695–0.761) 0.745 (0.676–0.801)
median (range)
Decrease in RI after meal, median 0.055 (–0.023–0.093) 0.072 (0.057–0.110) 0.099 (0.078–0.146)
(range)
Antral emptying time, median 77 (40 to .100) 73 (59 to .100) 94 (75 to .100)
(range, min.)

The British Journal of Radiology, October 2006 805


Figure 1. Fasting resistance index (RI) in the superior
mesenteric artery in eight patients with chronic pancreatitis
and exocrine insufficiency, eight patients with chronic
pancreatitis without manifest insufficiency and in eight
healthy control persons.
had significantly lower fasting RI in the SMA than
controls, 0.82 in both groups. This suggests a more
dilated vascular bed peripherally in the SMA territory in
the fasting state in patients with chronic pancreatitis. To
our knowledge this has not been previously reported,
and further studies are needed to confirm this incidental
observation. It could be speculated that in chronic
pancreatitis the mechanisms that regulate the flow in
the SMA are adapted to a weaker stimulus from the
digested food components.
Figure 2. Resistance index (RI) in the superior mesenteric
artery before (baseline) and 0–80 min after test meal in eight
patients with chronic pancreatitis and exocrine insufficiency
NN
( – ), eight patients with chronic pancreatitis without
manifest insufficiency (#–#), and in eight healthy control
persons (m–m).
806
M Hornum, S Larsen, O Olsen and J F Pedersen
Figure 3. Difference between fasting and postprandial
resistance index (RI) in the superior mesenteric artery in
eight patients with chronic pancreatitis and exocrine
insufficiency, eight patients with chronic pancreatitis with-
out manifest insufficiency and in eight healthy control
persons.
After the meal, RI decreased less (downstream resis-
tance decreased less) in patients with exocrine insuffi-
ciency compared with patients with preserved exocrine
function and with healthy controls (Figure 2). This lower
effect of a meal on the blood flow in SMA in patients with
exocrine insufficiency has not been demonstrated before.
The test meal contained protein, fat and carbohydrate. All
three components increase the blood flow in SMA after
isocaloric and iso-osmotic loads into the duodenum [13].
In a recent study, instillation into the duodenum of free
fatty acids (product of hydrolysis of triglycerides)
increased SMA blood flow in normal subjects [1] to the
same degree as we observed in the present study using a
mixed test meal. Our results indicate that digestive
products are more active in modulating SMA blood flow
than undigested food. This could be due to increased
release of vasoactive gastrointestinal hormones such as
glucagon-like peptide-2 (GLP-2) during intestinal absorp-
tion of digestive products [14].
Since fat in the duodenum provokes gallbladder
emptying [15], another explanation for our findings
could be that the bile rather than the food causes the
vascular effects. However, intravenous cholecystokinin
in doses known to induce gallbladder contraction causes
no flow response in SMA [16], so luminal bile does not
seem to be responsible for the increased flow. Also, it has
been reported that the postprandial gastric motility in
chronic pancreatitis is altered [17, 18] so that a difference
in gastric emptying of the test meal could influence our
results. We therefore monitored antral emptying and
observed no significant difference in antral emptying
rate between patients and controls.
Any practical consequences of our findings remain to
be defined. The modestly lower fasting RI in chronic
pancreatitis is not impressive, the difference between
medians in patients and controls being 0.03, but the
majority of patients (11 of 16) actually had RI levels
below the lowest level in our control persons (Figure 1).
The British Journal of Radiology, October 2006
Mesenteric blood flow and pancreatic insufficiency
If other studies confirm this observation, it might be
considered worthwhile to determine RI in the SMA
routinely in all fasting upper abdominal studies, or at
least in studies without an obvious diagnosis. An
unusually low RI would suggest undetected abdominal
pathology like inflammatory bowel disease or chronic
pancreatitis.
The study confirms our hypothesis of a reduced
postprandial increase in SMA flow in patients with
exocrine pancreatic insufficiency. Further studies are
needed to evaluate whether the meal induced shift in RI
in the SMA can be used in the diagnosis and monitoring
of chronic pancreatitis. Such a test would have the virtue
of being non-invasive, in contrast to the duodenal-tube-
based measurement of meal-induced release of lipase.
Acknowledgments
This study was supported by grants from the Danish
Hospital Foundation for medical research. Region of
Copenhagen, The Faroe Islands and Greenland (15/03).
Special thanks to Tove Laursen for technical support.
References
1. Andersen TC, Pedersen JF, Nordentoft T, Olsen O. Fat and
mesenteric blood flow. Scand J Gastroenterol 1999;34:894–7.
2. Dauzat M, Lafortune M, Patriquin H, Pomier-Layrargues G.
Meal induced changes in hepatic and splanchnic circula-
tion: a noninvasive Doppler study in normal humans. Eur J
Appl Physiol Occup Physiol 1994;68:373–80.
3. Qamar MI, Read AE. Effects of ingestion of carbohydrate,
fat, protein, and water on the mesenteric blood flow in man.
Scand J Gastroenterol 1988;23:26–30.
4. Sarner M, Cotton PB. Classification of pancreatitis. Gut
1984;25:756–9.
5. Worning H, Mullertz S. pH and pancreatic enzymes in the
human duodenum during digestion of a standard meal.
Scand J Gastroenterol 1966;1:268–83.
6. Van de Kamer JH, ten Bokkel Huinink H, Weyers HA.
Rapid method for determination of fat in feces. J Biol Chem
1949;177:347–55.
The British Journal of Radiology, October 2006
7. Spencer JA, Giussani DA, Moore PJ, Hanson MA. In vitro
validation of Doppler indices using blood and water. J
Ultrasound Med 1991;10:305–8.
8. Pedersen JF. A modified sonographic technique for assess-
ment of gastric emptying of liquid. Acta Radiol
2003;44:340–2.
9. Siegel S, Castellan NJ. Nonparametric statistics for the
behavioural sciences. 2nd edn. Boston: McGraw-Hill,
1988:216–22.
10. Maulik D, Arbeille P, Kadado T. Hemodynamic foundation
of umbilical arterial Doppler waveform analysis. Biol
Neonate 1992;62:280–9.
11. Maconi G, Imbesi V, Porro GB. Doppler ultrasound
measurement of intestinal blood flow in inflammatory
bowel disease. Scand J Gastroenterol 1996;31:590–3.
12. Yekeler E, Danalioglu A, Movasseghi B, Yilmaz S, Karaca C,
Kaymakoglu S, et al. Crohn disease activity evaluated by
Doppler ultrasonography of the superior mesenteric artery
and the affected small-bowel segments. J Ultrasound Med
2005;24:59–65.
13. Sieber C, Beglinger C, Jager K, Stalder GA. Intestinal phase
of superior mesenteric artery blood flow in man. Gut
1992;33:497–501.
14. Guan X, Stoll B, Lu X, Tappenden KA, Holst JJ, Hartmann
B, et al. GLP-2-mediated up-regulation of intestinal blood
flow and glucose uptake is nitric oxide-dependent in TPN-
fed piglets 1. Gastroenterology 2003;125:136–47.
15. Olsen O, Schaffalitzky de Muckadell OB, Cantor P,
Erlanson-Albertsson C, Hansen CP, Worning H. Effect of
trypsin on the hormonal regulation of the fat-stimulated
human exocrine pancreas. Scand J Gastroenterol
1988;23:875–81.
16. Sieber C, Beglinger C, Jaeger K, Hildebrand P, Stalder GA.
Regulation of postprandial mesenteric blood flow in
humans: evidence for a cholinergic nervous reflex. Gut
1991;32:361–6.
17. Layer P, Ohe MR, Holst JJ, Jansen JBMJ, Grandt D, Holtman
G, et al. Altered postprandial motility in chronic
pancreatitis: role of malabsorption. Gastroenterology
1997;112:1624–34.
18. Vu MK, Vecht J, Eddes EH, Biemond I, Lamers CBHW,
Masclee AAM. Antroduodenal motility in chronic pancrea-
titis: are abnormalities related to exocrine insufficiency? Am
J Physiol Gastrointest Liver Physiol 2000;278:G458–66.
807
 The British Journal of Radiology, 79 (2006), 808–817

Accuracy and precision of an external-marker tracking-system for

radiotherapy treatments
1 2
E M DONOVAN, PhD, P BRABANTS, 1P M EVANS, DPhil,
1
J R N SYMONDS-TAYLER, MSc and
3
R WILKS, PhD

1
Joint Department of Physics, Institute of Cancer Research and Royal Marsden Foundation Trust,
Downs Road, Sutton, Surrey SM2 5PT, 2QADOS Ltd, 5 Lakeside Business Park, Swan Lane, Sandhurst,
Berkshire GU47 9DN and 3Osiris Systems Ltd, Burraton Bungalow, Broadclyst, Exeter, Devon EX5
3DB, UK

ABSTRACT. The purpose of this work was to determine the accuracy and precision of a
real-time motion-tracking system (Osiris+) for the monitoring of external markers used
on patients receiving radiotherapy treatments. Random and systematic errors in the
system were evaluated for linear (1D), circular (2D) and elliptical (3D) continuous
motions, and for a set of static positions offset from an origin. A Wellhofer beam data
measurement system and a computer controlled platform (which could be
programmed to give motion in 3D) were used to move a hemi-spherical test object. The
test object had four markers of the type used on patients. Three markers were aligned
in the central plane and a fourth was positioned out of plane. Errors were expressed as
deviations from the planned positions at the sampled time points. The marked points
on the test object were tracked for the linear motion case with a variation from the
true position of less than ¡1 mm, except for two extreme situations. The variation was
within ¡2 mm when the lights were dimmed and when the amplitude of the
movement was ¡5.0 cm. The 2D circular motion was tracked with a standard deviation
of 1 mm or less over four cycles. The sampling rates of the system were found to be 0.3–
0.4 s when it was monitoring actively and 1.5–1.6 s otherwise. The recorded Osiris+
measurements of known static positions were within ¡1 mm of the value from the Received 22 July 2005
computer controlled platform moving the test object. The elliptical motions in 3D were Revised 22 December 2005
tracked to ¡1 mm in two directions (Y,Z), and generally to within ¡2 mm for the third Accepted 24 January 2006
direction (X); however, specific marked points could display an error of up to 5 mm at
DOI: 10.1259/bjr/24917728
certain positions in X. The overall displacement error for the 3D motion was ¡1 mm
with a standard deviation of 2.5 mm. The system performance is satisfactory for use in ’ 2006 The British Institute of
tracking external marker motion during radiotherapy treatments. Radiology

Portal imaging protocols for treatment verification and
data analysis have been derived, and used to determine
random and systematic errors in patient set up, based on
bony anatomy and internal markers for a range of
treatment sites, for example, breast [1, 2], prostate [3–5]
and lung [6]. The data thus measured has been used to
determine appropriate margins around organs at risk for
radiotherapy planning [7–9]. This type of protocol is well-
established and interest has moved towards real time
monitoring of both external and internal organ movement.
Information about external motion throughout the
duration of the treatment may be used in combination
with electronic imaging of internal structures to improve
further the accuracy and precision of radiotherapy
treatment. This may be of particular value in situations
where the radiation delivery is gated to a physiological
function such as breathing [10–12], or where there is
potential for the treated organ to be mobile, an example
being the breast.
There is considerable interest in respiratory gating for
thoracic tumours treated with radiotherapy. The clinical
problem is significant as the large margins necessary to
create a planning target volume (typically 1.5 cm) limit
the prescribed dose to the tumour so that normal tissue
tolerances are not exceeded. If these margins could be
reduced by some control of the tumour motion, it might
be possible to escalate the dose to the tumour [13, 14].
Systems which have been used to investigate thoracic
tumour motion and respiratory gating include those
based on external devices [15] and internal markers [16].
Some authors have evaluated the correlation of the
signals from devices placed on the abdomen and thorax
with tumour motion monitored with kilovoltage imaging
[17, 18]. Others have explored the variation of internal
markers in the tumour with imaging [19]. It appears
from these studies that it is difficult to generalize on any
relationship between external and internal motion and
that this needs to be determined on an individual patient
basis [16]. One component of any system required to
investigate this is a means of tracking external markers,
either to correlate to internal motion, or to provide a
signal for gating of the linac.
We have investigated the tracking capability of a
system called Osiris+ (QADOS, Sandhurst, Berkshire)

808 The British Journal of Radiology, October 2006

Assessment of radiotherapy motion tracking system
[20]. The Osiris+ equipment is a simple camera-based
system which enables the user to check patient
contours at the time of treatment and to monitor skin
marks. Patients are positioned for treatment using skin
tattoos, often highlighted by external, cross-shaped pen
markers on their skin. The system has a function to
monitor such marks. This function can be used to ensure
that the patient has not moved during the radiation
delivery and has the potential to provide a signal to be
used in gating. It does not require any monitoring
equipment to be fixed to the patient. It is one of a number
of systems which provide a motion tracking function
[15, 21, 22]. The measurement errors of any tracking
system must be evaluated prior to any potential use in
monitoring movement. For any tracking system to
fulfil its purpose, the system measurement errors must
be much lower than the movement to be measured. The
purpose of this work was to determine the accuracy
and precision of the Osiris+ system when tracking
markers.
Methods and materials
Osiris+ system
The Osiris+ equipment may be installed in radio-
therapy simulator and treatment rooms without mod-
ifications to the simulator or treatment equipment, or
direct patient contact [20]. It consists of a set of wall
mounted cameras which are used with the in-room
alignment lasers to acquire patient contours at multiple
levels along the patient, or alone, to track markers on the
patient’s skin. When the system is used for acquiring
external outlines, images of the patient in the set-up
position are captured and the patient outline and the
reference marks may be generated by means of an
automatic outlining function. The Osiris+ system also
has a real time movement monitoring function which
allows skin marks to be tracked during irradiation using
stereoscopy – the mathematical combination of two
images taken using different camera positions to create
three-dimensional information. The real time movement
monitoring uses the marks present on the patient’s skin,
or immobilization shell, and does not require any
external markers to be adhered to the patient. Our
evaluation of the system concentrated on this monitoring
function.
The system was installed in a linac treatment room. It
had four cameras mounted on the walls, all along a line
of sight to the linac isocentre. The user indicates the
points to be tracked on still images of the target object.
Each point must be seen by two cameras in order to be
tracked in 3D. After the points have been selected, the
system displays real time images from the camera
acquired at a slow frame rate. The active motion tracking
is initiated manually and the system acquires images at a
faster frame rate for the duration of the tracking period.
This is referred to as the ‘‘linac on’’ state in the software
although the acquisition of motion data does not
currently gate to the linac pulse production. The tracking
time is determined by the user. A point tracking
algorithm determines the coordinates of the selected
points in space and time and stores these in a simple
The British Journal of Radiology, October 2006
spreadsheet format. The image acquisition rates and the
accuracy and precision of the system were all evaluated
in the experiments described.
Description of point tracking algorithm
A pattern matching function has been implemented in
the Osiris+ system to enable the tracking of points. The
function is a grey level and colour pattern matching
library. The system is trained on a reference pattern and
afterwards locates its occurrences in other images. The
library works by superimposing the pattern over the
image and comparing them by computing a (normal-
ized) correlation score, i.e. measuring discrepancies
between the pattern and the target image.
For each point selected, a search region of interest 100
pixels square is defined, and a central region of interest
of 15 pixels square is stored as a search pattern. When
locking on points, a match is searched for only in the
search region. This typically allows about ¡40 mm of
search area – depending upon the calibration values for
that camera. (If the whole image were searched for a
match, the process would take too long, especially if
several points are being tracked).
If all positions had to be tried for a match, this would
lead to an unacceptable running time. To alleviate this, a
coarse-to-fine approach is used. This means that several
search stages called reductions are performed. At the
coarsest reduction, an approximate location is found
quickly. Then the location is improved, using the next
reductions, and working in a close neighbourhood. This
arrangement drastically reduces the number of positions
to be tried. At the final stage, additional processing can
be done to achieve sub-pixel accuracy.
Points are not ‘‘lost’’ if they are temporarily obscured,
but the data associated with those points when obscured
or out of range will be set to a null.
Point matching enables a point selected from the
image obtained from a camera to be ‘‘found’’ whenever
the image is updated. The location of each point found is
indicated by its x, y screen coordinates. For this location
to be translated from ‘‘screen coordinates’’ (pixels) to a
position in space (world coordinates – mm) it is
necessary for the same point to be seen by at least two
cameras. Osiris+ uses an x, y, z world coordinate system
where x is lateral, y is vertical and z is longitudinal. The
centre of each camera’s field of view is the isocentre of
the machine. At the isocentre ‘‘z’’ is always considered to
be 0, i.e. ‘‘Z0’’. ‘‘X ’’ and ‘‘Y’’ coordinates are considered
to be in this Z0 plane.
The calibration of the Osiris+ system results in the
position of each camera being known relative to the
isocentre and so, any x, y screen coordinate of a point
lying in the Z0 plane seen by any one camera can
mathematically be translated to its true world coordi-
nates, X, Y mm in the Z0 plane only. However, if the
same point is seen by another camera then it is possible
to translate the screen coordinates for that point (e.g. x1,
y1 for camera 1 and x2, y2 for camera 2) to world
coordinates of X,Y and particularly, Z mm.
The solution is to solve for vector or ‘‘skew’’ lines,
which are lines in space that are not parallel. If the same
points can be seen by camera 1 and camera 2, then for
809
 E M Donovan, P Brabants, P M Evans et al

each point selected from camera 1 it is necessary to

correlate that point with one of the points selected from
camera 2. This is done by generating a skew line which
passes through x1, y1 between Z50 and Z5100 (camera
1). For each point from camera 2, skew lines are similarly
generated and the one which passes closest to that
from camera 1 (,5 mm ) is considered to be derived
from the same point, say, x2, y2 on camera 2. These two
points are then correlated. The process is repeated for
all points selected from camera 1 against all points
selected from camera 2, ignoring those already corre-
lated. It is then possible using direction cosines to
determine the X, Y and Z coordinates for each pair of
correlated points.
The correlation process described occurs only once
when the points to be tracked are defined. After that,
every time a fresh set of images is acquired, a pattern
match is attempted for each point from each camera. This
matched point will have screen coordinates of x1’, y1’ etc.,
which may or may not be the same as the last time
around. The actual position/movement of the point can
then be determined in world coordinates. The repetition
rate is dependant upon how many points are being
tracked and how long it takes to find a match.

Figure 1. Photograph of the test object with markers. The

Determination of accuracy and precision
A rigid hemi-spherical test object was used to evaluate
the Osiris+ system in all of the experiments. The test
object was marked, using a permanent ink pen, with
three markers in the central plane and one out of plane
marker, as shown in Figure 1. The markers were cross
shaped of 1.5 cm length and 0.2 cm width.
The Osiris+ system was used to track continuously
throughout each movement set. Each mark on the test
object was seen by at least two cameras and was
manually indicated by the user on a static camera image.
For the marker on the top of the test object, there was a
systematic discrepancy in the position indicated by the
user on the two camera views of 2 mm. The system
recorded and tracked all marked positions separately,
hence the data are recorded for five points where points
3 and 4 refer to the same marker.
Once the positions are marked by the user, the system
begins monitoring at a slow frame rate. Active monitor-
ing at a faster frame rate is initiated by the user (this
corresponds to the ‘‘linac on’’ state). In all experiments
the active mode of Osiris+ was activated prior to the
movements driven by two motorized systems. The
output from the Osiris+ system uses a binary tag to
indicate whether passive or active monitoring is being
recorded. The active monitoring state is referred to as
‘‘linac on’’ in the system software. It is initiated manually
rather than being gated to the linac pulse production. All
experiments were performed without radiation.
The two motorized systems used were a clinical beam
data measuring system and an experimental program-
mable platform. The advantage of the clinical system was
that its measurement integrity had been well validated.
However, it was difficult to use this system to create
more complex motions in 3D. The programmable plat-
form allowed the input of any function and hence
provided additional flexibility.
photograph is taken from above the test object. The height
of the object is 8.0 cm. There are three markers in a plane
and one out-of-plane marker. One camera pair viewed the
left-side, top and out-of-plane marker; a second camera pair
viewed the top and right-side markers.
Using Wellhofer beam measurement system
A Wellhofer beam measurement system (Wellhofer
Scanditronix, Schwartzenbruck, Nurnberg, Germany)
was used for the first set of experiments. This system
was in routine clinical use and the positional accuracy of
the detector carriage had been verified previously to be
0.1 mm. The test object was fitted into the detector holder
of the system and hence could be moved around in 3D
space. It was aligned with the sagittal and lateral room
set up lasers using the three central plane markers, and
that position was defined as the origin. The Wellhofer
system was used to drive the test object in each of the
three cardinal directions from the origin at the ampli-
tudes and speeds given in Table 1 in a linear movement
that was repeated five times. The Linear 1 conditions
were taken as the reference conditions and one para-
meter was changed to give each of the linear movement
conditions labelled 2 to 7.
The coordinate system of the Wellhofer equipment
was related to the Osiris+ and Elekta linac treatment
room coordinate system in the following way:
Z5Gun2Target (GT) (longitudinal) direction,
Y5vertical, X5transverse (lateral) direction. The
Osiris+ system is designed to be used with maximum
room illumination; one of the tests was carried out with
dimmed lighting to test the ability of the system to track
marks under more extreme conditions.
A further experiment moved the phantom in the Z–Y
(GT/vertical) plane in an approximation to a circle. The
coordinates of movement are shown in Figure 2 and the

810 The British Journal of Radiology, October 2006

Assessment of radiotherapy motion tracking system

Table 1. Summary of the experimental parameters for the linear motion using the Wellhofer beam data measurement system
Experiment number Direction Speed (mm s21) Amplitude (cm) Number of points Room light
tracked conditions

Linear 1 Z (GT) 10.5 ¡1.0 5 On

Linear 2 Z (GT) 10.5 ¡1.0 5 Off
Linear 3 Z (GT) 10.5 ¡1.0 3 On
Linear 4 Z (GT) 0.87 ¡1.0 5 On
Linear 5 Z (GT) 10.5 ¡5.0 5 On
Linear 6 X (Lateral) 10.5 ¡1.0 5 On
Linear 7 Y (Vertical) 10.5 ¡1.0 5 On
GT, gun–target axis of the linear accelerator.

Wellhofer system was programmed to drive between the
positions in a straight line at a speed of 10.5 mm s21 with
no pause between sections. This sequence of movements
was repeated four times.
Using programmable computer controlled platform
with motion in X,Y,Z
The computer-controlled platform (Time and
Precision, Basingstoke, UK) has three orthogonal axes
driven by leadscrews and stepper motors. Each axis has
an incremental encoder which allows the position to be
known at a resolution of 0.5 mm, even if a motor skips
steps. The motors are microstepped to give a drive
resolution of 0.5 mm. A versatile three-axis motion
controller (Model DMC-2130; Galil Motion Control,
Rocklin, CA) generates the step and direction signals to
the three motor drives, according to the specified
trajectory. The controller also keeps track of the actual
position of the axes via the encoders, but the steppers are
run open-loop. The difference between commanded
position and actual position is less than 0.02 mm.
The test object was fixed on the platform and aligned
to the room lasers using the three markers in the central
plane to define the origin position. The platform was
used to drive the test object to 3 points in X,Y,Z from this
origin: ¡1 cm, ¡2 cm, ¡4 cm. After each set of move-
ments in the three cardinal directions, the test object was
returned to the origin. This experiment was designed to
assess the absolute positional accuracy of the system. The
platform was used to drive the test object in four
elliptical trajectories in three-dimensional space. To
generate the ellipses, the controller was run in its
‘‘electronic cam’’ mode, in which the trajectories of the
three axes are cyclically synchronised to a virtual cam
axis. Commands were sent to the controller from a
laptop via a serial link. These trajectories corresponded
to an ellipse of eccentricity 0.6 tilted at 45 ˚ in both
horizontal and vertical planes. The details are listed in
Table 2.
Results
Using the Wellhofer beam measurement system

Linear motion: positional data

Figure 2. Coordinates of circular motion in Y–Z plane
(Vertical: gun–target (GT) plane) using the Wellhofer beam
data system.
These data were analysed by evaluating the difference
between the true position and time, as given by the speed
and period of the Wellhofer equipment, with that
recorded by the Osiris+ system. Figure 3 provides a
summary of the standard deviation of the positional data
difference for the seven experiments. For each experi-
ment the data are separated into the X,Y,Z components
of movement and are presented as the average over all of
the tracked points and over all five cycles. Each of these
experiments involved motion in one of the cardinal
directions only. The direction of motion of the test object
is indicated by the black bar on the chart. Data were
collected for all three directions. The largest standard
deviations are found for experiment 2, when the system
was tracking the points under dimmed lighting, and for
experiment 5 where the motion had a large amplitude of
¡5.0 cm. Neither of these two conditions resulted in
standard deviations of greater than 2 mm; the other
experiments showed that the system could track the
points with standard deviations of less than 1 mm. The
conditions in which the system tracked the motion with
the greatest precision were when the motion was slow
(0.87 mm s21) and here the standard deviations did not
exceed 0.4 mm. Table 3 shows the standard deviation

The British Journal of Radiology, October 2006 811

 E M Donovan, P Brabants, P M Evans et al

Table 2. Summary of elliptical motion tracking experiment Table 3. SD (mm) averaged over all tracked points and each
parameters carried out using the computer controlled plat- direction of motion for each of the linear motions of the test
form to drive the test object object driven by the Wellhofer beam data system
Experiment Eccentricity Semi-major axis Period (s) Experiment number SD (mm)
number amplitude (cm)
Linear 1 0.64
Ellipse 1 0.6 2.0 6 Linear 2 1.43
Ellipse 2 0.6 1.0 2 Linear 3 0.55
Ellipse 3 0.6 1.0 4 Linear 4 0.33
Ellipse 4 (circle) 1.0 1.0 4 Linear 5 1.14
Linear 6 0.49
Linear 7 0.42
data averaged over all directions and for all points for all
the experimental conditions.

Circular motion: positional data
The data of Figure 4 show the standard deviation for the
2D circular motion in the Y–Z plane (GT, vertical) for each
of the repeat cycles. The data are given for each tracked
point averaged over the Y and Z positions recorded by the
Osiris+. The standard deviation values were determined
from the difference between the recorded values at each
Osiris+ monitoring time point and the expected Y, Z
values interpolated from the Wellhofer driven movements
between points. The standard deviation does not exceed
1 mm and the average value over all cycles and points is
0.66 mm. Values of the X coordinate were recorded also,
although the movement was not in this direction. The
average value of one standard deviation over all cycles and
points is 0.5 mm.
Temporal data: linear and circular motion
The data from both linear and circular movements of the
test objects have been used to determine the sampling rates
of the Osiris+ system. Figure 5 shows that these are 0.3–
0.4 s whilst the system is monitoring actively (i.e. ’’linac
on’’ state) and 1.5–1.6 s in the ‘‘linac off’’ state.
Using the programmable computer controlled
platform with motion in X,Y,Z
Stationary point measurement
Figure 6 summarizes the results obtained from the
Osiris+ system when the test object was driven to
specific, known displacements from its origin using the
computer controlled platform. The Osiris+ system was
assessed by determining the deviation of the recorded
values from the absolute position coordinates, as given
by the platform. Figure 6 gives the absolute differences
in millimetres between the platform values and the
recorded data from Osiris+. All the differences were
within ¡1 mm. The test object was returned to the origin
between each set of measurements and no variation
in this recorded value was found throughout the
experiment.
Elliptical motion in 3D
Data from the analysis of elliptical movements 1 and 3
are used as an example of the ability of the Osiris+
system to track motion in 3D. The results for the ellipse
experiments 2 and 4 follow the same pattern; as these
add no further information, the data are not given. As an
example, Figure 7 shows the movement in the X
direction (transverse) for ellipse 1. The true, calculated
trajectory and the trajectories recorded by the Osiris+
system are shown for each of the five points. The
absolute difference between the exact positions on the
elliptical trajectories and the recorded positions were
determined for each point for each of the X,Y,Z
directions. These data are given in Figure 8 for ellipse
1. Figure 8b,c shows that the absolute errors in Y and Z
are within ¡1 mm for the majority of the positions, but
the data in Figure 8a show that for points 1,2 and 3 the
absolute error values are only within ¡4 mm for ellipse
motion 1 in the X direction.

Figure 3. Summary of errors for

linear motion. The plot shows the
deviation from the exact to tracked
position averaged over all five cycles
of the linear movement.

812 The British Journal of Radiology, October 2006

Assessment of radiotherapy motion tracking system
Table 4 summarizes four parameters from the analysis
of the elliptical movements 1 and 3. For each tracked
point: the average displacement; the standard deviation
of the variation and the minimum and maximum
displacements are given. These are all averaged over
the three cycles of the motion. The average value of the
displacement is within ¡1 mm with an average stan-
dard deviation within 2.5 mm. The largest magnitudes of
the minimum and maximum displacements are
24.72 mm and +4.0 mm, both for the ellipse 1 motion.
In this experiment, the test object was moving at a fast
speed and with the largest amplitude.
Discussion
The tracking of patient movement during a fraction of
a radical radiotherapy treatment requires a monitoring
system with sufficient accuracy and precision. One
potential application of such a system is to use the
signal as a trigger to gate a linac for delivery based on the
respiratory cycle as described by Berson et al [10]. Real-
time monitoring of patient movements also has an
application in situations where control of breathing is
used to reduce an organ at risk dose, as suggested by Lu
et al [23] and reported by Remouchamps et al [24]. We
The British Journal of Radiology, October 2006
Figure 4. Summary of errors for
circular motion. The values for each
point are the average of the devia-
tions from the calculated to tracked
position.
have investigated the Osiris+ system in order to
determine whether its performance characteristics were
sufficient for this type of patient monitoring.
The systems used for respiratory controlled studies as
described in the literature tend to be either of the Active
Breathing Control (ABC) device [11] (now from Elekta
Oncology Systems, Crawley, UK) or the Varian Medical
Systems RPMTM equipment (Varian Medical Systems,
Palo Alto, CA) [10, 12]. These use a flow volume (the
ABC device) or a marker on the patient’s chest (Varian
RPM) to monitor respiratory motion and breath-hold
over time, in one dimension. The studies by Berson et al
[10] and Pedersen et al [12] show that the trace of the
external marker on a patient’s chest has an amplitude of
between 5 mm and 19 mm. It would be realistic to expect
a monitoring system to detect movements of the external
surface of a patient down to 2 mm.
Yan et al [22] examined the Novalis Body system
(BrainLAB Inc., Germany) which uses a combination of
infrared and X-ray imaging for tracking markers. In an
experiment similar to the absolute position checking, they
moved a phantom known distances (2 mm, 5 mm, 10 mm
and 20 mm) and determined the positional error of the
system in X, Y and Z. They quote the average errors in the
lateral direction as 0.6¡0.3 mm; those in the vertical
direction as 0.7¡0.2 mm and those in the longitudinal
Figure 5. Summary of sampling
intervals. Data are from linear and
circular motion.
813

direction as 0.5¡0.2 mm. The results from the Osiris+
system given in Figure 6 compare well with these data.
Yan et al did not investigate the continuous tracking of
markers on a moving object. Our assessment of the Osiris+
system covered a range of movements in one, two and
three dimensions and different speeds. Figure 3 and 4
demonstrate that the errors in tracking for the sawtooth
and 2D circular motions were within ¡1 mm for most
conditions. Two situations, which were more extreme,
were the linear 2 and 5 experiments. The former used the
Osiris+ with the lights dimmed and the latter had a ¡5 cm
amplitude, which is much larger than would be found in
practice. Here the errors rose above 1.5 mm. Shirato et al
[21] evaluated the performance of a Real-Time Tumour-
Tracking system (EXL-20DP; Mitsubishi Electronics Co.,
Ltd., Tokyo, Japan) using a phantom moving in a circle
over a range of speeds from 6 mm s21 to 40 mm s21. They
Figure 7. Data recorded from Osiris+ for displacement in X axis with time compared with programmed displacement for
platform movement Ellipse 1.
814
E M Donovan, P Brabants, P M Evans et al
Figure 6. Summary of errors for all
points at all positions for the test
object moved using the program-
mable platform.
quote the accuracy of the system as better than 1.5 mm; a
similar level of accuracy to the Osiris+.
The largest displacement errors were in the tracking of
the points when the test object was driven in the elliptical
movements 1 and 3, but only in the X direction
(transverse) and only for three of the points. Here the
error could reach ¡5 mm for ellipse 1 and ¡3 mm for
ellipse 3 at the extremes of the travel. The marker at the
top of the test object was seen by two different pairs of
cameras and recorded twice as point 3 and point 4. Only
the point 3 X direction errors were large for this marker.
The reason is unclear and this is under further
investigation. The differences in the phase of the data
(shown in Figure 7) may indicate that the cameras were
measuring at different time points as data points 1, 2 and
3 were recorded by one camera pair and the data from
points 4, 5 were from the other camera pair. The mean
The British Journal of Radiology, October 2006
Assessment of radiotherapy motion tracking system

Figure 8. (a) Absolute error in posi-

tion in the X direction for ellipse 1
for all points. (b) Absolute error in
position in the Y direction for ellipse
1 for all points. (c) Absolute error in
position in the Z direction for ellipse
1 for all points.

displacement error in the X direction was 20.23 mm ¡ performance was sufficient to meet the requirement to
1.7 mm for Ellipse 1. The mean error in Y was 0.12 mm track movement with a minimum threshold of 2 mm and
¡0.5 mm and that in the Z direction 20.09 mm ¡ the imaging rate of one every 0.4 s was sufficient for
0.6 mm. For most situations investigated, the system good monitoring accuracy.

The British Journal of Radiology, October 2006 815

 E M Donovan, P Brabants, P M Evans et al

Table 4. Summary of error analysis data for elliptical movement 1 and 3 using the programmable platform
Average (mm) SD (mm) Minimum (mm) Maximum (mm)

Ellipse number 1 3 1 3 1 3 1 3

Direction/point

X1 20.3 0.48 2.37 1.54 24.72 21.88 3.61 3.15

Y1 20.01 20.43 0.52 0.37 20.93 21.35 1.27 0.41
Z1 0.02 0.76 0.6 0.39 21.47 20.21 1.19 1.47
X2 20.34 0.56 2.32 1.43 24.52 21.68 3.31 3.05
Y2 0.19 20.58 0.39 0.35 20.6 21.19 1.17 0.21
Z2 0.06 0.45 0.73 0.51 21.47 20.81 1.19 1.37
X3 20.07 0.7 2.28 1.6 24.52 21.98 4.0 3.38
Y3 0.03 20.51 0.32 0.39 20.81 21.39 0.92 0.26
Z3 20.04 0.54 0.6 0.41 21.47 20.45 0.95 1.47
X4 20.12 0.56 0.81 0.69 22.12 20.54 1.5 2.03
Y4 0.26 20.36 0.59 0.55 20.99 21.81 1.89 0.45
Z4 20.35 0.51 0.6 0.57 21.89 20.52 0.78 1.65
X5 20.31 0.39 0.56 0.62 21.52 20.45 0.74 1.83
Y5 0.14 20.54 0.47 0.45 20.72 21.81 1.11 0.44
Z5 20.15 0.43 0.65 0.62 21.43 20.72 0.93 1.75
The errors are derived from the difference between the calculated position from the input function to the platform and that
recorded by the Osiris+ tracking software at the sampled time points.

Conclusion 5. Chung PWM, Haycocks T, Brown T, Cambridge Z, Kelly V,

Assessment of the Osiris+ system found that contin-
uous linear, circular and elliptical motions were tracked
with a standard deviation of 1 mm in the difference
between measured and true positions of the test object
markers. Under more extreme conditions, such as
dimmed lighting and large amplitude of motion, this
increased to 2 mm. The recorded values of static, known
displacements by the system were also within ¡1 mm.
The sampling rates of the system were 1.4–1.5 s prior to
the command to track and 0.3–0.4 s during the tracking
time. It is recommended that the system be used with the
lights on and that regular calibrations are carried out to
maintain its performance.
The system performance is considered satisfactory to
enable its use (i) in assessing external motion based on
skin markers; (ii) in conjunction with electronic portal
imaging to determine the nature, if any, of external-to-
internal organ movement correlation and (iii) to provide
a signal for respiratory gating.
References
1. Das IJ, Cheng C-W, Fosmire H, Kase KR, Fitzgerald TJ.
Tolerances in setup and dosimetric errors in the radiation
treatment of breast cancer. Int J Radiat Oncol Biol Phys
2000;26:883–90.
2. Lirette A, Pouliot J, Aubin M, Larochelle M. The role of
electronic portal imaging in tangential breast irradiation: a
prospective study. Radiother Oncol 1995;37:241–5.
3. Alasti H, Petric MP, Catton CN, Warde PR. Portal imaging
for evaluation of daily on-line setup errors and off-
line organ motion during conformal irradiation of carci-
noma of the prostate. Int J Radiat Oncol Biol Phys
2001;49:869–84.
4. Aubry JF, Beaulieu L, Girouard LM, Aubin S, Tremblay D,
Laverdiere J, et al. Measurements of intrafraction motion
and interfraction and intrafraction rotation of prostate by
three-dimensional analysis of daily portal imaging with
radiopaque markers. Int J Radiat Oncol Biol Phys
2004;60:30–9.
Alasti H, et al. On-line aSi portal imaging of implanted
fiducial markers for the reduction of interfraction error
during conformal radiotherapy of prostate carcinoma. Int J
Radiat Oncol Biol Phys 2004;60:329–34.
6. Van de Steene J, Van den Heuvel F, Bel A, Verellen D, De
Mey J, Noppen M, et al. Electronic portal imaging with
on-line correction of setup error in thoracic irradiation:
clinical evaluation. Int J Radiat Oncol Biol Phys
1998;40:967–76.
7. McKenzie A, van Herk M, Mijnheer B. Margins for
geometric uncertainty around organs at risk in radio-
therapy. Radiother Oncol 2002;62:299–307.
8. Ekberg L, Holmberg O, Wittgren L, Bjelkengren G,
Landberg T. What margins should be added to the clinical
target volume in radiotherapy treatment planning for lung
cancer? Radiother Oncol 1998;48:71–7.
9. De Boer HCJ, Van Os MJH, Jansen PP, Heumen BJM.
Application of the no action level (NAL) protocol to correct
from prostate motion based on electronic portal imaging of
implanted markers. Int J Radiat Oncol Biol Phys
2005;61:969–83.
10. Berson AM, Emery R, Rodriguez L, Richards GM, Ng T,
Sanghavi S, et al. Clinical experience using respiratory
gated radiation therapy: comparison of free-breathing and
breath-hold techniques. Int J Radiat Oncol Biol Phys
2004;60:419–26.
11. Wong JW, Sharpe MB, Jaffray DA, Kini VR, Robertson JM,
Stromberg JS, et al. The use of active breathing control
(ABC) to reduce margin for breathing motion. Int J Radiat
Oncol Biol Phys 1999;44:911–9.
12. Pedersen A, Korreman S, Nystrom H, Specht L. Breathing
adapted radiotherapy of breast cancer: reduction of cardiac
and pulmonary doses using voluntary inspiration breath-
hold. Radiother Oncol 2004;72:53–60.
13. Starkschall G, Forster KM, Kitamura K, Cardenas A, Tucker
SL, Stevens CW. Correlation of gross tumor volume
excursion with potential benefits of respiratory gating. Int
J Radiat Oncol Biol Phys 2004;60:1291–7.
14. Engelsman M, Sharp GC, Bortfeld T, Onimaru R, Shirato H.
How much margin reduction is possible through gating or
breath hold? Phys Med Biol 2005;50:477–90.
15. Denissova SI, Yewondwossen MH, Andrew JW, Hale ME,
Murphy CH, Purcell SR. A gated deep inspiration

816 The British Journal of Radiology, October 2006

Assessment of radiotherapy motion tracking system
breath-hold radiation therapy technique using a linear
position transducer. J Appl Clin Med Phys 2005;6:61–70.
16. Berbeco RI, Nishioka S, Shirato H, Chen GT, Jiang SB.
Residual motion of lung tumours in gated radiotherapy
with external respiratory surrogates. Phys Med Biol
2005;50:3655–67.
17. Ford EC, Mageras GS, Yorke E, Rosenzweig KE, Wagman
R, Ling CC. Evaluation of respiratory movement during
gated radiotherapy using film and electronic portal ima-
ging. Int J Radiat Oncol Biol Phys 2002;52:522–31.
18. Hoisak JDP, Sixel KE, Tirona R, Cheung PCF, Pignol J-P.
Correlation of lung tumor motion with external surrogate
indicators of respiration. Int J Radiat Oncol Biol Phys
2004;60:1298–306.
19. Onimaru R, Shirato H, Fujino M, Suzuki K, Yamazaki K,
Nishimura M, et al. The effect of tumor location and
respiratory function on tumor movement estimated by real-
time tracking radiotherapy (RTRT) system. Int J Radiat
Oncol Biol Phys 2005;63:164–9.
The British Journal of Radiology, October 2006
20. Wilks RJ. An optical system for measuring surface shapes
for radiotherapy planning. Br J Radiol 1993;66:351–9.
21. Shirato H, Shimizu S, Kunieda T, Kitamura K, van Herk M,
Kagel K, et al. Physical aspects of a real-time tumor-tracking
system for gated radiotherapy. Int J Radiat Oncol Biol Phys
2000;48:1187–95.
22. Yan H, Yin FF, Kim JH. A phantom study on the
positioning accuracy of the Novalis Body system. Med
Phys 2003;30:3052–60.
23. Lu H-M, Cash E, Chen MH, Chin L, Manning WJ, Harris J,
et al. Reduction of cardiac volume in left-breast treatment
fields by respiratory maneuvers: a CT study. Int J Radiat
Oncol Biol Phys 2000;47:895–904.
24. Remouchamps VN, Vicini FA, Sharpe MB, Kestin LL,
Martinez AA, Wong JW. Significant reductions in heart and
lung doses using deep inspiration breath hold with active
breathing control and intensity-modulated radiation ther-
apy for patients treated with locoregional breast irradiation.
Int J Radiat Oncol Biol Phys 2003;55:392–406.
817
 The British Journal of Radiology, 79 (2006), 818–827

Normalized data for the estimation of fetal radiation dose from

radiotherapy of the breast
1 2 2,3
B BRADLEY, BSc, A FLECK, BSc, MSc and E K OSEI, BSc, MSc, PhD

1
Department of Systems Design Engineering, University of Waterloo, 200 University Avenue West,
Waterloo, Ontario, 2Department of Medical Physics, Grand River Regional Cancer Center, 835 King
Street West, Kitchener, Ontario and 3Department of Physics, University of Waterloo, 200 University
Avenue West, Waterloo, Ontario, Canada

ABSTRACT. There can be several reasons why a pregnant patient may receive a
radiological examination. It could have been a planned exposure, or the exposure
might have resulted from an emergency when a thorough evaluation of pregnancy was
impractical. Sometimes the pregnancy was unsuspected at the time of the examination
and, with younger women being diagnosed with breast cancer, the likelihood of this
will increase in radiotherapy departments. Whatever the reason, when presented with
a pregnant patient who has received a radiological examination involving ionizing
radiation, the dose to the fetus should be assessed based on the patient’s treatment
plan. However, a major source of uncertainty in the estimation of fetal absorbed dose is
the influence of fetal size and position as these change with gestational age.
Consequently, dose to the fetus is related to gestational age. Various studies of fetal
dose during pregnancy have appeared in the literature. Whilst these papers contain
many useful data for estimating fetal dose, they usually contain limited data regarding
the depth and size of the fetus within the maternal uterus. We have investigated doses
to the fetus from radiation therapy of the breast of a pregnant patient using an
anthropomorphic phantom. Normalized data for estimating fetal doses that takes into
account the fetal size (gestational age: 8–20 weeks post-conception) and depth within
the maternal abdomen (4–16 cm) for different treatment techniques have been
provided. The data indicate that fetal dose is dependent on both depth within the
maternal abdomen and gestational age, and hence these factors should always be
considered when estimating fetal dose. The data show that fetal dose can be Received 24 January 2006
underestimated up to about 10% or overestimated up to about 30% if the dose to the Revised 13 April 2006
uterus is assumed instead of the actual fetal dose. It can also be underestimated up to Accepted 25 April 2006
about 23% or overestimated up to about 12% if a mean depth of 9 cm is assumed,
DOI: 10.1259/bjr/16416346
instead of using the actual depth of the fetus within the maternal abdomen. Multi-
segments sMLC technique showed consistently lower fetal doses compared with all the ’ 2006 The British Institute of
wedged plans employed. Radiology

Peripheral dose (PD), which is an inevitable conse-
quence of radiotherapy, is mainly due to the radiation
that is scattered within the patient and also the scattered
and leakage radiation from the head of the machine and
the collimator assembly. The magnitude of the scattered
radiation from within the patient will mainly depend on
the beam energy, distance from the radiation field edge,
field size and depth. For a pregnant patient, such dose is
of greater concern due to the fact that the developing
fetus is more susceptible to the adverse effects of ionizing
radiation, especially when the absorbed dose to the fetus
may be high. However, using additional shielding can
minimize the radiation dose due to scatter and leakage
radiation from the machine head and collimator assem-
bly [1, 2].
Address correspondence to: Ernest K Osei, Department of Medical
Physics, Grand River Regional Cancer Center, 835 King Street West,
Kitchener, Ontario, Canada. E-mail: ernestkwaku.osei@grhosp.on.ca.
Although radiotherapy treatment during a known
pregnancy (especially during the 8–15 weeks gestational
age period) should be avoided [3], it is not always
possible to postpone treatment for the full duration of the
pregnancy. Under such circumstances, it is essential to be
able to estimate the absorbed dose (and hence risk) to the
fetus if the patient must undergo treatment.
Modifications to the plan and shielding techniques can
then be incorporated where necessary to achieve an
optimum balance between risk to the fetus and benefit to
the mother, although the fetus also receives an indirect
benefit. There are also occasions when pregnancy may
not be known before the onset of a therapeutic treatment,
and sometime later the patient realises she is pregnant.
Under such a situation too, the fetal absorbed dose (and
hence risk) as a result of the radiation treatment received
should be assessed by a medical physicist. With younger
women being diagnosed with breast cancer, the like-
lihood increases that a patient may be pregnant when

818 The British Journal of Radiology, October 2006

Estimation of fetal dose from radiotherapy of the breast
treatment is indicated or that inadvertent treatment may
take place for women who subsequently learn that they
are pregnant.
Fetal dose estimation is a difficult and time-consuming
procedure that requires extensive dosimetric measure-
ments and sometimes the construction of a specific
phantom to simulate the patient’s geometry [3, 4]. A
review of the scientific literature regarding the measure-
ment of peripheral dose for the estimation of dose to the
fetus from radiotherapy procedures has revealed an
extensive collection of data [1, 5–14]. Mazonakis devel-
oped a method similar to the one described here for
estimating fetal dose during brain radiotherapy. Using
an anthropomorphic phantom, fetal absorbed dose was
determined at two gestational ages and depths [4].
Rincon [9] has also estimated the dose to the fetus using
peripheral dose to the uterus for two standard breast
radiotherapy treatments (a non-wedged plan and a 15 ˚
physical wedged plan). Point dose measurements were
taken in a phantom from the field edge to the uterus, and
the fetal depth was assumed to be 15 cm. However, there
are very limited data in the literature that take into
account the different fetal positions (depth within the
maternal abdomen) and gestational age (size) when
estimating dose, although it has been shown that these
fetal parameters are significant factors when estimating
dose [2, 15]. Therefore, a source of data, which accounts
for the different fetal parameters (i.e. fetal depth and
size), would be very useful when estimating the fetal
absorbed dose and hence risk.
This paper presents the experimental procedure and
results of fetal dose estimation from radiotherapy
treatment of the breast. Data are presented for two
different photon beam energies (6 MV and 15 MV), four
different physical wedge angles (15 ˚, 30 ˚, 45 ˚ and 60 ˚),
and a segmented multi-leaf collimator (sMLC) plan. Fetal
sizes (gestational age) considered range from 8 weeks to
20 weeks post-conception, and fetal depth within the
maternal abdomen ranges from 4 cm to 16 cm. By
treating the fetus as a three-dimensional volume at
various gestational ages and depths from the anterior
surface of the mother’s abdomen, data are generated for
a more accurate estimation of the mean dose to the fetus
from radiotherapy of the breast. Fetal dose is assumed to
be a whole body exposure of the fetus within the
maternal abdomen.
There are three scenarios where the data presented in
this paper can be of benefit to the medical and
paramedical personnel involved with fetal dose estima-
tion. For the case of a patient who has already completed
radiotherapy treatment and discovers later that she was
pregnant during the time of treatment, the data would be
useful in estimating the dose to the fetus and hence risk
using the patient’s treatment plan parameters. Such
inadvertent fetal exposures would be very rare if modern
patient safety legislation is implemented well in radi-
ological departments. For the case of a patient who,
during the course of treatment, discovers that she is
pregnant and there is clinical indication for her to
complete treatment, the data would be useful to estimate
the fetal dose from the treatment already received. The
patient’s treatment plan could then be modified and
optimized in order to minimize the fetal dose.
Appropriate abdominal shielding could also be
The British Journal of Radiology, October 2006
employed for the rest of the treatment to further reduce
the fetal dose and hence risk and the actual dose
measured for the rest of the treatment. Lastly, for the
case of a patient who is pregnant and wishes to have
radiotherapy treatment and there is also clinical indica-
tion for her to undergo treatment, the data would be
useful to estimate the fetal dose without taking into
account any form of shielding (maximum dose). A
detailed study of the actual treatment plan could then
be conducted, and the plan might be modified and
optimized to reduce the fetal dose. Furthermore, appro-
priate abdominal shielding could then be employed to
further reduce the fetal dose and risk from the treatment.
Actual fetal dose should then be measured using the
optimized treatment plan for risk assessment.
Materials and methods
Experimental set-up
The experimental set-up used for the estimation of fetal
dose from radiation therapy of the breast is shown in
Figure 1. An Alderson adult Rando anthropomorphic
phantom (Alderson Research Labs, Stamford, CT) was
used to represent the patient. The phantom consists of a
human skeleton encased in tissue-equivalent material,
with 33 transverse sections. It has an AP thickness of
20 cm, a width (shoulder-to-shoulder) of about 34 cm and
a height of 90 cm. Radiographic film (Kodak X-OMAT TL,
Rochester, NY) was used for all dosimetry measurements.
The films were placed between each of the slices spanning
the fetal region. The films spanned from slice 27 to 33
depending on the length of the fetus at the gestational age
being studied (Figures 1 and 2). The film calibration curve
was performed in a homogeneous medium. A sensito-
metric curve was generated for each batch of film at the
time of measurement and used to convert optical densities
into relative doses. The phantom was irradiated using the
plans generated (6 MV and 15 MV (all wedges), and
segmented MLC plan). The irradiated films were devel-
oped, scanned and analysed using a Vidar VXR-16
Dosimetry Pro scanner (Vidar, Herndon, VA) and
RIT113 (v4) Radiation Therapy Dosimetry Software.
Metal-oxide-semiconductor field-effect transistors
(MOSFETs) were also used to measure point doses in
order to verify the film dosimetry. The pegs along a
central axis throughout the fetal region of the Rando
phantom were removed and five high-sensitivity
MOSFET dosemeters were placed with the sensitive
region covering the holes (Figure 2). They were at a
depth of approximately 11 cm from the anterior surface
of the phantom and ranged from a distance of
approximately 22.5 cm to 32.5 cm from the field edge.
The phantom was placed in a supine position on the
treatment couch in accordance with department protocol
(Figure 1) and was ‘‘treated’’ with the same set-up for
each treatment plan described below.
Treatment planning
The phantom underwent a treatment planning CT
scan. At the time of scanning, the phantom was placed in
819

the treatment position for tangential breast radiation
therapy adopted at the centre (i.e. supine position). A
helical AcQSim CT scanner (Philips Medical Systems
Cleveland Inc., Ohio, USA) was used to acquire
contiguous 3 mm CT axial images. The CT data set was
then transferred to a 3D treatment planning workstation
(Pinnacle3; ADAC Laboratories, Milpitas, USA) for
treatment planning. Using our departmental protocol
for beam arrangements, a typical two-field tangential
breast plan was developed. This standard plan was
then modified to incorporate different photon energies
(6 MV and 15 MV), wedge angles (15 ˚, 30 ˚, 45 ˚ and 60 ˚),
and a multi-segment static MLC technique. The MLC
leaves were fully retracted in all wedge plans. In total,
10 different treatment plans were generated for the
study and they spanned the different plans usually used
in the radiotherapy centre for the treatment of breast
cancer.
Fetal parameters and dose calculation
A chart of fetal dimensions was created for all
gestational ages ranging from 8 weeks to 20 weeks
post-conception, which are based on the mean of a series
of measurements taken from the literature [16–19]. They
gave the fetal crown–rump-length (CRL), biparietal
diameter (BPD), abdominal circumference (AC), head
circumference (HC) and fetal mass (wt) at the various
820
B Bradley, A Fleck and E K Osei
Figure 1. The experimental setup
for a typical two-field tangential
breast treatment. The Rando phan-
tom is placed in a supine position on
the treatment couch and setup as
per departmental protocol, and was
‘‘treated’’ with the same set-up for
each treatment plan.
Figure 2. The films and Metal-
oxide-semiconductor field-effect
transistors (MOSFETs) placement
within the phantom showing the
positioning of the films and
MOSFETs. MOSFETs were located
along a central axis within the fetal
region. When measurements are
being taken with film, the
MOSFETs are removed and likewise,
when measurements are being
taken with the MOSFETs for com-
parison, the films are removed.
gestational ages, with statistical variation as 2 standard
deviations from the mean. These data were used to
create a three-dimensional rectangular volume to repre-
sent the fetus at the different gestational ages (Table 1).
The density (r) of the fetus was assumed to be
1.061023 kg m23. For gestational ages ranging from
8 weeks to 12 weeks post-conception, the CRL was used
for the length (l), the width (w) was estimated from the
mean of the abdominal circumference (i.e. AC/p) and the
anteroposterior (AP) thickness was calculated from:
wt
AP~
l|w|r
For gestational ages ranging from 13 weeks to 20
weeks post-conception, the biparietal diameter was used
for the width, AP thickness was estimated from the mean
abdominal circumference and the length was estimated
in a similar way as above. Using the RTT113 software, a
rectangular region of interest (ROI) (Figure 3) was
selected on each film using the width and AP thickness
dimensions of the fetus at the depth of interest. The mean
dose and standard deviation for this region are recorded.
This process is repeated for each subsequent film
spanning the length of the fetus at the same depth and
gestational age dimensions. From the complete set of
mean doses for each film, the mean dose for the volume
(gestational age) is estimated. The whole process is then
The British Journal of Radiology, October 2006
Estimation of fetal dose from radiotherapy of the breast

Table 1. Fetal dimensions at different gestational ages

Gestational age Length (cm) Width (cm) AP thickness (cm) Mass (g) Calculations
(weeks)

8 1.6 0.7 0.9 1.0 Length5CRL

9 2.3 0.8 1.1 2.0 Width5AC/p
10 3.3 1.0 1.2 4.0 AP5calculated
11 4.1 1.4 1.2 2.0
12 5.4 1.9 1.4 14.0
13 5.9 2.1 2.3 28.0 Width5BPD
14 6.5 2.5 2.6 43.0 AP5AC/p
15 8.0 2.9 3.0 70.0 Length5calculated
16 9.2 3.2 3.4 100.0
17 10.7 3.5 3.7 140.0
18 11.8 3.9 4.1 190.0
19 12.5 4.3 4.5 240.0
20 13.5 4.6 4.8 300.0
CRL, crown–rump length; AC, abdominal circumference; AP, anteroposterior; BPD, biparietal diameter.

repeated for each gestational age (8–20 weeks) and depth
(4–16 cm) of the fetus, and for all the films exposed from
each treatment plan delivered.
Results and discussion
In this paper, fetal dose is assumed to be a whole body
exposure of the fetus within the maternal abdomen. The
data collected and presented here were taken without any
modification to the treatment plan setup to minimize the
dose to the fetus. They would be useful for estimating dose
to the fetus when a patient realises she is or was pregnant
during or after the course of treatment. If there is a clinical
indication for a pregnant patient to complete treatment
when pregnancy is known, the data could be used to
estimate the fetal dose without shielding and the setup
could then be modified by the use of external shielding
over the abdominal region of the pregnant patient during
treatment to significantly reduce the fetal dose.
The dimensions of the rectangular shape representing
the fetus at different gestational ages are shown in
Table 1. Measured fetal doses normalized to the pre-
scribed dose at the isocentre of the plan as a function of
fetal depth within the maternal abdomen for different
breast treatment plans are given in Tables 2–6. Figure 4
shows the relationship between the normalized fetal
dose as a function of depth for all gestational ages using
the multi-segment static MLC treatment plan and 6 MV
photon beam. A similar relationship comparing the
various treatment plans (sMLC, 15 ˚, 30 ˚, 45 ˚ and 60 ˚)
and photon beam energies (6 MV and 15 MV) is shown
in Figure 5. One of the most important parameters in the
measurement of fetal dose from radiotherapy procedures
is the distance from the radiation field edge to the point
of measurement. The fetal dose decreases approximately
exponentially with distance from the field edge.
Published data [2, 4] show that the change in the
peripheral dose with depth is small, although other data
[1, 8] show greater change for a Co-60 treatment unit.
The results of fetal dose estimation from radiotherapy of
the breast show that fetal dose is dependent on both fetal
depth within the maternal abdomen and gestational age
(Figure 4). In general for all the techniques used and for

Figure 3. Region of interest repre-

senting a fetus at a gestational age
of 20-weeks post-conception and at
a depth of 8 cm from the abdominal
surface of the maternal abdomen.

The British Journal of Radiology, October 2006 821

 B Bradley, A Fleck and E K Osei

Table 2. Normalized mean fetal dose, as a percentage of prescribed dose (PD), for a typical two-field tangential beam plan
using multi-segment static MLC for both 6 MV and 15 MV photon energies. Data are presented for various gestational ages (GA)
and fetal depth within the maternal abdomen
Photon energy GA (weeks) Fetal depth (cm)

4 5 6 7 8 9 10 11 12 16

8 0.103 0.102 0.100 0.099 0.097 0.095 0.093 0.090 0.087 0.068
9 0.104 0.103 0.101 0.099 0.097 0.096 0.093 0.091 0.088 0.069
10 0.106 0.105 0.103 0.100 0.099 0.097 0.094 0.092 0.089 0.070
11 0.107 0.106 0.104 0.102 0.100 0.098 0.096 0.093 0.090 0.071
12 0.110 0.109 0.107 0.105 0.103 0.100 0.098 0.095 0.093 0.073
13 0.112 0.111 0.109 0.106 0.104 0.101 0.099 0.096 0.094 0.073
6 MV
14 0.114 0.113 0.111 0.108 0.106 0.103 0.100 0.098 0.095 0.074
15 0.119 0.118 0.116 0.113 0.110 0.108 0.105 0.101 0.098 0.076
16 0.125 0.124 0.121 0.118 0.115 0.112 0.109 0.105 0.101 0.079
17 0.133 0.131 0.129 0.126 0.122 0.119 0.115 0.110 0.106 0.081
18 0.140 0.138 0.135 0.132 0.128 0.124 0.119 0.114 0.109 0.083
19 0.144 0.142 0.139 0.136 0.132 0.128 0.123 0.117 0.111 0.085
20 0.151 0.149 0.146 0.142 0.138 0.133 0.128 0.122 0.115 0.087
8 0.088 0.086 0.084 0.082 0.080 0.077 0.075 0.072 0.069 0.054
9 0.089 0.087 0.085 0.083 0.081 0.078 0.076 0.073 0.070 0.054
10 0.090 0.089 0.087 0.085 0.083 0.080 0.078 0.075 0.072 0.055
11 0.091 0.090 0.088 0.086 0.084 0.082 0.079 0.076 0.073 0.056
12 0.093 0.092 0.090 0.088 0.086 0.084 0.081 0.079 0.075 0.057
13 0.095 0.093 0.091 0.089 0.087 0.085 0.082 0.079 0.076 0.057
15 MV 14 0.095 0.094 0.092 0.090 0.088 0.086 0.083 0.080 0.077 0.058
15 0.099 0.097 0.095 0.093 0.091 0.088 0.086 0.083 0.079 0.060
16 0.102 0.100 0.098 0.096 0.093 0.091 0.088 0.085 0.081 0.061
17 0.105 0.103 0.101 0.099 0.096 0.093 0.091 0.087 0.084 0.063
18 0.109 0.106 0.103 0.101 0.098 0.095 0.093 0.089 0.085 0.064
19 0.111 0.108 0.105 0.102 0.100 0.097 0.094 0.090 0.086 0.065
20 0.115 0.110 0.107 0.105 0.102 0.099 0.096 0.092 0.088 0.067

Table 3. Normalized mean fetal dose, as a percentage of prescribed dose (PD), for a typical two-field tangential beam 15 ˚
wedged plan for both 6 MV and 15 MV photon energies. Data are presented for various gestational ages (GA) and fetal depth
within the maternal abdomen
Photon energy GA (weeks) Fetal depth (cm)

4 5 6 7 8 9 10 11 12 16

8 0.325 0.323 0.321 0.319 0.315 0.308 0.301 0.292 0.280 0.218
9 0.331 0.330 0.328 0.326 0.322 0.315 0.307 0.298 0.287 0.221
10 0.342 0.341 0.338 0.336 0.332 0.325 0.317 0.308 0.296 0.227
11 0.350 0.349 0.346 0.344 0.340 0.333 0.325 0.316 0.304 0.232
12 0.365 0.364 0.361 0.358 0.354 0.347 0.339 0.329 0.317 0.241
13 0.370 0.369 0.367 0.364 0.359 0.353 0.344 0.334 0.322 0.245
6 MV
14 0.378 0.377 0.374 0.371 0.366 0.360 0.351 0.340 0.328 0.250
15 0.396 0.395 0.393 0.389 0.384 0.377 0.368 0.357 0.344 0.263
16 0.412 0.411 0.409 0.405 0.399 0.392 0.382 0.371 0.357 0.274
17 0.433 0.432 0.430 0.425 0.419 0.411 0.401 0.389 0.374 0.289
18 0.450 0.448 0.446 0.441 0.434 0.426 0.415 0.402 0.387 0.300
19 0.460 0.459 0.456 0.451 0.444 0.435 0.424 0.411 0.395 0.308
20 0.476 0.474 0.471 0.466 0.459 0.449 0.438 0.423 0.406 0.319
8 0.205 0.203 0.198 0.193 0.188 0.181 0.175 0.167 0.157 0.120
9 0.209 0.207 0.203 0.198 0.192 0.186 0.180 0.172 0.162 0.122
10 0.216 0.214 0.209 0.204 0.198 0.192 0.186 0.179 0.169 0.125
11 0.221 0.219 0.215 0.209 0.204 0.198 0.191 0.184 0.174 0.128
12 0.230 0.228 0.224 0.218 0.213 0.207 0.200 0.193 0.183 0.134
13 0.234 0.232 0.228 0.222 0.216 0.210 0.204 0.196 0.186 0.137
15 MV 14 0.239 0.236 0.232 0.227 0.221 0.215 0.208 0.200 0.190 0.140
15 0.251 0.248 0.244 0.238 0.232 0.226 0.219 0.211 0.200 0.147
16 0.261 0.258 0.254 0.248 0.242 0.235 0.228 0.219 0.208 0.153
17 0.275 0.271 0.266 0.261 0.254 0.247 0.239 0.230 0.219 0.162
18 0.286 0.281 0.276 0.270 0.264 0.256 0.248 0.238 0.227 0.169
19 0.293 0.288 0.283 0.277 0.270 0.262 0.254 0.243 0.231 0.173
20 0.304 0.297 0.292 0.286 0.279 0.271 0.262 0.251 0.238 0.180

822 The British Journal of Radiology, October 2006

Estimation of fetal dose from radiotherapy of the breast

Table 4. Normalized mean fetal dose, as a percentage of prescribed dose (PD), for a typical two-field tangential beam 30 ˚
wedged plan for both 6 MV and 15 MV photon energies. Data are presented for various gestational ages (GA) and fetal depth
within the maternal abdomen
Photon energy GA (weeks) Fetal depth (cm)

4 5 6 7 8 9 10 11 12 16

8 0.430 0.431 0.430 0.428 0.424 0.417 0.407 0.395 0.381 0.302
9 0.438 0.440 0.439 0.437 0.432 0.425 0.416 0.404 0.390 0.307
10 0.451 0.453 0.452 0.449 0.445 0.438 0.428 0.416 0.402 0.314
11 0.462 0.464 0.463 0.460 0.455 0.448 0.441 0.427 0.413 0.320
12 0.481 0.482 0.482 0.479 0.474 0.467 0.457 0.445 0.430 0.332
13 0.488 0.490 0.489 0.486 0.481 0.474 0.464 0.452 0.437 0.337
6 MV
14 0.497 0.499 0.499 0.496 0.491 0.483 0.473 0.460 0.445 0.344
15 0.522 0.524 0.524 0.521 0.515 0.507 0.497 0.484 0.467 0.361
16 0.544 0.546 0.545 0.542 0.536 0.528 0.517 0.503 0.485 0.376
17 0.573 0.575 0.574 0.571 0.564 0.555 0.544 0.529 0.509 0.396
18 0.595 0.597 0.596 0.592 0.586 0.576 0.564 0.548 0.527 0.412
19 0.610 0.611 0.610 0.607 0.600 0.590 0.575 0.560 0.539 0.422
20 0.632 0.633 0.632 0.628 0.621 0.611 0.597 0.579 0.556 0.437
8 0.295 0.292 0.286 0.281 0.275 0.267 0.259 0.249 0.238 0.184
9 0.301 0.298 0.293 0.287 0.281 0.273 0.265 0.255 0.244 0.188
10 0.309 0.307 0.302 0.297 0.290 0.283 0.274 0.265 0.253 0.194
11 0.316 0.314 0.309 0.304 0.298 0.290 0.281 0.272 0.261 0.199
12 0.329 0.327 0.322 0.317 0.311 0.302 0.293 0.284 0.272 0.207
13 0.334 0.331 0.327 0.322 0.315 0.307 0.298 0.288 0.276 0.211
15 MV 14 0.340 0.337 0.333 0.328 0.321 0.313 0.305 0.293 0.281 0.215
15 0.356 0.353 0.349 0.343 0.336 0.327 0.318 0.307 0.294 0.225
16 0.370 0.366 0.362 0.356 0.348 0.339 0.329 0.318 0.304 0.233
17 0.388 0.384 0.379 0.372 0.364 0.354 0.344 0.331 0.317 0.244
18 0.403 0.397 0.391 0.384 0.375 0.366 0.354 0.342 0.326 0.252
19 0.412 0.406 0.400 0.392 0.383 0.373 0.361 0.348 0.332 0.257
20 0.426 0.419 0.412 0.403 0.394 0.383 0.371 0.357 0.341 0.265

Table 5. Normalized mean fetal dose, as a percentage of prescribed dose (PD), for a typical two-field tangential beam 45 ˚
wedged plan for both 6 MV and 15 MV photon energies. Data are presented for various gestational ages (GA) and fetal depth
within the maternal abdomen
Photon energy GA (weeks) Fetal depth (cm)

4 5 6 7 8 9 10 11 12 16

8 0.334 0.330 0.324 0.318 0.312 0.304 0.297 0.287 0.273 0.208
9 0.341 0.337 0.331 0.324 0.318 0.311 0.303 0.293 0.279 0.211
10 0.350 0.347 0.341 0.334 0.328 0.320 0.312 0.302 0.288 0.217
11 0.359 0.355 0.349 0.343 0.336 0.328 0.320 0.310 0.296 0.222
12 0.373 0.370 0.364 0.357 0.350 0.342 0.333 0.322 0.308 0.231
13 0.379 0.375 0.370 0.363 0.355 0.347 0.338 0.327 0.313 0.236
6 MV
14 0.386 0.382 0.377 0.370 0.362 0.354 0.344 0.333 0.319 0.241
15 0.405 0.401 0.395 0.388 0.380 0.371 0.361 0.349 0.335 0.254
16 0.421 0.417 0.411 0.403 0.395 0.385 0.375 0.363 0.348 0.266
17 0.443 0.438 0.432 0.424 0.414 0.405 0.394 0.381 0.366 0.281
18 0.459 0.454 0.447 0.439 0.429 0.419 0.408 0.395 0.378 0.293
19 0.470 0.465 0.458 0.449 0.439 0.429 0.417 0.403 0.386 0.301
20 0.486 0.480 0.473 0.464 0.454 0.443 0.431 0.416 0.399 0.313
8 0.299 0.294 0.287 0.279 0.271 0.262 0.254 0.244 0.233 0.183
9 0.304 0.299 0.292 0.284 0.276 0.267 0.259 0.250 0.239 0.187
10 0.312 0.308 0.301 0.293 0.284 0.275 0.267 0.258 0.248 0.193
11 0.319 0.315 0.308 0.299 0.291 0.282 0.274 0.265 0.255 0.198
12 0.331 0.326 0.320 0.311 0.302 0.293 0.285 0.276 0.266 0.206
13 0.335 0.331 0.324 0.315 0.306 0.298 0.289 0.280 0.269 0.209
15 MV 14 0.341 0.336 0.330 0.321 0.312 0.303 0.295 0.286 0.274 0.213
15 0.356 0.351 0.344 0.335 0.326 0.317 0.308 0.299 0.287 0.224
16 0.370 0.364 0.356 0.347 0.338 0.329 0.320 0.309 0.297 0.232
17 0.387 0.380 0.372 0.363 0.353 0.344 0.334 0.323 0.310 0.244
18 0.401 0.392 0.384 0.375 0.365 0.355 0.345 0.333 0.320 0.252
19 0.410 0.400 0.392 0.383 0.373 0.362 0.352 0.339 0.326 0.258
20 0.424 0.413 0.404 0.394 0.384 0.373 0.362 0.349 0.334 0.267

The British Journal of Radiology, October 2006 823

 B Bradley, A Fleck and E K Osei

Table 6. Normalized mean fetal dose, as a percentage of prescribed dose (PD), for a typical two-field tangential beam 60 ˚
wedged plan for both 6 MV and 15 MV photon energies. Data are presented for various gestational ages (GA) and fetal depth
within the maternal abdomen
Photon energy GA (weeks) Fetal depth (cm)

4 5 6 7 8 9 10 11 12 16

8 0.381 0.377 0.370 0.362 0.355 0.348 0.341 0.331 0.317 0.245
9 0.389 0.384 0.378 0.370 0.363 0.355 0.348 0.338 0.324 0.251
10 0.400 0.395 0.389 0.381 0.374 0.366 0.358 0.348 0.334 0.259
11 0.409 0.404 0.398 0.390 0.383 0.375 0.366 0.356 0.343 0.266
12 0.425 0.421 0.415 0.407 0.399 0.391 0.382 0.371 0.358 0.278
13 0.431 0.427 0.421 0.413 0.405 0.397 0.388 0.377 0.363 0.283
6 MV
14 0.439 0.435 0.429 0.421 0.413 0.405 0.395 0.384 0.371 0.289
15 0.460 0.456 0.450 0.443 0.434 0.425 0.415 0.404 0.389 0.305
16 0.478 0.474 0.468 0.460 0.452 0.443 0.433 0.420 0.405 0.319
17 0.502 0.497 0.491 0.484 0.475 0.466 0.455 0.442 0.426 0.336
18 0.520 0.516 0.510 0.502 0.493 0.484 0.473 0.459 0.442 0.351
19 0.532 0.527 0.522 0.514 0.505 0.495 0.484 0.470 0.452 0.360
20 0.550 0.545 0.539 0.532 0.523 0.513 0.501 0.486 0.467 0.373
8 0.339 0.333 0.326 0.318 0.311 0.303 0.294 0.285 0.271 0.213
9 0.345 0.339 0.331 0.324 0.316 0.308 0.300 0.291 0.277 0.218
10 0.353 0.348 0.340 0.332 0.325 0.317 0.309 0.300 0.287 0.225
11 0.361 0.355 0.347 0.339 0.332 0.324 0.316 0.308 0.294 0.231
12 0.373 0.368 0.360 0.352 0.344 0.336 0.329 0.320 0.307 0.241
13 0.378 0.372 0.365 0.357 0.349 0.341 0.333 0.324 0.311 0.245
15 MV 14 0.384 0.378 0.371 0.363 0.355 0.349 0.339 0.329 0.317 0.250
15 0.401 0.400 0.388 0.379 0.371 0.363 0.354 0.344 0.331 0.263
16 0.416 0.409 0.401 0.393 0.384 0.376 0.367 0.356 0.343 0.273
17 0.435 0.427 0.420 0.411 0.402 0.393 0.383 0.372 0.358 0.286
18 0.450 0.436 0.434 0.425 0.416 0.406 0.396 0.383 0.369 0.296
19 0.460 0.451 0.443 0.434 0.425 0.415 0.404 0.391 0.376 0.303
20 0.475 0.465 0.457 0.448 0.438 0.427 0.415 0.402 0.386 0.313

all gestational ages, the fetal dose decreases with increase
in depth. The data show that fetal dose estimations based
on using a constant fetal depth of 9 cm would either
overestimate or underestimate the dose, depending on
the depth of interest. For example, dose could be
underestimated up to about 10% at a depth of 4 cm
and overestimated as high as about 30% at a depth of
16 cm (using 6 MV, sMLC plan) if a depth of 9 cm is
used instead of the actual fetal depth.
Fetal dose is also dependent on the gestational age of
the fetus. For all fetal depths, as gestational age increases,
fetal dose also increases. The data show that if fetal dose
were to be estimated as the dose to the uterus, the fetal
dose would again be either overestimated or under-
estimated at any given depth depending on the gesta-
tional age of interest. A uterine mass of 66.3 g and a
length, width and AP thickness of 7.5 cm, 3.5 cm and
2.5 cm, respectively, correspond most closely to the
dimensions of a fetus at 15 weeks post-conception.
Therefore, for dose estimates below 15 weeks post-
conception (i.e. uterus dose), the fetal dose would be
overestimated by up to about 12%, and for estimates
above 15 weeks, the fetal dose would be underestimated
by up to about 23% (using 6 MV with the sMLC plan).
The degree to which the uterus dose differs from fetal
dose is dependent upon the difference in gestational age,
as well as other parameters such as depth and orienta-
tion of the fetus within the maternal abdomen.
According to the literature [1, 6, 7], the presence of a
wedge in the beam could increase the fetal dose by a
factor of about 2–4 and our data give a similar result
ranging from a factor of about 2–5 for all the wedges. The
data show that the multi-segment sMLC technique
results in consistently lower fetal doses compared with
all the wedged techniques and could significantly reduce
fetal doses. This reduction in the fetal dose with the use
of a multi-segment (sMLC) plan as against that of a
wedged plan may be as a result of a combination of a
decrease in scatter and improved collimation when MLC
leaves are used in conjunction with the jaws to shape the
treatment field. The MLC leaves were fully retracted in
all plans employing a wedge. The data indicate that fetal
dose is dependent on both depth within the maternal
abdomen and gestational age, and hence these factors
should always be considered when estimating fetal dose.
The measured fetal doses are in good agreement with
other data obtained from the literature [1, 2, 4, 5, 9].
Comparison was made between these data and others
previously published, although most of the published
data refer to fetal doses at one specific gestational age
and depth. Rincon et al [9] reported a fetal dose
(assuming dose to the uterus) of about 40 mGy at a
depth of 15 cm using a 6 MV photon beam and a
prescription of 50 Gy at the isocentre. Assuming a fetal
size of 15 weeks post-conception to be that of the uterus
(as used in the Rincon study), 6 MV photon beam and
sMLC plan, we estimated a fetal dose of about 38 mGy at
a depth of 16 cm. The difference in the estimated doses
may be as a result of differences in fetal depth and
irradiation geometry used. For a fetus at 2–6 weeks

824 The British Journal of Radiology, October 2006

Estimation of fetal dose from radiotherapy of the breast
Figure 4. Normalized mean fetal dose, as a percentage of prescribed dose (PD), for a typical two-field tangential beam plan
using multi-segment static MLC for 6 MV photon beam as a function of fetal depth within the maternal abdomen. Data are
presented for various gestational ages (GA).
gestational age at a depth of 9 cm, Antypas et al [5]
reported relative fetal doses in the range of 0.079% to
0.085% of the prescribed dose. Using the same fetal
depth and 6 MV photon beam and sMLC plan, we
estimated a fetal dose of about 0.095% of the prescribed
dose for a gestational age of 8 weeks. The difference in
the estimated doses may be due to differences in
gestational age and irradiation geometry employed.
Conclusion
Normalized data for converting the prescribed dose at
the isocentre from medical exposure of a pregnant
woman undergoing radiation treatment of the breast, to
absorbed dose to the fetus, have been presented. These
data may be useful for estimating absorbed dose to the
fetus from breast cancer radiotherapy of the mother and
take into account the dependence of fetal age on
gestational age (fetal size), fetal depth, wedge dose and
photon beam energy. The results show that fetal dose is
dependent on both gestational age and depth within the
maternal abdomen, and hence these factors should be
The British Journal of Radiology, October 2006
taken into account when estimating dose. The data
collected and presented here were taken without any
modification to the treatment setup to minimize the dose
to the fetus. They would be useful for estimating dose to
the fetus when a patient realises she is or was pregnant
during or after the course of treatment. Again, if there is
a clinical indication for a pregnant patient to complete
treatment when pregnancy is known, the data could be
useful for estimating the fetal dose and the treatment
setup can then be modified and optimized, and also
external shielding over the abdominal region of the
pregnant patient could be employed during treatment to
significantly reduce the fetal dose. The sources of
radiation dose to the fetus are mainly due to the
radiation that is scattered within the patient and also
the scattered and leakage radiation from the head of the
machine and the collimator assembly. Whereas little can
be done to reduce the internal scatter within the patient
contributing dose to the fetus, the other sources of
radiation exposure contributing dose to the fetus could
be reduced significantly by the application of external
shielding over the abdominal region of the pregnant
patient.
825
 B Bradley, A Fleck and E K Osei

Figure 5. Normalized mean fetal dose, as a percentage of prescribed dose (PD), for a typical two-field tangential beam plan as a
function of fetal depth within the maternal abdomen. Data are presented for various treatment plans (sMLC, 15 ˚, 30 ˚, 45 ˚ and
60 ˚ wedged plans) and for 6 MV and 15 MV photon energies. The gestational age is 15 weeks post-conception.

References
1. Stovall M, Blackwell CR, Cundiff J, Novack DH, et al. Fetal
dose from radiotherapy with photon beams: Report of
AAPM Radiation Therapy Committee Task Group No. 36.
Med Phys 1995;22:63–82.
2. Islam MK, Saeedi F, Al-Rajhi N. A simplified shielding
approach for limiting fetal dose during radiation therapy of
pregnant patients. Int J Radiat Oncol Biol Phys 2001;49:1469–73.
3. ICRP-60. International Commission on Radiological
Protection (ICRP). 1990 Recommendations of the ICRP.
ICRP publication 60, Annals of the ICRP 21, No 1-3,
Pergamon Press, Oxford, 1991.
4. Mazonakis M, Damilakis J, Varveris H, Theoharopoulos N,
Gourtsoyiannis N. A method of estimating fetal dose
during brain radiation therapy. Int J Radiat Oncol Biol
Phys 1999;44:455–9.
5. Antypas C, Sandilos P, Kouvaris J, Balafouta E, et al. Fetal
dose evaluation during breast cancer radiotherapy. Int J
Radiat Oncol Biol Phys 1998;40:995–9.
6. Fraass B, Van de Geijn J. Peripheral dose from megavolt
beams. Med Phys 1983;10:809–18.
7. Diallo I, Lamon A, Shamsaldin A, Grimaud E, de Vathaire
F, Chavaudra J. Estimation of the radiation dose delivered
to any point outside the target volume per patient treated
with external beam radiotherapy. Radiother Oncol 1996;38:
269–71.
8. Van der Giessen P. A simple and generally applicable
method to estimate the peripheral dose in radiation
teletherapy with high energy x-rays or gamma radiation.
Int J Radiat Oncol Biol Phys 1996;35:1059–68.
9. Rincon CM, Jerez Sainz I, Farree IM, Espana Lopez ML,
Franco PL, et al. Evaluation of the peripheral dose to uterus
in breast carcinoma radiotherapy. Radiat Prot Dosim
2002;101:469–71.
10. Nuyttens JJ, Prado KL, Jenrette JM, Williams TE. Fetal
dose during radiotherapy: clinical implementation and
review of the literature. Cancer (Radiothérapie) 2002;6:
352–7.
11. Prado KL, Nelson SJ, Nuyttens JJ, Williams TE, Vanek KN.
Clinical implementation of the AAPM Task Group 36
recommendations on fetal dose from radiotherapy with
photon beams: a head and neck irradiation case report. J
Appl Clin Med Phys 2000;1:1–7.
12. Sneed PK, Albright NW, Wara WM, Prados MD, Wilson
CB. Fetal dose estimates for radiotherapy of brain tumors
during pregnancy. Int J Radiat Oncol Biol Phys
1995;32:823–30.

826 The British Journal of Radiology, October 2006

Estimation of fetal dose from radiotherapy of the breast
13. Antolak JA, Strom EA. Fetal dose estimates for electron-
beam treatment to the chest wall of a pregnant patient. Med
Phys 1998;25:2388–91.
14. Ngu SL, Duval P, Collins C. Fetal radiation dose in
radiotherapy for breast cancer. Aust Radiol 1992;36:321–2.
15. Osei EK, Faulkner K. Fetal position and size data for dose
estimation. Br J Radiol 1999;72:363–70.
16. Doublet PM, Benson CD, Nadel AS. Improved birth weight
table for neonates developed from gestations dated by early
ultrasonography. J Ultrasound Med 1997;16:241.
The British Journal of Radiology, October 2006
17. Hadlock FP, Shah YP, Kanon DJ, Lindsey JV. Fetal crown-
rump length: reevaluation of relation to menstrual age (5–18
weeks) with high-resolution real-time US. Radiology
1992;182:501–5.
18. Chitty LS, Altman DG, Henderson A, Campbell S. Charts of
fetal size: 3. Abdominal measurements. Br J Obstet
Gynaecol 1994;101:125–31.
19. Hadlock FP, Deter RL, et al. Estimating fetal age: computer-
assisted analysis of multiple fetal growth parameters.
Radiology 1984;152:497–501.
827
 The British Journal of Radiology, 79 (2006), 828–836

Dosimetric and treatment planning considerations for

radiotherapy of the chest wall
M M ASPRADAKIS, PhD, MIPEM, H M McCALLUM, PhD, MIPEM and N WILSON, DCR(T)

Regional Medical Physics Department, Newcastle General Hospital, Westgate Road, Newcastle
upon Tyne NE4 6BE, UK

ABSTRACT. Radiotherapy treatment planning calculations of the chest wall are

complex due to missing tissue, the thin chest wall and the presence of lung. The
accuracy of calculated dose is dependent on the type of algorithm employed. This work
evaluates a collapsed cone (CC) and a pencil beam (PB) convolution model for
radiotherapy planning of the chest wall. Various irradiation geometries simulating the
chest wall have been examined and calculations were compared with measurements
with an ionization chamber in epoxy resin water substitute and in low-density lung
substitute blocks. A retrospective treatment planning study comprising 6 patients was
carried out to evaluate the differences in the dose distributions and monitor units
predicted by the two algorithms. The calculated dose in unit density medium was
within ¡1% for the CC model and up to ¡2% for PB. In low density medium and under
full scatter conditions, CC overestimated the dose by 1% whereas PB overestimated the Received 21 October 2005
dose by 9%. In the tangential irradiation geometry with water and lung media, the PB Revised 14 February 2006
overestimated dose to the isocentre by up to 10%, whereas the dose from CC was Accepted 20 February 2006
within 3%. From the treatment planning study calculated monitor units (MU) and doses
DOI: 10.1259/bjr/26575438
were consistent with the experimental findings. The CC model is more accurate for
radiotherapy treatment planning of the chest wall and especially when there is ’ 2006 The British Institute of
significant involvement of lung tissue. Radiology

Breast cancer affects up to 1 in 10 women in the
Western World and is the most common female cancer in
the UK in terms of incidence and mortality [1]. Adjuvant
radiotherapy to the chest wall has an established role in
reducing the risk of locoregional recurrence following
mastectomy. A meta-analysis of all randomized con-
trolled trials published in 1987 showed that post
mastectomy radiotherapy was associated with a 66%
reduction in the risk of locoregional recurrence [2]. Post
mastectomy radiotherapy may be indicated either due to
inadequate surgical resection of tumour margins due to
lymphovascular involvement or due to a tumour size
greater than 5 cm.
Radiotherapy treatment planning calculations for the
chest wall are complex due to missing tissue and the
presence of lung tissue within the treatment field.
Accurate calculation of the dose distribution is important
as doses to neighbouring organs at risk, such as heart,
lungs and contralateral breast, needs to be minimized.
Treatment planning systems (TPS) would, ideally,
provide both accurate relative dose distributions and
monitor unit settings for such irradiation geometries. The
Helax-TMS (version 6.1a) TPS (Nucletron B.V., The
Netherlands) has two dose calculation algorithms avail-
able for external photon beam planning, the pencil beam
(PB) and the collapsed cone (CC) algorithms. Dose
Current address for M M Aspradakis: Klinik für Radio-Onkologie,
Universitäts Spital Zürich, Rämistrasse 100, Zürich 8091,
Switzerland.
Address correspondence to: Dr Hazel M McCallum.
calculations for patients on Helax-TMS use density
information from CT based on the conversion of
Hounsfield numbers (HN) to material density, thus
accounting for the attenuation properties of different
tissues, as well as using information on the material
composition. The two algorithms differ primarily in how
they model radiation transport and calculate dose in
heterogeneous media, with the CC algorithm better
approximating the dose directly to the medium [3–5].
Their characteristics and dosimetric accuracy in simple
and complex phantom irradiation geometries have been
well documented [6–9]. It has been shown that there is
little or no clinically significant difference in the
calculation of dose by the two algorithms when no large
non-unit density heterogeneities are present and under
full scatter conditions, such as in the irradiation of the
pelvis [9]. This excludes cases of pelvic irradiation with
high density metal prosthesis [10].
In cases of treatment planning in the thorax region (for
oesophagus and lung), differences between CC and PB in
the calculation of dose (predicted monitor units (MU)),
planning target volume (PTV) coverage and minimum
dose to the PTV has been reported [11]. Such differences
can potentially be clinically significant and affect tumour
control. At tangential beam irradiations for a range of
beam energies (4 MV, 6 MV, 15 MV) on a homogeneous
medium, experimental verification of the Helax-TMS
algorithms using a 15 cm615 cm square field had
shown the CC algorithm to be accurate to within ¡2%
of the measured dose and to model closely the reduction
in dose due to the missing tissue [9]. From PB the dose

828 The British Journal of Radiology, October 2006

Photon dose calculations in chest wall radiotherapy
close to phantom boundaries was overestimated by 4–5%
at 4 MV, but less so at higher energies.
The difficulties encountered with treatment planning
of the chest wall in particular have not been reported, nor
has the dosimetric accuracy of treatment planning
calculations for such extreme irradiation geometries with
the presence of lung and/or the plan normalization point
within thin tissue or within lung been investigated. It
was the purpose of this work to investigate the
dosimetric accuracy of the Helax-TMS convolution/
superposition algorithms for treatment planning of the
chest wall at the Northern Centre for Cancer Treatment
(NCCT), in Newcastle upon Tyne, UK. The dosimetric
differences between the Helax-TMS CC and PB algo-
rithms were examined for clinically relevant scenarios of
chest wall irradiations. A planning study comprising six
patients was also undertaken to support changes in
clinical practice.
Materials and methods
Dosimetric verification of Helax-TMS algorithms for
chest wall irradiation
Experimental set-up
All measurements were carried out with a 6 MV
photon beam on a Siemens Primus accelerator (Siemens,
Figure 1. Experimental geometries employed for measurements and/or calculations in this work. All measurements were carried
out with a 6 MV photon beam, gantry angle of 270 ˚ and collimator angle of 90 ˚.
The British Journal of Radiology, October 2006
Erlangen, Germany). This is the most frequently used
beam energy at NCCT for treatment of the breast and
chest wall. Three experimental geometries relevant to
chest wall irradiation were examined. The WT1-tissue
and lung substitute blocks (Epoxy resin water substitute
blocks with elemental composition: H(8.09); C(67.22);
N(2.4); O(19.84); Ca(2.32); Cl(0.13) and relative electron
density of 1.02 (mass density 1.04 g cm23) [12]. Lung
equivalent blocks of elemental composition: H(8.38);
C(60.50); N(1.68); O(17.28); Cl(0.15); Si(0.84); Mg(11.17)
and relative electron density of 0.25 (mass density of
0.26 g cm23) [13, 14]) used in the measurements were
manufactured by ScanPlus (St Bartholomews, London).
The measurement system used was an NE2571 0.6 cm3
cylindrical ionization chamber and its associated electro-
meter with a valid calibration factor (traceable to the
primary standard at the National Physical Laboratory) to
convert the ionization reading to dose in water.
The irradiation geometries used are shown in Figure 1.
The ionization chamber was positioned in all cases at the
isocentre. In Figure 1a, the lateral extent of the phantom
normal to the beam central axis varied from 1 cm to 3 cm
and a measurement was also made at full scatter
conditions with a lateral thickness of 10 cm. In
Figure 1b the phantom comprised lung and tissue
equivalent blocks. Measurements were carried out with
the chamber in the tissue equivalent block at 1 cm,
1.5 cm and 2 cm from the edge of the lung heterogeneity
and the phantom boundary varied from 1 cm to 3 cm
829

and at full scatter conditions at 10 cm in the lateral
direction. Figure 1c was analogous to Figure 1b, but the
calculation point was in lung equivalent material.
The position of the normalization/prescription point
with respect to the phantom boundary or the lung was
chosen to be similar to that often selected in patient
treatment plans. Simulations on Helax-TMS were carried
out for the slice separations and dose matrix resolutions
of 0.25 cm.
Low density dosemeter correction factor
For the measurement of dose in a low density medium
using the NE2571 ionization chamber and its absorbed
2571
dose to water calibration factor (Nwater,6MV ), the chamber
reading is converted to dose in an infinitesimally small
volume of water within the low density medium (which
is what is also modelled by the treatment planning
system), without accounting for the high replacement
perturbation effect caused by the chamber and its high
density graphite wall in the low density medium. The
NE2571 chamber perturbs fluence differently in water
and in low density lung and this difference should be
accounted for when converting the ionization reading to
dose.
The application of a perturbation correction to the
measurement in lung in this work was based on the work
of Krieger and Sauer [15]. The replacement perturbation
factor prepl(A) at field size A was defined as the ratio of
dosemeter reading with zero chamber wall thickness to
the reading at normal wall thickness. Krieger and Sauer
[15] measured doses in Styrofoam (with mass density of
0.035 g cm23) for variable chamber wall thicknesses
made of plastic (PMMA) and at different field sizes,
using a PTW 31003 ionization chamber which has a
PMMA wall and wall thickness of 0.0655 cm (or
0.078 g cm22) and mass density of 1.19 g cm23. From
an exponential curve fit to this data, dose values were
calculated at zero wall thickness. For the purpose of our
work, we assumed that the atomic compositions of
graphite and PMMA are similar, that there is a linear
relationship for the perturbation correction with medium
density and that differences in the geometric character-
istics of the chambers have a small influence on the
perturbation in low density medium. The dose values in
Styrofoam measured using a chamber with a PMMA
wall were converted to dose values in Styrofoam
measured with a chamber with a graphite wall. The
replacement perturbation factor for the NE2571 in lung
was further derived from linear interpolation between
replacement perturbation factors for this chamber in
Styrofoam and in water (prepl(A) in water is unity). Using
this replacement perturbation correction, ionization
2571
readings (Mlung ðAÞ) measured for a field size A with
the NE2571 in a homogeneous lung medium, were
converted to dose to water in an infinitesimally small
volume of water within the low density lung using:
D2571 2571 2571
homogeneous-lung ðAÞ~Mlung,repl ðAÞ|Nwater,6MV
ð1Þ
2571
|Mhomogeneous-lung ðAÞ
In the case of a measurement in a heterogeneous
medium comprising lung and water where the chamber
830
M M Aspradakis, H M McCallum and N Wilson
is placed within lung, a replacement perturbation
correction was applied to the dose from the part of the
field irradiating the lung medium (D2571 homogeneous lung ). Thus,
if D2571
meas,heterogeneous is the measured dose in lung in the
heterogeneous medium and p2571 lung,repl (Aeq ) is the replacement
perturbation correction for the equivalent field size Aeq
irradiating the low density medium, the dose to water in an
infinitesimally small volume of water within the lung in the
heterogeneous medium is calculated from [15]:
D2571 2571
heterogeneous (A)~Dmeas,heterogeneous (A)z
! ð2Þ
1
1{ |D2571
homogeneous-lung (A eq )
p2571
lung,repl (Aeq )
Thus for the derivation of the perturbation correction
in the heterogeneous geometry, the dose in a homo-
geneous lung medium under full scatter conditions was
also measured.
Treatment planning study
Standard breast planning technique at NCCT
All patients at NCCT undergoing external beam
radiotherapy treatment planning for the breast have a
three-dimensional (3D) planning study based on images
from helical CT scanning using a 5 mm slice reconstruc-
tion. Patients are immobilized in a reproducible position,
lying supine on a PosiboardTM-2 breast backrest (manu-
factured by Sinmed BV) and with couch position
registration. Both arms are positioned above the patient’s
head, facilitating the tangential field arrangements of the
treatment plan. Ball bearings are placed at stable points
of the patient to define a CT reference point to aid
positional verification during treatment delivery. A
planning CT scan incorporating these markers is
acquired with a 0.5 cm slice separation and this dataset
is transferred to Helax-TMS. The patient external contour
(skin) and organs at risk are outlined, namely involved
lungs and heart, if appropriate.
The isocentric breast planning technique at NCCT
treats the breast with two tangential fields and, in some
cases, a single supraclavicular field. Non-divergent beam
edges are produced at the match plane with the
supraclavicular field and at the posterior edge of the
tangential fields [16–18]. The plan normalization/pre-
scription point is always set at the isocentre. The
generation of a dose distribution within Helax-TMS is a
two stage process. A dose plan is generated using fast
interactive optimization of open and wedged beams.
Interactive dose calculations (in the Beam Modelling
module on Helax-TMS) are possible only with the PB
algorithm. A final dose calculation is generated once the
plan is placed for non-interactive computation (in the
Evaluation module). Non-interactive dose computations
are also possible with the CC algorithm, but this
algorithm cannot be employed during manual optimiza-
tion. Although a full 3D image set is used in the
computation of dose distributions, current clinical
practice at NCCT for breast planning dictates that
The British Journal of Radiology, October 2006
Photon dose calculations in chest wall radiotherapy
optimization and evaluation of treatment plans are
carried out only on the slice containing the normalization
point.
Definition of the planning target volume (PTV)
It is not standard practice at NCCT to define a
planning target volume (PTV) for planning of the breast
or chest wall. For the purpose of this study, however, a
PTV was defined to enable the comparison of dose–
volume histograms (DVHs) from calculations employing
different calculation algorithms. The PTV was defined by
the geometrical beam limits of the superior and inferior
field borders and the pleura, and drawn 5 mm inside the
external outline. With Helax-TMS, a structure cannot be
defined once a beam geometry has been added to the
study dataset. In order to define a PTV retrospectively
and for the purpose of this work, each patient study set
(images and structures) was exported in DICOM RT
format to the Exomio (Version 2.0) virtual simulation
software. PTVs were added on Exomio and the patient
structure set was re-imported into Helax-TMS.
Dose calculations on Helax-TMS
All six patient studies were planned with a 6 MV beam
using the standard breast planning technique. The PTV
for each patient was drawn based on the criteria
described previously. Dose calculations were carried
out with both algorithms, both with and without
inhomogeneity correction. The option of no-inhomo-
geneity correction simply assigns all tissue in the patient
geometry to unit mass density. This was carried out in
order to exclude the influence of the lung tissue on the
dose calculations. The dose calculation grid was 0.5 cm.
Evaluation of plans
For the evaluation of treatment plans, isodose dis-
tributions, monitor units (to deliver 1 Gy at the normali-
zation point) and data from DVHs were compared. To
aid comparisons and relate the changes observed for the
individual chest wall geometry and to current clinical
practice, geometric measurements were taken from all
CT studies on the central planning slice containing the
normalization point. These measurements, as shown in
Figure 2 were: the distance from the normalization point
to the skin surface dskin-norm point, the thickness of breast,
i.e. the distance from the lung-tissue interface to the skin
surface dbreast, and the maximum thickness of lung tissue
involved in this central slice dmax lung. For three patients,
the normalization point was in breast tissue (patients 1,
2, 3) and for the others, in the lung (patients 4, 5, 6). The
differences in monitor units from open and wedged
beams for the PB and CC algorithms were analysed and
the minimum, maximum and mean percentage doses in
the PTV and lung were recorded from the DVH data.
Results
Experimental verification
The CC algorithm modelled the dose at the isocentre in
a homogeneous unit density medium under tangential
irradiation (Figure 1a) with varying lateral thickness of
The British Journal of Radiology, October 2006
Figure 2. Distances measured on the central planning slice
containing the isocentre. dskin-norm point is the distance from
the normalization point (isocentre) to the skin of the patient,
dbreast is the thickness on the chest wall at the level of the
isocentre and dmax lungthe maximum lung involvement in the
field.
tissue more accurately than PB. The uncertainties in the
measured values did not exceed 0.5%. Deviations from
measurement for CC did not exceed ¡1%, whereas for
PB and for lateral thickness equal to or less than 2.5 cm,
these were greater than +1% (maximum of +2.3%). These
results were consistent with previous findings and
generally the differences between the two algorithms
for this beam energy are not large [9].
Figure 3 shows the variation of prepl with field size for
the NE2571 in lung at 6 MV for collimator settings
greater than 5 cm65 cm (the use of this chamber for
dosimetry in smaller field sizes is not appropriate due to
its size in relation to the size of the beam). The
uncertainty in the derivation of prepl, as quoted by
Krieger and Sauer, is 2.3% at 5 cm65 cm and is reduced
to 1.5% at 15 cm615 cm [15]. In the case of a
homogeneous lung medium irradiated with a
10 cm610 cm field under full scatter conditions, it was
found that the CC model predicted the dose to within
0.9% of the measurement, whereas the PB overestimated
dose by +8.8%. The combined uncertainty in the dose
measurements did not exceed 2.1%.
In Figure 4a–c, it can be seen how the measured dose
varies with distance from lung tissue and phantom
boundary, when the chamber is in unit density medium
(experiment shown in Figure 1b). As the distance from
the phantom boundary decreases, the dose decreases
and it decreases even more (by 1.6%) once its proximity
to the low density inhomogeneity is equal to or less than
1.5 cm. In this case, the overall deviation from measure-
ment for CC is ¡0.6%, where errors by PB vary from
+2.5% to +7%.
For tangential irradiation with the normalization point
in lung, the combined uncertainties in the measurement
are expected to be of the order of 3% due to the
assumptions in the derivation of the perturbation
831

correction [15]. The dose predicted by the PB model
varied between +10.7% to +8.2%, depending on the
effective lateral thickness from the calculation point
(Figure 4d). The CC predicted values between +3.6% and
+2.3% for increasing effective thickness. This is in
agreement with other workers who, in heterogeneous
media and within low density and close to an interface,
have reported the CC model to be more accurate than PB
Figure 4. Dose in water with varying distance from phantom edge and lung. (a) 10 mm; (b) 15 mm (c) 20 mm. (d) Dose in lung
with varying distance from phantom edge. &, pencil beam; m, collapsed cone; , measurement.
832
M M Aspradakis, H M McCallum and N Wilson
Figure 3. Perturbation correction
factors for the NE2571 ionization
chamber (with its graphite cap) in
lung and styrofoam. The data from
Krieger and Sauer [15] for the PTW
31003 chamber with its PMMA cap in
Styrofoam are shown for comparison
(with permission). ¤, PTW31003
with PMMA wall in styrofoam
(Krieger & Sauer, with permission);
&, NE2571 with graphite wall in
N
styrofoam; , NE2571 with graphite
in wall in lung.
[3, 15, 19]. Here, the CC model predicted a higher dose
than the measured dose, but one has to consider the
uncertainty in the derivation of the perturbation
correction for the measurement in lung. In the case of
the heterogeneous medium with the point of measure-
ment in lung and under complex missing tissue irradia-
tion geometry, the CC predicts dose values close to
measurement and within 3% in the case of lateral
N
The British Journal of Radiology, October 2006
Photon dose calculations in chest wall radiotherapy

Figure 5. Typical dose distributions on the central planning slice generated by (a) the pencil beam (PB) and (b) the collapsed
cone (CC) dose algorithms. These data are for patient 5 of this study.

electronic equilibrium (effective lateral thickness greater
than the 15 mm).
Treatment planning study
Figure 5 shows isodoses for patient 5 of this study on
the CT slice containing the normalization point within
lung. The isodose distribution calculated with CC
follows the lung contour, predicting an overall lower
dose to lung tissue and a relatively higher dose in tissue.
This is because the energy released in the lung is
scattered laterally to the adjacent breast and lung tissue
and PB does not model this.
Table 1 summarizes the patients included in this study
and their measurements. The effective lateral distance
from the normalization point to the skin is defined as the
distance of tissue scaled with mass density. Table 2
summarizes the results from the two algorithms in terms
of absolute difference in MU per beam, the average
percentage difference in MU for the open fields only and
the percentage difference in minimum, maximum and
mean doses to the PTV and lung extracted from DVHs.
The analogous results for the calculations without an
inhomogeneity correction are shown in Table 3. Figure 6
illustrates how the percentage differences in mean doses
to PTV and lung calculated by PB and CC vary with
breast tissue thickness, maximum lung thickness and
distance from the normalization point to the skin.
As seen from Table 2, the CC model predicted higher
MU than the PB, the latter overestimating dose, and this
was consistent with our experimental findings. It can be
seen (Figure 6) that the greater the thickness of breast
tissue the smaller the differences between the two
algorithms. For patient 3 a thickness of 4.4 cm of breast
tissue, resulted in, on average, 3 MU difference per beam
between algorithms, a 4.1% difference to the dose at the
isocentre from open fields and a 0.4% difference in the
mean PTV dose. For patient 5 the thickness of breast
tissue was 1.0 cm and this resulted in 5 MU difference
between algorithms, thus 16.5% difference to the dose
contributions at the isocentre from the open beams and
11% difference at the mean dose to the PTV. For patients
with the normalization point placed in breast tissue
(patients 1, 2 and 3) the differences between the two
algorithms were the lowest.

Table 1. Details and geometric measurements for patients included in this study
Patient Chest wall/ Breast tissue Maximum lung Normalization Effective distance from normalization point to
isocentre depth depth (cm): point to skin (cm): skin (g cm23). For the normalization point in
location (cm): dbreast dmax lung dskin-norm point lung: dbreast + (dskin-norm point2dbreast)60.26

1 Left/breast 1.9 1.8 1.3 1.3

2 Left/breast 2.6 2.0 2.1 2.1
3 Left/breast 4.4 2.5 2.5 2.5
4 Right/lung 1.1 2.4 1.2 1.1
5 Right/lung 1.0 2.8 1.8 1.2
6 Right/lung 1.5 2.3 1.7 1.6

The British Journal of Radiology, October 2006 833

 M M Aspradakis, H M McCallum and N Wilson

Table 2. The difference in monitor units calculated between a pencil beam (PB) and collapsed cone (CC), their average
percentage differences for the open fields and the percentage difference in dose to the planning target volume (PTV) and lung,
with the inhomogeneity correction switched on
Patient Beam Difference in MU PB-CC % average difference % Difference in dose from DVH ((CC2PB)/PB)6100
modulation in MU for open fields
Medial Lateral ((CC2PB)/PB)6100 PTV Lung
field field Min. Max. Mean Min. Max. Mean

1 Open 22.8 22.9 +5.3 27.4 1.8 +1.4 0.0 22.8 22.2
Wedged 20.7 20.4
2 Open 23.5 23.3 +5.2 26.4 0.4 +1.4 16.7 26.2 0
Wedged – –
3 Open 22.4 23.0 +4.1 1.3 0.9 +0.4 23.1 28.6 +1.9
Wedged 20.4 –
4 Open 27.4 27.4 +16.5 14.7 10.4 +11.2 214.3 3.6 +10.7
Wedged 21.5 21.5
5 Open 29.1 28.9 +16.5 10.9 9.3 +11.0 222.2 6.1 +5.8
Wedged – –
6 Open 25.8 24.8 +9.6 2.6 1.5 +4.5 8.3 0.8 +1
Wedged 2 21.0
DVH, dose–volume histogram.

Figure 6 shows that with increasing lung involvement for the presence of lung in the dose calculation and when
in the plan, the mean dose to the PTV and lung increases. a lung heterogeneity is present, differences between
Patients 4 and 5 had maximum lung involvement of algorithms increase with increasing involvement of lung.
2.4 cm and 2.8 cm, and for these the average percentage In Table 2 it is also seen that the minimum dose
difference in MU was the highest, 16.5% and 16.6%, difference to the PTV between the algorithms was small
respectively. For patient 3, with similar lung involve- in some cases and in particular for patient 3. As pointed
ment (2.5 cm), an analogous result was expected. But for out previously, the optimization of these plans was
this patient the normalization point was within 4.4 cm of carried out based on the distribution calculated on the
breast tissue and the percentage difference in MU and central slice of the study and this is why the results from
the difference to the mean dose in the PTV between DVHs are not optimal in terms of the minimum dose to
algorithms was the lowest observed. This would indicate the PTV. In addition to this, these PTVs were drawn
that when there is sufficient breast tissue laterally, the close to the patient skin and close to the penumbral
influence of lung on the dose at the normalization point region of both fields that would contribute to small dose
is reduced. values.
With the inhomogeneity correction switched off, the
differences in MU between PB and CC were smaller
(Table 3). For patient 5, for example, the difference in Discussion and conclusions
MU was 2 (3%) when the lung was not accounted for, as
opposed to 9 (16.6%) when accounted for in the The dose at the normalization point in the case of
calculation. This confirms that it is important to account isocentric breast irradiation using a 6 MV beam, with no
lung involvement and in a region of lateral electronic
equilibrium, is modelled to within +3% by the PB
Table 3. The difference in monitor units calculated between algorithm and within ¡1% by CC.
pencil beam (PB) and collapsed cone (CC) and their average For breast cancer patients who have undergone
percentage differences for the open fields with the inho- mastectomy and are also to be treated with external
mogeneity correction switched off beam radiotherapy, the presence of lung tissue and often
Patient Beam Difference in MU % average difference the small thickness of the chest wall present in the
modulation PB-CC in MU for open fields treatment volume (less than 2.5 cm), is a concern due to
((CC2PB)/PB)6100 the known limitations of dose calculation algorithms in
Medial Lateral
field field modelling dose in lung, and at lung-tissue interfaces, due
to the missing tissue geometry and due to patient
1 Open 22.1 22.1 +3.9
movement during treatment delivery. In terms of
Wedged 20.3 20.4
2 Open 22 22 +3.0 treatment planning, the CC algorithm calculates dose
Wedged – – more accurately under most of these conditions (to
3 Open 21.6 21.8 +2.7 within ¡1%), with the exception of positions very close
Wedged – – to media interfaces. The PB algorithm can underestimate
4 Open 21 21 +2.1 the required MU by up to 7% in some cases and in all
Wedged 20.2 20.2 cases generates a misleading dose distribution, over-
5 Open 22 21.6 +3.0 estimating the dose in lung and underestimating the
Wedged – – dose to the PTV. The differences in MU between PB and
6 Open 22 21.7 +3.3
Wedged – 20.2
CC depend on the thickness of chest wall and the
position of the normalization point with respect to the

834 The British Journal of Radiology, October 2006

Photon dose calculations in chest wall radiotherapy
Figure 6. Percentage differences in mean dose to the planning target volume (PTV) and lung between pencil beam (PB) and
collapsed cone (CC) for patients included in this study plotted in terms of: (a) thickness of chest wall, (b) maximum thickness of
lung on the central slice and (c) the distance from the normalization point to the skin. m, mean dose to PTV; & mean dose to
lung.
lung. Our findings indicate that for a chest wall thickness
equal or less than 2.5 cm and with 2 cm or more lung
tissue involvement, the CC algorithm should be used
instead of PB.
For cases when the plan normalization point has to be
placed in the lung and close to an interface with breast
tissue, one needs to be aware of the limitations of the TPS
algorithms in such regions. It is preferable to avoid
placement of the plan normalization point close to
interfaces, but as seen in some of the clinical cases
examined here, the thin chest wall often leaves no other
option. In our study, in some cases the breast tissue
received higher than the prescribed dose (at least 5%
higher) and the clinician should be advised to prescribe
to the clinically significant isodose level covering the
chest wall tissue.
This work investigated the dosimetric accuracy of
isocentric chest wall irradiation using 6 MV photon
beams, which is the energy used for such irradiations at
our centre. A previous comparison between the PB and
CC algorithms with the plan normalization point
(isocentre) in homogeneous water under a tangential
irradiation geometry and for a range of beam energies
(4 MV, 6 MV, 15 MV) has shown that the difference from
measurement is greater at the lowest energy and in
The British Journal of Radiology, October 2006
particular for the PB algorithm. At higher beam energies,
for a tangential irradiation with the normalization point
in low density medium, it is expected that the perfor-
mance of both calculation algorithms against measure-
ment could worsen, because at higher energies the range
of travel of secondary particles is greater and modelling
the deposition of their energy becomes more complex,
especially in the lateral direction and close to media
interfaces.
Routine clinical implementation of the CC dose
calculation algorithm on Helax-TMS is hindered because
it is not possible to implement this interactively.
However, in the special case of chest wall irradiation it
is advised that an optimum treatment plan is produced
using the PB algorithm and the same plan is also
calculated with the CC algorithm for the clinician to
make an informed decision on the prescription to the
chest wall tissue. At NCCT the Helax-TMS system will
soon be replaced with the Oncentra (OTP) TPS
(Nucletron B.V.). Both PB and CC dose calculation
engines are available on OTP, and CC could be used
interactively. Therefore, CC would be the algorithm of
choice for treatment planning of the chest wall, and other
sites with significant involvement of non-unit density
heterogeneity.
835

The isocentric breast planning technique at NCCT will
develop further with the optimization and evaluation of
plans based on distributions generated on all CT slices.
New approaches have been investigated for plan
optimization for the breast, such as electronic compensa-
tion using field-in-field (forward IMRT) [20, 21]. For
these to be implemented clinically it is required that a
PTV is defined on all slices of the CT study. This
necessitates additional input by the clinician in treatment
planning of the breast and/or the revision of guidelines
for breast planning to be followed by experienced
dosimetrists. New, complex treatment planning techni-
ques for breast planning also necessitate a revision of the
methods used for the independent check of monitor
units generated by TPSs. Patient movement during
treatment has not been addressed here. Future efforts
in improving the treatment of chest wall patients would
also need to account for chest wall movement [22, 23].
Acknowledgments
The authors are thankful to Dr Otto Sauer (Department
of Radiotherapy, Julius-Maximilians-University of
Würzburg, Germany) for discussions on the derivation
of chamber perturbation correction for measurements in
lung. Mr Geoff Lambert (Head of Radiotherapy Physics,
Regional Medical Physics Department, Newcastle upon
Tyne) and Ms Gill Lawrence (Consultant Clinical
Scientist in Radiotherapy Physics, Regional Medical
Physics Department) are gratefully acknowledged for
their support and helpful comments to the manuscript.
Dr Ujjal Mallick (Consultant Clinical Oncologist,
Northern Centre for Cancer Treatment) is acknowledged
for his clinical input to this work.
References
1. Swerdlow A, Dos Santos Silva I, Doll R, editors. Cancer
incidence and mortality in England and Wales: trends and
risk factors. Oxford University Press, 2001:71–196.
2. Johnson S. Incidence and mortality in breast cancer. In:
International handbook of breast cancer. Euromed
Publications, 2002.
3. Ahnesjö A. Collapsed cone convolution of radiant energy
for photon dose calculation in heterogeneous media. Med
Phys 1989;16:577–92.
4. Ahnesjö A, Saxner M, Trepp A. A pencil beam model for
photon dose calculation. Med Phys 1992;19:263–73.
5. Ahnesjö A, Knöös T, Montelius A. Application of the
convolution method for calculation of output factors for
therapy photon beams. Med Phys 1992;19:295–301.
6. Knöös T, Ceberg C, Weber L, Nilsson P. The dosimetric
verification of a pencil beam based treatment planning
system. Phys Med Biol 1994;39:1609–28.
836
M M Aspradakis, H M McCallum and N Wilson
7. Engelsman M, Damen EMF, Kiken PW, van’t Veld A, Ingen
KM, Mijnheer BJ. Impact of simple tissue inhomogeneity
correction algorithms on conformal radiotherapy of lung
tumours. Radiother Oncol 2001;60:299–309.
8. Weber L, Nilsson P. Verification of dose calculations with a
clinical treatment planning system based on a point kernel
dose engine. J Applied Clin Med Phys 2002;3:73–87.
9. Aspradakis MM, Morrison RH, Richmond ND, Steele A.
Experimental verification of convolution/superposition
photon dose calculations for radiotherapy treatment plan-
ning. Phys Med Biol 2003;48:2873–93.
10. Wieslander E, Knöös T. Dose pertubation in the presence of
metallic implants: treatment planning system versus Monte
Carlo simulations. Phys Med Biol 2003;48:3295–305.
11. Irvine C, Morgan A, Crellin A, Nisbet A, Beange I. The
clinical implications of the collapsed cone planning algor-
ithm. Clin Oncol 2004;16:148–54.
12. Constantinou C, Attix FH, Bhudatt RP. A solid water
phantom material for radiotherapy x-ray and gamma-ray
beam calibrations. Med Phys 1982;9:436–41.
13. White DR, Martin RJ, Darlison R. Epoxy resin based tissue
substitutes. Br J Radiol 1977;50:814–21.
14. White DR, Constantinou C, Martin RJ. Technical Note:
Foamed epoxy resin-based lung substitutes. Br J Radiol
1986;59:787–90.
15. Krieger T, Sauer O. Monte Carlo- versus pencil beam-/
collapsed -cone-dose calculation in a heterogeneous multi-
layer medium. Phys Med Biol 2005;50:859–68.
16. Siddon RL. Solution to treatment planning problems using
coordinate transformations. Med Phys 1981;8:766–74.
17. Casebow MP. Matching of adjacent radiation beams for
isocentric radiotherapy. Br J Radiol 1984;57:735–40.
18. Wilkinson J. Consensus management - a departmental
study of the external beam therapy of carcinoma breast.
Radiography Today 1992:17–20.
19. Carrasco P, Jornet N, Duch MA, Weber L, Ginjaume M,
Eudaldo T, et al. Comparison of dose calculation algorithms
in phantoms with lung equivalent heterogeneities under
conditions of lateral electronic disequilibrium. Med Phys
2004;31:2899–911.
20. Richmond ND, Turner R, Dawes PDK, Lambert GD,
Lawrence GL. Evaluation of the dosimetric consequences
of adding a single asymmetric or MLC shaped field to a
tangential breast radiotherapy technique. Radiother Oncol
2003;67:165–70.
21. Wilkinson M, Aspradakis MM. A study of electronic
compensation methods in breast radiotherapy planning.
In: IPEM Annual Scientific Meeting. York, UK: IPEM,
2004.
22. Pedersen AN, Korreman S, Nystrom H, Specht L. Breathing
adapted radiotherapy of breast cancer: reduction of cardiac
and pulmonary doses using voluntary inspiration breath-
hold. Radiother Oncol 2004;72:53–60.
23. Korreman SS, Pedersen AN, Nottrup TJ, Specht L, Nystrom
H. Breathing adapted radiotherapy for breast cancer:
comparison of free breathing gating with the breath-hold
technique. Radiother Oncol 2005;76:311–8.
The British Journal of Radiology, October 2006
The British Journal of Radiology, 79 (2006), 837–842

Patient and staff radiation doses from early radiological

examinations (189921902)
1
C J KOTRE, PhD and 2B G LITTLE, MSc

1
Regional Medical Physics Department, Newcastle General Hospital, Newcastle-upon-Tyne NE4 6BE
and 2Radiotherapy Physics Department, North Glasgow University Hospitals, Dumbarton Road,
Glasgow G11 6NT, UK

ABSTRACT. A source of data on radiographic and fluoroscopic examinations, including

radiographic technique factors, was used in conjunction with information about cold-
cathode X-ray apparatus to estimate patient and staff radiation doses for the years
1899 to 1902 at the Forth Banks Infirmary, Newcastle-upon-Tyne. Physical evidence Received 3 January 2006
from representative apparatus of the period was used with a beam spectral simulation Revised 22 February 2006
program to characterize the X-ray beam, and information about the electrical supply Accepted 8 March 2006
waveform was produced by experimental operation of a contemporary induction coil.
DOI: 10.1259/bjr/16982267
Results are given in terms of skin entrance dose, and these are compared with modern
values. An estimate of the annual dose received by the radiographer known to have ’ 2006 The British Institute of
carried out all of the examinations within this period is also given. Radiology

A collection of historical artefacts associated with early
radiology in the north-east of England is on display in
the Radiology Department of the Royal Victoria
Infirmary, Newcastle-upon-Tyne. This collection,
assembled by former Newcastle radiologist Dr C K
Warrick, includes a number of very early cold-cathode X-
ray tubes and an example of the type of induction coil
high voltage apparatus which would have been used to
power them. Also included is an original log-book listing
the radiological examinations carried out at the Forth
Banks Infirmary (later to become the Royal Victoria
Infirmary) in the years 1899 to 1902. The log-book also
contains entries on the exposure time in minutes and the
distance used for the radiographic examinations (69% of
the total), or denotes that the examination was carried
out by fluoroscopy. Due to the coincidence of both
original documentary and physical evidence relating to
early radiographic practice in this collection, it was
decided to use it to attempt an estimation of patient and
staff radiation doses for comparison with modern values.
taken as sodium glass as lead glass would have
produced a blue glow [1]. The examination log quotes
distances from the tube to the patient entrance surface
for each exposure. For the calculations below, these are
taken as distances from the outside of the tube envelope
as some of the distances were very short and it is thought
unlikely than the focal spot would be taken as the source
of the radiation in routine work of the time. The distance
from the centre of the target to the outside of the tube
envelope was measured at 3 cm.
Induction coil
The induction coil examined for this work (Figure 2) is
also dated c. 1899 and was used in conjunction with cold-
cathode X-ray tubes at a pharmacist’s establishment in
Sunderland. It was actually used with the X-ray tube
detailed above, but not at the Forth Banks Infirmary. The

Evidence and assumptions

X-ray tube
A cold-cathode X-ray tube from around 1899
(Figure 1) was used to formulate the specification of
the ‘‘typical’’ X-ray tube which would have been used at
the Forth Banks Infirmary during the 1899–1902 period
under study. The anode material was taken to be
platinum, as tungsten anodes were not yet in use [1].
The target angle was measured at 35 ˚ and the thickness
of the glass envelope estimated at 2 mm. From the
evidence of Warrick [2], who quotes an early Newcastle
radiologist remembering the ‘‘beautiful apple green Figure 1. The cold-cathode X-ray tube of c. 1899 used to
fluorescence which always pleased me’’ given off by formulate the specification for the X-ray tube used in the
this type of equipment in operation, the glass type can be simulations.

The British Journal of Radiology, October 2006 837

 C J Kotre and B G Little

example to hand, so documentary evidence was relied

upon. Warrick [2], quoting a letter from an early
Newcastle radiologist, mentions that, ‘‘The coil was a
large one of conventional pattern with a platinum make
and break...I think the spark gap was ten inches’’. The
potential corresponding to this 10-inch spark gap
depends on the shape of the spark gap electrodes, being
lowest for point electrodes and much higher for spherical
electrodes. A 10-inch spark gap corresponds to approxi-
mately 120 kV for point electrodes, which seems rather
high, although it is possible that the 10-inch recollection
refers just to the size of the induction coil itself as the
length of the coil was often quoted in advertising
literature.
In early radiography, part of the skill of the radio-
grapher was selection of the best cold-cathode X-ray tube
for the examination. The tubes had a short lifetime, and
Figure 2. Induction coil of c. 1899 under test with a low- the hardness of the vacuum increased during use.
voltage power supply. Younger ‘‘soft’’ tubes were therefore preferred for
examinations requiring less penetrating radiation and
coil is 15 cm in diameter and 30 cm in length, and stands older ‘‘hard’’ tubes were selected for large body sections.
on a hollow wooden plinth supporting the sprung An early guide to tube selection, quoted by Burrows [3],
contact-breaker arrangement and containing a capacitor uses three classifications of tube and gives examples of
composed of flat foil electrodes insulated with waxed their use. Table 1 gives these three classifications,
paper. Despite being the size of a telephone book, this together with an estimate of the required kVp based on
capacitor measured only 1 mF. The capacitor insulation more modern experience [4]. These latter values of kVp
was found to be intact and some evidence of re-wiring are assumed for the tube types used at the Forth Banks
was found inside the plinth suggesting that restoration, Infirmary in 1899–1902. The values lie at the low end of
possibly for teaching purposes, had taken place at some the modern diagnostic kVp range, but this is consistent
stage. The primary winding was found to be functional, with the finding reported in a number of the early X-ray
but not the very fine secondary winding. Breakdown of examinations that the image was poor and under-
penetrated for thicker body sections.
the secondary winding insulation was also suspected.
The use of a mechanical contact-breaker with an
Various methods of deriving the probable tube
induction coil produces an intermittent high voltage
potential supplied by the induction coil were investi-
waveform, and some estimate of this waveform was
gated, but rejected due to the large number of unknowns
needed. After some hesitation, the authors gave in to
related to the tube loading and current/voltage relation-
their curiosity and experimentally operated the 1899
ships of cold-cathode tubes at various states of evacua-
induction coil from a high current capacity variable low
tion. It was also suspected that the induction coil used at
voltage power supply. The supply voltage was increased
the Forth Banks Infirmary was rather larger than the
until the contact-breaker operated with a regular,
reasonably even frequency. Despite the discontinuity of
Table 1. Classifications and uses of cold-cathode X-ray tubes the secondary winding and suspected breaks in its
from an early guide [3] together with assumed tube insulation, high voltages were produced across a chain
potentials for these examinations based on more modern of large resistors and a short spark-gap. Oscilloscope
experience [4] traces of the primary voltage waveform showed a very
Tube Description Applications Assumed uneven pattern due to the arcing of the contact-breaker,
type tube but traces of the secondary high-voltage waveform
potential showed a regular train of pulses (Figure 3) from which
(kVp)
the beam-on time was estimated at 9.6% of the recorded
1 ‘‘Soft’’ Tendons 50 exposure time. It was assumed that this figure would be
Limbs in children typical for apparatus of the time.
Epiphyses in very The final essential electrical parameter was the tube
young subjects current. The unknown electrical characteristics of a
Less dense structures
typical cold-cathode X-ray tube made this also difficult
2 ‘‘Medium’’ Thorax 60
Renal calculi in thin to estimate from a physical basis. Burrows [3] gives
patients values of 1–2 mA for early cold cathode tubes, noting
Limbs in adults that thin sheet platinum targets would melt at the higher
Fluoroscopy values of current which later became available as heavier
3 ‘‘Medium–Hard’’ Thorax 70 induction coils were developed. Reynolds [4] (who
Renal calculi in stout himself designed a modified cold cathode X-ray tube in
patients 1900) gives values of 2–5 mA. For this early tube and
Spine induction coil, a compromise value of 2 mA was
Pelvis assumed for the Type I tube (50 kVp), with the current
Fluoroscopy
scaled down to a constant power output from the

838 The British Journal of Radiology, October 2006

Radiation doses from early radiological examinations

reader is referred to Birch and Marshall [7]. The software

limits the input parameters to those appropriate to
modern day X-ray tubes, therefore some of the para-
meters from the 1899 X-ray tube could not be simulated
directly. The parameters used were as follows: tungsten
target, 22 ˚ target angle, 50 kVp, 60 kVp and 70 kVp for
tube types 1, 2 and 3, respectively (Table 2), voltage
ripple of 0% (for the actual beam on time) and beam
filtration of 2 mm of glass. Where the parameters used
were not that assumed from Table 2, various corrections
were made and this is discussed later in more depth. The
logged examination types were classified into groups
thought to match up with tube types 1, 2 and 3 based on
the applications given in Table 1.
The simulation software produced values of air kerma
per mAs at 75 cm from the target, which were then
converted to ESDs for the historical examinations by
Figure 3. Oscilloscope trace of the high-voltage secondary multiplying them by the recorded time of exposure, tube
waveform obtained with the induction coil in operation. current, backscatter factor and taking account of the
actual distance from the tube. As the distances given are
induction coil for the harder tubes, giving 1.7 mA for the generally quite short (down to 3 cm) it is assumed that
Type 2 tube (60 kVp) and 1.4 mA for the Type 3 tube these are measured from the surface of the glass
(70 kVp). envelope rather than from what we would now call the
The specification of the X-ray generating apparatus focal spot. An extra 3 cm was therefore added to account
arrived at from the above considerations is summarized for the distance from the focus to the outside of the tube
in Table 2. envelope. A realistic backscatter factor of 1.16 was
determined from tabulated data of backscatter factor
for a glass window tube with no external filtration at a
potential of 60 kV [8]. An output correction between the
Calculation of patient entrance skin doses maximum target angle in the simulation package, 22 ˚,
A useful quantity to estimate for comparison with and that measured for the cold-cathode tube target, 35 ˚,
modern day dose levels is the entrance surface dose was made using a simple model of the self attenuation of
(ESD). The first step in this calculation was to simulate the anode [7] by fitting parameters to this model from the
the output spectrum from the tube used over the time relationship between output and target angle produced
period noted in the log-book. For this, a commercially by the software package at 50 kVp, 60 kVp and 70 kVp.
available software package was used [6]. The input Finally a correction was made for the difference in air
parameters required are peak tube potential, target kerma due to the difference in target materials. It is
angle, target material and filtration in the beam. The known that for Bremsstrahlung X-ray production the air
intensity of photons at a given energy is calculated for kerma produced is proportional to the atomic number of
Bremsstrahlung and characteristic radiation via empiri- the target material [9]. The atomic number of platinum is
cal relationships. For a description of the theory, the 78 as compared with 74 for tungsten and the ESD was
therefore modified by the ratio of the atomic numbers. It
was assumed that changes in the energy of the
Table 2. Summary of quantities used to specify the X-ray characteristic radiation made a negligible difference to
tube and operating conditions the air kerma and no modification was made for this.
Quantity Value measured (M) or assumed
from indirect evidence (A)
Calculation of annual staff dose
Anode material Platinum (A) [1]
Anode angle 35 ˚ (M) The radiographer known to have carried out the
Envelope material Sodium glass (A) [1, 2] examinations listed in the 1899–1902 record was Mr
Envelope thickness 2 mm (M) Thomas Dodd. When the Forth Banks Infirmary acquired
Tube potential Type 1: 50 kVp (A) [3, 4] its first X-ray equipment in 1899, Mr Dodd, the older
Type 2: 60 kVp (A) [3, 4]
brother of the head porter, was employed as technician
Type 3: 70 kVp (A) [3, 4]
Tube current Type 1: 2.0 mA when beam on to operate it and keep the log-book [3]. He worked at the
(A) [3, 5] old Infirmary and later the new Royal Victoria Infirmary
Type 2: 1.7 mA when beam on for over 20 years. In the process, he suffered severe
(A) [3, 5] radiation injuries, losing three fingers (one of which is
Type 3: 1.4 mA when beam on preserved in the historical collection) and ultimately
(A) [3, 5] dying as a result of his injuries in 1929 [10].
Duty cycle (% of time 9.6% (M) A later c. 1920 photograph of Mr Dodd in action [2, 3]
beam on) shows him positioned between the patient and the high
Distance from focus to 3 cm (M)
voltage generating equipment with the X-ray tube at his
exterior of tube
head height. It seems likely that this was his normal

The British Journal of Radiology, October 2006 839


working position. Although this later photograph shows
what might be crude shielding around the tube, the early
tubes were not shielded. The distance between his body
centreline to the tube is estimated from the photograph
to be 1.4 m. It is assumed that he was present throughout
even the long exposures to monitor the induction coil
operation and check for patient movement. In order to
estimate Mr Dodd’s annual whole-body dose, the dose to
a point at which a modern waist-level personal dose-
meter would be worn was calculated. This point is taken
as 1.6 m from the tube focus and 1.4 m from the beam
entry point on the patient. For simplicity, the tube axis is
taken to be parallel to the table with Mr Dodd alongside.
No correction for the possible shielding effect of the
anode was therefore applied.
In order to estimate the annual staff dose, the radio-
graphic exposure times and number of fluoroscopic
examinations were totalled over the 2.8 years recorded in
the log-book. The exposure time for the fluoroscopic
exposures was estimated at 5 min per examination, to
include the setting up of the tube and recording of the
image by sketching. This latter practice was used in
the 1920s when more technicians were employed in the
department [3] and it is assumed here that a similar
method was used in the early days when Mr Dodd
worked alone. Using the model of the X-ray apparatus
described above, the annual dose to a waist-level
personal dosemeter, including backscatter, was esti-
mated. Tube type 2 was assumed for all fluoroscopic
examinations, and the same classification of tube type as
used above was employed for the radiographic exami-
nations. The scatter contribution was estimated using
modern dose–area product to scatter dose conversion
factors [11] and assuming (in the absence of any field
limiting device) a scattering patient surface of
25 cm625 cm. For the fluoroscopic examinations, any
attenuating effect of the patient on the primary radiation
was neglected on the grounds that the patient section
would only partially shield the radiographer from the
uncollimated primary beam and unshielded tube.
Results and discussion
Breakdown of examination types
The frequency of examination types (as named in the
record) is shown in Table 3 for the radiographic
examinations and Table 4 for the fluoroscopic examina-
tions. Over the 2.8 year time period in the log book there
were 507 examinations noted. The top three examina-
tions overall (for both fluoroscopy and radiography
examinations) were investigation of fracture (or bone
disorder), foreign bodies and renal calculi with propor-
tions of 50%, 35% and 8%, respectively.
Some patterns within this group are also of interest. Of
the 174 foreign body examinations, the most frequently
occurring (50%) were cases of needle and pin injuries
including needle in throat, hands, hat-pin in stomach,
etc. A further significant proportion of the foreign body
examinations (18%) involved bullet injuries to subjects
between the ages of 12 years and 26 years, with one-third
of these being described as ‘‘bullet in hand’’.
840
C J Kotre and B G Little
Table 3. Breakdown of examination types for radiography
Radiographic examination type No. in record
Fracture 189
Foreign body 83
Renal calculus 38
Pulmonary tuberculosis 8
Sarcoma 4
Aneurysm of aorta 3
Ricketts 3
Others 12
Total radiographic 340
Table 4. Breakdown of examination types for fluoroscopy
Fluoroscopic examination type No. in record
Foreign body 101
Fracture 53
Aneurysm of aorta 6
Sarcoma 1
Others 6
Total fluoroscopic 167
Patient ESDs compared with modern values
Due to the high frequency and variety of peripheral
and foreign body examinations in the historical data set,
it is quite difficult to make meaningful comparisons with
modern ESD data for the majority of the examinations.
Three sub-groups were, however, extracted for compa-
rison. These were adult abdomen/pelvis examinations
(40 exposures), adult non-penetrated chest radiographs
for diagnosis of tuberculosis and sarcoma of rib (10
exposures) and adult penetrated/lateral chest examina-
tions associated with investigations of foreign bodies (14
exposures). The distributions of ESD for these examina-
tions are shown in Figure 4 (abdomen/pelvis), Figure 5
(chest) and Figure 6 (penetrated/lateral chest). The
median ESD values for these three sub-groups are
189 mGy for the abdomen/pelvis examinations, 68 mGy
for the chest examinations and 354 mGy for the pene-
trated/lateral chest examinations. These can be compared
with modern median ESD values [12] of 4.1 mGy for
abdomen anteroposterior (AP), 3.2 mGy for pelvis AP,
0.14 mGy for chest AP views and 0.51 mGy for lateral
chest. This gives a ratio approximately 52 times higher
ESD for the abdomen/pelvis examinations, 486 times
higher for the chest examinations and 694 times higher for
the penetrated/lateral chest examinations.
In order to cross-check the magnitude of these results
and elucidate the dose implications of technique changes
on the patient dose results, the simulation software [6]
was further used to calculate the expected ratio of ESD
between the historical and modern techniques [12] for
these three examinations, normalized to a constant value
of receptor dose. The differences in kVp, distances,
filtration, grid use (abdomen/pelvis only), target angle
and X-ray spectrum were considered. The effects of these
technique differences result in an expected ratio of ESD
historical/modern of 2.0 for the abdomen/pelvis, 5.2 for
the AP chest examinations and 8.6 for the lateral/
penetrated chests. From the results above, this leaves a
The British Journal of Radiology, October 2006
Radiation doses from early radiological examinations
Figure 4. Histogram of entrance surface dose frequency for
abdomen/pelvis examinations.
Figure 5. Histogram of entrance surface dose frequency for
posteroanterior chest examinations.
Figure 6. Histogram of entrance surface dose frequency for
lateral/penetrated chest examinations.
The British Journal of Radiology, October 2006
ratio of 26 for the abdomen/pelvis, 93 for the (few) chest
examinations and 80 for the lateral/penetrated chests to
be accounted for by the improvements in radiographic
speed offered by modern screen–film systems compared
with photographic emulsion alone. These ratios are in
the region of what would be expected from the
intensification factor of modern screen–film systems
(30–100 [13]).
The highest radiographic ESD for a single exposure
estimated for the whole log was 1.8 Gy for an exposure
for ‘‘steel in neck’’ carried out at a focus–skin distance of
20 cm and requiring an exposure time of 20 min.
Annual staff dose
The annual whole-body dose for Mr Dodd during the
period of the record from both primary and scattered
radiation is estimated to be 940 mSv year21 (97% from
primary irradiation, 3% from scatter), a figure which is
47 times the annual dose limit for workers under current
UK regulations. Mr Dodd was associated with radio-
graphy at the Forth Banks Infirmary, and later the Royal
Victoria Infirmary which replaced it, for 20 years [3].
During this period, the need for radiation protection was
becoming recognized, and the 1920s photograph of Dodd
[2, 3] does appear to show protective plates either side of
the X-ray tube, but equally it shows him working
without any personal protective equipment. The work-
load of the Forth Banks and later Royal Victoria
Infirmary Departments increased rapidly from the early
level of approximately 200 examinations per year
reported here, to almost 14 000 per year by 1918 [3].
Mr Dodd was the sole radiographer until around 1910
when an assistant radiographer, Cornelius McMeekin,
was appointed [2, 3], and he had some assistance from
two R.A.M.C privates during the war years [2], but with
the documented workload expansion during his career it
seems likely that Mr Dodd’s personal radiation exposure
continued to be considerable. He died of cancer in
December 1929 [10].
Conclusions
This work has attempted to use modern methods to
estimate patient and staff radiation doses from X-ray
examinations performed over a century ago. The actual
values obtained show some inconsistencies and hang by
a very long chain of stated assumptions, but physical
evidence has been used where practicable to add weight
to these assumptions. As would be expected, the results
show patient doses of a magnitude unacceptable by
modern standards, but not perhaps as high as might be
guessed at merely by noting the very long exposure
times in the original record. The dangerous level of staff
irradiation for the earliest radiographers is quantitatively
confirmed.
Acknowledgments
We would like to thank the Radiology Department,
Royal Victoria Infirmary, Newcastle-upon-Tyne for
841

access to its historical collection, Dr C K Warrick for
useful discussions on the origins of the material, and Mr
Mike Feeney for discussions on early high-voltage
generating equipment and assistance with the operation
of the induction coil.
References
1. Mould RF. The early history of x-ray diagnosis with
emphasis on the contributions of physics 189521915. Phys
Med Biol 1995;40:1741–87.
2. Warrick CK. Notes on the history of the Department of
Radiology of the Royal Victoria Infirmary, Newcastle upon
Tyne. Radiography 1977;43:190–4.
3. Burrows EH. Pioneers and early years: a history of British
radiology. Channel Islands: Colophon Ltd, 1986.
4. Goldman M, Cope D. A radiographic index, 4th edn.
Bristol: John Wright and Sons, 1970.
5. Reynolds RJ. Sixty years of radiology. Br J Radiol
1956;29:238–45.
842
C J Kotre and B G Little
6. IPEM. Catalogue of Diagnostic X-ray Spectra and other
data, CDROM Report No. 78. York: Institute of Physics and
Engineering in Medicine, 1997.
7. Birch R, Marshall M. Computation of Bremsstrahlung X-ray
spectra and comparison with spectra measured with a
Ge(Li) detector. Phys Med Biol 1979;24:505–17.
8. BJR Supplement 25. Central axis depth dose data for use
in radiotherapy. London: British Institute of Radiology,
1996.
9. Dendy PP, Heaton B. Physics for diagnostic radiology, 2nd
edn. Bristol and Philadelphia: IoP Publishing, 1999.
10. Obituary, Thomas Dodd. Br J Radiol 1930;3:95.
11. Sutton DG, Williams JR. Radiation shielding for diagnostic
X-rays. London: British Institute of Radiology, 2000.
12. Hart D, Hillier MC, Wall BF. Doses to patients from medical
X-ray examinations in the UK – 2000 Review. National
Radiological Protection Board Report NRPB-W14. Didcot:
Health Protection Agency, 2002.
13. Farr RF, Allisy-Roberts PJ. Physics for medical imaging.
London: WB Saunders, 1998.
The British Journal of Radiology, October 2006
The British Journal of Radiology, 79 (2006), 843–849

PICTORIAL REVIEW

Ultrasound spectrum in intraductal papillary neoplasms of breast

S GANESAN, MD, G KARTHIK, DNB, M JOSHI, MD, DNB and V DAMODARAN, MS, FRCS

Department of Radiology and Imaging, G.K.N.M Hospital and Research Centre, PN Palayam,
Coimbatore – 641037, India

ABSTRACT. Intraductal papillary neoplasms (IPNs) of breast form a wide spectrum of

pathological changes with benign intraductal papilloma occupying one end of the
spectrum and papillary carcinoma at the other end. Intraductal papillomas are known
to occur anywhere within the ductal system and are broadly classified into central and
peripheral types. Intraductal papillary carcinoma is an uncommon ductal malignancy
forming papillary structures, and these lesions characteristically lack the myoepithelial
layer present in benign papillary neoplasms. Three basic patterns of IPNs are recognized
on ultrasound – intraductal mass with or without ductal dilatation, intracystic mass and
a predominantly solid pattern with the intraductal mass totally filling the duct. Benign
papillomas are known to exhibit calcifications which tend to be extremely dense and
Received 24 November
coarse. IPNs are highly vascular tumours and have a propensity to bleed spontaneously. 2004
A distinct vascular pedicle is identified within the central core of IPNs, with branching Revised 16 May 2005
vessels arborising within the mass. In an older age group, presence of a large solid Accepted 1 June 2005
component and evidence of spontaneous intracystic bleed are more suggestive of
DOI: 10.1259/bjr/69395941
papillary carcinomas than benign papillomas. We have serially studied 42 cases of
intraductal papillary neoplasms with sonomammography and mammography from ’ 2006 The British Institute of
2001 to 2004. Radiology

Intraductal papillomas are common neoplasms with a
relative incidence of 2–3% [1]. In elderly patients,
intraductal papillomas are often asymptomatic and are
seen commonly as an incidental finding in biopsy
specimens [2]. Even though intraductal papillomas are
primarily benign, these lesions can pose problems in
view of their similarity to intraductal papillary carci-
noma clinically, on ultrasound and histologically [3].
Intraductal papillary carcinoma (IPC) is a rare ductal
carcinoma forming papillary structures, with reported
incidence of 1–4% of breast carcinomas [4]. These
neoplasms have certain characteristic imaging features
which help to differentiate these lesions from other focal
breast abnormalities.
Pathological observations
Papillomas are essentially benign proliferations of
ductal epithelium. They may occur at any age between
30 years and 77 years [4], but are commonly seen
between 30 years and 55 years [3]. They are known to
occur anywhere within the ductal system and are
broadly classified into central and peripheral types.
Central types are usually solitary, while the peripherally
located papillomas tend to be multiple within the
terminal duct lobular unit. Lesions are often confined
Address correspondence to: Dr Karthik Ganesan, Department of
CT and MRI, Jaslok Hospital and Research Centre, 15, Dr G
Deshmukh Marg, Mumbai – 400026, Maharashtra, India.
to a single breast, while bilateral lesions are reported in
up to 14% of cases [4].
Central papillomas are subareolar in location within a
major duct. On macroscopic examination, a papilloma
appears as a round to oval, small mass usually
measuring a few millimetres in size within a dilated
duct. Larger lesions dilate the duct more and extend
along the long axis of the duct presenting a spheroidal
shape. With ductal obstruction, the dilated duct with a
papillary lesion may resemble a cyst with an intracystic
solid component, this variant being termed as an
intracystic papilloma. Histologically, papillomas show
hyperplastic proliferation of ductal epithelium, having
an arborescent growth pattern with branching fibromus-
cular core of myoepithelial and epithelial cells. Lesions
may be pedunculated or broad based [3, 4].
Multiple peripheral papillomas are a rare entity in
which the lesions are located in the peripheral duct
system within the terminal ductal lobular unit. Several
adjacent ducts are involved with segmental dilatation of
the ducts, often resulting in a peripherally located mass.
The incidence of nipple discharge is lower in these
patients compared with the papillomas in larger ducts.
There is an increased risk of carcinoma in peripheral
papillomas which is directly related to the degree of
cellular atypia. Peripheral papillomas are often asso-
ciated with coexisting malignancy with a reported
incidence of 10–30% [3–6].
Intraductal papillary carcinoma (IPC) is an uncommon
ductal malignancy forming papillary structures.
Histologically papillary carcinoma shows multilayered

The British Journal of Radiology, October 2006 843

 S Ganesan, G Karthik, M Joshi and V Damodaran

papillary projections with microscopic frond formations

extending from the vascularized stalks. These lesions
characteristically lack the myoepithelial layer present in
benign lesions. IPCs are reported in patients from 25 years
to 89 years of age with a peak incidence between 40 years
and 75 years [4]. IPCs have a wide spectrum of presenta-
tions varying from a focally invasive lesion with micro-
scopic frond formation to a large mass located within a
cystically dilated duct. Multiple lesions tend to occur
within the same duct with papillary configurations.

Ultrasound features
Ultrasound features of intraductal papillary neo-
plasms (IPN) primarily depend on the gross macroscopic
appearance of the lesion. Three basic patterns of IPNs are
recognized on ultrasound – intraductal mass with or
without ductal dilatation, intracystic mass and a pre-
dominantly solid pattern with the intraductal mass
totally filling the duct [3, 7]. If the tumour is small, a
focally dilated duct may be the only observation. A
solitary dilated duct, even in the absence of a demon-
strable intraductal mass, is highly suggestive of an
intraductal papilloma, especially, if the patient is Figure 2. Papillary carcinoma. A moderately large mass is
presenting with a serosanguinous nipple discharge [3]. seen to almost totally fill the entire dilated duct. Relatively
Dilated duct with an intraductal mass or a cyst with an hypoechoic debris is seen to fill the peripheral duct adjacent
intracystic solid mass is the hallmark of intraductal to the mass. A short segment of proximal duct is noted at 10–
papillomas (Figure 1). The ductal component may vary 11 o’clock position.
in size from a minimally dilated duct to a large cystically
dilated, obstructed duct. Similarly the intraductal soft
relationship between the mass and the duct on ultra-
tissue component may range in size from a very small
sound and classified the masses into four categories: type
lesion which may be impossible to image to a large mass
I – intraluminal mass; type II – extraductal mass; type III
completely filling the dilated duct or the cyst obscuring
– purely solid mass; type IV – mixed variety [9]. Benign
the ductal or cystic component simulating other
papillomas are known to exhibit calcifications. These
solid masses (Figure 2) [3, 8]. Han et al analysed the
calcifications tend to be dense and coarse (Figures 3–5).

Figure 3. Benign intraductal papilloma with calcification.

Figure 1. A focal mass arising from the ductal wall with Focal dilatation of a solitary duct with intraluminal echo-
relatively narrow base of attachment is present. Note the genic debris. Note small focal mass with dense, coarse
branching pattern and peripheral fronding typical of calcifications in the proximal duct, with ductal obstruction
intraductal papilloma/papillary carcinoma (D-Duct). (small arrows).

844 The British Journal of Radiology, October 2006

Pictorial review: Intraductal papillary neoplasms of breast

(a) (b)

Figure 4. Benign calcified intraductal papilloma with adjacent oil cyst. (a) Small focal mass with coarse, irregular and dense
calcifications (small arrows) adjacent to a cystic mass (C). Echogenic floating debris within the cyst with floating fat-fluid level
(long arrows). (b) Colour flow studies – focal increase in flow within the mass.

Small IPNs are often mammographically negative. A
minimal to moderate duct dilatation may be observed on
mammography as a progressively tapering band-like
density extending from the nipple-subareolar region
towards the breast parenchyma for a variable distance
(Figure 6). Larger lesions in a dilated duct may resemble
any other focal well-circumscribed dense mass on
mammography (Figure 7). Calcified IPNs exhibit dense,
central, peripheral or combined form of coarse calcifica-
tion similar to those seen in cases of calcified fibro-
adenomas (Figure 8). Boonjuwetat et al described the
mammographic appearances of papillary neoplasms in a
series of 15 cases. They reported that most lesions
presented as solitary dense masses with no evidence of

(a) (b)

Figure 5. Calcified giant intraductal papilloma. (a) Ultrasound and (b) mammography demonstrate a large, bilobed, densely
calcified mass with distal shadowing (arrows).

The British Journal of Radiology, October 2006 845

 S Ganesan, G Karthik, M Joshi and V Damodaran

(a) (b)

(c)

Figure 6. Benign intraductal papilloma. (a) Mammography – oblique band like density along inferolateral quadrant of the left
breast. (b) Ultrasound – focal dilatation of a solitary duct with an intraluminal mass arising from the ductal wall. (c) Doppler
studies – distinct vascular pedicle within the central core with branching vessels arborising within the mass.

calcification in any of these lesions. A few lesions were
mammographically negative either due to the size of the
lesion or due to the dense parenchymal pattern [10].
IPNs are highly vascular and have a propensity to bleed
spontaneously. Spontaneous haemorrhage into a dilated
duct characteristically produces a fluid–debris level due to
the denser cellular components settling down to the
dependant position. The supernatant serum is anechoic
while the dependant cellular debris is echogenic. Presence
of fluid–debris level in a cyst, representing spontaneous
bleed into the cyst, is virtually suggestive of a mural
proliferative lesion (Figures 9 and 10) [11]. IPNs have a
characteristic flow pattern on colour flow studies. A
distinct vascular pedicle is identified in IPNs within the
central core with branching vessels arborising within the
mass. Colour flow studies are sensitive in identifying even
very small IPNs, in view of its characteristic vascularity
(Figure 11). Intraductal papillomas and papillary carcino-
mas have considerable overlap in imaging features and it
may not be possible to differentiate them on ultrasound. In

846 The British Journal of Radiology, October 2006

Pictorial review: Intraductal papillary neoplasms of breast
(a)
Figure 7. (a) Mammography – focal well circumscribed dense mass along the retroareolar region of the left breast.
(b) Ultrasound – large cystic mass with echogenic debris totally filling the cyst.
an older age group presence of a larger solid component
and evidence of spontaneous intracystic bleed are more
suggestive of papillary carcinomas than benign papillo-
mas (Figure 12) [9].
The differential diagnosis of IPNs depends upon the
basic imaging appearances. Presence of sectoral dilata-
tion of ducts with no demonstrable intraductal mass has
to be differentiated from mammary duct ectasia, which is
a chronic inflammatory condition. Bleeding into a duct,
inspissated material in mammary duct ectasia and ductal
carcinoma in situ may produce dilated ducts with
intraductal filling defects resembling IPNs. Mammary
ductectasia is often bilateral and tends to affect multiple
ducts. In intraductal carcinoma, ductal dilatation is
unilateral, sectoral and irregular with ductal wall
thickening. Colour flow studies reveal lack of flow in
inspissated intraductal debris. Increased or variable
periductal flow may be present in intraductal carcinomas
while the IPNs reveal the characteristic arborescent
The British Journal of Radiology, October 2006
(b)
vascularity. A cystically dilated duct may resemble a
simple cyst when the intracystic component is very
small. This has to be differentiated from other cystic
masses like a simple cyst, complex cyst, haematoma,
abscess and fat necrosis. A dilated duct with an
intraductal solid component consisting of a central core
and peripheral fronds, with characteristic flow on colour
flow studies, is virtually diagnostic of IPNs. When the
mass is large enough to fill the dilated duct or the cyst, it
may not be possible to delineate the peripheral ductal or
cystic component. These lesions have to be differentiated
from other solid masses [3].
Fine needle aspiration cytology (FNAC) or core biopsy is
required in all cases to arrive at a definitive diagnosis even
though the imaging findings are suggestive of IPNs.
FNAC from non-palpable small masses and from the solid
component in large cystic lesions can be performed under
ultrasound control. At our institution, small papillary
lesions within a minimally dilated duct, observed as
847

Figure 8. Mammography. Well circumscribed peripherally
calcified lesion within a progressively tapering band-like
density extending from the nipple-subareolar region
towards the breast parenchyma.
incidental findings in an asymptomatic patient on sono-
mammography are not subjected to FNAC or core biopsy.
These patients are advised serial follow up with sono-
mammography. Larger lesions, lesions with atypical
characteristics and lesions in symptomatic patients are
subjected to FNAC and core biopsy. The criteria for calling
Figure 9. Intracystic papillary carcinoma. A moderately large
cystic mass in the central breast region with large intracystic
solid component is present. The mass is attached to the wall
by a broad base and shows irregular branching pattern with
peripheral fronding. (D – Cystically dilated duct).
848
S Ganesan, G Karthik, M Joshi and V Damodaran
Figure 10. Colour flow studies in intraductal papillary
neoplasms (IPNs). Distinct vascular pedicle within the central
core with branching vessels arborising within the mass.
a core B3 are the following: papillary lesion, atypical
intraductal epithelial proliferations, radial scar, lobular
neoplasia and fibroepithelial lesions. Lee et al reported that
the B3 core group is a more heterogeneous group and has a
lower rate of malignancy on further biopsy. However, they
concluded that the majority of lesions categorised as B3
required excision [12]. Agoff et al evaluated the need for
surgical excision in intraductal papillary neoplasms and
suggested that all such lesions with atypical ductal
hyperplasia required excision owing to the high rate of
associated neoplasia [13]. Although some lesions cate-
gorised as B3 lesions on core biopsy may be re-categorised
Figure 11. Benign intraductal papilloma in a cystically
dilated duct. Cystically dilated duct with a small focal mass
attached to the ductal wall with a narrow base between 10
and 11 o’clock position. A short segment of dilated proximal
duct is identified with dependant echogenic debris with
layering effect forming a fluid-debris level.
The British Journal of Radiology, October 2006
Pictorial review: Intraductal papillary neoplasms of breast
on excision biopsy as B2 lesions, all lesions categorised as
B3 and above are subjected to excision biopsy at our
institution.
References
1. Cilotti A, Bagnolesi P, Napoli V, et al. Solitary intraductal
papilloma of breast. An echographic study of 12 cases.
Radiol Med (Torino) 1991;82:617–20.
2. Kramer WM, Rush BF. Mammary duct proliferation in the
elderly: a histopathologic study. Cancer 1973;31:130–7.
3. Tohno E, Cosgrove DO, Sloane JP. Benign processes-
tumors. In: Tohno E, Cosgrove DO, Sloane JP, editors.
Ultrasound diagnosis of breast diseases. Edinburgh:
Churchill Livingstone, 1994:94–7.
4. Dahnert W. Breast disorders. In: Dahnert W. Radiology
review manual, 4th edn. Philadelphia, PA: Williams and
Wilkins, 1999:458–74.
5. Haegenson CD. Diseases of the breast, 3rd edn.
Philadelphia, PA: W.B. Saunders, 1986:136–75.
6. Murad TM, Contesso G, Mouriesse H. Papillary tumors of
large lactiferous ducts. Cancer 1981;48:122–33.
7. Kasumi F. Ultrasound of breast diseases. Shinohara –
shuppan Co., Tokyo, 1983. Cited in Benign processes –
Tumors, Intraductal Papilloma, In: Tohno E, Cosgrove DO,
The British Journal of Radiology, October 2006
Sloane JP, editors. Ultrasound diagnosis of breast diseases.
Edinburgh: Churchill Livingstone, 1994:94.
8. Yang WT, Suen M, Metrewell C. Sonographic features of
benign papillary neoplasms of the breast: review of 22
patients. J Ultrasound Med 1997;16:161–8.
9. Han BK, Choe YH, Ko YH, Yang JH, Nam SJ. Benign
papillary lesions of the breast: sonographic-pathologic
correlation. J Ultrasound Med 1999;18:217–23.
10. Boonjunwetwat D, Prathombutr A. Imaging of benign
papillary neoplasm of the breast: mammographic, galacto-
graphic and sonographic findings. J Med Assoc Thai
2000;83:832–8.
11. Tohno E, Cosgrove DO, Sloane JP. Diagnostic features on
ultrasound. In: Tohno E, Cosgrove DO, Sloane JP, editors.
Ultrasound diagnosis of breast diseases. Edinburgh:
Churchill Livingstone, 1994:58–9.
12. Lee AH, Denley HE, Pinder SE, Ellis IO, Elston CW, Vujovic
P, et al. Excision biopsy findings of patients with breast
needle core biopsies reported as suspicious of malignancy
(B4) or lesion of uncertain malignant potential (B3).
Histopathology 2003;42:331–6.
13. Agoff SN, Lawton TJ. Papillary lesions of the breast with
and without atypical ductal hyperplasia: can we accurately
predict benign behaviour from core needle biopsy? Am J
Clin Pathol 2004;122:440–3.
849
BJR
The British Journal
of Radiology

November
2006
Volume 79
Issue 947
November 2006, Volume 79, Issue 947

● Government reform of the National Health Service: implications

for radiologists and diagnostic services

● Reporting overexposures and unintended exposures in

diagnostic procedures

● Are targeted contrast agents realistically going to reach the

clinic? Recent regulatory experience with targeted MRI contrast
agents

● The development and optimization of high spatial resolution MRI

for imaging the oesophagus using an external surface coil

● Pancreatic adenocarcinoma: signs of vascular invasion

determined by multi-detector row CT

● Conversion factor for CT dosimetry to assess patient dose using a

256-slice CT scanner

● Equivalent dose to organs and tissues in hysterosalpingography

calculated with the FAX (Female Adult voXel) phantom

● Comparison of radiation doses to patients undergoing standard

radiographic examinations with conventional screen–film
radiography, computed radiography and direct digital
radiography

● Dose escalation to combat hypoxia in prostate cancer: a

radiobiological study on clinical data

● Scatter from radiotherapy beams emerging from primary

barriers: an aid to bunker design
 ● The use of megavoltage cone-beam CT to complement CT for
target definition in pelvic radiotherapy in the presence of hip
replacement

● CT of thoracic lymph nodes. Part I: anatomy and drainage

● Book reviews

● Various holes and lesions

● Book reviews
The British Journal of Radiology, 79 (2006), 861–865

COMMENTARY

Government reform of the National Health Service: implications

for radiologists and diagnostic services
G W L BOLAND, MRCP, FRCR

Department of Radiology, Massachusetts General Hospital and Harvard Medical School,

Massachusetts General Hospital, Boston, MA 02114, USA

ABSTRACT. Demand for radiology services within the National Health Service (NHS)
continues unabated and current NHS operations cannot keep up with demand.
Therefore, to meet this demand, the government has decided to outsource a significant
number of investigations to the independent sector and will actively promote patient
referrals to the new government sponsored Treatment Centres as they become
available. This presents opportunities to patients, but threatens existing public sector Received 13 February 2006
providers (including doctors) as competition for radiology services may result in both Revised 25 May 2006
loss of patient referrals and revenue to these providers. This article is a personal opinion Accepted 14 June 2006
and will focus on the current challenges facing the provision of radiology services in the
DOI: 10.1259/bjr/80900968
NHS. I will suggest the possible negative outcomes for providers (NHS hospitals and
staff alike) and will offer strategies, tactics and tools that can be employed to counter ’ 2006 The British Institute of
the threat to their existing services. Radiology

The founders of the National Health Service (NHS)
intended that it was the responsibility of government to
provide free healthcare to all British citizens at the point
of delivery. However, they could not have foreseen the
dramatic medical innovations over the last 50 years, nor
their cost. Nor could they have predicted the massive
increase in demand for these services. Whether they
would have advocated the same mandate with this
knowledge in hindsight is open to debate, but the
relentless and increasing demands on clinical services
have undoubtedly demonstrated profound weaknesses
in the structure of the NHS. While this is true for most
clinical services, radiological services have been parti-
cularly exposed. Massive increases in demand have
meant prolonged waiting lists for essential radiological
services, with some patients having to wait months for
basic diagnostic and therapeutic procedures [1]. The
Audit Commission in 2002, for instance, demonstrated
average waiting times for outpatient MRI at 20 weeks
and for CT at over 6 weeks [1]. Although waiting lists for
these services are coming down, most consider waiting
lists for radiological services within the NHS to be
unacceptably long. Worse, the demand for services
continues unabated due to an ageing population and
because patients and their doctors see an ever increasing
value from these radiological services. Newer and better
imaging techniques combined with newer imaging
applications means that imaging is now increasingly
Address correspondence to: Dr Giles Boland, Department of
Radiology, White 270C, Massachusetts General Hospital, 55 Fruit
Street, Boston, MA 02114, USA. E-mail: gboland@partners.org.
used as a primary diagnostic tool [2, 3]. This in turn has
raised the stakeholder expectations from radiology for
hospital administrations, referring physicians and
patients alike. However, as the number of radiology
investigations performed today cannot meet demand
(and is likely to deteriorate further), an unsustainable
tension has arisen between the NHS providers of
radiology services (radiologists and radiographers)
and their customers, the referring physicians and
patients.
The cause of this crisis within radiology, however, is
not so clear. Many stakeholders, particularly the provi-
ders, advocate that the NHS has been significantly
and chronically underfunded, leading to too few radi-
ologists, radiographers and equipment (i.e. resources).
Undoubtedly there is some truth to this despite the
recent expansion in the number of trainee radiologists
[4]. Britain has some of the fewest radiologists per head
of population when compared with other Western
nations, approximately half that of Germany and one
quarter that of France, although it is interesting that
Britain also has one of the fewest number of examina-
tions performed per head of population in Western
nations, suggesting it would need fewer radiologists (the
UK performs 52 investigations per 100 head of popula-
tion annually compared with a European annual average
of approximately 100 investigations per 100 head of
population) [5]. Furthermore, government itself has
recognized that the NHS has been underfunded and is
currently injecting massive increases in cash into the
NHS in order to bring its expenditure on healthcare in
line with other Western European nations [6, 7]. In

The British Journal of Radiology, November 2006 861


England, the cash budget for the NHS has doubled since
1999 [6] with large increases to radiological services.
However, the real and sometimes perceived lack of
resources is too often used as the explanation for much of
the NHS’ problems in general, and particularly for
radiology, due to the high cost of its resources. Indeed,
attempts by recent governments and others to seek better
value out of existing resources have often been met with
resistance [8, 9–11].
However, without precise data to benchmark the
utilization of current resources, it is not realistically
possible to accurately argue one way or the other. The
Audit Commission of NHS radiological services in 2002
gave some insight as to what analysis of existing
resource utilization might show – that a significant
component of the shortfall in radiological capacity is due
to the poor use of its resources. This report demonstrated
major differences in the hours of operations of existing
equipment and of radiologist productivity [1]. Most CT
and half of MRI scanners were operational for less than
9 h day21 and examination throughput for the same
scanners can be less than 50% of that seen in the USA
[12]. Productivity rates for radiologists (the number of
investigations they interpret) vary significantly, with
some consultants being three times more productive
than others [1]. It has also been demonstrated that other
clinical services within the NHS also fail to use their
resources effectively [8]. As more accurate data become
available it will be increasingly evident to most informed
healthcare professionals (and probably the public) that
public sector NHS reform is desperately needed. Such
data will likely continue to point to marked variations in
resource utilization. It will therefore become increasingly
hard to deny the widespread inefficient use of existing
resources, an unacceptable situation considering the high
cost of these resources and that they are ultimately paid
for from public funding.
What is controversial, however, is what reform is
required, who should pay for it and what are the goals,
incentives and achievable outcomes? Since the 1980s
there has been a number of significant government
policy agendas that have attempted to address this
problem, but many have either run their course or been
overturned by succeeding governments. However, the
general concept of NHS reform and particularly market
reform have stuck, and the current government has been
actively seeking means to reduce waste and increase
capacity through more efficient use of limited resources
[6, 7, 13, 14].
This in turn has led to further tension amongst
healthcare providers, particularly the clinical staff, as
reform in healthcare is usually viewed with suspicion [8–
11]. Reform inevitably requires an evaluation and
measurement of current practices to establish the
strengths and weaknesses of those services. National
Health Service physicians have been generally resistant
to having their practices measured [8]. Not only
have they seen themselves historically as autonomous
professionals, some may fear that any measurement of
their services may expose significant discrepancies
(either personal or organizational). Indeed, some may
argue that one of the primary reasons why such little
data on resource utilization is available is because it
might expose such discrepancies and therefore give
862
G W L Boland
the doctor’s adversaries the data they need to force
change.
All this is, of course, understandable considering the
organizational structure of the NHS. It is generally
understood that it is not the role of doctors to measure
resource utilization including their own productivity;
rather it is left to the myriad of hospital managers that
have sprung up over the last 20 years [7, 9]. Furthermore,
if confronted with productivity data by their managers,
physicians may feel that these managers do not fully
understand the clinical issues that drive the healthcare
process and therefore consider their data of little
relevance. The result is that a situation of general
mistrust and lack of morale pervades the system,
whereby physicians feel over appraised and over
inspected by professionals who either do not have their
interest at heart, or more importantly, do not have the
patient’s interest at heart [8]. A relative stalemate has
therefore arisen over the last 10 years whereby the
interests of physicians do not necessarily meet those
of other stakeholders, particularly managers and
government civil servants. Interestingly, both sides
passionately feel that they are defending the interests
of the patients.
This lack of clear vision from NHS doctors to engage
and address the variances in resource productivity and
utilization in the public sector has forced the government
to address the problems for them. The government has
embarked upon an ambitious attempt to enforce better
utilization of its resources, which in turn is hoped to
increase the value of its healthcare system to its
population. First, the government insisted on a new
physician contract (albeit agreed to with significant
difficulty), which was partly an attempt to align
productivity with incentives [7, 8, 11]. Now it is
attempting to reduce waiting lists by outsourcing large
swathes of low risk outpatient surgery and diagnostics to
the independent sector, either through development of
specialized Treatment Centres currently under construc-
tion or directly to existing independent providers [15].
Indeed, the NHS is planning to purchase £1 billion of
diagnostic services directly from the independent provi-
ders and hundreds of thousands of MRI services are
already being outsourced to independent providers [6].
It is predicted that as much as 15% of the NHS funding
may be shifted to the independent sector over the next 2–
3 years as the government attempts to meet its political
objectives [7, 8].
Perversely, these events may initially be strongly
welcomed and encouraged by NHS healthcare provi-
ders, both managers and physicians alike, as they
attempt to manage the prolonged waiting lists and meet
the government targets imposed to help reduce patient
lists. Some physicians may actually feel that this scenario
may be a vindication of their position after all – namely,
the government is effectively agreeing to their long
standing opinion that more resources are needed – and
bolstered by the fact that the government has now
provided the necessary funding, albeit partially to the
independent sector.
While these beliefs may bring satisfaction to some
doctors in the short term, they may find that the
government has the ‘‘last laugh’’. Unless physicians
are willing to critically and objectively look at their
The British Journal of Radiology, November 2006
Commentary: Government reform of the NHS
organizations and departments and make significant
attempts to work collaboratively to improve perfor-
mance and productivity, their influence and even income
may be marginalized. Under the government’s propo-
sals, the funds to pay for these clinical services will
follow the patient to the institution that provides the
services – the independent sector [7]. If the government’s
intentions are played out, a large proportion of out-
patient imaging (perhaps up to 15% or more) may be
outsourced to the independent sector. While this may
ease the current waiting list crisis within hospitals, it will
inevitably deprive these hospitals of much needed
income down the road.
The effect of these government initiatives will equate
to competition in the conventional business sense. A
potential outcome for public sector hospitals that do not
undergo significant reform and compete may be loss of
valuable market share, which in turn will lead to lost
revenue. In any traditional business, competition, which
threatens its survival, is usually countered with restruc-
turing and reform in order to become more productive
and in turn provide similar services or products for less
cost (i.e. they become more competitive). Ideally, in the
process, restructuring may yield an increase in quality or
value in the services, a very favourable circumstance.
Within radiology, witness the impact that voice recogni-
tion technology can do to replace traditional transcrip-
tional services. Within a few months of implementation,
the financial costs are significantly less, but the value of
the radiologist services (the ability to deliver a timely
report) has greatly increased [16, 17]. The result is that
customers (i.e. patients and referring physicians) have
now gained a better product for less cost. It should come
as no surprise to anyone within the NHS that this is
exactly what the current government is trying to do.
While many healthcare professionals may not like to
think of themselves as conventional businesses, it is now
increasingly evident that they need to think and act that
way.
While any changes may be good news for patients,
they may not be good news for many doctors. Through
an idiosyncratic historical contractual arrangement,
doctors have been able to earn significant income
outside their primary work place (the NHS) within the
independent sector [18]. Indeed, this contractual
arrangement may be seen by some as having presented
a conflict of interest to senior doctors in the NHS,
in which longer waiting lists in their NHS practice
served to increase the likelihood that those patients
who could afford to (an increasing proportion) would
move to their more lucrative private practice. However,
this potential opportunity is under threat. If the
government’s reforms are successful, NHS waiting lists
may start to shorten significantly. This could adversely
affect many doctors who perform private practice.
Doctors are likely to see a dwindling of their private
practice revenues as more and more patients are either
treated or evaluated in the government sponsored
independent sector institutions or existing independent
providers (paid for from government revenue). Any
doctor thinking that they may benefit from the newly
outsourced business to the independent sector is
likely to be disappointed. The government has stipulated
that most of the new opportunities in the independent
The British Journal of Radiology, November 2006
sector may not be available to NHS employees or if
they are, there will be significant limitations to the
opportunities.
What are radiologists and other doctors to do in this
unpredictable environment? One option is to do little,
assuming that this is yet another government attempt at
healthcare market reform that will run its course. While
this is possible, the evidence is against this outcome.
The government has allocated vast new sums of money
towards this agenda and is determined for it to succeed.
Those doctors who fail to respond creatively and
positively risk being left behind. Not only will the
doctor’s influence be marginalized, it may be too late
for them to save their private practice income should
the government reforms be successful. Besides, even if
the government initiatives fail, positive action by
doctors to improve productivity and efficiency in the
public sector should bring strong benefits to their
organization and therefore the patients they serve. No
one can argue that delivering timely efficient high
quality healthcare to patients is not something worth
striving for.
A more sensible option therefore would be for
doctors to rapidly understand what changes are happen-
ing, why they are happening and what they can do to
respond to these changes. First and foremost, they need
to recognize and accept that this is a serious effort by the
government to increase the value of the healthcare
system to its population. While doctors may not
necessarily agree with these initiatives, they need to
replace their ambivalence and inertia by recognizing that
health systems in 2006 effectively have to act along
efficient business lines. Those who remain entrenched,
believing that clinical care and business do not mix,
may have a lot to lose. Doctors need to rapidly gain an
understanding of the business fundamentals required
to efficiently operate their specialty. For radiologists, it
is vital they develop an understanding of effective
resource utilization and mechanisms to improve depart-
mental productivity. After all, they are partly responsible
for managing and operating some of the highest cost
items of capital equipment within their organizations.
Typically this will require analysis of their existing
assets to determine current operational productivity.
Once this is known, it should then be possible to
benchmark their operation against the best practices in
the industry. The analysis will inevitably require opera-
tional changes to increase productivity from their
existing assets, i.e. greater patient throughput. While
some additional resources may be needed to make
this happen, the costs are minimal compared with the
overall benefit to the organization. For example, in
radiology, two radiographers can perform more than
twice as many CT scans as a single radiographer (and
possibly even three times) [19, 20]. The number of
patients scanned each day per CT scanner could increase
from approximately 25 per day to over 90 patients per
day (current volume of the busiest scanners at the
Massachusetts General Hospital, Boston, MA). The cost
of an additional radiographer is minimal compared
with the thousands of additional CT scans that can be
performed annually. Furthermore, increasing operations
into the weekend and evening hours will further
increase capacity and help to significantly shorten
863

waiting lists. Analyses of many other parameters should
also be performed including length of waiting lists, time
to perform the procedure, report turnaround time,
radiologist’s productivity and costs per examination,
amongst others [19].
Such an analysis of the operations will then permit
radiologists to devise strategies and the tactics to counter
the new threats and position their organization to
compete with the independent sector. While radiologists
may prefer to delegate analysis of their operations to
their managers, there is unlikely to be broad acceptance
by doctors for any changes that are recommended by
other groups. Besides, because doctors are on the ‘‘front
line’’ of the healthcare delivery cycle, they should be
intimately involved in developing solutions to improve
their operations. Doctors therefore need to become
more business-like and understand their operations
through careful data analysis and then cooperate with
managers to maximize capacity and utilization of their
fixed assets.
Inevitably this will require physician leadership,
which will in turn require greater knowledge, which
for many will require further education and training. It
should not, of course, be assumed that managers under-
stand appropriate asset management either and further
training may be needed by them to understand how they
can help maximize their organization’s resources (staff
and equipment). Doctors and managers will both need to
learn to better cooperate with each other to improve the
productivity and efficiency of their department, thereby
creating a more valuable product for their organization.
While this process is not easy (and typically change for
some doctors has been difficult), it is essential that
physician leaders respond to the challenge and convince
their peers that it is in their best interest to change. This
will require particular skill, particularly as many radi-
ologists will immediately demand to know who is going
to read the additional studies generated by any increase
in productivity, particularly due to the relative shortage
of radiologists.
The answer to this conundrum is not at all clear given
the current contractual arrangement within the NHS.
There has structurally been little incentive for doctors in
the NHS to work towards increasing productivity and
services within the NHS. Some may, of course, argue that
this is one of the NHS’ strengths; a system that provides
financial incentives to doctors to perform more investi-
gations on a fee for service basis is likely to be more
costly [21]. However, it is extremely hard to significantly
change organizational behaviour without offering some
form of incentive. Unfortunately, attempts by some NHS
physicians to help improve productivity and services
have frequently been met with resistance from their
managers who fear the increase in costs that may occur,
even if marginal. However, Trusts are becoming more
financially autonomous, so it is quite conceivable that
radiologists could negotiate additional pay for the
additional studies interpreted, perhaps on a modified
fee for service basis. Besides, this financial arrangement
has existed in the independent sector for many years.
Furthermore, radiologists will be unlikely to participate
in a service improvement program without some form of
incentive, even if it is the right thing to do. It is therefore
864
G W L Boland
strongly recommended that managers seriously
consider incentive schemes for radiologists who add
value to their organizations and help to shorten patient
waiting lists.
In summary, in order to bring greater value from the
health service to its citizens, the British government is
actively creating a publicly funded independent or part-
independent sector that is likely to directly compete with
the public sector NHS. A likely intention by the
government, perhaps covertly, is to force public sector
NHS organizations and particularly doctors to adapt to
this new competitive threat. Those doctors who ignore or
even resist these changes may have a lot to lose both
in influence and financial reward. However, those
doctors who respond creatively with knowledge, skill
and leadership to maximize the productivity and value
of radiology services within their organizations are
likely to receive greater respect, recognition and prob-
ably financial reward. This in turn should increase the
value that radiologists bring to their department, their
organization, but most importantly, to the patients they
serve.
References
1. Radiology. Review of National Findings. Audit
Commission – United Kingdom July 2002 no. 7.
2. Wolbarst AB, William R, Hendee WR. Evolving and
experimental technologies in medical imaging. Radiology
2006;238:16–39.
3. Thrall JH. Reinventing radiology in the digital age part III.
Facilities, work processes, and job responsibilities.
Radiology 2005;237:790–3.
4. Board of the Faculty of Clinical Radiology. The Royal
College of Radiologists. Changing working lives. London:
Royal College of Radiologists, 2005.
5. Board of the Faculty of Clinical Radiology. The Royal
College of Radiologists Clinical Radiology: A workforce in
crisis. London: Royal College of Radiologists, 2002.
6. High Anxiety. The NHS’s financial difficulties. Economist
Magazine Dec 20th 2005.
7. Pollock AM. NHS plc: The privatisation of our health care.
London: Verso, 2004.
8. Klein R. Britain’s National Health Service revisited. N Engl J
Med 2004;350:937–42.
9. Dawson P. The British National Health Service. J Am Coll
Radiol 2004;1:287–92.
10. Hargreaves S. A quarter of consultants ready to resign over
new contract. BMJ 2003:326:569.
11. Doctors warn new contract will herald the death of the
NHS. Guardian Newspaper, London. May 21, 2003.
12. Boland GWL, Palumbo D, Tabor-McEwen K, Harlem P,
Pathak A. CT productivity differences between the UK and
USA: opportunities for the National Health Services. RSNA
Annual Scientific Meeting 2004;231(D):459.
13. Shapiro J. Can Labour take the NHS to market? BMJ
2005;331:359.
14. Steven S. Reform strategies for the English NHS. Health
Affairs 2004;23:37–44.
15. Private treatment centres to expand. Guardian Newspaper,
London. January 8, 2005.
16. Amit Mehta. Voice-Recognition Technology. The primary
benefits of voice recognition have been decreased
turnaround time and enhancement of the capabilities of
PACS. In: Imagingeconomics.com [Accessed 13 September
2006].
The British Journal of Radiology, November 2006
Commentary: Government reform of the NHS
17. Mardini M, Mehta A. Voice recognition. In: PACS:
A guide to the digital revolution. Springer Publications,
2005.
18. Bulstrode C. Embarrassing greed. BMJ 1995;310:198–9.
19. Boland GWL. Stakeholder expectations for radiologists:
obstacles or opportunities. J Am Coll Radiol (In press -
JACR-D-05-00165).
The British Journal of Radiology, November 2006
20. Boland GWL, Meehan M, Doncaster R. Three technologist
model for CT: cost implications and benefits. RSNA Annual
Scientific Meeting 2004;231(D):365.
21. Maitino AJ, Levin DC, Parker L, Rao VM, Sunshine JH.
Practice patterns of radiologists and nonradiologists in
utilization of noninvasive diagnostic imaging among the
Medicare population 1993–1999. Radiology 2003;228:795.
865
 The British Journal of Radiology, 79 (2006), 866–869

COMMENTARY

Reporting overexposures and unintended exposures in

diagnostic procedures
P R CLEWER, MPhil, MSRP, MIPEM and P C JACKSON, BSc, PhD, FIPEM

Department of Medical Physics and Bioengineering, Southampton General Hospital, Southampton,

UK

ABSTRACT. Legislation requires the reporting to regulatory authorities of incidents in

which patients have been exposed to ionizing radiation to an extent ‘‘much greater
than that intended’’. The authorities have published guidance on what is considered to
meet this requirement. However, there is still some confusion regarding, particularly,
the necessity to report some unintended doses. It is believed that there is a
disproportionate amount of resource spent investigating some unintended exposures Received 24 January 2006
because all such exposures will have an effective overexposure factor of infinity, Revised 5 June 2006
irrespective of the magnitude of the dose and the associated risk. This paper proposes Accepted 16 June 2006
changing the definitions of ‘‘overexposure’’ and ‘‘unintended exposure’’ and the
DOI: 10.1259/bjr/68149575
adoption of a reporting process based upon risk assessment. All records and data would
be collected and, if required, reported, but investigation of individual incidents would ’ 2006 The British Institute of
take place only for incidents carrying a greater risk than 1 in 10 000. Radiology

For many years now, the medical radiation protection
community has lived with the legal requirement that
when it is suspected that a patient has been exposed to
ionizing radiation to an extent ‘‘much greater than that
intended’’ due to equipment malfunction or defect, the
employer must notify the Health and Safety Executive
(HSE) forthwith. The requirement was stated in the
Ionising Radiation Regulations 1985 (regulation 33(2))
and in the replacement Ionising Radiations Regulations
1999 (regulation 32(6)). The HSE published guidance [1]
on what should be considered to be ‘‘much greater than
intended’’. The HSE’s document included a table of
‘‘guideline multiplying factors’’ and stated that if the
ratio of the suspected exposure to the intended exposure
is greater than or equal to the appropriate multiplying
factor, it should be presumed that HSE expects to be
notified of the incident. The multiplying factors for
diagnostic examinations were 3, 10 or 20, depending on
the magnitude of the intended effective dose. For
instance, an incident where a patient undergoing a chest
X-ray receives an exposure of 20 times or more than
intended, the HSE should be notified. For any CT
examination, the ‘‘reporting threshold’’ was set at three
times the intended exposure. The factors also applied to
situations where equivalent multiples of patients
received doses greater than intended since such situa-
tions might arise with equipment malfunctions.
In 2000 the introduction of the Ionising Radiation
(Medical Exposure) Regulations 2000, known as
Address correspondence to: Mr Philip Clewer, Medical Physics and
Bioengineering, Southampton General Hospital, Tremona Road,
Southampton SO16 6YD, UK. E-mail: philip.clewer@suht.
swest.nhs.uk.
IR(ME)R, brought a similar requirement for notification
of incidents when a patient is exposed to ionizing
radiation to an extent ‘‘much greater than intended’’,
but due to any reason other than equipment malfunction
or defect (regulation 4(5)). In this case, the notification is
to the Department of Health (DoH). The guidance [2]
issued with IR(ME)R stated that application of the HSE
guidance on which doses are likely to be much greater
than intended is appropriate. Therefore, for example,
procedural failures that lead to a patient receiving 20
times the intended exposure from a chest X-ray should
be reported to the DoH.
Paragraph 6.8.2 of the same guidance stated that
‘‘patients who undergo a procedure that was not
intended ... should be considered as having received an
unintended dose of radiation’’. The introduction of the
phrase ‘‘unintended dose of radiation’’ has caused some
confusion because it appears to imply that unintended
doses are separate to events described as ‘‘much greater
than intended’’. This has raised some anomalies that lead
to the question, ‘‘what is the purpose of reporting these
incidents?’’ – a question to be addressed later in this
paper.
It appears that the DoH intended to publish guidelines
that would be aimed specifically at incidents reportable
under IR(ME)R rather than reportable under IRR99.
These guidelines have not been forthcoming, but are still
expected [private communication with DoH, 2005 and
2006].
For the purposes of this paper the authors will take the
intended effective dose to be the ‘‘typical effective dose’’
published in 1999 by the National Radiological
Protection Board (now the Radiation Protection
Division of the Health Protection Agency) [3].

866 The British Journal of Radiology, November 2006

Commentary: Reporting overexposures and unintended exposures
Overexposures and unintended exposures
If a patient is referred for a plain film chest X-ray, then
the intended dose is 0.02 mSv.
The guideline reporting threshold for an adverse event
for such an examination is 20 times the intended dose.
Therefore if the patient received 19 times the intended
dose there is no requirement to report the incident,
although the Employer may wish to do so. This means
that the patient could receive an effective dose of
0.36 mSv over and above the intended 0.02 mSv, giving
a risk of a fatal cancer of around 1 in 55 000, using the 5%
per Sv risk factor model [4].
If a patient who is not intended to undergo any X-ray
examination actually undergoes a chest X-ray for some
reason, e.g. misidentification, the effective dose to that
patient is 0.02 mSv with a risk of around 1 in 1 000 000 of
a fatal cancer. Although the overexposure factor could be
said to be infinity, the actual risk to the individual
patient is much lower than the 19 times-the-intended-
dose incident, but this incident is theoretically reportable
because it was an ‘‘unintended exposure’’.
From a patient perspective this may seem confusing in
that the error with a lower risk of harm is reportable,
whereas the error with a relatively higher risk of harm is
not subject to that requirement.
It is clear that an error is reportable if the overexposure
factor is above the relevant threshold for reporting, e.g.
someone has a pelvic radiograph (effective dose 0.7 mSv)
when a chest radiograph (effective dose 0.02 mSv) was
intended. In this example the overexposure is 0.72/0.02
(536) (assuming they go on to have the intended
examination), i.e. greater than 20 and therefore repor-
table. However, if the sequence of events was the other
way around the overexposure factor would be 0.72/0.7
(51.03). This would not be over the reporting threshold
as far as an overexposure is concerned, but the chest
radiograph would have been unintended and so the
incident is reportable for this reason rather than over-
exposure.
For instance, if a patient undergoes an incorrect
examination, in practice this will lead to a ‘‘greater than
intended’’ dose rather than an unintended dose since
both the ‘‘incorrect’’ and ‘‘correct’’ examinations will
have been performed. In this case, reporting to the DoH
would be dependent on the magnitude of the over-
exposure.
The DoH currently regards any case of wrong
examination, such as ‘‘left ankle’’ instead of ‘‘right
ankle’’, to be an ‘‘unintended exposure’’ and therefore
reportable, regardless of magnitude. This means that
hospitals have to report unintended doses even if they
result in a dose and risk much lower than the dose and
risk to patients who experience overexposures that do
not exceed the reporting threshold. It appears that some
hospitals are unclear as to the reporting requirements –
some do report the ‘‘wrong ankle’’ examination whereas
others do not.
The future
In 2003, the HSE published proposals for a document,
HS(G)226, to replace PM77. One of the significant
The British Journal of Radiology, November 2006
proposals was that the guideline multiplying factor for
all diagnostic procedures should be reduced to 1.5.
Therefore all occasions when a repeat X-ray exposure is
required due to equipment defect or malfunction would
require reporting to HSE. This proposal attracted
considerable comment and when the final document
was published in March 2006 [5], the guideline multi-
plying factors for diagnostic radiology and nuclear
medicine were divided into three bands as follows:
N 1.5 times intended dose – interventional radiology,
radiographic and fluoroscopic procedures involving
contrast agents, nuclear medicine with intended
effective dose . 5 mSv and CT examinations
N 10 times intended dose – mammography, nuclear
medicine with intended effective dose # 5 mSv but
. 0.5 mSv, everything else not referred to elsewhere
N 20 times intended dose – radiography of extremities,
skull, dentition, shoulder, chest, elbow, knee and
nuclear medicine with intended effective dose
# 0.5 mSv.
Hence the factors for middle and low dose examina-
tions will remain unchanged, as do those for therapeutic
procedures, but the higher dose diagnostic procedures
have been given a guideline multiplying factor of 1.5.
Therefore any repeat exposures in this category are
reportable. The DoH has indicated that, certainly for the
time being, they will not adopt the 1.5 times factor, but
will continue with the 3 times factor for high dose
diagnostic procedures [private communication with
Department of Health, 2006].
The above arrangements are appropriate to situations
where there has been equipment failure, but should the
same approach be applied to the individual patient?
Why report?
Perhaps it would be useful to return to the question
‘‘why do we report incidents?’’.
IR(ME)R implement most of the requirements of the
Medical Exposures Directive [6], but the Directive does
not require member states to implement a reporting
system. However, it does require in Article 11 that
member states should ensure that ‘‘all reasonable steps
to reduce the probability and the magnitude of acci-
dental or unintended doses of patients from radiological
practices are taken...’’. Perhaps this is in part a solution to
the confusion; incident reporting can lead to prevention
of similar occurrences by building up statistical evidence
of where problems lie and where attention should be
focused.
The Directive also states that ‘‘the main emphasis in
accident prevention should be on the equipment and
procedures in radiotherapy, but some attention should
be paid to accidents with diagnostic equipment’’.
This implies a need to concentrate on preventing
radiotherapy incidents because of their relatively high
doses, and therefore risks, compared with diagnostic
procedures.
In practice there are two reasons for reporting
incidents, be they overexposures or unintended
exposures.
867

The first of these, as mentioned above, is to produce a
national database of statistics on incidents, showing
where the problems lie and which areas of practice
require more attention to reduce the likelihood or
magnitude of incidents in future. However, it is believed
that this could be done more efficiently by employers
being required to supply their own statistical informa-
tion regularly, perhaps annually.
The second reason for reporting incidents is that
significant errors can be fully investigated by an
independent body and appropriate actions considered
in relation to employees and/or employers. These may
lead to an investigation by the HSE or DoH, perhaps
leading to an improvement notice or prosecution. Such
incidents obviously require an individual notification to
the appropriate regulatory body.
The following is proposed as a way of combining ease
of reporting for statistical purposes and the need for
‘‘enforcement’’ reporting:
N Requirements for reporting should be redefined in
terms of effective dose. Exposure is too vague a term
for use in this context. Although there may be some
controversy regarding the relationship between risk
and effective dose, this would at least take into
account the tissues and organs of the body that are
irradiated. In fact, the PM77 does use effective dose in
its table of guideline multiplying factors while the
regulations refer to overexposures.
N ‘‘Overexposure’’ should be redefined as ‘‘overdose of
radiation’’ and related to the occasion of a patient who
it is intended should receive a dose receiving more
than intended. By using the effective dose quantity, an
overdose would include incidents where the patient
initially received an examination of the wrong part of
their body and then the correct examination. It would
be intended that they receive an effective dose, but
because they receive an additional dose they have
been given a greater dose than intended. Therefore a
patient who attends for a radiograph of the pelvis but
receives a chest radiograph in error, followed by the
intended pelvis radiograph, or vice versa, is the
subject of an overdose of radiation, whichever way
round the events occur – although the magnitude of
the overexposure factor is very different.
N The system of published guideline multiplying factors
for reporting of overdoses of radiation could be
retained, but the authors suggest that an alternative
system is adopted for reporting unintended doses as
outlined below.
N ‘‘Unintended exposure’’ should be redefined as
‘‘unintended dose’’ and related to the occasion of a
patient who it is intended should not receive a
radiation dose receiving a dose. This would clearly
identify those patients who should not have received
any dose of radiation but who have for some reason,
usually mistaken identity, done so.
N A threshold should be set, in terms of effective dose
above which unintended doses should be reported to
the regulatory authority individually and urgently.
Below the threshold the incident should be logged by
the employer and information sent to the regulatory
authority at regular intervals as nationally agreed.
Setting a threshold would enable appropriate effort to
868
P R Clewer and P C Jackson
be directed towards investigating and preventing re-
occurrence of the more serious incidents in a timely
manner. Of course, incidents involving doses below
the threshold could still be reported individually if the
employer believes there is merit in doing so. It is
accepted that the public may view the setting of a
threshold as condoning error and employers not being
held to account to a higher authority. However, this
situation might be addressed by the Healthcare
Commission having access to the adverse event
reporting of radiation events.
If this philosophy is accepted, the next consideration is
the setting of the reporting threshold for unintended
doses and reviewing the threshold for reporting over-
doses. Many values could be proposed, but before doing
so it may be instructive to look at risks associated with
other areas of life. In ‘‘Living with Radiation’’ [7], the
National Radiological Protection Board published com-
parative risk data for the annual risk of death in the UK
from some ‘‘common’’ causes. This included a 1 in
100 000 annual risk of being murdered, a risk of any
40-year-old individual dying from any cause as 1 in 700
and the annual average risk of death from cancer being 1
in 400.
If the guideline multiplying factor of 20 for low dose
examinations, which includes intended effective doses
up to 0.5 mSv as described in PM77 [1], were to be
retained then the threshold for reporting unintended
doses could be set at 2060.5 mSv510 mSv. However,
this does seem relatively rather a high dose for reporting
individual incidents of this nature and is above the dose
level of a CT chest examination. Another approach
would be to set the threshold at 20 times a chest X-ray
dose. This would set the threshold at 0.4 mSv and
require individual reports of incidents involving many
more types of diagnostic examination. This type of
approach is subject to change in the effective dose of
examinations over time and subjective assessment of the
acceptability of a threshold. Alternatively, we could
consider who is actually receiving this dose of radiation.
A hospital patient receiving an unintended dose should
not have undergone a medical exposure and is no
different to any other member of the public. In 1993, the
National Radiological Protection Board published gui-
dance [8] relating to the latest international recommen-
dations and included a recommendation for a public
dose constraint for a single new source [of ionizing
radiation] of 0.3 mSv per year. This dose, equivalent to a
lateral lumber spine X-ray examination [4] and with a
risk of a fatal cancer of 1 in 67 000 using the 5% per
Sievert model, could be suggested as an appropriate
reporting threshold for incidents where someone who
should not have received a dose at all has been
irradiated. This approach is perhaps less than acceptable
in that it appears that a patient should not accept any
greater risk of harm than a member of the public. This
approach is not acceptable as, de facto, all forms of
medical intervention have some risk. Apart from the
unavoidable risk associated with the intended interven-
tion, there is also a risk of an adverse event involving the
patient but unrelated to the patient’s condition.
The HSE’s approach to risk and benefit in the nuclear
power industry may give some guidance. The HSE has
The British Journal of Radiology, November 2006
Commentary: Reporting overexposures and unintended exposures
looked at the boundary between a tolerable risk (of
death) and an unacceptable risk to both workers and the
public from the use of a nuclear power station. For the
public who have the risk imposed on them ‘‘in the wider
interest of society’’, the boundary is said to be 1 in 10 000
per annum [9]. It would be difficult to maintain that the
use of X-rays for medical diagnosis is not ‘‘in the wider
interests of society’’ and so there will always be a risk of
unintended exposures. Using the 5% per Sievert model, a
risk of 1 in 10 000 is equivalent to an effective dose of
2 mSv, similar to that from a CT scan of the head and not
very different from the average annual ‘‘background’’
dose to a citizen of the UK. This is still a reasonably high
level and is likely to mean that no plain film unintended
doses would require reporting. However, putting this
into context, a lifetime risk of 1 in 10 000 of a fatal cancer
is still very small in comparison with the natural risk of a
fatal cancer of about 1 in 4 [10].
Table 1 summarizes the possible unintended dose
reporting thresholds and reasons for their adoption.
The authors propose a threshold of 2 mSv, above
which employers should make a prompt individual
report regarding unintended doses of radiation. Below
this threshold the employer should retain statistical data
for forwarding to the regulatory authority as required.
By choosing this threshold, society is not viewing the
risks associated with medical examinations involving
ionizing radiations as being higher than those from the
nuclear industry, rather that adopting such an approach
has harmonized a consistent approach to radiation risks.
Conclusion
The present system of reporting overexposures and
unintended exposures can lead to confusion and the
benefits are unclear. The current use of the linear no
threshold dose–risk model means that all radiation doses
should be as low as reasonably achievable. However, it
does seem an inappropriate use of resources to give the
same effort in investigating and reporting unintended
doses to, say, the ankle as a CT scan of the pelvis. This
paper proposes alternative definitions to ‘‘overexpo-
sure’’ and ‘‘unintended exposure’’ and the adoption of a
reporting threshold for unintended doses. Unintended
doses above this threshold would be the subject of an
individual report to the relevant regulatory authority.
Unintended doses below the threshold would be
The British Journal of Radiology, November 2006
Table 1. Possible unintended dose reporting thresholds
Reporting Argument for adoption
threshold (mSv)
10 20 times the 0.5 mSv intended dose level
from PM77
0.4 20 times the chest X-ray effective dose
0.3 Public dose constraint
2 Risk of 1 in 10000 of a fatal cancer
recorded by the employer and a statistical report made
to the regulatory authority at appropriate intervals.
References
1. Health and Safety Executive. Fitness of Equipment Used for
Medical Exposure to Ionising Radiation, HSE PM77.
London, UK: HSE, first edition 1992, second edition 1998.
2. Department of Health. The Ionising Radiation (Medical
Exposure) Regulations (2000 together with notes on good
practice). London, UK: available in electronic format only
at www.dh.gov.uk/assetRoot/04/05/78/38/04057838.pdf
[Accessed 3 July 2006].
3. National Radiological Protection Board. Guidelines on
Patient Dose to Promote the Optimisation of Protection
for Diagnostic Medical Exposures, Documents of the NRPB,
vol. 10. no. 1. Didcot, UK: NRPB, 1999.
4. International Commission on Radiological Protection. 1990
Recommendations of the International Commission on
Radiological Protection, ICRP Publication 60. Oxford, UK:
Pergamon Press, 1991.
5. Health and Safety Executive. Equipment used in
Connection with Medical Exposure. Guidance Note PM77
(third edition) March 2006 on HSE website: www.hse.
gov.uk/pubns/guidance/pm77.pdf [Accessed 3 July 2006].
6. Council of the European Union. Council Directive 97/43/
Euratom of 30 June 1997 on health protection of individuals
against the dangers of ionising radiation in relation to
medical exposure. The Council of the European Union,
1997.
7. National Radiological Protection Board. Living with
Radiation. National Radiological Protection Board. Didcot,
UK: NRPB, 1998.
8. National Radiological Protection Board. Board Statement on
the 1990 Recommendations of ICRP, Documents of the
NRPB, vol. 4, no. 1. Didcot, UK: NRPB, 1993.
9. Health and Safety Executive. Reducing Risks, Protecting
People. Norwich, UK: HMSO, 2001.
10. Cancer Research UK website: http://info.cancerresearchuk.
org/cancerstats/mortality/ [Accessed 3 July 2006].
869

COMMENTARY
Are targeted contrast agents realistically going to reach the
clinic? Recent regulatory experience with targeted MRI contrast
agents
A DZIK-JURASZ, PhD, FRCS, FRCR
Cambridge, Massachusetts, USA
Developments in the visualization of molecular events
in vivo are progressing at an unprecedented rate, a
characteristic of rapidly developing or emergent tech-
nologies. The need to target the molecular mechanisms
of disease has been the driving force behind this rapid
growth in knowledge whilst the bridging of disciplines,
such as imaging science, chemistry and biology, in some
cases with a good deal of entrepreneurship, has made
such technological advances possible. At the vanguard is
the field of targeted molecular imaging agents. Targeted
molecular imaging agents are now used in biological,
medical and pharmaceutical research where they are
proving to be powerful tools in the investigation of
experimental in vivo systems. It is almost taken for
granted that these technologies will translate readily into
clinical practice, providing unique molecular informa-
tion on disease. But will targeted molecular contrast
agents reach the clinic and deliver the benefits expected
of them? Journals, conferences and societies are now
dedicated to molecular imaging and targeted contrast
agents. But to make an agent available to widespread
clinical use requires regulatory approval from either the
Food and Drug Administration (FDA) in the USA or the
European Agency for the Evaluation of Medicinal
Products (EMEA) in Europe. The main purpose of these
agencies is to ensure that safe and efficacious agents
reach the market while denying access to agents that do
not meet these standards.
In defining targeted molecular imaging agents I am
specifically referring to agents that will be used widely in
the clinic and be prescribed because of their role in
guiding clinical management. This is quite different from
how such agents are currently used in, for example,
probing cell biology or validating drug targets. Positron
emission tomography (PET) tracers are making an
impact in drug development and are likely to increase
in use, but the widespread translation of these
870
The British Journal of Radiology, 79 (2006), 870–872
Received 6 April 2006
Revised 25 July 2006
Accepted 31 July 2006
DOI: 10.1259/bjr/68807002
’ 2006 The British Institute of
Radiology
techniques into the clinic is unlikely to occur. Also
excluded are the niche applications such agents might
have in the manner of in vitro biomarkers [1].
Nevertheless, the medical community has high expecta-
tions of targeted molecular imaging. Witness the invest-
ment by national, private and business institutions across
Europe and the USA. But if targeted contrast agents are
increasingly viewed as potentially important clinical
biomarkers, the role and value of using these agents
remains to be established, and in this commentary I
discuss in brief whether there is a realistic prospect of
this occurring.
Clinically, molecular imaging has been successfully
applied for several decades via radionuclide imaging
and, more recently, via modalities such as magnetic
resonance [2, 3] and ultrasound. To address whether
targeted contrast agents will reach the clinic requires an
understanding of the process behind the development of
a contrast agent. A contrast agent, despite rarely being
administered more than once, is a drug and is treated as
such by regulatory authorities. Prior to approval for
clinical use, a contrast agent would have been guided
through the discovery and development phases by a
biotechnology or large pharmaceutical company, often in
collaboration with academia. The processes require the
coordination and cooperation of multidisciplinary teams.
The financial outlay in bringing a contrast agent from the
laboratory to the clinic is in the order of $100–150 million.
This is a small sum compared with a therapeutic drug
whose development costs, depending on the manner in
which the data are interpreted, are reported to be on the
order of $800 million [4]. However, the time required to
develop such a contrast agent is no less than a
therapeutic; it takes at least 8–10 years to bring any drug
to market. More importantly, diagnostic imaging agents
are very unlikely to return the type of profits seen with
therapeutic drugs [5]. It is difficult to imagine how
The British Journal of Radiology, November 2006
Commentary: Targeted contrast agents in the clinic?
molecularly targeted agents are going to bring an equal
or greater return than the non-specific imaging agents
currently available in the clinic. The cost of targeted
imaging agents may therefore be higher than expected,
and this effect could be compounded by increasing price
controls being imposed on drugs by governments.
Any drug must demonstrate safety and efficacy in
order to gain regulatory approval, and much of the drug
development process is geared toward achieving this
goal. The most recent regulatory review of a MRI agent
was that of EPIX Pharmaceuticals, Inc. (Cambridge, MA).
An NDA (new drug application) had been filed in
December 2003 with the FDA, supported by four phase 3
international multicentre trials. The drug (VasovistTM) is
a blood pool agent targeted to human serum albumin,
and is intended for use as an MR angiography agent. The
FDA’s fundamental concern was with the efficacy data.
My personal view (the author was the head of imaging at
EPIX following the original approvable letter) is that the
conflict between the FDA and the company can be
distilled to two issues, namely (1) to what level is it
reasonable and necessary to deconstruct a clinical
experiment (in the manner of a laboratory bench top
experiment)? And (2) what is a reasonable level of
experimental standardization? These two issues are
intimately related and affect how widely applicable the
results are in clinical practice. The company unsuccess-
fully argued that clinical practice at individual sites or
the equipment manufacturer’s sequence recommenda-
tions are a de facto basis of standardization (and showed
that the variation in acquisition techniques was in fact
very small and did not significantly affect the results).
The FDA declined to accept this argument on the
grounds that the efficacy results might still be biased in
favour of the drug and therefore did not approve the
compound (the actual term was ‘‘approvable’’, but the
result is that the compound is not approved for use in
man in the USA). Further details can be found in the
press release section of EPIX’s website at www.
epixpharma.com.
The role of the FDA is to protect the health of the USA
by approving, on a scientific basis, safe and effective
drugs. But it also has the clear duty of not withholding
beneficial drugs from the public. More often than not
there is a fine balance between these two alternatives,
especially at a time of heightened public and political
scrutiny of the drug development business. The
European authorities approved the drug in September
2005 after a 15-month review utilizing the same clinical
data submitted to the FDA.
This is not to say that there is a fundamental difference
in the goals of EMEA and the FDA. Although organiza-
tionally different, EMEA differs little from the FDA with
respect to the challenge of evaluating new targeted
contrast agents. One fundamental difference is that
EMEA’s decisions tend to be more reliant on the review
of impartial outside experts for the technical assessment
of an application. EMEA, for example, has a document
analogous to the FDA’s guidance document on imaging
agents ‘‘Points to consider on the evaluation of diag-
nostic agents’’. Of note is the European position that
‘‘…there are no generally accepted principles for the
evaluation of efficacy of diagnostic agents’’, indicating it
The British Journal of Radiology, November 2006
is for the sponsor to provide convincing clinical scientific
evidence of efficacy.
I remain troubled that the American public has been
denied a valuable diagnostic agent. No one will deny the
importance of experimental standardization, and stan-
dardization is straightforward when dealing with the
physical parameters that define radionuclide or X-ray
based imaging. Standardization is subtler when dealing
with MRA where there is no meaningful standard
comparator, and there are several different, but equally
representative ways to approach imaging inflowing
blood. In the EPIX trials, VasovistTM was compared
against blood flow techniques such as time-of-flight and
phase contrast angiography. EPIX’s argument was that
the drug was effective in clinical practice, which is a
legitimate means by which to seek approval. Indeed,
adopting arbitrary standardized parameters, in the case
of flow-sensitive techniques when there is no proven
superiority among several approaches, might itself result
in unexpected bias. Those in the field of MRA and
particularly those who are involved in drug-related
studies should oppose where appropriate the unilateral
imposition of views that do not take into account the
complexities of the modality being studied.
Whatever the merits of this case, this series of events
provides insight into the regulatory hurdles to the
delivery of a drug/imaging agent to market. Such
setbacks are ultimately resolvable, but remain a concern
for industry strategists and decision makers who finance
investments in research and development (R&D).
The recent review of another MRI contrast agent
(Combidex; Advanced Magnetics, Inc., Cambridge, MA;
www.advancedmagnetics.com) highlights additional
regulatory challenges faced by an imaging agent.
Combidex, an iron oxide nanoparticulate, was issued
an ‘‘approvable’’ letter by the FDA following review by
an advisory committee, but for very different reasons
from EPIX. An advisory committee consists of a panel of
experts called by the FDA to make a recommendation,
usually on one or a small handful of questions that
guides the Agency’s decision making. Details of the
FDA’s advisory committee (which sat in April 2005)
on the subject can be found at http://www.fda.
gov/ohrms/dockets/ac/05/transcripts/2005-4095T1.pdf.
Bearing in mind that submission for regulatory approval
encompasses a substantial body of work including
chemistry, manufacturing, toxicity, pre-clinical and clinical
data that are typically the sum of 8–10 years of work,
Combidex remains unapproved because both the FDA
and the independent advisory committee criticised aspects
of its safety profile, requesting too broad an indication and
in providing limited efficacy data. These two cases as
indices of the regulatory climate for imaging agents in man
have left a sense of considerable unease in the world of
diagnostic imaging agent development.
Efficacy, safety and regulatory issues not withstand-
ing, additional technical issues for consideration in the
development of a targeted contrast agent include
identifying a target of interest, recognizing its biological
significance and assessing the likelihood of discovering
and synthesising an agent that will bind specifically to
and be at sufficient concentration at the target site to
generate a recordable signal. Any of these issues could
prove to be as problematic as any regulatory, efficacy or
871

safety issue. But in part, the role of the translational
radiologist in collaboration with scientific colleagues will
be to navigate and guide those imaging studies that will
define the clinical efficacy and relevance of a targeted
imaging agent to satisfy both regulators and consumers.
These studies will influence whether the commercial
argument can be made that will stimulate investment in
developing targeted contrast agents. Ultimately, the sales
of an agent will be an index of its clinical value. A
successful agent will be rapidly assimilated into clinical
practice where further refinements in dosage, indication
and safety profile will be determined.
Can targeted agents be brought into the clinic, or will
they simply remain laboratory tools divorced from
clinical application? The answer is, of course, that any
efficacious and safe agent can be brought to the clinic,
but at what price? More important is the question of
whether such agents will make an impact on clinical
practice, and this is a more difficult answer to predict.
Regulators are not in agreement as to the requirements
for demonstrating likely clinical impact before approval.
The need for development both in clinical trial designs
and in appropriate criteria for regulatory requirements
may override any other limiter on the adoption of
molecular imaging in the clinic. Indeed it is surprising
that the imaging community, despite its resolve to
promote molecular imaging, has been almost mute in
its response to the current regulatory climate. I strongly
encourage the radiological community to give it some
thought.
As end-users, radiologists and medical-imaging scien-
tists have influence over contrast agent development.
Their technical expertise is sought to guide the practical
application of imaging strategies, but the radiological
community has limited exposure of the drug develop-
ment world in terms of the requirements, needs and
execution of imaging trials. Imaging experts could
become more engaged in the development of new
imaging agents, particularly of targeted contrast agents
by:
N Understanding the needs and requirements of the
regulatory process behind drug approval
872
A Dzik-Jurasz
N Promoting an improved understanding of clinical trial
design, execution and requirements including the
standardization of experimental design in multicentre
trials
N Fostering an initiative to advance and promote new
outcome measures, particularly in respect of targeted
contrast agents
N Developing realistic strategies for the development of
targeted contrast agents based on clinical need
The opportunity is there and need only be taken up by
the radiological community. It would be sad if the only
legacy of targeted imaging agents were a wealth of good
science, but shameful if patients were not to benefit from
its findings.
Acknowledgments
I would like to express my gratitude to Dr J A
Newmyer, Dr E Parsons and Dr A Uprichard for their
critical review of the manuscript.
Disclaimer
The opinions expressed in this manuscript are those of
the author and do not necessarily reflect those of other
parties or institutions. No financial conflict is declared.
References
1. Ludwig JA, Weinstein JN. Biomarkers in cancer staging,
prognosis and treatment selection. Nat Rev Cancer
2005;5:845–56.
2. Padhani AR, Leach MO. Antivascular cancer treatments:
functional assessments by dynamic contrast-enhanced mag-
netic resonance imaging. Abdom Imaging 2005;30:324–41.
3. Jaffer FA, Weissleder R. Molecular imaging in the clinical
arena. JAMA 2005;293:855–62.
4. DiMasi JA, Hansen RW, Grabowski HG. The price of
innovation: new estimates of drug development costs. J
Health Econ 2003;22:151–85.
5. Nunn AD. The cost of developing imaging agents for routine
clinical use. Invest Radiol 2006;41:206–12.
The British Journal of Radiology, November 2006
The British Journal of Radiology, 79 (2006), 873–879

The development and optimization of high spatial resolution MRI

for imaging the oesophagus using an external surface coil
A M RIDDELL, BSc, FRCS, FRCR, C RICHARDSON, DCRr, E SCURR, BSc, DCRr and G BROWN, MD, MRCP, FRCR

Department of Radiology, Royal Marsden Hospital, Fulham Road, London SW3 6JJ, UK

ABSTRACT. This paper describes the development and optimization of an innovative

technique using an external surface coil to obtain high resolution, thin section MR
images of the oesophagus using volunteers. T2 weighted fast spin echo sequences were
performed with and without cardiac gating. The field of view (FOV), matrix size, slice
thickness, number of signal averages (NSA), and repetition time (TR)/echo time (TE)
were altered to optimize signal to noise ratio (SNR) whilst maintaining spatial
resolution. The effect of cardiac gating was also investigated. Workstation images were
evaluated on the ability to visualize: individual oesophageal wall layers;
perioesophageal fat; the azygos vein and wall of the descending aorta, giving
qualitative assessment of image clarity. The optimum sequence enabled the layers of
the oesophageal wall and perioesophageal tissues to be demonstrated in an acceptable
scan time of 7.07 min. A FOV of less than 250 mm degraded image quality so that Received 9 February 2006
individual oesophageal wall layers could not be depicted and noise within the image Revised 27 March 2006
impaired visualization of posterior mediastinal structures. The results indicate that high Accepted 24 April 2006
resolution imaging of the oesophagus using an external surface coil can depict
DOI: 10.1259/bjr/36989440
anatomic structures clearly and that the use of cardiac gating improves image clarity.
The technique offers an alternative, non-invasive method of detailed imaging of the ’ 2006 The British Institute of
oesophagus. Radiology

Endoscopic ultrasound is currently used for local
staging of oesophageal cancer, but has technical limita-
tions, with a failure rate in up to 17–24% of patients due to
stenotic tumours preventing the passage of the endoscope
[1, 2]. It is also operator dependent, with a recognized
learning curve [3]. The high frequency probe (12 MHz)
used to achieve high spatial resolution, has a limited
sonographic range and, as a consequence, the relationship
of the oesophagus to the surrounding perioesophageal
tissues can be difficult to visualize with clarity.
There have been few studies that have evaluated MRI
of the oesophagus. Early in vivo studies used an
integrated body coil and a low field strength magnet
(0.35–0.5 T). As a consequence, the images generated had
a low in-plane spatial resolution [4, 5]. More recent
studies have concentrated on the use of an endoluminal
surface coil attached to the tip of a modified endoscope.
Preliminary results have been encouraging, with con-
firmation of the ability to depict the layers of the
oesophageal wall [6–8]. However, the technique has
recognized limitations, namely: the inability of the coil to
traverse strictures, limiting the use for the evaluation of
oesophageal cancers; a short radius for receiving signal
(3–4 cm), which necessitates repositioning of the endo-
scope for evaluation of long tumours therefore increasing
the overall scan time; and the action of peristalsis,
resulting in motion artefact within the images acquired
and possible coil migration.
The only study to date using a 1.5 T magnet and an
external surface coil examined oesophageal specimens
in vitro [9]. The study concluded that individual layers of
the oesophageal wall were more clearly depicted on T2
weighted rather than T1 weighted images.
The aim of this study was to use healthy volunteers to
develop the technique of high resolution thin slice MRI,
using an external surface coil in vivo, as a non-invasive
method of imaging the oesophagus and perioesophageal
tissues; and to both describe the methods used to
optimize the signal to noise ratio (SNR) and to assess
the effect of cardiac gating.
Methods and materials
Four volunteers were imaged using a 1.5 T magnet
(Philips Intera, software version 9.5.2, Andover, MA) and
an external 5 channel surface coil (Philips Sense
CardiacTM). The study evaluated images of the lower
oesophagus, as clinically the majority of patients at our
institution who are considered for surgical resection are
those with adenocarcinoma of the lower oesophagus and
gastro-oesophageal junction.
The study objectives were to:
1. Maximize spatial resolution by alterations to the feld
of view (FOV), matrix size and slice thickness, whilst
maintaining acceptable image quality.
2. Maximize SNR by alterations to the echotime (TE)
and number of signal averages (NSA), whilst main-
taining acceptable image quality and scan time.
3. Minimize the effects of cardiac motion, by the use of a
saturation band placed over the heart and also the use
of cardiac gating.

The British Journal of Radiology, November 2006 873

 A M Riddell, C Richardson, E Scurr and G Brown

4. Maintain an acceptable scan time. The maximum limit

for an individual sequence was set at 7 min 30 s for 22
axial images.

T2 weighted sequences were used as it had previously

been established that these are optimum for delineating
the layers of the oesophageal wall [9, 10]. A sagittal
sequence was performed initially to localize the oeso-
phagus and to ensure optimum coil placement. Axial
images were acquired perpendicular to the long axis of
the oesophagus.
A standard fast spin echo T2 weighted sequence used
for pelvic imaging [11] was modified and used as a
baseline sequence for the studies without cardiac gating,
and the baseline parameters used when cardiac gating
was employed were modified from a standard cardiac T2
weighted sequence, with a reduction in the FOV. In both
instances, the baseline sequence was given a relative
SNR of 1. Changes in the SNR due to alterations in
sequence parameters were expressed as a ratio from this
baseline.

Assessment of image quality

For each of the study objectives, the images acquired Figure 1. The margins of acceptable signal to noise,
using specific sequence parameters were viewed on a confined within the white lines. A saturation band is placed
over the heart (X) to reduce artefact from cardiac motion.
workstation (eFilm workstationTM Version 1.9.3) and
The oesophagus is clearly seen anterior to the vertebral
were scored objectively by one radiologist (AMR) as column (white arrows).
good, moderate or poor, on the basis of the ability to
visualize individual layers of the oesophageal wall, the
achieved with an acceptable SNR was obtained using a
perioesophageal fat, the azygos vein and wall of the
256 mm6256 mm matrix, with 240 mm FOV, 3 mm slice
descending aorta. This provided a qualitative evaluation
thickness and 6 NSA. These parameters gave an in-plane
of image clarity. Using the same parameters, the effect of
cardiac gating on image quality was assessed by resolution of 0.94 mm60.94 mm and a voxel size of
comparing equivalent images with and without cardiac
gating.

Results
The sagittal T2 weighted sequence provided valuable
information to optimize coil placement. Figure 1 illus-
trates the boundary of acceptable signal and marked
signal drop off outside this region. The individual layers
of the oesophageal wall could be depicted clearly on the
axial high resolution T2 weighted images. The mucosa
returned intermediate to low signal. This was sur-
rounded by high signal intensity submucosa and the
outer low signal intensity muscularis propria, as shown
in Figure 2. The perioesophageal fat returned high
signal. The volume of fat varied between subjects and
for those subjects with less fat, the clarity of the
oesophageal wall layers and perioesophageal fat was
graded as moderate, compared with a grading of good
for subjects with a greater volume of perioesophageal fat.

Maximizing spatial resolution

For the T2 weighted sequence without cardiac gating a
relative SNR, when compared with the initial baseline
sequence parameters, of less than 0.6 resulted in unaccept-
able image quality. The highest spatial resolution
Figure 2. The main oesophageal wall layers: low signal
mucosa (white arrow), surrounded by high signal submucosa
(black arrow) and low signal muscularis propria (arrow-
heads). The descending thoracic aorta (A) and vertebral body
(V) are marked.

874 The British Journal of Radiology, November 2006

MRI of the oesophagus

Table 1. The optimum sequences for the T2 weighted sequences without and with cardiac gating
Sequence type TR (ms) TE (ms) Matrix (mm) FOV (mm) NSA Slice thickness SNR Voxel size
(mm) (mm3)

No cardiac gating 3328 120 2566256 240 6 3 1 2.63

No cardiac gating 5027 80 2566256 225 6 3 1.4 2.32
Cardiac gating 2250 90 3016512 250 8 4 0.44* 1.62
*This value for SNR is relative to the baseline for the sequence with cardiac gating.
SNR, signal to noise ratio; TR, repetition time; TE, echo time; FOV, field of view; NSA, number of signal averages.

2.63 mm3. A matrix size of 512 mm6256 mm produced
higher resolution images, but unacceptable noise within
the image. Image quality was also significantly impaired
by reducing the FOV to less than 240 mm. The optimum
sequence parameters are shown in Table 1.
Using cardiac gating, the highest achievable spatial
resolution with acceptable signal to noise was achieved
by employing a 250 mm FOV, a 301 mm6512 mm
matrix and a slice thickness of 4 mm, giving an in-plane
spatial resolution of a 0.5 mm60.8 mm and a voxel size
of 1.62 mm3, as shown in Table 1.
Maximizing SNR whilst maintaining spatial
resolution
The optimized sequence parameters, without cardiac
gating, achieved from the first study were used as the
baseline for technique refinement. A decrease in TE
resulted in an increase in the relative signal to noise
(Figure 3), whist maintaining the T2 weighting. As a
consequence, by utilizing a shorter TE, a smaller field of
view could be used to improve spatial resolution.
Increasing the NSA improved the SNR (Figure 4); an
NSA of 6 was considered optimum as the scan time was
increased to 9.28 min for 22 slices with an NSA of 8,
exceeding the maximum limit. The optimum sequence
parameters are shown in Table 1.
Minimizing the effects of cardiac motion
The images acquired with a saturation band alone
remained subject to motion artefact, primarily due to
cardiac pulsation. Image quality was superior for the
sequence with cardiac gating, when directly compared
with the equivalent image without cardiac gating
(Figure 5). Specifically, the clarity of the perioesophageal
tissues and wall of the descending thoracic aorta was
improved with the use of cardiac gating, with the
majority of images with cardiac gating being scored as
good quality. Depiction of the individual oesophageal
wall layers was also superior with cardiac gating,
although the improvement was not as marked.
Initially, the RF pulse was set to trigger on every third
R wave (3 beats), giving an effective TR of 2250 ms. This
gave adequate T2 weighting to the image. Increasing the
effective TR to 4000 ms (6 beats) resulted in a reduction
in image quality, in particular the clarity of the
perioesophageal tissues (Figure 6).

(a) (b)

Figure 3. Images at the same level of the oesophagus in one volunteer. The echo time (TE) is reduced from (a) 120 ms to
(b) 80 ms. All other parameters are maintained (matrix 256 mm6256 mm, repetition time (TR) 5446 ms, field of view (FOV)
225 mm, number of signal averages (NSA) 6, turbo spin echo (TSE) 16). The signal is increased, improving the conspicuity of the
oesophagus (white arrow), the wall of the aorta (arrowhead) and the perioesophageal fat (double arrow heads) with the
shorter TE.

The British Journal of Radiology, November 2006 875

 A M Riddell, C Richardson, E Scurr and G Brown

(a) (b)

Figure 4. Images at the same level of the oesophagus in one volunteer. The number of signal averages (NSA) is increased from
(a) 4 to (b) 8. All other parameters are maintained (matrix 256 mm6256 mm, repetition time (TR) 5446 ms, field of view (FOV)
225 mm, echo time (TE) 80 ms, turbo spin echo (TSE) 16). The higher NSA improves image quality by improving conspicuity of the
oesophageal wall layers: the high signal submucosa (white arrow) and lower signal muscularis propria (arrowhead); the
perioesophageal tissues (black arrow) and wall of the aorta.

When the T2 weighted sequence without cardiac Scan duration

gating was modified to utilize the sequence parameters
optimized with cardiac gating, there was a 37% drop in
relative signal to noise and the scan time increased from
7.07 min to 10.0 min for 22 slices, exceeding the
acceptable scan time limit.
Using the optimized sequence parameters without
cardiac gating, an acceptable scan time of 7.07 min for 22
slices was achieved. The use of cardiac gating did not
significantly increase the scan duration. The scan time

(a) (b)

Figure 5. Images taken at the same level of the oesophagus in one volunteer, using the optimized high resolution T2 weighted
sequence (a) without and (b) with cardiac gating. The sequence with cardiac gating provides improved image quality. This is
illustrated by the clarity of the oesophageal wall submucosa (black arrow), the aortic wall (white arrow) and right pleural
reflection (fine black arrow).

876 The British Journal of Radiology, November 2006

MRI of the oesophagus
(a)
Figure 6. Images taken at the same level of the oesophagus in one volunteer. Increasing the repetition time (TR) from (a) 3
beats (effective TR 2250 ms) to (b) 6 beats (effective TR 4000 ms) caused blurring within the image. The oesophageal wall
submucosa (black arrowhead), muscularis propria (black arrow), and right pleural reflection (white arrow) are more clearly seen
with an effective TR of 2250 ms (3 beats).
using cardiac gating varied slightly between volunteers
due to differences in resting heart rate, but for the
optimized sequence parameters did not exceed the scan
time limit.
Discussion
High resolution thin slice MRI of the oesophagus using
an external surface coil is potentially challenging due to
the relatively low signal to noise returned from the
structure due to its small size and location within the
thorax. This study shows that the individual layers of
the oesophageal wall and surrounding anatomical
structures can be delineated using the technique. The
signal returned from the individual wall layers in this
study is in accordance with the description given by
Yamada et al [9, 10]. Endoscopic ultrasound can also
delineate the layers of the oesophageal wall. However,
the ability of the technique to delineate the perioesopha-
geal tissues is limited by the relatively short ultrasound
range of the probe (in the order of 3–4 cm). The axial
images acquired of the posterior mediastinum obtained
using high resolution MRI clearly delineate the relation-
ship of the oesophagus to all its surrounding anatomical
structures, providing improved detail regarding poten-
tial resection margins for those patients with oesopha-
geal carcinoma who are being considered for surgery.
The variation in the amount of perioesophageal fat
between subjects resulted in a reduction in image clarity
in subjects with a small amount of perioesophageal fat,
which could be considered a limitation of the technique.
The British Journal of Radiology, November 2006
(b)
The non-invasive nature of the technique will allow
optimum visualization of stenotic tumours, not amen-
able to endoscopic evaluation. In many instances these
bulky tumours are not considered surgically resectable.
However, a technique which allows for detailed initial
evaluation of the local extent of tumour will provide the
most accurate method of local staging of the tumour. It
would also provide a basis for alternative treatment
planning and a baseline for future imaging to fully assess
treatment response.
The perioesophageal tissues are visualized with CT,
but recent studies have shown that the ability of the
technique to determine resectability is relatively low.
Markland et al reported a sensitivity of between 0% and
66.7% [12] and a further study of 51 patients showed that
59% had more advanced disease than was appreciated
on pre-operative CT, reinforcing the need to refine the
techniques for the local staging of oesophageal carci-
noma to improve patient selection for radical therapy
[13]. The superior contrast resolution achieved with MRI
would suggest that the technique is likely to be useful in
differentiating tumour abutting surrounding structures
(stage T3) from direct invasion of tumour into the
surrounding tissues (stage T4).
Spatial resolution
Previous attempts to achieve high resolution, thin slice
imaging of the oesophagus have been limited to in vitro
studies of resected oesophageal specimens from patients
with known oesophageal carcinoma [9, 10]. The studies
877

used a 4 cm loop surface coil and a 1.5 T or 4.5 T magnet,
with FOV of between 40 mm and 60 mm, and as such
were able to achieve a voxel size of 0.08–0.16 mm3. Given
this high resolution, 6–8 layers of the oesophageal wall
were visualized and the technique was able to identify
tumour which was confined to the mucosa. Prior in vivo
studies using surface coils used low field strength
magnets (0.35–0.5 T) and acquired images with a slice
thickness of 8–10 mm, to achieve adequate signal to
noise. As a consequence, these studies concluded that
MRI was no better than CT at providing detailed analysis
of the oesophageal wall and surrounding anatomical
structures [4, 5]. Investigations using an endoluminal coil
have achieved an in plane resolution of 0.469 mm
60.625 mm (voxel size 1.17 mm3), enabling visualiza-
tion of the three main mural layers (mucosa, submucosa
and muscularis) [14]. Our study shows that using fast
spin echo sequences, high resolution images can be
acquired with an external surface coil, with a voxel size
of 1.62 mm3. At this resolution the main component
layers of the oesophageal wall were readily delineated,
indicating that it would be possible to use the technique
for the local staging of oesophageal tumours.
Signal to noise
The benefit of using an endoluminal surface coil is that
the coil is closer to the region of interest, thus improving
signal to noise. As a consequence, the spatial resolution
can be maximized. There is, however, rapid signal decay
beyond a 3–4 cm radius of the endoluminal coil and
therefore the technique has limited ability to assess the
perioesophageal tissues and determine tumour resect-
ability. Previous MRI studies with an external coil used a
body coil integrated into the bore of the magnet. The
introduction of multi-channel external surface phased
array coils, as used in our study, has improved the
achievable SNR, by the use of parallel imaging techni-
ques. The technology allows for faster sequences
enabling higher resolution imaging to be acquired with
increased signal to noise.
Cardiac gating
Several of the previous studies using endoluminal and
external surface coils have used cardiac gating, but no
previous study has directly compared the effect of using
the technique on image quality. The acquisition of
imaging data when using cardiac gating is synchronized
with and limited to specific times or phases of the cardiac
cycle. The R wave on the electrocardiogram (ECG) is
used to trigger each radiofrequency (RF) pulse. The RF
signal is therefore generated at the same point of
subsequent cardiac cycles. Since the R to R interval is
controlled by the patient’s heart rate – the TR (and as a
consequence the image weighting and number of slices)
depends on the heart rate. To enable variation in the TR
value, the system can be adapted to trigger an RF pulse
at every second or third R wave; in this way the
‘‘effective TR’’ is lengthened, allowing for T2 weighted
imaging. Our study would indicate that the use of
cardiac gating does improve image quality.
878
A M Riddell, C Richardson, E Scurr and G Brown
Scan time
Our experience is that, in general, patients are able to
lie still for an individual sequence which lasts for
between 7 min and 7K min. Our study used 7K min
as the maximum acceptable scan time for 22 axial
images. Although image quality was improved by
increasing the NSA from 6 to 8, for the T2 weighted
sequence without cardiac gating, this exceeded our
upper time limit. These scan times compare favourably
with endoluminal MRI studies, which have documented
scan times of between 7 min and 10 min for 8–12 images.
No previous study has specifically assessed the affect of
cardiac gating on scan time; these results show that its
use does not significantly increase the overall scan
duration.
We conclude that our study indicates that the
technique of high resolution, thin slice MRI using and
external surface coil is a reproducible, non-invasive
method of imaging the oesophagus, which could
potentially provide an alternative method for the local
staging of oesophageal carcinoma.
References
1. Wakelin SJ, Deans C, Crofts TJ, Allan PL, Plevris JN,
Paterson-Brown S. A comparison of computerised tomo-
graphy, laparoscopic ultrasound and endoscopic ultra-
sound in the preoperative staging of oesophago-gastric
carcinoma. Eur J Radiol 2002;41:161–7.
2. Preston SR, Clark GW, Martin IG, Ling HM, Harris KM.
Effect of endoscopic ultrasonography on the management
of 100 consecutive patients with oesophageal and junctional
carcinoma. Br J Surg 2003;90:1220–4.
3. Schlick T, Heintz A, Junginger T. The examiner’s learning
effect and its influence on the quality of endoscopic
ultrasonography in carcinoma of the esophagus and gastric
cardia. Surg Endosc 1999;13:894–8.
4. Quint LE, Glazer GM, Orringer MB. Esophageal imaging by
MR and CT: study of normal anatomy and neoplasms.
Radiology 1985;156:727–31.
5. Lehr L, Rupp N, Siewert JR. Assessment of resectability of
esophageal cancer by computed tomography and magnetic
resonance imaging. Surgery 1988;103:344–50.
6. Inui K, Nakazawa S, Yoshino J, Yamao K, Yamachika H,
Wakabayashi T, et al. Endoscopic MRI: preliminary results
of a new technique for visualization and staging of
gastrointestinal tumors. Endoscopy 1995;27:480–5.
7. Kulling D, Feldman DR, Kay CL, Hoffman BJ, Reed CE,
Young JW, et al. Local staging of esophageal cancer using
endoscopic magnetic resonance imaging: prospective com-
parison with endoscopic ultrasound. Endoscopy 1998;30:
745–9.
8. Dave UR, Williams AD, Wilson JA, Amin Z, Gilderdale DJ,
Larkman DJ, et al. Esophageal cancer staging with endo-
scopic MR imaging: pilot study. Radiology 2004;230:
281–6.
9. Yamada I, Izumi Y, Kawano T, Yoshino N, Tetsumura A,
Ohashi K, et al. Superficial esophageal carcinoma: an in
vitro study of high-resolution MR imaging at 1.5T. J Magn
Reson Imaging 2001;13:225–31.
10. Yamada I, Murata Y, Izumi Y, Kawano T, Endo M, Kuroiwa
T, et al. Staging of esophageal carcinoma in vitro with 4.7-T
MR imaging. Radiology 1997;204:521–6.
11. Brown G, Daniels IR, Richardson C, Revell P, Peppercorn D,
Bourne M. Techniques and trouble-shooting in high spatial
The British Journal of Radiology, November 2006
MRI of the oesophagus
resolution thin slice MRI for rectal cancer. Br J Radiol
2005;78:245–51.
12. Markland CG, Manhire A, Davies P, Beggs D, Morgan WE,
Salama FD. The role of computed tomography in assessing
the operability of oesophageal carcinoma. Eur J Cardiothorac
Surg 1989;3:33–6.
The British Journal of Radiology, November 2006
13. Sariego J, Mosher S, Byrd M, Matsumoto T, Kerstein M.
Prediction of outcome in ‘‘respectable’’ esophageal carci-
noma. J Surg Oncol 1993;54:223–5.
14. Feldman DR, Kulling DP, Hawes RH, Kay CL, Muckenfuss
VR, Cotton PB, et al. MR endoscopy: preliminary experi-
ence in human trials. Radiology 1997;202:868–70.
879
 The British Journal of Radiology, 79 (2006), 880–887

Pancreatic adenocarcinoma: signs of vascular invasion

determined by multi-detector row CT
1 2 2 3
H LI, MD, M S ZENG, MD, Pro, K R ZHOU, MD, Pro, D Y JIN, MD, Pro and 3W H LOU, MD, Pro

1
Department of Radiology, The Affiliated First People’s Hospital, Shanghai Jiao Tong University, 85
Wujin Road, 200080, Shanghai, Departments of 2Radiology and 3General Surgery, The Affiliated
Zhongshan Hospital, Fudan University, 180 Fenglin Road, 200032, Shanghai, P R China

ABSTRACT. The purpose of this study was to analyse multi-detector row CT (MDCT)
signs of peripancreatic arterial and venous invasion in pancreatic carcinoma. Among
101 patients with pancreatic carcinoma examined by MDCT, 54 candidates for surgery
were pre-operatively evaluated for vascular invasion based on MDCT signs. The
peripancreatic major vessels (including superior mesenteric artery, coeliac artery,
common hepatic artery, superior mesenteric vein and portal vein) were examined
carefully by surgeons during the operation. At surgical exploration, 78 of 224 vessels
were invaded by tumour. The invaded peripancreatic major arteries (n529) and veins
(n549) presented different MDCT signs: 43% of invaded veins (18/42, except for 7
occluded veins) were surrounded by tumour less than 50% of the vessel circumference
compared with 97% (28/29) of the invaded arteries, which were surrounded by tumour
more than 50% of the vessel circumference or were embedded in tumour (p,0.001). Received 28 July 2005
69% (34/49) of the invaded veins had vascular stenosis or obliteration, compared with Revised 17 November 2005
41% (12/29) of the invaded arteries (p,0.05). Irregularity of the vein wall, 74% (31/42, Accepted 2 May 2006
except for 7 occluded veins); occurred more often than that of the artery wall, 45%
DOI: 10.1259/bjr/19684199
(13/29) (p,0.05). In conclusion, the MDCT signs of peripancreatic arterial and venous
invasion have different characteristics, which should be considered in pre-operative ’ 2006 The British Institute of
evaluation. Radiology

One decisive factor for the irresectability of pancreatic
carcinoma is the major vascular invasion [1]. Multi-
detector row CT (MDCT) scanning of the pancreas
enables multiphasic thin collimation scanning, with
excellent spatial resolution, especially in the z-plane [2].
These volume data sets can be easily manipulated with
three-dimensional imaging, potentially providing addi-
tional information to conventional axial display [3–6].
The purpose of our study was to assess the MDCT
signs of arterial and venous invasion in pancreatic
carcinoma.
MDCT images of the 101 patients were all prospec-
tively evaluated for resectability: including vascular
invasion and the presence of metastatic disease. Of the
54 candidates for surgery (37 males and 17 females, 40–
79 years old, average 61.2 years), 17 patients underwent
pancreaticoduodenectomy, and the others were found to
be irresectable during surgical exploration. The time
interval between MDCT examination and surgery was
#2 weeks.
Examination techniques

Materials and methods
Patients
Between December 2001 and February 2004, 255
consecutive patients with presumed pancreatic carci-
noma underwent pancreas examinations by triphasic
MDCT. 101 patients with pancreatic carcinoma were
confirmed by surgical-pathology (n554) or clinical
follow-up (n547). 54 patients were pre-operatively
judged to be resectable or requiring surgical bypass
due to jaundice. Pancreatic adenocarcinoma was patho-
logically confirmed by biopsy or fine-needle aspiration.
The remaining 47 patients deemed irresectable received
chemotherapy or radiotherapy, and all of them survived
less than 2 years.
All MDCT examinations of the pancreas were per-
formed on Mx8000 (Philips, Best, The Netherlands; four-
slice spiral CT scanners). 600–800 ml water was routinely
administered 10–20 min before the examination to
distend the stomach, duodenum and proximal jejunum.
Each patient received 120 ml of non-ionic contrast
material (iopromide 350 mg ml21; Ultravist; Schering,
Berlin, Germany) via intravenous injection at the rate of
5 ml s21.
Unenhanced and triphasic (arterial phase, pancreatic
phase and hepatic phase) enhanced scans were per-
formed. Unenhanced scan images were obtained from
the top of the diaphragm to the caudal level of the
uncinate process. The scanning parameters were:
120 kVp, 250 mA, a 0.875 pitch and 5 mm collimation.
After a 20 s delay from the start of the intravenous

880 The British Journal of Radiology, November 2006

Signs of vascular invasion
infusion, the arterial phase was obtained from the level
of the hepatic hilum through the entire pancreas. The
collimation was altered to 2.5 mm, and the other
parameters were the same as those of the unenhanced
scan. The delay time of the pancreatic phase and the
hepatic phase was 45 s and 80 s, respectively. The
scanning range and parameters were the same as those
of the arterial phase and the unenhanced scans,
respectively.
After scanning, the data were transferred to the image
server linked to an Mxview workstation with software
version 3.5. This workstation was used for three-
dimensional analysis of the local anatomy utilizing
volume rendering (VR) combined with maximum
intensity projection (MIP) and multiplanar reconstruc-
tion (MPR).
Image analysis and surgical correlation
CT signs of peripancreatic major vessels (including
coeliac artery (CA); common hepatic artery (CHA);
superior mesenteric artery (SMA); portal vein (PV);
superior mesenteric vein (SMV)) were determined pre-
operatively at consensus reading of the axial and 3D
images by two observers (MSZ and HL, 10 years and
5 years of experience reading pancreatic imaging,
respectively):
1. Contiguity of tumour with the adjacent vessel was
graded A,D: (grade A, fat plane or normal pancreatic
tissue visible between tumour and vessel; grade B and
grade C, tumour surrounding of less than and more
than 50% of the vessel circumference; grade D, arterial
embedment in tumour or venous occlusion.)
2. Detailed vascular anatomic deformation: vessel ste-
nosis presented a semi-circular or concentric smaller
contour of the vessel. But a straight contour on one
side was regarded as flattened, not as stenosed. Vessel
wall infiltration presented an irregular and indented
shape at the vascular margin abutting tumour.
(a)
Figure 1. Pancreatic body carcinoma. (a,b) Consecutive axial images showed that the superior mesenteric artery (SMA) was
surrounded to more than 50% of the vessel circumference by tumour and the vessel wall appeared infiltrated. Pre-operatively,
this case was judged irresectable. At surgical exploration, the SMA was found not to be invaded (only fibrotic infiltration) by
tumour and was successfully resected.
The British Journal of Radiology, November 2006
224 vessels were examined carefully at surgery by two
experienced surgeons (DYJ, WHL, 12 years and 8 years
of experience operating pancreatic carcinoma, respec-
tively) together. The surgical criterion of vascular
ingrowth was that the vessel could be observed, or
found by palpation, to be infiltrated or occluded at
surgery. It was not considered vascular invasion if
tumour was adherent to but could be separated from
the vessel due to inflammatory or fibrotic reaction.
Statistical analysis
Chi-square tests were performed on the CT signs:
(a) tumour surrounding more than 50% of the vessel
circumference; (b) vessel stenosis or occlusion; (c) vessel
wall irregularity, to check for any significant difference
between the invaded arteries and veins. p , 0.05 was
required to show statistical significance.
Results
43 of the 54 tumours at surgery were located in the
pancreatic head or uncinate process, 11 in the body or
tail. 17 tumours were surgically resectable, of which 16
were pre-operatively correctly diagnosed by MDCT. One
was incorrectly judged irresectable because MDCT had
shown tumour ingrown in the SMA (Figure 1). Of 37
patients who received by-pass palliative surgery or
surgical exploration, 6 tumours were incorrectly diag-
nosed resectable due to underdiagnosed invaded vessels
(n54), missed hepatic metastases (n51) and peritoneal
metastases (n51).
Of the 224 observed vessels, 146 vessels were found
not to be invaded at surgery and 78 invaded. Except for
the above-mentioned SMA, which MDCT incorrectly
judged to be locally infiltrated by tumour on MDCT, the
calibre of the other 145 not invaded vessels was
unchanged and the vessel wall was regular. On MDCT,
the correlation between the surgically confirmed vessels
that were not invaded and tumour was:
(b)
881
 H Li, M S Zeng, K R Zhou et al

(a) (b)

Figure 2. Tumour in the uncinate process of the head of the pancreas. (a,b) Consecutive axial images showed the superior
mesenteric artery (SMA) to be surrounded more than 50% of the vessel circumference by tumour. It was found not to be invaded
by tumour at exploration and tumour was successfully resected.

1. Grade A of contiguity of tumour with the adjacent
vessel (fat plane or normal pancreatic tissue pre-
served) (n5129, 22 SMAs, 29 CAs, 30 CHAs, 18 SMVs
and 30 PVs)
2. Grade B of contiguity of tumour with the adjacent
vessel (,180 ˚) (n59, 3 SMAs, 1 CA, 3 SMVs, 2 PVs)
3. Grade C of contiguity of tumour with the adjacent
vessels (.180 ˚) (n58, 5 SMAs, 2 CAs, 1 CHA)
(Figure 2)
The MDCT appearances of surgically confirmed
invaded arteries (n529) and veins (n549) are shown in
Tables 1 and 2, respectively. Statistical analysis related to
invaded arteries and veins is shown in Table 3.
Circumferential involvement
43% of invaded veins (18/42, except for 7 occluded
veins) were surrounded by less than 50% of the vessel

Table 1. Multi-detector CT (MDCT) appearances of surgically confirmed invaded arteries

SMA CA CHA

Grade D & irregular wall & stenosed vessel 4 1 1

Grade D & irregular wall & unchanged calibre 2 – 1
Grade D & regular wall & stenosed vessel 2 1 1
Grade D & regular wall & unchanged calibre 2 2 2
Grade C & irregular wall & stenosed vessel 2 – –
Grade C & irregular wall & unchanged calibre 1 1 –
Grade C & regular wall & stenosed vessel – – –
Grade C & regular wall & unchanged calibre 4 – 1
Grade B & regular wall & unchanged calibre 1 – –
Grade A,D: the grade of contiguity of tumour with the adjacent vessel.
SMA, superior mesenteric artery; CA, coeliac artery; CHA, common hepatic artery.

Table 2. Multi-detector CT (MDCT) appearances of surgically confirmed invaded veins

SMV PV

Grade D 4 3
Grade C & irregular wall & stenosed vessel 11 6
Grade C & irregular wall & unchanged calibre 2 3
Grade C & regular wall & stenosed vessel 1 1
Grade C & regular wall & unchanged calibre – –
Grade B & irregular wall & stenosed vessel 4 2
Grade B & irregular wall & unchanged calibre 2 1
Grade B & regular wall & stenosed vessel 1 1
Grade B & regular wall & calibre was ‘‘teardrop’’ 3 –
Grade B & regular wall & unchanged calibre 2 1
Grade A & regular wall & unchanged calibre – 1
Grade A,D: the grade of contiguity of tumour with the adjacent vessel.
SMV, superior mesenteric vein; PV, portal vein.

882 The British Journal of Radiology, November 2006

Signs of vascular invasion

Table 3. Statistical analysis between invaded arteries and veins

CT signs Artery Vein Chi2 p-value

CA CHA SMA Total SMV PV Total

1* 5/5 6/6 17/18 28/29 14/26 10/16 24/42 13.6 ,0.001

2 2/5 2/6 8/18 12/29 21/30 13/19 34/49 5.9 ,0.05
3* 2/5 2/6 9/18 13/29 19/26 12/16 31/42 6.1 ,0.05
1: Tumour surrounding more than 50% of the vessel circumference.
2: Vessel stenosed or occlusive.
3: Vessel wall irregularity.
*Seven veins (4 SMVs, 3 PVs) occluded by tumours were not included.
CA, coeliac artery; CHA, common hepatic artery; SMA, superior mesenteric artery; SMV, superior mesenteric vein; PV, portal vein.

circumference by tumour. 97% (28/29) of invaded Stenosis or occlusion

arteries were surrounded by more than 50% of the
vessel circumference by tumour or were embedded in
tumour. Statistical analysis showed a significant differ-
ence between invaded arteries and veins (except for
seven occluded veins) utilizing circumferential involve-
ment of more than 50% of the vessel circumference by
tumour (p,0.001).
69% (34/49) of invaded veins presented stenosed or
occluded (Figure 3). Eight veins (5 SMVs, 3 PVs), which
were surrounded less than 50% of the vessel circumfer-
ence by tumour, also appeared stenosed (Figure 4).
41% (12/29) of invaded arteries did not appear
stenosed, and the difference was statistically significant

(a) (b)

(c) (d)

Figure 3. Tumour in the uncinate process of the head of the pancreas. (a,b) Axial images and (c,d) 3D images showed that the
tumour in the uncinate process of the head of the pancreas eroded a side of the superior mesenteric vein (SMV) vessel wall and
penetrated it to form tumour thrombus (arrow). Axial images also showed that the superior mesenteric artery (SMA) was
embedded. Surgery confirmed tumour invasion of the SMV and the SMA.

The British Journal of Radiology, November 2006 883

 H Li, M S Zeng, K R Zhou et al

(a)

(b)

Figure 4. Pancreatic head carcinoma. (a,b) Consecutive axial

images showed the portal vein (PV) vessel calibre (arrow) was
changed, although it was surrounded less than 50% of the
vessel circumference by tumour. (c) Volume rendering (VR)
3D image showed that a segment of PV was stenosed. The
axial images also showed that the common hepatic artery
(CHA) was embedded in tumour and the coeliac artery (CA)
was surrounded by more than 50% of the vessel circumfer-
ence. PV, CHA and CA were proven to be infiltrated at
(c) surgical exploration.

884 The British Journal of Radiology, November 2006

Signs of vascular invasion
(a)
(b)
(c)
Figure 5. Tumour in the uncinate process of the head of the
pancreas. (a,b) Axial images showed coeliac artery (CA)
(short arrow) and common hepatic artery (CHA) (long arrow)
embedded in tumour with regular vessel walls, although (c)
volume rendering (VR) 3D image showed that the vessel
calibre of CA and CHA was unchanged. Surgery confirmed
tumour invasion of CA and CHA.
(p,0.05). Nine arteries (2 CAs, 3 CHAs, 4 SMAs),
embedded in tumour, did not appear stenosed
(Figure 5).
The British Journal of Radiology, November 2006
Figure 6. Pancreatic head carcinoma. The confluence of the
portal vein (PV) and the superior mesenteric vein (SMV) was
shown to be surrounded by less than 50% of the vessel
circumference by tumour, with unchanged vessel calibre. But
a focal area of vessel wall (arrow) was irregular and
infiltrated. The confluence of the PV and the SMV was
proven to be infiltrated at surgical exploration.
Infiltration
74% of invaded veins (31/42, except for 7 occluded
veins) appeared infiltrated with an irregular wall. Three
veins (2 SMVs, 1 PV) were found to be invaded only
from their irregular wall, were surrounded by less than
50% of the vessel circumference and were not stenosed
(Figure 6). 45% (13/29) of invaded arteries appeared
infiltrated less often than invaded veins (except for 7
occluded veins) (p,0.05).
Other MDCT appearances of arterial and venous
invasion
24% (7/29) of invaded arteries appeared deformed
and seemed to be stretched or stiffened in three-
dimension reconstruction images, such as VR
(Figure 7). 10% (3/30) of invaded SMVs were diagnosed
correctly because they appeared as ‘‘teardrop’’ in axial
images (Figure 8). When invaded veins were highly
stenosed or occluded, 20% (10/49), there were multiple
venous collaterals.
Discussion
According to previous reports, using single-detector
row CT the diagnostic accuracy of pancreatic irresect-
ability was 95% and of pancreatic resectability was
70,80% [7–10]. The most important reason for inaccu-
rate assessment of resectability is underestimation of the
vascular invasion.
To improve the accuracy of estimating invaded
vessels, it is necessary to evaluate the MDCT signs of
arterial and venous invasion separately because we have
found clinically that the features of peripancreatic
arterial and venous invasion on CT are different.
885

(a)
(b)
(c)
Figure 7. Pancreatic head carcinoma. (a,b) Consecutive axial
images showed that the superior mesenteric artery (SMA)
was embedded in tumour. (c) Volume rendering (VR) 3D
image showed a segment of the SMA to be stiffened (arrow)
where the vessel calibre was low-grade stenosed. Surgery
confirmed tumour invasion.
In our opinion, the major reason for the different CT
signs of arterial and venous invasion is that the vein wall
is thinner and weaker than the artery wall. When veins
are surrounded and infiltrated by tumour, the wall tends
to be irregular and the calibre becomes narrowed. At the
same time, the flow rate in veins is slow, and tumour
often penetrates the vein wall to form tumour thrombus,
causing vascular occlusion (Figure 3). As the artery wall
is thicker and more flexible than the vein wall, and the
artery calibre is smaller, even when the arteries are
886
H Li, M S Zeng, K R Zhou et al
Figure 8. Tumour (T) in the uncinate process of the head of
the pancreas tethered the superior mesenteric vein (SMV)
(arrow) into a teardrop shape, although the SMV was
surrounded by less than 50% of the vessel circumference
and vessel wall remained regular. The superior mesenteric
artery (SMA) (arrowhead) was also shown to be nearly
embedded in tumour. The SMV and the SMA were proven to
be infiltrated at surgical exploration.
embedded in tumour, the calibre is not easily changed
and the wall remains regular. In our study, some arteries
were found not to be invaded at surgical exploration,
although they were surrounded by more than 50% of the
vessel circumference, and their calibre remained
unchanged and the wall was regular.
Hough et al [11] found that tumours in the head of the
pancreas could cause a tethered, teardrop appearance of
the SMV in axial images. They believed that the teardrop
SMV sign was a reliable indicator of irresectability. In
their retrospective study, teardrop SMV was the only
sign of irresectability in 13 of 17 patients. This sign
presumably results from either direct tumour infiltration
or peritumoural fibrosis adherent to the vessel that
retracts or tethers the vessel, changing its normal round
shape. In our series, there were three invaded SMVs
confirmed by surgical exploration exhibiting the tear-
drop sign on axial images in which the vessels were
surrounded by less than 50% of the vessel circumference,
and the wall of which was regular.
Since pancreatic carcinoma may be accompanied by
focal tissue fibrosis, the invaded arteries may appear
stretched. In our study, seven stretched arteries were
appreciated on three-dimensional reconstruction MDCT
angiography (MDCTA) images [3].
Lepanto et al [12] found that CTA significantly
increased the ability to identify venous invasion but
did not improve the detection of arterial invasion. There
are few MDCTA cases reported. As our experience has
shown, the image quality of MDCTA exceeds that of
CTA. MDCTA in particular allows the three-dimensional
display of the relationship between tumour and vessel,
but conventional axial images should also be reviewed as
axial images appear better able to demonstrate the
contiguity of tumour to vessel and change in calibre of
the vessel wall.
A major limitation of our study is our gold standard
of surgical palpation, because irresectable tumours
The British Journal of Radiology, November 2006
Signs of vascular invasion
could not be histologically assessed, although the surgi-
cal margins of resected tumours were confirmed by
pathology.
In conclusion, invaded peripancreatic arterial and
venous MDCT signs are different. It is important to
pay attention to these differences in order to improve the
accuracy of diagnosing vascular invasion and pancreatic
resectability.
References
1. Hommeyer SC, Freeny PC, Crabo LG. Carcinoma of the
head of the pancreas: evaluation of the pancreaticoduode-
nal veins with dynamic CT-potential for improved accuracy
in staging. Radiology 1995;196:233–8.
2. McNulty NJ, Francis IR, Platt JF, Cohan RH, Korobkin M,
Gebremariam A. Multi-detector row helical CT of the
pancreas: effect of contrast-enhanced multiphasic imaging
on enhancement of the pancreas, peripancreatic vascula-
ture, and pancreatic adenocarcinoma. Radiology
2001;220:97–102.
3. Horton KM, Fishman EK. Multidetector CT angiography of
pancreatic carcinoma: part 1, evaluation of arterial involve-
ment. AJR Am J Roentgenol 2002;178:827–31.
4. Horton KM, Fishman EK. Multidetector CT angiography of
pancreatic carcinoma: part 2, evaluation of venous involve-
ment. AJR Am J Roentgenol 2002;178:833–6.
5. Horton KM, Fishman EK. 3D CT angiography of the celiac
and superior mesenteric arteries with multidetector CT data
The British Journal of Radiology, November 2006
sets: preliminary observations. Abdom Imaging
2000;25:523–5.
6. Nino-Murcia M, Jeffrey RB, Beaulieu CF, Li KCP, Rubin
GD. Multidetector CT of the pancreas and bile duct system:
value of curved planar reformations. AJR Am J Roentgenol
2001;176:689–93.
7. Hommeyer SC, Freeny PC, Crabo LG. Carcinoma of the
head of the pancreas: evaluation of the pancreaticoduode-
nal veins with dynamic CT-potential for improved accuracy
in staging. Radiology 1995;196:233–8.
8. Raptopoulos V, Steer ML, Sheiman RG, Vrachliotis TG,
Gougoutas CA, Movson JS. The use of helical CT and CT
angiography to predict vascular involvement from pan-
creatic cancer: correlation with findings at surgery. AJR Am
J Roentgenol 1997;168:971–7.
9. Diehl SJ, Lehmann KJ, Sadick M, Lachmann R, Georgi M.
Pancreatic cancer: value of dual-phase helical CT in
assessing resectability. Radiology 1998;206:373–8.
10. Kaneko K, Honda H, Hayashi T, Fukuya T, Irie H, Masuda
K. Helical CT evaluation of arterial invasion in pancreatic
tumors: comparison with angiography. Abdom Imaging
1997;22:204–7.
11. Hough TJ, Raptopoulos V, Siewert B, Matthews JB.
Teardrop superior mesenteric vein: CT sign for unresect-
able carcinoma of the pancreas. AJR Am J Roentgenol
1999;173:1509–12.
12. Lepanto L, Arzoumanian Y, Gianfelice D, Perreault P,
Dagenais M, Lapointe R, et al. Helical CT with CT
angiography in assessing periampullary neoplasms: identi-
fication of vascular invasion. Radiology 2002;222:347–52.
887
 The British Journal of Radiology, 79 (2006), 888–892

Conversion factor for CT dosimetry to assess patient dose using a

256-slice CT scanner
S MORI, PhD, MPH, RT, K NISHIZAWA, PhD, MPH, M OHNO, MSc, MPH and M ENDO, PhD, MPH

Department of Medical Physics, National Institute of Radiological Sciences, Chiba 263-8555,

Japan

ABSTRACT. Recent rapid progress in CT technology has yielded equipment with large
numbers of detector rows and standard computed tomography dose index (CTDI) is
therefore no longer an adequate integration range. An integration range of 300 mm is
necessary to accurately measure dose under a nominal beam width of 128 mm due to
scattered radiation. However, such a long phantom is inconvenient to use routinely in
cone-beam CT patient dose checking. To assess patient dose accurately with standard
dosimetry methods, we determined a conversion factor (CF) which was calculated from
the weighted dose profile integral (DPIw) for the 300 mm integration range with a
Received 14 November
300 mm long CTDI phantom using a 300 mm long ionization chamber divided by that 2005
for the 100 mm integration range with a standard CTDI phantom (140 mm long) with a Revised 14 March 2006
100 mm long chamber. CF values increase with increasing nominal beam width and Accepted 14 March 2006
effective energy in the range from 1.5 to 2.0. CF values can also be adapted for use with
DOI: 10.1259/bjr/66519303
other CT systems as their dose profiles are thought to be analogous to those for the
300 mm phantom and are useful in any hospital situation to assess accurate patient ’ 2006 The British Institute of
doses using standard dosimetry methods. Radiology

State-of-the-art commercially based CT represents a We propose CF values obtained by calculating ratios of

marked improvement over conventional multislice CT
(MSCT), especially in cardiac imaging. Its craniocaudal
coverage without gantry movement, however, is typi-
cally only 20–40 mm and this has limited the width of
coverage for cine imaging in the craniocaudal direction.
To overcome this disadvantage, we developed a 256-slice
CT scanner. Volumetric cine imaging is realised by
scanning continuously at the same position (without
table movement); this provides a large amount of
diagnostic information and solves some of the limitations
of present helical CT methods [1–5].
However, since the maximum nominal beam width is
128 mm, which is three or four times larger than the
latest MSCT in common use, the dose is increased in
proportion to the scan time. Therefore, it is very
important to assess the dose for volumetric cine imaging.
For conventional CT dose measurement, a standard
computed tomography dose index (CTDI) phantom
(140 mm long) [6] and 100 mm long ionization chamber
have generally been used. However, the conventional CT
dose measurement is not sufficient in 256-slice CT
because its beam width is larger than the 100 mm long
ionization chamber. We therefore extended the FDA-
recommended CTDI phantom [6] to a length of 300 mm,
on the basis of our group’s previous report that phantom
length and integration range for dosimetry needed to be
at least 300 mm to represent more than 90% of the line
integral dose with a beam width between 20 mm and
138 mm [7]. Since a 300 mm CTDI phantom is quite
inconvenient, determining a conversion factor (CF) is
useful for all hospitals to assess patient dose accurately
using standard dosimetry methods.
the dose profile integrals (DPIs) with the standard CTDI
and 300 mm long phantoms.
Materials and methods
Second model of the 256-slice CT-scanner
The second model of our 256-slice CT scanner was
based on the design of the first model [8, 9] which used a
wide-area cylindrical 2D detector incorporating current
CT technology. The second model was mounted on the
gantry frame of a 16-slice CT [10] (Aquilion; Toshiba
Medical Systems, Otawara, Tochigi). The 256-slice CT
has 912 (transverse) 6 256 (craniocaudal) elements, each
approximately 0.5 mm60.5 mm at the centre of rotation.
The 128 mm total beam width allows the continuous use
of several collimation sets (e.g. 25660.5 mm,
12861.0 mm, 6462.0 mm). Large and small filters are
shaped to compensate for the variable path length of
each patient across the scan field of view (FOV). The
small filter is used for an object under 240 mm field of
view (FOV), and the large filter is used for over 240 mm-
FOV (e.g. chest and abdomen). The second model of the
256-slice CT scanner incorporates several improvements
over the first model, including the acceleration of
rotation time from 1.0 s to 0.5 s per rotation, and
elongation of the detector dynamic range from 16 to 18
bits [11]. The detector element consists of a Gd2O2S
ceramic scintillator and single-crystal silicon photodiode,
as used in conventional multislice CT. Maximum X-ray
exposure time per scan is 60 s and exposure can be

888 The British Journal of Radiology, November 2006

Conversion factor using 256-slice CT
repeated in a series of scans according to a pre-
determined program.
The x–y coordinate plane is parallel to the transverse
direction, and the z-coordinate axis is parallel to the
craniocaudal direction.
Phantoms
The length of the FDA-recommended CTDI phantom
[6] is at least 140 mm. This conventional phantom
contains holes just large enough to accept a pencil-
shaped ionization chamber. For cone-beam CT dose
measurement, the phantom length should be longer,
because the nominal beam width of 128 mm is longer
than that of the conventional CTDI dosimetry method.
Therefore, a CTDI phantom of 300 mm length designed
for cone-beam CT (Kyoto-kagaku, Kyoto, Japan) [7] was
used. The phantoms, 140 mm and 300 mm long, were
made of PMMA (polymethylmethacrylate) with dia-
meters of 160 mm for head examinations and 320 mm
for body examinations. Holes of 10 mm diameter for the
pencil-shaped ionization chamber were located parallel
to the rotation axis, and the centres of the holes were
located at the cylinder centre and also 10 mm below the
cylinder surface at 90 ˚ intervals.
Dosemeter probe
A pencil-shaped ionization chamber with an active
length of 100 mm (CT-10; Applied Engineering Inc.,
Tokyo, Japan) or 300 mm (CT-30; Applied Engineering
Inc., Tokyo, Japan) was connected to a dosemeter (AE-
132; Applied Engineering Inc., Tokyo, Japan). The
300 mm long chamber was an extended form of the
100 mm long pencil-shaped ionization chamber. The
dosemeter was calibrated at NMIJ (National Metrology
Institute of Japan) for the appropriate radiation qualities
by comparison with the secondary radiation standards.
Effective energy
Effective energy was calculated from the attenuation
curve of X-ray intensity. The 300 mm-length ionization
chamber was positioned at the horizontal plane that
passed through the rotational axis centre plane. The X-
ray tube was fixed under the ionization chamber, and X-
ray irradiation was initiated. The X-ray intensity was
measured by setting aluminium attenuators of various
thicknesses between the X-ray tube and the ionization
chamber [12]. The effective energy was derived from an
attenuation length that gave half of the X-ray intensity
produced without aluminium attenuators (half-value
layer (HVL)), calculated from the attenuation curve of
the beam intensity [13]. Effective energy was measured
at 0 mm, 50 mm, 100 mm and 150 mm along the
transverse direction for 80 kV, 100 kV, 120 kV and
135 kV with the nominal beam width of 128 mm.
The effective energy at the periphery was averaged from
x50 mm to x5100 mm for the small filter and from
x50 mm to x5150 mm for the large filters. Those
measurement ranges almost completely covered the head
The British Journal of Radiology, November 2006
and body phantoms, respectively. Average effective
energy (Eave) was estimated using the following equation:
1 2
Eave ~ Ec z Ep ð1Þ
3 3
where Ec and Ep were defined as the effective energy at
the centre and periphery, respectively.
Dose profile integral (DPI)
The DPIs were measured with the 300 mm long
chamber for the 300 mm long phantoms (extended
dosimetry) and with the 100 mm long chamber for the
140 mm long phantoms. The DPI was given as the
output of the ionization chamber. The weighted DPI
(DPIw) for x, y coordinates was given by:
1 2
DPIL,w ðvÞ~ DPIL,c ðvÞz DPIL,p ðvÞ mGy:mm ð2Þ
3 3
where DPIc and DPIp denote DPIs in the measurement
range L at the centre and periphery of the phantom. v
denotes the beam width.
Then we calculated the DPIw ratio as the CF value:
DPI300,w
CF~ ð3Þ
DPI100,w
Again, we omitted the variable v from the notation of
DPI. Using the above equations we calculated the CF for
various beam sizes.
The phantom was placed on the patient table and its
centre was aligned at the isocentre. An ionization
chamber was inserted into either the central or one of
the peripheral cavities of the phantom (other cavities
were filled with PMMA rods). The exposure (expressed
as C kg21) was obtained with the ionization chamber
dosemeter and converted to the values of the absorbed
dose for PMMA. The exposure length-integral
(expressed as C kg21 cm) was obtained with the
ionization chamber dosemeter and converted to the
values of DPI for PMMA with the f-factors (0.898 cGy
(C kg21)21) [14]. Absorbed doses at the centre and
periphery of the phantom were calculated using the
effective energy at the isocentre (x50 mm) and the
periphery, respectively.
All scans for the dose measurement were made in the
axial scan (non-helical) mode using the 256-slice CT.
Scan conditions were tube voltage of 80 kV, 100 kV,
120 kV, or 135 kV; 100 mAs; and nominal beam width of
8 mm, 32 mm, 64 mm, 96 mm, or 128 mm. The results
were averaged for 8 repeated DPI measurements.
Results
The effective energies for 80 kV, 100 kV, 120 kV and
135 kV, which were used in converting the exposure to
the values of absorbed dose, are summarized in Table 1.
The effective energy for the large filter is slightly higher
889
 S Mori, K Nishizawa, M Ohno and M Endo

Table 1. Effective energies for CT dose index (CTDI) for the 135 kV tube voltage. Other tube voltages have
phantoms similar values of linear correlation coefficient. DPI100,w
shows reasonably good linearity between the beam widths
Phantom Voltage (kV) Effective energy (keV)
of 8 mm and 128 mm (e.g. the linear correlation coefficient
Centre Periphery Average is 0.9910 for 135 kV). Since the 140 mm long phantom
could not cover primary dose and scattered dose in the
Head 135 47.9 54.9 52.5
(small filter) wider beam width such as 128 mm, second term poly-
120 40.6 49.9 46.8 nomial fitting was better for DPI100,w (e.g. the fitting
100 36.5 44.0 41.5 correlation is 0.9991 for 135 kV). DPI100,w and DPI300,w
80 33.3 40.8 38.3 were represented as follows:
Body 135 48.7 61.1 56.9
(large filter)
DPI100,w ðvÞ~av2 zbvzc
120 46.2 56.0 52.8 ð4Þ
100 43.3 50.3 48.0 DPI300,w ðvÞ~dvze
80 39.6 43.8 42.4

than that for the small filter due to the different filter
shapes.
Figure 1 shows DPI300,w and DPI100,w which are normal-
ized against 100 mAs for the body phantom. The relation-
ship between DPI300,w and beam width shows good
linearity and the linear correlation coefficient is 0.9996
These coefficients are summarized in Table 2.
The DPIws also increase with increasing tube voltage;
in particular, DPI300,w for 135 kV for the body phantom is
4.5 times higher than at 80 kV. DPIs for the head
phantom (results not shown) are similar to those for
the body phantom.

(a) (b)

Figure 1. The relationships between weighted dose profile integral (DPIw) at the centre of the body phantom and the beam
width. (a) DPI300,w. (b) DPI100,w.

Table 2. Coefficients for the fitting equations for DPI100,w, and DPI300,w
Phantom Tube voltage Averaged Coefficient
(kV) effective energy
(keV) a b c d e

Head 135 52.5 20.0355 18.4 81.2 25.4 210.0

120 46.8 20.0263 13.9 56.1 18.9 141.7
100 41.5 20.0168 9.1 32.6 12.2 83.7
80 38.3 20.0085 5.0 17.3 7.0 34.6
Body 135 56.9 20.0086 8.4 63.2 13.8 120.5
120 52.8 20.0084 6.5 32.0 10.1 70.4
100 48.0 20.0052 4.1 16.2 6.2 36.5
80 42.4 20.0026 2.1 6.8 3.1 16.1
DPI, dose profile integral.

890 The British Journal of Radiology, November 2006

Conversion factor using 256-slice CT

(a) (b)

Figure 2. The relationships between the conversion factor and the beam width for (a) the body and (b) the head phantoms.

Figure 2 shows the relationships between CF value Discussion

and the nominal beam width for each tube voltage. CF
value increases gradually up to the nominal beam width
of 128 mm with increasing nominal beam width as well
as tube voltage. CF values range from 1.5 to 2.0 for the
beam widths and tube voltages.
The relationships between CF value and the effective
energy are shown in Figure 3. The CF increases with
increasing effective energy, and it has the same tendency
as in Figure 2.
DPIws were measured using the standard CTDI and
the 300 mm long phantoms for nominal beam widths of
9–128 mm. We believe that the standard 100 mm
ionization chamber and 140 mm long phantom are not
reliable, even for the 8 mm beam width [7, 15].
Although the standard phantoms measured DPIws for
the nominal beam width less than 100 mm, CF was not
1.0. In particular, CF value ranged from 1.5 to 1.8 for the

(a) (b)

Figure 3. The relationships between the conversion factor and the effective energy for (a) the body and (b) the head phantoms.

The British Journal of Radiology, November 2006 891


nominal beam widths of 8 mm and 32 mm, which can be
selected by the MSCT. These results can be explained by
scattered radiation which depends on the phantom
length and which affects DPI strongly [7]. Therefore,
CF value is important to both MSCT and 256-slice CT in
order to assess doses accurately.
CF should be adaptable to MSCT systems because the
scatter tails of their dose profiles are thought to be
analogous to those for the 256-slice CT, if the effective
energies are similar [16].
Conclusion
Use of the standard 100 mm ionization chamber and
140 mm phantoms instead of the inconvenient 300 mm
long alternatives is made possible by use of the CF values
determined in our paper. It was found that the standard
system was unreliable even at 8 mm beam width. The
relationship demonstrated between the CF and effective
energy will allow this method to be adapted to other CT
systems.
References
1. Mori S, Endo M, Obata T, Murase K, Fujiwara H, Susumu
K, et al. Clinical potentials of the prototype 256-detector
row CT-scanner. Acad Radiol 2005;12:148–54.
2. Mori S, Kondo C, Suzuki N, Yamashita H, Hattori A,
Kusakabe M, et al. Volumetric cine imaging for cardiovas-
cular circulation using prototype 256–detector row com-
puted tomography scanner (4-dimensional computed
tomography) a preliminary study with a porcine model. J
Compt Assist Tomogr 2005;29:26–30.
3. Mori S, Obata T, Kishimoto R, Kato H, Murase K, Fujiwara
H, et al. Clinical potentials for dynamic contrast-enhanced
hepatic volumetric cine imaging with the prototype 256-
MDCT scanner. AJR Am J Roentgenol 2005;185:253–6.
4. Mori S, Obata T, Nakajima N, Ichihara N, Endo M.
Volumetric perfusion CT using prototype 256-detector
row computed tomography scanner: preliminary study
with healthy porcine model. Am J Neuroradiol 2005;26:
2536–41.
892
S Mori, K Nishizawa, M Ohno and M Endo
5. Kondo C, Mori S, Endo M, Kusakabe K, Suzuki N, Hattori
A, et al. Real-time volumetric imaging of human heart
without ECG gating by 256-detector row computed
tomography: initial experience. J Compt Assist Tomogr
2005;29:694–8.
6. US FDA Code of Federal Regulations. Diagnostic x-ray
systems and their major components. Govt. Printing Office,
August 1984;21CFR 11020.23.
7. Mori S, Endo M, Nishizawa K, Tsunoo T, Aoyama T,
Fujiwara H, et al. Enlarged longitudinal dose profiles in
cone-beam CT and the need for modified dosimetry. Med
Phys 2005;32:1061–9.
8. Endo M, Mori S, Tsunoo T, Kandatsu S, Tanada S, Aradate
H, et al. Development and performance evaluation of the
first model of 4D CT-scanner. IEEE Trans Nucl Sci
2003;50:1667–71.
9. Mori S, Endo M, Tsunoo T, Kandatsu S, Tanada S, Aradatte
H, et al. Physical performance evaluation of a 256-slice CT-
scanner for four-dimensional imaging. Med Phys
2004;31:1348–56.
10. Saito Y, Aradate H, Igarashi K, Ide H. Large area 2-
dimensional detector for real-time 3–dimensional CT (4D
CT). Proc SPIE 2001;4320:775–82.
11. Endo M, Mori S, Kandatsu S, Tanada S, Sugihara N, Saito Y,
et al. Development of real 4D CT with real-time reconstruc-
tion and display. IEEE NSS/MIC Conf Record
2006;M11–252.
12. Edyvean S, Lewis M, Britten A. CT scanner dose survey:
measurement protocol Ver. 5. http://www.impactscan.
org/1997 [Accessed 5 May 2006].
13. Hubbell J, Selzer S. Tables of x-ray mass attenuation
coefficients and mass energy-absorption coefficients.
National Institute of Standard and Technology (NIST)
physical reference data 1996.
14. International Commission on Radiation Units and
Measurements. Radiation dosimetry: X rays generated at
potentials of 5 to 150 kV: ICRU. Publication. 17,
Washington D.C.:ICRU Publications, 1970.
15. Nakaonechny K, Fallone B, Rathee S. Novel methods of
measuring single scan dose profiles and cumulative dose in
CT. Med Phys 2005;32:98–109.
16. Felmlee JP, Gray JE, Leetzow ML, Price JC. Estimated fetal
radiation dose from multislice CT studies. AJR Am J
Roentgenol 1990;154:185–90.
The British Journal of Radiology, November 2006
The British Journal of Radiology, 79 (2006), 893–898

Equivalent dose to organs and tissues in hysterosalpingography

calculated with the FAX (Female Adult voXel) phantom
1 1 1
R KRAMER, PhD, H J KHOURY, PhD, C LOPES, PhD and 2J W VIEIRA, PhD

1
Departamento de Energia Nuclear, Universidade Federal de Pernambuco, Avenida Prof. Luiz
Freire, 100, Cidade Universitária, CEP 50740-540, Recife, PE and 2Escola Politécnica, UPE, Recife, PE,
Brazil

ABSTRACT. Hysterosalpingography (HSG) is a radiological examination indicated for

investigating infertility or uterine and tubal pathologies. Women who undergo HSG
are relatively young, typically between 20 years and 40 years, and equivalent doses to
the ovaries are usually reported to be around 4 mSv per examination. A review of
studies on patient dosimetry in HSG revealed that almost all absorbed doses to organs
and tissues had been calculated with conversion coefficients (CCs) based on
hermaphrodite versions of MIRD5-type phantoms. The CCs applied had been taken
from data sets for abdominal or pelvic examinations because CCs for HSG examination
were not available. This study uses the FAX (Female Adult voXel) phantom in order to
calculate equivalent doses to radiosensitive organs and tissues especially for exposure
Received 26 December
conditions used in HSG. The calculations were also performed for the MIRD5-type EVA 2005
phantom to demonstrate the influence of anatomical differences on organ equivalent Revised 5 April 2006
dose. The results show organ and tissue equivalent doses as a function of the variations Accepted 12 April 2006
of the exposure conditions. At 4.56 mSv the ovarian equivalent dose calculated for the
DOI: 10.1259/bjr/61874578
FAX phantom is about 21% greater than the average ovarian equivalent dose reported
in the literature, which reflects the anatomical differences between the FAX and the ’ 2006 The British Institute of
MIRD5-type phantoms. Radiology

Hysterosalpingography (HSG), a radiological exam- voxel-based phantoms, which are a true-to-nature

ination which delivers relatively high equivalent doses to
the ovaries and the uterus, is used to examine the uterine
cavity and the patency of the Fallopian tubes. Common
indications for HSG are primary and secondary inferti-
lity, assessment of tubal patency following reversal of
sterilization, of tubal blockage following a difficult
sterilization and of the uterine cavity following division
of an intrauterine septum.
Equivalent doses to the ovaries from HSG exami-
nations are usually around 4 mSv, which triggered a
series of investigations on patient dosimetry and on
possibilities for the reduction of patient exposure.
Table 1 summarizes typical values for the entrance
surface air kerma (ESAK), including the contribution
from backscattered radiation, and the equivalent dose
to the ovaries reported in recent publications [1–5],
in which the equivalent dose to the ovaries was
calculated by multiplying the ESAK by a conversion
coefficient (CC) determined by Monte Carlo methods
for hermaphrodite MIRD5-type phantoms, i.e. male
bodies with female organs such as ovaries, uterus and
breasts [6–8].
However, during the last decade in radiological
protection the mathematical MIRD5-type phantoms
have increasingly been replaced by tomographic or
This work was funded by the Conselho Nacional de Desenvolvimento
Cientı́fico e Tecnológico (CNPq) and the Fundação de Amparo à
Ciência do Estado de Pernambuco (FACEPE).
representation of the human body. The recently devel-
oped FAX (Female Adult voXel) phantom [9], together
with the traditional MIRD5-type EVA phantom [10],
have been selected to simulate HSG examinations in
order to see the influence of the different anatomies on
organ and tissue equivalent doses, and also to demon-
strate the effect of varying exposure parameters, tube
voltage, filtration, field size, field orientation, focus-to-
film distance and projection on these doses.
Materials and methods
The phantoms
The recently developed FAX (Female Adult voXel)
phantom has been segmented from CT images of
patients [9]. Organ and tissue masses correspond to the
anatomical specifications recommended by the
International Commission on Radiological Protection
(ICRP) in its Publication 89 for the female reference
adult [11], while tissue compositions and densities are
based on data published by the International
Commission on Radiation Units (ICRU) in its report
No. 44 [12].
The EVA phantom has been developed from the first
hermaphrodite MIRD5 phantom, which already had
ovaries and a uterus, by scaling down the male body to
the height of the female reference adult from ICRP
Publication 23 and by introducing female breasts [10].

The British Journal of Radiology, November 2006 893


Table 1. Entrance surface air kerma and ovarian equivalent doses reported in references [1–5]
Reference ESAK (mGy) Ovarian dose (mSv)
Fife et al [1] 13.3 2.8
Fernandez et al [2] 23.4 4.6
Gregan et al [3] 13.1 3.1
Calicchia et [4] 24.6 4.6
Khoury et al [5] 19.2 2.9
Average 18.7 3.6
Tissue compositions and densities of the EVA phantom
have been taken from an early MIRD5 publication [13].
The EGS4 Monte Carlo code
The EGS4 Monte Carlo code [14] simulates coupled
electron-photon transport through arbitrary media. The
default version of EGS4 applies an analogous Monte
Carlo method, which was used for the calculations of
this investigation. Rayleigh scattering has been taken
into account, but secondary electrons have not been
transported. With respect to the simulation of radio-
logical examinations, a special user code has been
developed that outputs absorbed dose to radiosensitive
organs and tissues normalized to the ESAK. The X-ray
spectra have been taken from the IPEM spectra catalogue
[15].
Exposure conditions
Based on a review of the exposure conditions reported
in references [1–5], the following representative irradia-
tion parameters have been identified for the simulation
of the HSG examination:
X-ray generator: constant potential
Target: tungsten, 17 ˚
Voltage: 70–120 kV at a constant potential (kVcp)
Filtration: 2.0–4.0 mm Al
Projection: anterior–posterior (AP), posterior–anterior
(PA)
Field size: 18 cm624 cm, 24 cm630 cm in the detec-
tor plane
Field position: Centred on uterus
Focus to skin distance (FSD): 70 cm, 80 cm, 90 cm
Focus to film distance (FFD): 100 cm, 110 cm, 120 cm
Figures 1 and 2 show silhouettes of the FAX and
the EVA phantoms with X-ray beams, field sizes and
FFDs, the uterus in grey and the ovaries in black,
respectively.
Results
The results will be presented as CCs between organ
equivalent doses and the ESAK as a function of the tube
voltage and the filtration. ESAK refers to the centre of the
radiation field at the surface of the patient’s body. As
tube voltage and filtration increase, so do the CCs.
However, the absolute value of the equivalent dose to
organs will decrease, because increasing the tube voltage
894
R Kramer, H J Khoury, C Lopes and J W Vieira
OD/ESAK (Sv /Gy) Average voltage (kVcp) Field size (cm6cm)
0.21 82 24630
0.20 80 24630
0.24 78 24630
0.19 72 24630
0.15 70 30624
0.20
or the filtration reduces the mAs necessary for constant
exposure to the detector system.
Conversion coefficients
Tube voltage
Figure 3 shows equivalent dose to the ovaries of the
FAX and the EVA phantoms normalized to the ESAK as
a function of the tube voltage for field sizes of
18 cm624 cm, and 24 cm630 cm, respectively. For the
whole range of tube voltages, the FAX ovarian equiva-
lent dose is ca. 25% greater than the EVA ovarian
equivalent dose for the smaller field size, while for the
large field this number is about 18%. The reason for
the differences is the different depth at which the ovaries
are located in the two phantoms – beginning at 6.5 cm
depth in the FAX phantom and beginning at 8.3 cm
depth in the EVA phantom.
The average CC between ovarian equivalent dose and
ESAK in Table 1 for references [1–3] is 0.22, and for
references [4, 5] it is 0.17. This should be compared with
the CCs for the EVA phantom in Figure 3 for a
24 cm630 cm field, for 80 kVcp and 70 kVcp, respec-
tively. For the EVA phantom one finds CCs of 0.24
for 80 kVcp, and of 0.20 for 70 kVcp. These EVA CCs
agree reasonably well with the CCs from Table 1, if one
takes into account the differences between the real
HSG examination and its simulation that is not known
here, like the position of the ovaries, patient thickness,
X-ray generator, filtration, etc. When the tube voltage
increases from 70 kVcp to 120 kVcp, the CCs in
Figure 3 increase by ca. 65% over this range of tube
voltage.
Figure 4 presents CCs for the uterus as function of the
tube voltage for both phantoms and the two field sizes
already mentioned. The uterus equivalent dose of the
EVA phantom is ca. 36% greater than the equivalent dose
to the uterus of the FAX phantom over the whole range
of photon energies. Again the explanation comes from
different depths at which the uteri are located in the two
phantoms.
When the tube voltage increases from 70 kVcp to
120 kVcp, the CCs in Figure 4 increase by ca. 55% over
this range of tube voltage.
Filtration
Figure 5 shows CCs for the ovaries for the two
phantoms as function of the filtration for the two field
The British Journal of Radiology, November 2006
Hysterosalpingography equivalent dose to the FAX phantom
Figure 1. The FAX phantom: hysterosalpingography (HSG) exposure set-up for field524 cm630 cm and FFD 5100 cm.
sizes. When the filtration increases from 2.0 mm Al to
4.0 mm Al, the CCs in Figure 5 increase by ca. 20% over
this range of filtration.
Field orientation
Taking the dimensions of the field given as width6
height, then one can see from Table 1 that HSG
examinations in Recife/Brazil are usually made with
the larger field side representing the width, while the
opposite is the case for the other references cited. In
order to get an idea of the effect of the field orien-
tation on the organ equivalent doses, the calcula-
tions shown so far have been repeated with the
larger field side representing the field width. For the
radiation fields A1524 cm618 cm, A2518 cm624 cm,
B1530 cm624 cm and B2524 cm630 cm, the following
results have been found for the range of tube voltages
between 70 kVcp and 120 kVcp:
Uterus – No change of the uterus equivalent dose has
been found when changing the field orientation from
A2 to A1, or from B2 to B1.
The British Journal of Radiology, November 2006
Ovaries – For the FAX phantom an increase of the
ovarian equivalent dose has been found to be between
5% and 8%, and between 2% and 4% for changing the
field orientation from A2 to A1, and from B2 to B1,
respectively.
For the EVA phantom an increase of the ovarian
equivalent dose has been found between 13% and 16%,
and between 4% and 6% for changing the field
orientation from A2 to A1, and from B2 to B1,
respectively.
Focus-to-skin distance (FSD)
Variations of the FSD between 100 cm and 120 cm did
not significantly change the equivalent dose to the
ovaries and the uterus.
Equivalent doses per radiograph
Although the CCs shown in Figures 3 and 4 increase
with the tube voltage, the absolute equivalent dose per
895

Figure 2. The EVA phantom: hysterosalpingography (HSG) exposure set-up for field524 cm630 cm and FFD5100 cm.
radiograph usually decreases with increasing tube
voltage, because for constant exposure to the detector
system the ESAK decreases with increasing tube voltage.
Figure 6 shows the ESAK measured with a PMMA
phantom as a function of the tube voltage. The reduction
Figure 3. Conversion coefficient between equivalent dose
to the ovaries and entrance surface air kerma as function of
the tube voltage for field sizes of 18 cm624 cm, and
24 cm630 cm.
896
R Kramer, H J Khoury, C Lopes and J W Vieira
is about 50% when the tube voltage increases from
70 kVcp to 80 kVcp, while for voltages above 80 kVcp
the reduction is about 10–15% per 10 kVcp increase.
Application of the ESAK data from Figure 6 to the CCs
from Figures 3 and 4 provides the organ equivalent
Figure 4. Conversion coefficient between equivalent dose
to the uterus and entrance surface air kerma as function of
the tube voltage for field sizes of 18 cm624 cm, and
24 cm630 cm.
The British Journal of Radiology, November 2006
Hysterosalpingography equivalent dose to the FAX phantom
Figure 5. Conversion coefficient between equivalent dose
to the ovaries and entrance surface air kerma as function of
the filtration for field sizes of 18 cm624 cm, and
24 cm630 cm.
doses per radiograph shown in Figure 7, which confirm
the reduction of exposure to the patient with increasing
tube voltage by up to 50%.
The influence of the projection on organ
equivalent dose
In the HSG studies cited in Table 1, anterior–posterior
(AP) was the most frequently used projection. However,
in some countries undercouch X-ray tubes are more
prevalent than overcouch ones, i.e. posterior–anterior
(PA) projections dominate the HSG examinations.
Therefore some of the calculations reported above for
the FAX phantom have also been done for PA projection.
Figure 8 presents ratios between equivalent doses for PA
and AP projection for the ovaries and the uterus.
The ratios show that using the PA projection can lead
to significant equivalent dose reductions compared with
the application of the AP projection. The main reason is
the presence of the pelvis in the radiation field, which
shields internal abdominal organs like the ovaries and
the uterus from the incident photon radiation. The
position of the ovaries in the frontal part of the FAX
abdomen causes an additional reduction of equivalent
dose because for PA incidence the ovaries are located at a
greater depth. For both field sizes used in this study, the
Figure 6. Entrance surface air kerma per radiograph
measured with a homogeneous PMMA phantom 19 cm thick
as function of the tube voltage.
The British Journal of Radiology, November 2006
Figure 7. Ovarian and uterine equivalent dose per radio-
graph in hysterosalpingography (HSG) as function of the
tube voltage.
effect of equivalent dose reduction is about 60–75% for
the ovaries, and about 30–43% for the uterus between
70 kVcp and 120 kVcp tube voltage.
Equivalent dose per HSG examination with AP
projection
The results of the previous sections have shown that
the CCs change with the variations of the exposure
parameters. However, a tube voltage of 80 kVcp, a
filtration of 3.0 mm Al, a field size of 24 cm630 cm, a
FSD of 100 cm and AP projection can be considered
typical for a HSG examination. For these settings, one
finds the following CCs for the FAX phantom in
Figures 3 and 4:
Ovaries: 0.244 Sv Gy21
Uterus: 0.205 Sv Gy21
From Table 1, one can find an average ESAK of
18.7 mGy. In absolute terms, this means that for a typical
HSG examination of the FAX phantom one gets:
Equivalent dose to the ovaries: 4.56 mSv
Equivalent dose to the uterus: 3.83 mSv
The average ovarian dose for the MIRD5-type phan-
tom from Table 1 is 3.6 mSv, i.e. 21% less than the value
for the FAX phantom, which is in agreement with the
findings of Figure 3.
Figure 8. Ratios between organ equivalent doses for postero-
anterior (PA) and anteroposterior (AP) projection as a function
of the tube voltage.
897

Conclusion
Equivalent dose to the ovaries and to the uterus was
calculated with the FAX/EGS4 exposure model for HSG
examinations as a function of tube voltage, filtration and
FSD. The results have been compared with similar data
for the MIRD5-type EVA phantom, which was often
used in recent studies.
The data have shown that ovarian equivalent doses are
18–25% greater in the FAX phantom compared with the
corresponding values for the EVA phantom due to a
2 cm difference between the location of this organ below
the surface in the two phantoms.
For similar reasons, the uterus equivalent dose of the
EVA phantoms was found to be 36% greater than the
corresponding value for the FAX phantom.
As for the variation with exposure parameters, the
calculations revealed an increase of 65% of the ovarian
CC, and of 55% of the uterus CC for an increase of tube
voltage from 70 kVcp to 120 kVcp. But as demonstrated,
the absolute values of ovarian and uterus equivalent dose
decrease by up to 50% when the tube voltage increases.
Therefore, increasing the tube voltage is usually recom-
mended as step to reduce the exposure to the patient.
This applies also to the increase of filtration, although
in absolute terms this effect was not shown here, and
especially also to the choice of the PA projection.
Finally it has to be pointed out that the orientation of
the field with regard to width and height can cause
differences of between 2% and 15% for the organ
equivalent doses discussed in this presentation.
The CCs presented can serve as a tool for patient
dosimetry. If the ESAK, tube voltage, filtration and field
size are known, one can multiply the ESAK by the
appropriate CC from Figures 3–5 to obtain an estimate of
the equivalent dose to the ovaries or to the uterus.
Acknowledgments
The authors would like to thank the Conselho
Nacional de Desenvolvimento Cientı́fico e Tecnológico
- CNPq and the Fundação de Amparo à Ciência do
Estado de Pernambuco - FACEPE for the financial
support.
References
1. Fife IAJ, Wilson DJ, Lewis CA. Entrance surface and
ovarian doses in hysterosalpingography. Br J Radiol
1994;67:860–3.
898
R Kramer, H J Khoury, C Lopes and J W Vieira
2. Fernandez JM, Vañó E, Guibelalde E. Patient doses in
hysterosalpingography. Br J Radiol 1996;69:751–4.
3. Gregan ACM, Peach D, McHugo JM. Patient dosimetry in
hysterosalpingography: a comparative study. Br J Radiol
1998;71:1058–61.
4. Calicchia A, Chiacchiarelli L, de Felice C, Gigliotti T,
Indovina PL, Mazzei F, et al. Evaluation of effective dose in
histerosalpingography. Radiat Prot Dosim 1998;80:159–61.
5. Khoury HJ, Maia A, Oliveira M, Drexler G, Kramer R.
Patient dosimetry in hysterosalpingography. IAEA
International Conference on Radiological Protection of
Patients in Diagnostic and Interventional Radiology,
Nuclear Medicine and Radiotherapy. Malaga, Spain, 2001.
6. Peterson LE, Rosenstein M. Computer program for tissue
doses in diagnostic radiology. Food and Drug
Administration, Centre for Devices and Radiological
Health, Rockville, Maryland, USA, 1989.
7. Jones DG, Wall BF. Organ doses from medical X-ray
examinations calculated using Monte Carlo techniques.
NRPB-R186. London: HMSQ, 1985.
8. Hart D, Jones DG, Wall BF. Estimation of effective dose in
diagnostic radiology from entrance surface dose and dose-
area product measurements. NRPB-R262. London: HNSQ,
1994.
9. Kramer R, Khoury HJ, Vieira JW, Loureiro ECM, Lima VJM,
Lima FRA, et al. All about FAX: a female adult voxel
phantom for monte carlo calculation in radiation protection
dosimetry. Phys Med Biol 2004;49:5203–16.
10. Kramer R, Zankl M, Williams G, Drexler G. ‘‘The calcula-
tion of dose from external photon exposures using reference
human phantoms and Monte Carlo methods. Part I: The
male (Adam) and female (Eva) adult mathematical phan-
toms’’. GSF-Report S-885. Institut für Strahlenschutz, GSF-
Forschungszentrum für Umwelt und Gesundheit,
Neuherberg-München, 1982.
11. ICRP 89. ‘‘Basic Anatomical and Physiological Data for Use
in Radiological Protection: Reference Values’’. ICRP
Publication 89, International Commission on Radiological
Protection. Oxford: Pergamon Press, 2003.
12. ICRU 44. ‘‘Tissue substitutes in radiation dosimetry and
measurement’’. ICRU Report 44. International Commission
on Radiation Units and Measurements, Bethesda, MD, 1989.
13. Snyder WS, Ford MR, Warner GG, Fisher HL. Estimates of
absorbed fractions for monoenergetic photon sources
uniformly distributed in various organs of a heterogeneous
phantom. Medical Internal Radiation Dose Committee
(MIRD) Pamphlet No. 5. J Nucl Med 1969;10(Suppl.):3.
14. Nelson WR, Hirayama H, Rogers DWO. ‘‘The EGS4 Code
System’’. SLAC-265 Stanford Linear Accelerator Center,
Stanford University, Stanford, California, 1985.
15. Cranley K, Gilmore BJ, Fogarty GWA, Desponds L.
Catalogue of diagnostic X-ray spectra and other data. The
Institute of Physics and Engineering in Medicine (IPEM),
Report No. 78, Electronic version prepared by D Sutton,
September 1997.
The British Journal of Radiology, November 2006
The British Journal of Radiology, 79 (2006), 899–904

Comparison of radiation doses to patients undergoing standard

radiographic examinations with conventional screen–film
radiography, computed radiography and direct digital
radiography
1
G COMPAGNONE, PhD, 2M CASADIO BALENI, MD,
1
L PAGAN, MSc,
2
F L CALZOLAIO, 2L BAROZZI, MD
and 1C BERGAMINI, PhD

1
Medical Physics Department and 2Accident and Emergency Department, S. Orsola-Malpighi
Hospital, Via Massarenti 9, 40138 Bologna, Italy

ABSTRACT. New flat-panel direct digital radiography equipment has recently been
installed in our Accident and Emergency Department; its characteristics and versatility
are well suited to the work undertaken in this environment. The aim of this study was
to compare radiation doses to patients undergoing standard radiographic
examinations using conventional screen–film radiography, computed radiography and
direct digital radiography; entrance surface dose and effective dose were calculated for Received 6 May 2005
six standard examinations (a total of 10 projections) using standard patient exposure Revised 13 March 2006
parameters for the three imaging modalities. It was found that doses for computed Accepted 12 April 2006
radiography (all examinations) were higher than the doses for the other two
DOI: 10.1259/bjr/57138583
modalities; effective doses for direct digital radiography were ,29% and ,43% lower
than those for screen–film radiography and computed radiography, respectively. The ’ 2006 The British Institute of
image quality met the criteria in the European guidelines for all modalities. Radiology

In Accident and Emergency (A&E) departments,
patients undergo radiological examinations to establish
their clinical condition. The importance of having a
radiologist to report the images is documented in the
literature; discrepancies of up to 58% have been reported
between primary care physicians and radiologists [1].
In our hospital, there are eight departments where X-
ray equipment is used. One of these is the A&E
department, comprising three X-ray rooms where
approximately 50 000 standard radiographic examina-
tions out of a total of 400 000 are performed annually.
In addition to outpatient casualties presenting at A&E,
the department also performs radiological examinations
for inpatients at night and at the weekend, when the
other radiology departments in the hospital are closed. It
is for these reasons that new technologies may be
installed in A&E departments first.
With digital radiography, namely computed radio-
graphy (CR) and direct digital radiography (DDR),
becoming a viable technology for acquiring X-ray
images, departments are looking to replace conventional
screen–film radiography (SFR).
Potential advantages of digital systems over conven-
tional radiography are well known [2], i.e. that they have a
greater dynamic range, wider exposure latitude, post-
processing facilities available, and that there is improved
access to images by clinicians and decreased film costs.
New flat-panel DDR equipment has recently been installed
in the A&E department in our hospital because its
characteristics and versatility are well suited to the work
undertaken in this environment, e.g. being able to manage
taking X-rays of critically ill patients with relative ease.
The aim of this study was to compare radiation
doses to patients undergoing standard radiographic
examinations using SFR, CR and DDR; entrance skin
dose (ESD) and effective dose (E) were calculated for six
standard examinations (a total of 10 projections) using
standard patient exposure parameters for the three
imaging modalities. ESD and E are considered to be
efficient and powerful parameters in the protection of
patients [3, 4].
Materials and methods
In 2003, the SFR system in the A&E department was
replaced with a CR system and in May 2004, the X-ray
equipment in X-ray Room 2 was replaced with a Siemens
Axiom Aristos FX radiography system with a flat panel
detector (Siemens, Erlangen, Germany). The Axiom Aristos
is a multifunctional system which enables virtually the
entire range of radiographic applications to be performed
in a single room. A brief description of the X-ray equipment
used before and after May 2004 is given in Table 1.
The SFR system comprised Kodak TMAT G/RA film
(Kodak, Rochester, NY) and Trimax Regular screen
(nominal speed class 400). The CR system comprises
two Kodak CR readers (CR-850 and CR-900) and three
sizes of general purpose image plates, i.e. 18 cm624 cm,
24 cm630 cm and 35 cm643 cm. The image plates
contain a europium-activated barium fluoro-halide
phosphor with 179262392, 204862500 and 204862500
matrix sizes, and 0.100 mm, 0.120 mm and 0.172 mm
pixel pitches, respectively.

The British Journal of Radiology, November 2006 899

 G Compagnone, M Casadio Baleni, L Pagan et al

Table 1. Description of the X-ray equipment used in the Accident and Emergency Department (before and after May 2004)
Room Before May 2004 After May 2004

1 IAE OVERCOUCH TUBE; PLANMECA ORTHOPANTOMOGRAPH; Unchanged

Tube output @ 80 kV:
77 mGy/mAs @ 75 cm; Focal
spot sizes: 0.6 mm, 1.2 mm;
Antiscatter grid ratio 12:1,
44 strips/cm
2 COMET OVERCOUCH TUBE;
Tube output @ 80 kV:
65 mGy/mAs @ 75 cm;
Focal spot sizes: 0.6 mm, 1.2 mm;
Antiscatter grid ratio 12:1,
36 strips/cm
3 COMET OVERCOUCH TUBE;
Tube output @ 80 kV:
69 mGy/mAs @ 75 cm;
Focal spot sizes: 0.6 mm, 1.3 mm;
Antiscatter grid ratio 12:1,
36 strips/cm
Tube output @ 76 kV:
167 mGy/mAs @ 35 cm;
Focal spot size: 1.2 mm
COMET CEILING-MOUNTED TUBE; SIEMENS AXIOM ARISTOS FX;
Tube output @ 80 kV: Tube output @ 81 kV:
78 mGy/mAs @ 75 cm; 118 mGy/mAs @ 75 cm;
Focal spot sizes: 0.6 mm, 1.2 mm; Focal spot sizes: 0.6 mm, 1.0 mm;
Antiscatter grid ratio 10:1, 36 strips/cm Antiscatter grids ratio 12:1, 40
strips/cm, fgd5115 cm and 180 cm
VARIAN CEILING-MOUNTED TUBE; Unchanged
Tube output @ 80 kV:
55 mGy/mAs @ 75 cm;
Focal spot sizes: 0.6 mm, 1.2 mm;
Antiscatter grid ratio 12:1, 36 strips/cm

The DDR system comprises a 0.5 mm thick thallium
doped caesium iodide phosphor (CsI:Tl). The readout
array, to which the phosphor is coupled, consists of a
43 cm643 cm amorphous silicon (a-Si) photodiode and
a thin film transistor (TFT) array (300063000 pixels and
0.143 mm pixel pitch). The X-rays interact with the
phosphor and release light photons. The light from the
phosphor promotes charge in the photodiode array with
the amount of charge being proportional to the intensity
of the incoming X-ray photons. Each TFT element is then
sequentially addressed so that the charge in the photo-
diodes is read out and digitized. The detector resolution
is limited by pixel pitch; this may be reduced by the
effect of light spread in the phosphor layer. The images
were finally reported by a consultant radiologist on a
Kodak Autorad workstation.
ESD were calculated for six standard radiographic
examinations (a total of 10 projections): anteroposterior
(AP) Abdomen, posteroanterior (PA) Chest, lateral (LAT)
Chest, AP Lumbar Spine, LAT Lumbar Spine, LAT
Lumbo-Sacral Joint, AP Pelvis, AP Skull, LAT Skull, AP
Urinary Tract.
To calculate E, first we measured the X-ray tube output
using a technique previously described [5]. For each X-ray
tube, 30 air kerma measurements were made with an
ionization chamber (model 90X6-6, connected to a
Radiation Monitor Controller model 9010; Radcal
Corporation, Monrovia, USA) held in a scatter-free support
on the central axis of the X-ray beam (Table 1). Instruments
are calibrated annually, with the calibration traceable to an
SIT (National Calibration Service in Italy) centre.
The Harpen mathematical model [6] was then used to
obtain the X-ray tube output for every tube voltage
(kVp), milliampere-seconds product (mAs) and focus-to-
skin distance (FSD) used in each of the clinical protocols
(provided by the radiologist). A backscatter factor of 1.35
was used to calculate the ESD from air kerma, as
suggested in the European Guidelines [7]. Finally, E
has been calculated from ESD using the NRPB conver-
sion coefficients [8].
In order to correctly apply these factors, it is necessary
to know the tube potential used in clinical protocols and
the total X-ray tube filtration. The total tube filtration
is measured annually as part of a quality assurance
programme; acceptance, status and constancy tests are
performed by the Medical Physics Department (the
department is certified to UNI EN ISO 9001-2000).
Assessment of image quality was undertaken by
Consultant Radiologists within A&E; images from the
three modalities, i.e. SFR, CR and DDR, were qualita-
tively evaluated to ensure that the quality criteria for
diagnostic radiographic images of the European
Guidelines [7] were met.
Results
The parameters used for standard radiographic exam-
inations on adult patients are reported in Table 2. It
should be noted that for some of the examinations
performed with the Siemens Axiom Aristos equipment,
the FSDs are different to those used with SFR and CR as
the system is self-positioning and self-centring; this is a
feature of the equipment.
Manufacturers of digital systems often recommend the
use of low tube potentials, but this will lead to less
penetrating beams and hence possibly higher doses. It is
important, therefore, to ensure that systems are optimized
[9]. For Kodak CR systems, the exposure index (EI) is the
average raw data pixel value within the anatomical area
exposed [10]: an EI ,2000 is indicative of optimum
exposure level. EI is directly related to patient dose by
the equation EI510006log10(exposure in mR)+2000 [10].
Although the use of low tube potentials is often
recommended, clinically diagnostic images can be
achieved using high tube potentials. An example of this
is the PA chest radiograph obtained at 125 kVp (Figure 1);
an EI very close to the value recommended by the
manufacturer was obtained, i.e. EI51970, and the quality
criteria of the European Guidelines [7] were satisfied. With
the digital systems we generally used a higher tube
potential and lower mAs to reduce the patient dose
(Table 2). Careful adjustment of these parameters will
allow patient doses to be optimized without sacrific-
ing image quality by taking full advantage of the

900 The British Journal of Radiology, November 2006

Comparison of SFR, CR and DDR radiation doses in standard examinations

Table 2. Parameters used for standard radiographic examinations using conventional screen–film radiography (SFR), computed
radiography (CR) and direct digital radiography (DDR)
Examination and projection kVp FSD (cm) Total filtration (mm Al)

SFR CR DDR SFR CR DDR SFR CR DDR

AP Abdomen 85 86 81 92 92 79 4.0 4.0 3.3

PA Chest 120 120 125 151 151 155 4.3 4.4 3.3
LAT Chest 130 135 125 140 140 144 4.3 4.4 3.3
AP Lumbar Spine 75 80 81 91 91 78 4.3 4.4 3.3
LAT Lumbar Spine 85 90 90 86 86 73 4.3 4.4 3.3
LAT Lumbo-Sacral Joint 85 90 90 86 86 73 4.3 4.4 3.3
AP Pelvis 75 85 81 90 90 77 4.3 4.4 3.3
AP Skull 75 72 73 92 92 79 4.3 4.4 3.3
LAT Skull 70 72 73 98 98 85 4.3 4.4 3.3
AP Urinary tract 80 86 N/A 90 90 N/A 4.0 4.0 N/A
FSD, focus-to-skin distance; AP, anteroposterior; PA, posteroanterior; LAT, lateral.

post-processing tools, because higher tube potential can systems of approximately 50% compared with a conven-
reduce contrast [11]. tional 200-speed SFR system [14], whilst others have
In this study, doses to patients were assessed by reported the same values of ESD and E for the LAT
calculating ESD for six standard radiographic examina- Lumbar Spine radiograph when a CR system replaced a
tions (a total of 10 projections), and E for complete 300-speed SFR system [15]. Other authors have reported
examinations (Tables 3 and 4). a dose increase in CR systems of 33–58% compared with
5 months after the installation of the DDR equipment, a 400-speed SFR system [16]. The findings of this study
a survey was undertaken on 175 A&E patients to are therefore neither completely unexpected nor in
establish the preferred imaging modality. contradiction with those of other trials. It would seem
Approximately 30% of examinations were necessarily reasonable to state that doses with the use of CR systems
performed using the CR imaging modality, e.g. patients are approximately the same as those for a 300-speed SFR
being X-rayed using mobile radiographic equipment. Of system, but it has been also reported that the speed class
the remaining patients, approximately 86% were X-rayed quoted by manufacturers does not predict the actual
using the DDR equipment. relative speeds of the SFR systems [17] and there are
wide differences in image quality for similar speed
systems [18]: therefore the importance of optimization,
Discussion whichever system is being used, must be stressed. The
CR system used in this study has been optimized
Dose comparisons between CR, SFR and DDR have together with the manufacturer engineers: nevertheless,
been reported for chest radiography [12, 13]; the results a full optimization may take a long time because it is
of this study confirm the findings of the two reported a dynamic process where radiologists, physicists,
studies, i.e. that CR generally results in higher ESDs than
those in SFR and DDR. Comparison between DDR and
SFR doses shows that, in general, DDR results in lower
ESDs than those in SFR; the AP Skull projection is the
only one where the ESD for DDR is higher than that for
SFR. A possible reason for this is that AP Skull is not
performed by high tube potential technique with DDR.
Table 2 shows that AP Skull and Abdomen are the only
examinations where both tube potential and FSD for
DDR are lower than tube potential and FSD for SFR,
respectively: the effect of these two factors is an overall
decrease in the mAs value for DDR equipment compared
with SFR system. This reduction in mAs is greater for
Abdomen than for AP Skull: as a consequence, the AP
Skull is the only projection where ESD for DDR is higher
than ESD for SFR. However, E for complete Skull
examination for DDR is still lower than that for SFR.
Table 4 shows that for complete examinations, the
values of E for DDR are approximately 29% and 43%
lower than those for SFR and CR, respectively. E has
been calculated from ESD using the NRPB conversion Figure 1. Posteroanterior (PA) chest radiograph obtained
coefficients [8]; the coefficients increase with applied with computed radiography: male, 45 years old, 125 kVp,
tube potential and to a lesser extent with tube filtration. focus to skin distance (FSD) 5151 cm, Exposure Index51970
The magnitude of any dose saving that can be (the Exposure Index recommended by the manufacturer is
achieved with different imaging modalities is controver- 2000). The image satisfies the quality criteria for PA chest
sial [13]: some authors found a dose reduction with CR radiographs given in the European Guidelines [7].

The British Journal of Radiology, November 2006 901

 G Compagnone, M Casadio Baleni, L Pagan et al

Table 3. Entrance surface dose (mGy) for standard radiographic examinations using conventional screen–film radiography
(SFR), computed radiography (CR) and direct digital radiography (DDR), and percentage differences between the systems
Examination and projection SFR CR DDR CR vs SFR DDR vs SFR DDR vs CR

AP Abdomen 1.88 2.40 1.64 +28% 213% 232%

PA Chest 0.07 0.11 0.06 +57% 214% 245%
LAT Chest 0.15 0.20 0.13 +33% 213% 235%
AP Lumbar Spine 1.77 2.54 1.16 +44% 234% 254%
LAT Lumbar Spine 4.27 5.39 1.72 +26% 260% 268%
LAT Lumbo-Sacral Joint 5.12 5.39 1.72 +5% 266% 268%
AP Pelvis 1.81 1.83 1.02 +1% 244% 244%
AP Skull 1.44 1.61 1.58 +12% +10% 22%
LAT Skull 1.10 1.11 0.89 +1% 219% 220%
AP Urinary tract 2.18 2.51 N/A +15% N/A N/A
FSD, focus-to-skin distance; AP, anteroposterior; PA, posteroanterior; LAT, lateral.

Table 4. Effective dose (mSv) for standard radiographic examinations using conventional screen–film radiography (SFR),
computed radiography (CR) and direct digital radiography (DDR), and percentage differences between the systems
Examination SFR CR DDR CR vs SFR DDR vs SFR DDR vs CR

Abdomen (AP) 0.280 0.358 0.223 +28% 220% 238%

Chest (PA & LAT) 0.029 0.041 0.023 +41% 221% 244%
Lumbar Spine (AP & LAT) 0.309 0.476 0.179 +54% 242% 262%
Pelvis (AP) 0.295 0.326 0.168 +11% 243% 248%
Skull (AP & LAT) 0.027 0.029 0.022 +7% 219% 224%
Urinary tract (AP) 0.168 0.193 N/A +15% N/A N/A
FSD, focus-to-skin distance; AP, anteroposterior; PA, posteroanterior; LAT, lateral.

(a) (b)

Figure 2. Posteroanterior (PA) chest radiographs obtained with computed radiography and direct digital radiography. (a) Direct
digital radiography: female, 47 years old, 125 kVp, focus to skin distance (FSD) 5155 cm. (b) Computed radiography: male,
28 years old, 125 kVp, FSD5151 cm. Both images satisfy the quality criteria for PA chest radiographs given in the European
Guidelines [7].

902 The British Journal of Radiology, November 2006

Comparison of SFR, CR and DDR radiation doses in standard examinations
radiographers and manufacturers are involved day after
day in the achievement of continuous improvement.
On the whole, CR increases doses for a single
exposure, but the larger dynamic range reduces the
number of examinations that have to be repeated [19, 20],
e.g. for stretcher patients or for patients in intensive care
where the FSD often cannot be standardized.
However, the increase in CR patient doses found in
this study should be put into perspective, e.g. chest
examinations. In this example, the additional E per
complete examination (1 PA projection + 1 Lateral
projection) using CR was approximately 0.01 mSv and
0.02 mSv compared with SFR and DDR, respectively.
The risk of developing a fatal or other cancer, or other
serious defect (including hereditary effects), during the
course of life for the population undergoing chest
examinations in the A&E Department (about 20 000
examinations/year) is the same as for the whole
population, i.e. 7.3% per Sievert [3]. Thus, nearly 1 million
patients would each have to undergo complete chest
examinations with the use of a CR system in the A&E
Department to produce 1 and 2 additional health defects
in this population compared with SFR and DDR systems,
respectively. Moreover, the probability of a fatal cancer
being induced in an individual patient is dependent on
the age of the patient: e.g. for patients exposed after the
age of 70 years, the risk can be assumed to be reduced to
one-third of the average risk as the patient is likely to die
before the cancer develops [21]. Similar calculations
could be made for all other examinations.
Potential advantages of digital systems over conven-
tional radiography are well known [2], i.e. that they have
a greater dynamic range, wider exposure latitude, post-
processing facilities available, and that there is improved
access to images by clinicians and decreased film costs.
The post-processing tool is a very important and useful
characteristic provided that the individual performing
this task is adequately trained [22]. The DDR equipment
has been shown to have the following advantages:
(a) the radiographer can stay close to the patient
throughout the examination;
(b) the equipment is self-positioning and self-centring
around the patient, and this characteristic is very
useful for stretcher patients;
(c) the image acquisition times are much shorter than in
SFR or CR;
(d) an increased patient throughput (a further survey
showed that out of a total of 175 patients, 60% of
examinations are now performed using DDR);
(e) the rapid availability of images for timely clinical
decision-making, which may be vitally important for
severely injured patients.
When the workload is heavy, the patient throughput
can be maximized by using the DDR equipment for
chest, abdomen, pelvis, spine and skull examinations,
and by using CR for extremities.
The image quality characteristics of flat-panel technol-
ogy are theoretically better than those of SFR and CR
because at a comparable resolution the DDR has a higher
detective quantum efficiency (DQE) [23]; this advantage
can be utilized in reducing the patient dose whilst
maintaining image quality and without changing the
The British Journal of Radiology, November 2006
signal-to-noise ratio. The appropriateness of this
approach would, however, require verification. All
images met the criteria in the European Guidelines for
all modalities and were used for reporting by the
Consultant Radiologists. Figure 2 shows chest images
obtained using CR and DDR. It shows that both systems
satisfied the quality criteria for diagnostic radiographic
images [7]: full inspiration; symmetrical reproduction of
the thorax; scapulae outside the lung field; reproduction
of the whole rib cage above the diaphragm; visualization
of the retrocardiac lung and the mediastinum; sharp
reproduction of the trachea and proximal bronchi, the
borders of the heart and aorta; visualization of the spine
through the heart shadow. Finally, the radiologists prefer
the appearance of the DDR images.
Conclusions
This study compared radiation doses to patients
undergoing standard radiographic examinations using
SFR, CR and DDR; ESD and E were calculated using
standard exposure parameters. Effective doses for DDR
were found to be ,29% and ,43% lower than those for
SFR and CR, respectively. The image quality met the
criteria in the European Guidelines for all modalities.
Even though in the A&E department most examinations
are now performed with the DDR system, we are aware
that the optimization effort has to also continue for the
CR equipment, because digital systems need to be
optimized very carefully by a dynamic process that
may take a long time. Conventional technologies are
soon to be replaced by digital ones and although DDR
equipment has been shown to have many advantages,
further work is required to specifically look at image
quality and dose for real patients.
Acknowledgments
We express our thanks to the Reviewers for their
valuable advice regarding this study.
References
1. Gatt ME, Spectre G, Paltiel O, Hiller N, Stalnikowicz R.
Chest radiographs in the emergency department: is the
radiologist really necessary? Postgrad Med J 2003;79:214–7.
2. Volk M, Strotzer M, Gmeinwieser J, Alexander J, Frund R,
Seitz J, et al. Flat-panel X-ray detector using amorphous
silicon technology. Invest Radiol 1997;32:373–7.
3. International Commission on Radiological Protection. 1990
Recommendations of the ICRP. Annals of the ICRP Vol. 21
Publication 60. Oxford, UK: Pergamon Press, 1991.
4. McCollough CH, Schueler BA. Calculation of effective dose.
Med Phys 2000;27:828–37.
5. Compagnone G, Pagan L, Bergamini C. Local diagnostic
reference levels in standard X-ray examinations. Radiat Prot
Dosim 2005;113:54–63.
6. Harpen MD. A mathematical spread sheet application for
production of entrance skin exposure nomograms. Med
Phys 1996;23:241–2.
7. European Commission. European guidelines on quality
criteria for diagnostic radiographic images. EUR 16260EN.
Office for Official Publ. of the Europ Commun. 1996.
903

8. Hart D, Jones DG, Wall BF. Estimation of effective dose in
diagnostic radiology from entrance surface dose and dose-
area product measurements. NRPB-R262. Chilton, Didcot:
National Radiological Protection Board, 1994.
9. Peters SE, Brennan PC. Digital radiography: are the
manufacturers’ settings too high? Optimisation of the
Kodak digital radiography system with aid of the com-
puted radiography dose index. Eur Radiol 2002;12:2381–7.
10. Bogucki TM, Trauernicht DP, Kocher TE. Characteristics of
a storage phosphor system for medical imaging. Technical
and Scientific Monograph 6. Rochester, USA: Kodak Health
Sciences, 1995.
11. Livingstone RS, Koshy CG, Ray DV. Evaluation of work
practices and radiation dose during adult micturating
cystourethrography examinations performed using a digital
imaging system. Br J Radiol 2004;77:927–30.
12. Weatherburn GC, Bryan S, Davies JG. Comparison of doses
for bedside examinations of the chest with conventional
screen-film and computed radiography: results of a
randomized controlled trial. Radiology 2000;217:707–12.
13. Herrmann KA, Bonél H, Stabler A, Kulinna C, Glaser C,
Holzknecht N, et al. Chest imaging with flat-panel detector at
low and standard doses: comparison with storage phosphor
technology in normal patients. Eur Radiol 2002;2:385–90.
14. Sandmayr H, Wallentin D. Computer integrated radiology
system: analogue goes digital. Eur Radiol 1997;7(Suppl.
3):S90–3.
15. Weatherburn GC, Bryan S. The effect of a picture archiving
and communication systems (PACS) on patient radiation
doses for examination of the lateral lumbar spine. Br J
Radiol 1999;72:534–45.
904
G Compagnone, M Casadio Baleni, L Pagan et al
16. Bragg DG, Murray KA, Tripp D. Experiences with
computed radiography: can we afford the cost? AJR Am J
Roentgenol 1997;169:935–41.
17. Brennan PC, Devereux SA. An assessment of the usefulness
of screen-film speed classifications. Eur Radiol 2002;12:
1577–83.
18. Guibelalde E, Fernandez JM, Vano E, Llorca A, Ruiz MJ.
Image quality and patient dose for different screen-film
combinations. Br J Radiol 1994;67:166–73.
19. Strotzer M, Gmeinwieser J, Spahn M, Volk M, Frund R,
Seitz J, et al. Amorphous silicon, flat-panel, X-ray detector
versus screen-film radiography. Invest Radiol 1998;33:
33–8.
20. Weatherburn GC, Bryan S, West M. A comparison of image
reject rates when using film, hard copy computed radio-
graphy and soft copy images on picture archiving and
communication systems (PACS) workstations. Br J Radiol
1999;72:653–60.
21. Cox R, McGibbon BH. Diagnostic medical exposures:
advice on exposure to ionising radiation during pregnancy.
Estimates of late radiation risks to the UK population.
Documents of the National Radiological Protection Board,
Vol.4,No.4, 1993.
22. Tsapaki V, Kottou S, Kollaros N, Dafnomili P, Koutelou M,
Vano E, et al. Comparison of a conventional and a flat-panel
digital system in interventional cardiology procedures. Br J
Radiol 2004;77:562–7.
23. Samei E, Hill JG, Frey GD, Southgate WM, Mah E, Delong
D. Evaluation of a flat panel digital radiographic system for
low-dose portable imaging of neonates. Med Phys
2003;30:601–7.
The British Journal of Radiology, November 2006
The British Journal of Radiology, 79 (2006), 905–911

Dose escalation to combat hypoxia in prostate cancer: a

radiobiological study on clinical data
1 2
J Z WANG, PhD, X A LI, PhD and 1N A MAYR, MD

1
Department of Radiation Medicine, The Ohio State University, Columbus, OH 43210 and
2
Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee, WI 53226, USA

ABSTRACT. Earlier studies have demonstrated that hypoxic regions exist in human
prostate cancer and the degree of hypoxia correlates with the treatment outcome of
radiotherapy. Using the concept of the clinical oxygen enhancement ratio (COER), the
linear-quadratic (LQ) model was extended to account for the effect of tumour hypoxia.
The clinical data collected at the Fox Chase Cancer Center for prostate cancer were
analysed based on the LQ model as well as the tumour control probability (TCP) model.
The LQ and TCP parameters (a50.15 Gy 21, a/b53.1 Gy and the number of clonogens
K5106,107 cells) determined in earlier studies were used to derive the COER for
prostate cancer: COER51.4 with a standard confidence interval (CI) of (1.2, 1.8). The
result is consistent with the in vitro OER measurements of human tumour cell lines
under chronic hypoxia conditions. This implies that a higher dose is needed to
overcome tumour hypoxia. For prostate tumours, the prescription dose required to
overcome tumour hypoxia is 165 Gy (CI: 153,186 Gy) for permanent 125I implants and Received 4 January 2006
88 Gy (CI: 74,118 Gy) in 2 Gy fractions for external-beam radiotherapy. The impact of Revised 2 June 2006
LQ parameters on the calculations of COER and dose escalation was discussed. This Accepted 5 June 2006
study provides a preliminary estimate of the dose escalation needed to overcome
DOI: 10.1259/bjr/18700614
tumour hypoxia based on clinical data. More clinical data with better statistics and
longer follow-up time are required to further tune the radiobiological modelling of ’ 2006 The British Institute of
hypoxia for prostate cancer. Radiology

Hypoxia has been identified in many human cancers,
including head and neck, lung, breast, cervix, bladder
and prostate cancer, and the adverse impact of hypoxia
on clinical tumour control has been well recognized
[1–10].
Recently, Movsas et al [1–3] and Parker et al [4] have
demonstrated that hypoxic regions do exist in human
prostate carcinoma and the treatment outcome of
radiation therapy correlates with the degree of hypoxia.
Oxygenation status (oxygen pressure, PO2) may thus be
another underlying biological parameter beyond the
classic prognostic factors (clinical stage, Gleason score
and prostate specific antigen (PSA)) predicting a critical
component of treatment failure in prostate cancer.
Movsas et al [1] reported a significant difference in the
biochemical control rate at 2 years (31% versus 92%)
between two patient groups stratified by the prostate/
muscle (P/M) PO2 ratio (,0.05 versus >0.05). Many
other investigators have also found similar correlations
between hypoxia and tumour progression [5–10]. It is
generally believed that high-risk tumours tend to exhibit
more severe hypoxia [3, 10]. Although it is well
established that hypoxic cells/regions are less sensitive
to radiation and that tumour hypoxia can result in
clinical radiation therapy failure, it remains unclear how
Address correspondence to: Jian Z Wang, Department of Radiation
Medicine, James Cancer Hospital and Solove Research Institute, The
Ohio State University, 300 W 10th Ave, Rm 094, Columbus, OH
43210, USA. E-mail: wang.993@osu.edu.
to overcome hypoxia in the clinical setting. Specific
quantitative dose guidelines are needed that can be
applied to the clinical setting, particularly to dose
escalation trials aimed to combat hypoxia in prostate
cancer.
The linear-quadratic (LQ) and tumour control prob-
ability (TCP) models extended to address tumour
hypoxia were used to analyse the clinical data.
Recently, the a/b ratio of the LQ model for prostate
cancer has become a highly debated topic in the
radiation therapy community [11–22]. By taking into
account the effect of tumour repopulation, Wang et al
[14, 15] and Kal and Van Gellekom [19] took two
different approaches, and obtained similar results of
a/b ratio (around 3.1 Gy). Based on the clinical data of
external-beam radiotherapy (EBRT) and low- and high-
dose rate (LDR/HDR) brachytherapy [12, 13, 23], Wang
et al [14, 15] have derived a new set of LQ parameters
(a50.15 Gy21, a/b53.1 Gy and repair half-time
Tr516 min) with estimated numbers of clonogens
around 106,107, depending on the patient risk level.
The new results provide reasonable estimates of radio-
sensitivity and the number of clonogens for human
prostate tumours [14, 15, 24–26]. The radiobiological
modelling with these parameters provides a consistent
interpretation for most currently available clinical data,
including EBRT, permanent brachytherapy, HDR bra-
chytherapy and their combinations [14, 15, 26].
Fowler et al [16, 21] argued that for slowly growing
tumours (e.g. prostate cancer), the onset of tumour

The British Journal of Radiology, November 2006 905


repopulation may be delayed significantly to over
200 days. Wang et al [17, 22] pointed out that there is
no strong biological evidence for such arguments. The
clinical results reported so far have been controversial. In
a recently reported study, Perez et al [27] did observe an
impact of elapsed treatment time on outcome of EBRT
for T2 prostate tumours, implying the onset time of
repopulation should be relatively short. Therefore in our
study, the original assumption of the onset time of
tumour repopulation will be followed for various sets of
LQ parameters. Recently, the effect of hypoxia has been
taken into account by Nahum et al [18] in their modelling
study. However, their model parameters (a, b and
oxygen enhancement ratio (OER)) were completely
based on in vitro measurements, which may not be fully
applicable to the clinical micro-environment of prostate
cancer in vivo [28]. Tumour heterogeneity of radio-
sensitivity has been studied [20] and due to the large
uncertainty existing in the current clinical data and the
large number of parameters used, it is not possible to
give a reasonable estimate of a/b ratio for any statistical
significance using the heterogeneous model. For this
reason, we chose the homogeneous model in this study.
The purpose of our study is to quantify the dose
escalation to address tumour hypoxia in prostate cancer.
The LQ and TCP models are used to interpret the clinical
data recently reported from Fox Chase [1] based on the
concept of OER. The specific aim of this analysis is to
derive the clinical OER (COER) and to determine the
dose escalation necessary to offset the hypoxia effect
using either permanent brachytherapy or EBRT.
Materials and methods
LQ and TCP models for prostate cancer
The general LQ model [29] was used in this study. In
this model, the surviving fraction S of cells irradiated to a
total dose D within an overall treatment time T is given
by:
S~ exp ({aD{bGD2 zcT) ð1Þ
where a and b characterize intrinsic radiosensitivity, G is
the dose-rate factor accounting for sub-lethal damage
repair, c is the effective tumour repopulation rate
[c5ln(2)/Td, and Td is the effective clonogen doubling
time]. A median potential tumour doubling time of
42 days measured in vivo from 7 prostate cancer patients
was used [30]. This repopulation rate corresponds to a
biological effective dose (BED) of 0.11 Gy per day (for
a50.15 Gy 21), compatible with the results of slow-
growing tumours presented in a repopulation-dose-
equivalent study by Jones et al [31]. The repopulation
time T can be replaced by (T–Tk), where Tk is the onset
time of tumour repopulation. Because the actual value of
this parameter is under debate [16, 17, 21, 22, 27] and no
clinical data are available, we followed the original
assumptions used by different investigators when they
derived the LQ parameters from clinical data, i.e. Tk50
for the LQ parameters by Wang et al [14, 15] and by Kal
and Van Gellekom [19], and Tk.200 days (i.e. ignoring
906
J Z Wang, X A Li and N A Mayr
the repopulation effect) for the LQ parameters by
Brenner et al [11, 12] and by Fowler et al [13].
For EBRT, the total dose D equals nd and the dose-rate
factor G equals 1/n, where n is the number of dose
fractions and d is the dose per fraction. The treatment
time T of EBRT can be approximated as the number of
treatment fractions multiplied by 1.4 (5 fractions per
week). For permanent brachytherapy, D, G and T have
been derived in many papers (e.g. [14, 29]) as follows:
R0

D~ 1{e{lT ,
l
 
2R2 1
1

G~ 2 0 1{e{2lT { 1{e{(mzl)T ; ð2Þ
D (m{l) 2l mzl
 
1 c
T&{ ln
l aR0
where m is the repair rate of tumour cells [m5ln(2)/Tr], R0
is the initial dose rate (R05D0?l, D0 is the prescribed
dose) and l is the decay constant for the implanted
isotopes [l5ln(2)/Ts, Ts is the half-life time of the
radioisotope]. Ts559.4 days for 125I implants.
The Poisson TCP model is often used to link cell killing
of a radiotherapy scheme to treatment outcome [32],
TCP~ exp ({KS) ð3Þ
where K is the initial number of tumour clonogens. We
consider it as an effective number of clonogens asso-
ciated with the patient risk level.
Two representative sets of LQ parameters derived
from clinical data [11–15, 19] and one set of generic LQ
parameters for tumour [10] were used in this study
(Table 1). The numbers shown in the parentheses of
Table 1 indicate the standard confidence intervals (CI)
[14] and they were used to estimate the uncertainties of
the COER and dose escalations. The clonogen number K
of 1.66106 (CI: 5.66104,8.86107) derived from clinical
data [14] was used in the calculations with the first
parameter set (see the following subsection ‘‘Fox Chase
clinical data’’). Due to limited data for the other two
parameter sets, the clonogen numbers were arbitrarily
set to match the clinical data and the CIs were not
considered. Our calculation was mainly based on the
first parameter set. The later two parameter sets were
used to test the model dependency of the final results
presented in this study.
LQ model extended for tumour hypoxia
OER is a dose-modifying parameter, which is defined
as the ratio of doses needed to produce the same
biological effect for cells without and with oxygen.
Traditionally, OER was commonly used to describe the
cell survival curves obtained in in vitro measurements.
The OER of X-ray irradiation for most mammalian cells
in vitro is about 3 at high doses and is possibly lower
(about 2) at doses below approximately 2 Gy [10]. In this
work, we apply the OER concept to clinical data (i.e.
in vivo) to characterize tumours with or without hypoxia
The British Journal of Radiology, November 2006
Hypoxia in prostate cancer
Table 1. Radiobiological parameters used in this study
21
LQ/TCP parameters a (Gy ) a/b (Gy)
a,b
Wang et al 0.15 (0.11–0.19)
Brenner et alb 0.04
Generic parametersb 0.3
a
The numbers in the parentheses show the standard confidence intervals (CIs).
b
See text for detailed references.
c
The repopulation effect is ignored. Td5‘ is equivalent to Tk.200 days.
LQ, linear-quadratic; TCP, tumour control probability.
[5, 33–35]. Compared with in vitro experiments, the
hypoxia in in vivo situations may be quite different. The
population for tumours in vivo always reflects a
spectrum of hypoxia. Even for well-identified hypoxic
tumours, the hypoxic cells may exist only in sub-
volumes of tumours. Therefore, the OER obtained
retrospectively from clinical data represents only a
statistical average and may be different from the
corresponding in vitro values for the same type of
tumours. Zolzer and Streffer [36] argued that tumours
in vivo should be dominated by chronic hypoxia and the
radiosensitivity for cells under chronic hypoxia condi-
tions would be higher than those under acute hypoxia
conditions. They reported that the OER measurements of
two human tumour cell lines under continued/chronic
hypoxia conditions (1.3 and 1.5 for MeWo and squamous
carcinoma 4451, respectively) were quite different from
the OER values measured under acute hypoxia condi-
tions. To distinguish this feature, we term herewith the
OER obtained from clinical data as clinical OER (COER).
For simplicity, the COER is represented by H throughout
this paper.
Given a certain amount of clonogens, to achieve the
same TCP for two patient groups, the cell killing
efficiency should be the same. That is:
S~ exp ({ah Dh {bh GD2h zcT)
~ exp ({ao Do {bo GD2o zcT)
where subscripts h and o label the parameters for
tumours with and without hypoxia, respectively.
Assuming Dh and Do represent the doses to achieve
the same TCP for the two patient groups, respectively,
we have Dh5H?Do. To be consistent with Equation (4),
we assume that ah5ao/H, bh5bo/H2, and they are valid to
apply to the LDR 125I brachytherapy and the EBRT with
low dose fractions (d52 Gy). It appears that the hypoxia
impacts more significantly on the quadratic term than on
the linear term. Such relationships of the LQ parameters
between aerated and hypoxic cells have been observed in
in vitro experiments of low dose irradiation (0.5–3 Gy)
for Chinese hamster V79-171 cells [37, 38]. Therefore, the
LQ model extended to the hypoxia effect can be
expressed as (Sh for hypoxic cells):
 paper [26] were used to automate these calculations.
Sh ~ exp ({aD=H{bGD2 H 2 zcT)
Similar formulae to describe the hypoxia effect can be
found in the literature [5, 33–35].
Because of the limited clinical data, we will focus on
the effect of hypoxia on radiosensitivity only, and neglect
The British Journal of Radiology, November 2006
Tr (min) Td (days)
3.1 (2.6–3.6) 16 (1–90) 42 (22–62)
1.5 114 ‘c
10 60 42
the influence of hypoxia on repair time or repopulation
rate. Hill et al presented data of animal and human
carcinoma of the cervix showing no evidence for a
difference in repopulation kinetics between aerated or
hypoxic tumours [39, 40]. It is generally assumed that the
transient hypoxia can be overcome by tumour reoxy-
genation by protracting radiotherapy (i.e. fractionated
EBRT or LDR brachytherapy). The reoxygenation effect
has been modelled in previous studies (e.g. [41]).
Therefore, it is believed that the oxygen effect on the
clinical outcome observed in the 125I data (see next sub-
section) should be mainly due to the chronic hypoxia
[36]. In this work, we use the above formulae to study the
effect of chronic hypoxia.
For patients with tumours that are only partially
hypoxic, i.e. containing oxygenated and hypoxic sub-
volumes, the two-compartment model was used to
calculate the overall surviving fraction [18, 33, 35],
S~(1{fh )So zfh Sh ð6Þ
where fh is the hypoxic fraction of cells in the tumour and
So is for aerated cells. Animal tumour data and human
tumour data indicate that the hypoxic fraction ranges
from 0 to 50%, with an average value of 15% [5, 10, 42].
This average value along with the range from 5% to 50%
was used in this study.
ð4Þ The uncertainty in estimating the COER originates
from the uncertainties of the LQ/TCP parameters (as
shown in Table 1), the hypoxia fraction and the clinical
data. The LQ/TCP parameters are strongly correlated to
each other and constrained by clinical data; therefore,
their uncertainties are not independent from each other
[14] and have been considered in this study. The code
autoEUD described in Wang and Li [26] was used to
propagate the different uncertainties to determine the
CIs of COER and dose escalation. For independent
parameters, their contributions to the final CIs were
calculated based on the root of the quadratic-summation.
For a given patient group of prostate cancer, based on
Equations (1) to (6), we could calculate the TCP for a
given radiation treatment or determine the required
radiation dose D to achieve a given TCP. The software
AutoEUD and the typical dose–volume histogram (DVH)
of EBRT and brachytherapy described in a previous
ð5Þ
Fox Chase clinical data
Since 1999, Movsas et al have published a series of
papers to address the hypoxia of prostate cancer [1–3].
907

For the first time, they demonstrated the existence of
hypoxia in human prostate cancer based on the in vivo
electrode measurements of oxygen levels, and that the
degree of hypoxia in prostate cancer correlated to the
treatment outcome of radiotherapy. 57 patients with
localized prostate cancer were included in their study [1].
Before radiotherapy treatments, custom-made
Eppendorf microelectrodes were used to measure the
PO2 in both pathologically involved regions of the
prostate and normal muscle (as an internal control).
Real-time ultrasound imaging was used to guide the
microelectrode measurement. For each patient, approxi-
mately 100 PO2 readings were obtained along 3 to 5
tracks in both prostate and muscle [3]. Following the
Eppendorf PO2 measurement, 48 patients received
permanent 125I brachytherapy. The dose prescribed to
the prostate was 145 Gy. Nine patients were treated on a
dose escalation protocol involving a 46 Gy (in 23 daily
fractions) of EBRT to the pelvic region plus two HDR
boost implants (8.75 Gy or 9.75 Gy per fraction).
Hormonal therapy prior to radiation therapy was given
to nine patients. In their study, the biological failure was
defined as two consecutive rises in PSA without a return
to baseline. With a median follow-up time of 19 months
(ranging from 4 months to 31 months), the overall 2 year
bNED (biologically no evidence of disease) was esti-
mated as 81% for the entire patient group (as shown in
Figure 1 of this paper and Figure 2 of [1]). In order to
study the impact of hypoxia, the P/M ratio of PO2, which
eliminates the potential interpatient and technical varia-
tions, was used to analyse the treatment outcome.
The Fox Chase experience demonstrated that the P/M
ratio strongly correlated to the bNED outcome. A
threshold analysis showed that the 2 year bNED differed
significantly at a P/M ratio of ,0.05 (hypoxic group)
versus >0.05 (aerated group): 31% versus 92%
(p,0.0001) (see Figure 1). There were 12 and 45 patients
in the two groups, respectively. The classic prognosis
factors, including clinical stage, Gleason score, pre-
treatment PSA and age at treatment were similar in the
two patient groups. Based on the numbers of all patients
and the patients at risk for each group, using the bi-
nominal distribution, we estimated the standard CI of
the 2 year bNED for the hypoxic tumour group to be
(15%, 47%). Because of the large numbers of patients in
the aerated tumour group, the uncertainty of their 2 year
bNED is negligible.
According to the patient characteristics of the three
prognosis factors, patients of the Fox Chase data were
stratified into the low- or intermediate-risk groups.
Because most patients presented low-risk prognosis
factors, the clonogen number (1.66106) of low-risk
patients was used for this study [14, 15]. Two radio-
therapy modalities (125I brachytherapy vs combined
EBRT and HDR) were used in the Fox Chase Study.
Since the patient number of the combined modality
(EBRT + HDR) is relatively small (9 patients) and the
treatment outcome for low-risk patients is close for the
two regimens, we ignored the difference of the two
modalities and combined the data in the following study.
For similar reasons, we also ignored the influence of
hormone therapy applied to nine patients out of the total
57 patients. A representative DVH obtained from
permanent 125I implant (Figure 1 of Ref. [26]) was used
908
J Z Wang, X A Li and N A Mayr
Figure 1. 2 year biologically no evidence of disease (bNED)
versus prescription dose of 125I brachytherapy. The symbols
represent the Fox Chase clinical data [1] and the curves
represent the model calculations. The three points (circle,
square and triangle), as well as the three curves (solid,
dashed and dash-dotted), are for patient groups with
hypoxic, mixed and aerated prostate tumours, respectively.
The prescription dose of the clinical data was 145 Gy of 125I
implant. The vertical dotted line indicates the prescription
dose for the hypoxic group to achieve a bNED of 81%. The
vertical error bar shows the standard confidence interval (CI)
of the clinical data for the hypoxic group (the CIs for the
other two groups are not shown) and the horizontal error
bar shows the CI of dose escalation for the hypoxic group to
achieve a target bNED of 81%.
to account for the dose inhomogeneity of the Fox Chase
data.
Results
Clinical OER for hypoxic prostate cancer
The clinical data and the model calculation are
summarized in Figure 1. It has been reported that, based
on LQ and TCP models, the 145 Gy 125I implant has a
TCP of 82% for low-risk prostate patients [14, 26]. This
calculated TCP value is consistent with the clinically
observed 2 year bNED (81%) for the mixed patient group
in the Fox Chase study (see ‘‘Fox Chase clinical data’’ in
the Materials and Methods section). Using the bNED
data for the aerated and hypoxic groups, we derived the
COER and the CI for the hypoxic tumours: H51.4 (CI:
1.2,1.8) (Table 2). This result is very consistent with the
in vitro OER measurements of two human tumour cell
lines (1.3 and 1.5 for MeWo and squamous carcinoma
4451, respectively) under chronic/continued hypoxia
conditions [36]. The large uncertainty is mainly related
to the limited information for the hypoxic fraction in the
hypoxic tumours. If the uncertainty of hypoxic fraction is
not considered, the uncertainty of COER is greatly
reduced to¡0.13, i.e., in the range of (1.27, 1.53).
Similar calculations, except for the uncertainty estima-
tion, were performed for other sets of LQ parameters. In
the COER derivation with the parameters of Brenner et al
(i.e. a50.04 Gy 21, a/b51.5 Gy, Tr51.9 h with repopula-
tion ignored), only the hypoxic clonogens were consid-
ered [43]. The COER obtained with this parameter set is
The British Journal of Radiology, November 2006
Hypoxia in prostate cancer
H51.5. Using the LQ parameters that are considered
generic for tumours (i.e. a50.3 Gy 21, a/b510 Gy, Tr51 h
and Td542 days), we obtained the COER: H51.2.
Interestingly, the result derived with the parameters of
Wang et al lies between the values obtained with these
two sets of LQ parameters, and the differences among
the COER values obtained with the three parameter sets
are smaller than the CIs determined by various uncer-
tainties. More clinical data with better statistics and
better knowledge of the hypoxic fraction are required to
further narrow down the CIs. To summarize the results,
the COERs, as well as the required dose escalations (see
the following subsection), derived with different LQ/
TCP parameters are listed in Table 2.
Dose escalation for hypoxic tumours
Based on the derived COER, the dose required to
compensate for tumour hypoxia was calculated using the
code autoEUD [26]. To bring the TCP for hypoxic
tumours to the same level as that for the general
population of prostate patients (TCP581%), the pre-
scription dose required for permanent 125I implant
should be 169 Gy. This dose may be used to target the
hypoxia region if it can be identified. Otherwise, a
reduced dose of 165 Gy (CI: 153,186 Gy) should be
prescribed to the entire prostate, including the non-
hypoxic regions, to achieve the same treatment outcome
(a TCP of 81%) as shown in Figure 1. This dose escalation
is practically feasible, as demonstrated by a recent study
[44].
The calculations performed so far provide a guide of
dose escalation for a cohort of patients with hypoxic
prostate tumours. If the hypoxic subvolume of indivi-
dual tumours can be detected, for example by biological/
functional imaging, the data presented may also be used
to target these subvolumes and design the individua-
lized radiation plan. Figure 2 shows the TCP and dose
prescription as a function of the hypoxic fraction fh for
individual prostate tumours. The TCP presented in
Figure 2a is for the dose prescription (145 Gy of 125I
implant) used in the current clinical practice. Figure 2b
shows the dose escalation required to overcome the
effect of hypoxia and to bring the TCP to 81% as that of
the general population of prostate cancer patients. For
individual patients with hypoxic fraction changing from
5% to 50%, the prescription dose is increased from
157 Gy to 181 Gy.
The British Journal of Radiology, November 2006
Figure 2. (a) Tumour control prob-
ability (TCP) as a function of the
fraction of tumour hypoxia (fh) for
individual patients of prostate can-
cer. The prescription dose is 145 Gy
for 125I brachytherapy. (b) Dose
escalation of 125I brachytherapy
required overcoming tumour
hypoxia (to achieve a TCP of 81%)
as a function of hypoxia fraction fh.
The above results obtained from the Fox Chase data
(brachytherapy) data set may be used to derive the dose
escalation of EBRT. Various clinical studies have shown
that a biological equivalence exists between EBRT and
permanent brachytherapy for prostate cancer [13, 26].
For patients with similar risk-levels, the 145 Gy bra-
chytherapy yielded a similar treatment outcome as 71 Gy
EBRT in 2 Gy fractions [13]. Based on the LQ and TCP
models and the COER derived in this study, we
calculated the EBRT dose required to compensate for
the effect of hypoxia. This EBRT dose is found to be
88 Gy (CI: 74,118 Gy) in 2 Gy daily fractions.
Similarly, based on the COERs derived with the
parameters of Brenner et al or with the generic LQ
parameters for tumours (see Table 1), we obtained the
dose escalation necessary to compensate for prostate
hypoxia. With Brenner et al’s parameters, the prescrip-
tion dose would be 193 Gy for 125I brachytherapy and
109 Gy for EBRT. With the generic tumour parameters,
these doses would be 155 Gy for the 125I brachytherapy
and 102 Gy for EBRT. Because the clinical data used to
derive the COERs are from 125I brachytherapy, the dose
escalations calculated for the 125I brachytherapy have
smaller CIs and they match the sequence of the a/b ratio
in the three parameter-sets. However, the extrapolated
EBRT doses show large uncertainty and the results based
on the parameters of Brenner et al and the generic
tumour parameters are unusually high; therefore, the
prediction of EBRT dose escalation has limited clinical
value.
Conclusion and discussion
In this study, we have analysed the reported clinical
data for hypoxic prostate cancer. Based on the LQ and
TCP models extended to account for tumour hypoxia,
the clinical OER for prostate tumour was obtained:
COER51.4 (CI: 1.2,1.8). The result agrees with the
in vitro OER measurements under chronic/continued
hypoxia conditions [36]. To overcome tumour hypoxia, a
dose escalation to 165 Gy (CI: 153,186 Gy) (instead of
145 Gy) is required for 125I permanent implants. A
similar calculation was performed for EBRT. The EBRT
dose needs to be escalated from 71 Gy to 88 Gy in 2 Gy
fractions in order to achieve a TCP of 81% for the hypoxic
tumour group.
There is in vivo evidence showing that increasing levels
of hypoxia in prostate cancer correlate significantly with
909

Table 2. Impact of radiobiological parameters on the calculation of COERs and dose escalations
LQ/TCP parameters COER
a,b
Wang et al 1.4 (1.2–1.8)
Brenner et alb 1.5
Generic parametersb 1.2
a
The numbers in the parentheses show the standard confidence intervals (CIs).
b
The corresponding LQ/TCP parameters used in this study are shown in Table 1.
LQ, linear-quadratic; TCP, tumour control probability; COER, clinical oxygen enhancement ratio; EBRT, external-beam
radiotherapy.
the increasing clinical stage and patient age [2]. The P/M
ratio of PO2 may be used as a prognostic factor for high-
risk patients. Dose escalation in both EBRT and
brachytherapy has been proven effective to treat high-
risk patients of prostate cancer [12, 23, 26, 44, 45]. This
study used similar approaches to address the unsatis-
factory treatment outcome of hypoxic tumours. The
required dose escalations (165 Gy for 125I implant and
88 Gy for EBRT) are practically deliverable [23, 44], and
clinical trials with similar dose levels or even higher are
being conducted in several institutions.
The results obtained in this work are based on the Fox
Chase clinical data [1] that, for the first time, shows the
existence of hypoxia in human prostate tumours and the
correlation of the degree of hypoxia with treatment
outcome. There are several limitations in the Fox Chase
study that may compromise the accuracy in the current
estimation of the COER and the corresponding target
doses. First, the PO2 measurement with Eppendorf
microeletrodes is a sampling method which measures
only a sampled portion of the prostate. It may not reflect
the overall PO2 distribution in the entire tumour. The
method also fails to discriminate cell type and viability,
resulting in PO2 readings from less significant tissues.
Second, this method cannot distinguish the hypoxia
types: chronic and transient. The transient hypoxia may
be reduced or overcome by providing enough time for
reoxygenation during the treatment course. Third, the
study contained only a small number of patient samples
with limited follow-up time, and furthermore mixed
treatment modalities (two radiation methods and hor-
mone therapy) were involved in the small data sample.
Because of the limited statistical power, only two patient
groups could be identified with an artificially deter-
mined threshold of the P/M ratio. More clinical data
with larger numbers of patients, longer follow-up time
and consistent treatment are certainly desired to further
refine our conclusions.
Because of the limited clinical data, it was not possible
to derive independent relationships between a, b and
COER in this study. Instead, the relationships, ah5ao/H,
bh5bo/H2, from the literature [5, 33–35] were used.
Although similar relationships have been observed in
in vitro measurements [37, 38], the results obtained in this
analysis would be affected if such relationships were not
applicable to prostate cancer. The COER calculation was
based on an average hypoxic fraction of 15% [5, 10, 42]. A
change of this value would change the results of the
COER as well as the dose escalation. There are also
concerns about the OER changing with the irradiation
dose fraction. In this study, we assume that the COER
keeps constant for the LDR brachytherapy and the
910
J Z Wang, X A Li and N A Mayr
125
I dose (Gy) EBRT dose (Gy)
165 (153–186) 88 (74–118)
193 109
155 102
low-dose-fraction (2 Gy) EBRT. The use of alternative
models, which include temporal variations in the
hypoxic fraction, might also lead to quite different
conclusions. Therefore, the results obtained in this study
may be limited by the approximations used in the
models and the uncertainties shown in the clinical data.
Caution needs to be exercised in using the presented
data for clinical decision-making purposes.
References
1. Movsas B, Chapman JD, Hanlon AL, et al. Hypoxic
prostate/muscle Po2 ratio predicts for biochemical failure
in patients with prostate cancer: preliminary findings.
Urology 2002;60:634–9.
2. Movsas B, Chapman JD, Greenberg RE, et al. Increasing
levels of hypoxia in prostate carcinoma correlate signifi-
cantly with increasing clinical stage and patient age. Cancer
2000;89:2018–24.
3. Movsas B, Chapman JD, Horwitz EM, et al. Hypoxic
regions exist in human prostate carcinoma. Urology
1999;53:11–8.
4. Parker CP, Milosevic M, Toi A, et al. Polarographic
electrode study of tumor oxygenation in clinically localized
prostate cancer. Int J Radiat Oncol Biol Phys 2004;58:750–7.
5. Okunieff P, Hoeckel M, Dunphy EP, et al. Oxygen tension
distributions are sufficient to explain the local response of
human breast tumors treated with radiation alone. Int J
Radiat Oncol Biol Phys 1993;26:631–6.
6. Höckel M, Knoop C, Schlenger K, et al. Intratumoral pO2
predicts survival in advanced cancer of the uterine cervix.
Radiother Oncol 1993;26:45–50.
7. Höckel M, Schlenger K, Mitze M, et al. Hypoxia and
radiation response in human tumors. Semin Radiat Oncol
1996;6:3–9.
8. Nordsmark M, Overgaard M, Overgaard J. Oxygenation
status predicts radiation response in advanced squamous
cell carcinoma of head and neck. Radiother Oncol
1996;41:31–9.
9. Brizel DM, Sibley GS, Prosnitz LR, et al. Tumor hypoxia
adversely affects the prognosis of carcinoma of the head
and neck. Int J Radiat Oncol Biol Phys 1997;38:285–9.
10. Hall EJ. Radiobiology for the radiologist. 5th edn.
Philadelphia, PA: Lippincott Williams and Wilkins,
2000:91–111.
11. Brenner DJ, Hall EJ. Fractionation and protraction for
radiotherapy of prostate carcinoma. Int J Radiat Oncol
Biol Phys 1999;43:1095–101.
12. Brenner DJ, Martinez AA, Edmundson GK, et al. Direct
evidence that prostate tumors show high sensitivity to
fractionation (low a/b ratio), similar to late-responding
normal tissue. Int J Radiat Oncol Biol Phys 2002;52:6–13.
13. Fowler J, Chappell R, Ritter M. Is a/b for prostate
cancer really low? Int J Radiat Oncol Biol Phys
2001;50:1021–31.
The British Journal of Radiology, November 2006
Hypoxia in prostate cancer
14. Wang JZ, Guerrero M, Li XA. How low is the a/b ratio for
prostate cancer? Int J Radiat Oncol Biol Phys
2003;55:194–203.
15. Wang JZ, Li XA, Yu CX, et al. The low a/b ratio for prostate
cancer: what does the clinical outcome of HDR brachyther-
apy tell us? Int J Radiat Oncol Biol Phys 2003;57:1101–8.
16. Fowler JF, Ritter MA, Fenwick JD, et al. How low is the a/b
ratio for prostate cancer? In regard to Wang et al. [letter]. Int
J Radiat Oncol Biol Phys 2003;57:593–5.
17. Wang JZ, Guerrero M, Li XA. Low a/b ratio for prostate
cancer: in response to Dr. Fowler et al. [letter]. Int J Radiat
Oncol Biol Phys 2003;57:595–6.
18. Nahum AE, Movsas B, Horwitz EM, et al. Incorporating
clinical measurements of hypoxia into tumor local control
modeling of prostate cancer: Implications for the a/b ratio.
Int J Radiat Oncol Biol Phys 2003;57:391–401.
19. Kal HB, Van Gellekom MPR. How low is the a/b ratio for
prostate cancer? Int J Radiat Oncol Biol Phys
2003;57:1116–21.
20. Carlone M, Wilkins D, Nyiri B, et al. Comparison of a/b
estimates from homogeneous (individual) and heteroge-
neous (population) tumor control models for early stage
prostate cancer. Med Phys 2003;30:2832–48.
21. Dasu A, Fowler JF. Comments on ’Comparison of in vitro
and in vivo a/b ratios for prostate cancer’. Phys Med Biol
2005;50:L1–4.
22. Wang JZ, Stewart RD, Carlson DJ, et al. Reply to
‘‘comments on ‘comparison of in vitro and in vivo a/b
ratios for prostate cancer’’’. Phys Med Biol 2005;50:
L5–8.
23. Levegrün S, Jackson A, Zelefsky MJ, et al. Fitting tumor
control probability models to biopsy outcome after three-
dimensional conformal radiation therapy of prostate cancer:
pitfalls in deducing radiobiologic parameters for tumors
from clinical data. Int J Radiat Oncol Biol Phys
2001;51:1064–80.
24. King CR, Mayo CS. Is the prostate a/b ratio of 1.5 from
Brenner and Hall a modeling artifact? [letter]. Int J Radiat
Oncol Biol Phys 2000;47:536–8.
25. Wyatt RM, Beddoe AH, Dale RG. The effects of delays in
radiotherapy treatment on tumor control. Phys Med Biol
2003;48:139–55.
26. Wang JZ, Li XA. Evaluation of external beam radiotherapy
and brachytherapy for localized prostate cancer using
equivalent uniform dose. Med Phys 2003;30:34–40.
27. Perez CA, Michalski J, Mansur D, et al. Impact of elapsed
treatment time on outcome of external-beam radiation
therapy for localized carcinoma of the prostate. Cancer, J
2004;10:349–56.
28. Wang JZ, Mayr NA, Li XA, et al. Modeling prostate cancer:
in regard to Nahum et al. [letter]. Int J Radiat Oncol Biol
Phys 2005;61:309–10.
29. Dale RG. The application of the linear-quadratic dose-effect
equation to fractionated and protracted radiotherapy. Br J
Radiol 1985;58:515–28.
The British Journal of Radiology, November 2006
30. Haustermans K, Fowler JF. A comment on proliferation
rates in human prostate cancer [letter]. Int J Radiat Oncol
Biol Phys 2000;48:303.
31. Jones B, Cominos M, Dale RG. Application of biological
effective dose (BED) to estimate the duration of sympto-
matic relief and repopulation dose equivalent in palliative
radiotherapy and chemotherapy. Int J Radiat Oncol Biol
Phys 2003;55:736–42.
32. Porter EH. The statistics of dose/cure relationships for
irradiated tumors. Br J Radiol 1980;53:210–27.
33. Buffa FM, West C, Byrne K, et al. Radiation response
and cure rate of human colon adenocarcinoma spheroids
of different size: the significance of hypoxia on tumor
control modeling. Int J Radiat Oncol Biol Phys 2001;49:
1109–18.
34. Hendry JH, Thames HD. Fractionation sensitivity and the
oxygen effect. Br J Radiol 1990;63:79–80.
35. Popple RA, Ove R, Shen S. Tumor control probability for
selective boosting of hypoxic subvolumes, including the
effect of reoxygenation. Int J Radiat Oncol Biol Phys
2002;54:921–7.
36. Zolzer F, Streffer C. Increased radiosensitivity with chronic
hypoxia in four human tumor cell lines. Int J Radiat Oncol
Biol Phys 2002;54:910–20.
37. Watts ME, Hodgkiss RJ, Jones NR, et al.
Radiosensitization of Chinese hamster cells by oxygen
and misonidazole at low X-ray doses. Int J Radiat Biol
1986;50:1009–21.
38. Skarsgard LD, Harrison I, Durand RE. The radiation
response of asynchronous cells at low dose: evidence of
substructure. Radiat Res 1991;127:248–56.
39. Speke AK, Hill RP. The effects of claiming and reoxygena-
tion on repopulation during fractionated irradiation. Int J
Radiat Oncol Biol Phys 1995;31:857–63.
40. Hill RP, Fyles W, Milosevic M, et al. Is there a relationship
between repopulation and hypoxia/reoxygenation? Results
from human carcinoma of the cervix. Int J Radiat Biol
2003;79:487–94.
41. Brenner DJ, Hlatky LR, Hahnfeldt PJ, et al. A convenient
extension of the linear-quadratic model to include redis-
tribution and reoxygenation. Int J Radiat Oncol Biol Phys
1995;32:379–90.
42. Jones B, Dale G. Estimation of tumor hypoxia fraction from
clinical data sets compatible with accelerated repopulation.
Acta Oncologica 1998;37:263–8.
43. Brenner DJ, Hall EJ. In response to Drs. King and Mayo:
Low a/b values for prostate appear to be independent of
modeling details [letter]. Int J Radiat Oncol Biol Phys
2000;47:2753–65.
44. Li XA, Wang JZ, Stewart RD, et al. Dose escalation in
permanent brachytherapy for prostate cancer: dosimetric
and biological considerations. Phys Med Biol
2003;48:2753–65.
45. Stock RG, Stone NN, Tabert A, et al. A dose-response study
for I-125 prostate implants. Int J Radiat Oncol Biol Phys
1999;41:101–8.
911
 The British Journal of Radiology, 79 (2006), 912–917

SHORT COMMUNICATION

Scatter from radiotherapy beams emerging from primary

barriers: an aid to bunker design
1
D C BENITO, MSc and 2A L MCKENZIE, DSc

1
H H Wills Physics Laboratory, University of Bristol, Tyndall Avenue, Bristol BS8 1TL and
2
Department of Medical Physics and Bioengineering, Bristol Haematology and Oncology Centre,
Horfield Road, Bristol BS2 8ED, UK

ABSTRACT. Charts to assist in linear accelerator bunker design have been produced
using the Monte Carlo framework, Geant4. These charts assess the amount of forward Received 10 January 2006
scatter produced at different angles to the beam axis by concrete and steel barriers Revised 24 May 2006
Accepted 30 May 2006
irradiated by 6 MV and 10 MV photon beams at normal incidence. These new charts
complement existing charts of broad-beam transmission through walls. This is because DOI: 10.1259/bjr/94993761
the existing charts give no indication of the amount of scatter emerging at large angles
from the beam axis, for example, towards the maze entrance. ’ 2006 The British Institute of
Radiology

Designers of linear accelerator bunkers commonly
refer to charts such as those given in references [1, 2].
These plot primary beam attenuation against wall
thickness at given X-ray energies. Experience shows that
such charts are sufficiently accurate for design purposes.
The relative intensity of back-scatter from the bunker
walls may also be determined from charts in the above
publications to a level of accuracy that is sufficient for
radiation protection calculations. Such back-scatter may
be used to estimate dose rates at maze entrances,
provided that the linear accelerator is not orientated in
the room in such a way that the primary beam can be
aimed at the inner maze wall. In such cases, if the inner
maze wall has been designed insufficiently thick, the
dose rate from wide-angle scatter from the X-ray beam
emerging from the wall can be more than an order of
magnitude greater than the dose rates, typically a few
mGy h21, arising from the multiple-order scatter that
eventually finds its way down the maze (Figure 1). An
example is discussed later.
Surprisingly, perhaps, no charts have been published
to cater for this situation. While primary-beam charts can
be used to calculate the intensity of broad-beam
transmission through an inner maze wall, no charts exist
to show the level of scatter that may emerge at discrete
angles and contribute to the dose at the maze entrance
(Figure 1).
Designers who have access to software packages based
on Monte Carlo programs may use these to find the
scatter from such beam configurations. However, such
packages are not widely available, and require a period
of training and familiarization. For those who do not
design bunkers regularly, these inconveniences militate
against their use.
In attempting to provide a simple, practical solution to
what is a relatively common problem we have calculated
a series of plots, illustrated in Figures 2–5. These show
the intensity of radiation scattered at different angles
from a primary beam as it emerges from a wall made of
concrete or steel. The data are plotted as ratios of
scattered dose rate at 1 m from the exit point relative to
the dose rate of the primary beam as it enters the wall,
where it is assumed to have an area of 100 cm2.
In practical use, the bunker designer will need to know
the scatter not at 1 m, but at several metres from the
centre of the area of the beam as it emerges from the wall.
This distance would typically be the distance to the maze
entrance. The dose rate there will be found by applying
the inverse-square law. The inaccuracies of using a
distance measured from a single point in the centre of a
finite beam area are minimized because the distance to
the maze entrance is generally significantly greater than
the size of the exit beam.
Method
In order to calculate the scatter charts, we used the
Geant4 Monte Carlo framework code [3] to simulate
scattering within the wall. We chose 6 MV and 10 MV
spectra [4] from Elekta (Elekta; Elekta Ltd, Crawley, West
Sussex, UK) and Varian (Varian; Varian Medical Systems
UK Ltd, Crawley, West Sussex, UK) linear accelerators to
create beams incident on concrete of density 2.35 tonnes
m23 and steel of density 7.85 tonnes m23. These
materials are commonly used in bunker walls.
Where practicable, we checked that our use of the code
generated results that were consistent with our experi-
ments or with published charts. Comparison with
existing charts was only indirect, since no charts of the
kind we have produced were previously available.
Experimental verification was similarly not without
difficulty, because dose rates were generally low and
the level of background leakage and scatter tended to be

912 The British Journal of Radiology, November 2006

Short communication: Aiding bunker design
of the same order as that of scatter from the maze wall.
This is discussed in more detail below.
Plots
The calculated plots are shown in Figures 2–5. The
statistics were such that the plots were smooth, apart
The British Journal of Radiology, November 2006
Figure 1. A simplified plan view of
a radiotherapy treatment room.
from data points at the largest barrier thicknesses at the
greatest scatter angles where random deviations from
smoothed data were in the order of 30%. Because of this,
some plots have been replaced with the best fitting
(exponential) dotted line. The position of the calculated
data points indicates the closeness of fit.
All four charts show the expected decrease in scatter
dose rate with increasing wall thickness at given angles
Figure 2. Dose ratio D1/Do, as illu-
strated in Figure 1 (concrete, 6 MV).
913

of scatter. The agreement between the gradients of these
charts and those of primary attenuation is striking. For
instance, referring to Figure 2, for 6 MV radiation
through concrete, the gradient of the intensity of scatter
exiting at 60 ˚ is three decades per metre of concrete. This
is the same as the gradient found in published charts
[1, 2] of the attenuation of primary 6 MV radiation
through concrete.
Similarly, for 6 MV radiation through steel (Figure 3),
the gradient of the intensity of scatter exiting at 60 ˚ is
three decades for approximately 300 mm of steel. After
correcting for the relative densities of steel and concrete,
this is essentially the same as the gradient in Figure 2 for
6 MV radiation through concrete.
Consideration of the physics shows that the similarity
of the gradients of scatter and primary plots is to be
expected. Suppose that the intensity of 60 ˚ scatter exiting
a concrete wall of thickness 0.2 m is measured. Suppose
that an extra metre of concrete is now added
‘‘upstream’’, making a total thickness of 1.2 m. Any
scatter from this extra metre will be so attenuated by the
final 0.2 m when it exits the wall that it will contribute a
negligible amount to the measured scatter, relative to
that arising from the 0.2 m nearest the detector.
However, after the primary beam has travelled the
added metre and enters the final 0.2 m, it is diminished
in intensity by three decades. Therefore, the scatter
arising within the final 0.2 m will also have diminished
by three decades. This, in broad terms, explains why
plots of both the exit scatter and primary attenuation
have the same gradient.
An example will serve to illustrate the use of the
charts.
Suppose that the inner maze wall in Figure 1 is made of
concrete of density 2.35 tonnes m23 and is 0.6 m thick. This
was a common thickness for secondary barriers in bunkers
914
D C Benito and A L McKenzie
Figure 3. Dose ratio D1/Do, as illu-
strated in Figure 1 (steel, 6 MV).
designed for cobalt treatment. Suppose that a 6 MV linear
accelerator is to be installed in the bunker. Assuming an
instantaneous dose rate of 4 Gy min21 at the isocentre, the
dose rate at the entrance to the maze wall, 4 m from the
source, is 15 Gy h21. For a collimator setting of
10 cm610 cm, the area at the wall is 1600 cm2.
Using Figure 2, the dose rate scattered at 60 ˚ to 1 m
from the centre of the exit beam is 561026615 Gy h21
61651200 mGy h21. The factor 16 arises because the
charts are presented for a beam area of 100 cm2 on entry
to the wall. Using the inverse-square law to find the dose
at the maze entrance, which also happens to be 4 m from
the centre of the exit beam, and using a factor of unity to
convert whole-body dose to effective dose, the effective
dose rate at the maze entrance is then 75 mSv h21.
Such a calculation shows that additional protection
might be needed for the inner maze wall in this case,
possibly provided by steel. From the discussion above,
the same Figure 2 could be used to determine the scatter
exiting from a wall of concrete and steel by finding the
equivalent thickness of an all-concrete wall.
Comparison with measurement
Before these plots can be used with confidence to
design bunkers, the level of scattered dose that they
predict must be compared with measurement. The
reason why direct comparison with experiment is
difficult is explained below. We were able to check that
elements of our implementation of the Geant4 code were
in accordance with existing charts by comparing our
predictions of backscatter with those of the experimental
data summarized in Figure 19 of Reference [2]. The
middle column of Table 1 shows our Geant4 estimates of
the intensity of 6 MV radiation back-scattered to 1 m
The British Journal of Radiology, November 2006
Short communication: Aiding bunker design

Figure 4. Dose ratio D1/Do, as illu-

strated in Figure 1 (concrete, 10 MV).

Figure 5. Dose ratio D1/Do, as illu-

strated in Figure 1 (steel, 10 MV).

The British Journal of Radiology, November 2006 915


Table 1. Comparison between the Geant 4 simulation and
data from reference [2] for back-scattered doses for three
different scattering angles
Scattering angle Relative back-scattered Data from
(degrees) dose estimated from Geant4 Reference [2]
180 6.6261025 ¡ 0.861025 7.4061025
150 6.3761025 ¡ 0.761025 6.8061025
120 4.5261025 ¡ 0.461025 5.1061025
from a concrete barrier irradiated by a 100 cm2 beam.
The figures are shown relative to the intensity of the
incident beam.
The statistical uncertainties in the Monte-Carlo results
are indicated by two standard deviations. The estimates
agree with published data to within about 10%, which
suggests that the code was functioning as intended, as
far as could be tested using results for back-scatter.
The difficulty in verifying the plots by direct experi-
ment is that (1) any barrier from which oblique scatter is
measured will also transmit head leakage from the linear
accelerator into the detector and (2) back-scatter from the
barrier tends to be multipally scattered around the
barrier, and again adds to the signal due to the oblique
scatter. The intensity of both the leakage and the multiple
scatter are generally of the same order as, or greater than,
that of the oblique scatter, and it is difficult to identify
the individual components in the detected signal. This
neutralizes the effectiveness of building temporary
barriers within a linear accelerator room in order to test
the oblique-scatter plots.
Experiments based on bunkers in clinical use run into
difficulty because dose rates from scatter from inner
maze walls in existing installations tend to be very low.
Inner maze walls may well be over-protected: maze
walls that were inadvertently designed too thin would
have had extra protection added at the first opportunity.
These low dose rates are not only of the same order as
the background from leakage and multiple scatter, but
they are also at the detection limit of most readily
available ionization chambers.
It might be perceived that the presence of head leakage
and multiple scatter down a maze would detract from
the usefulness of oblique-scatter plots. However, in
practice, it is the dose beyond the maze entrance that is
of interest, where the head leakage is insignificant
because of attenuation by the outside secondary wall,
and where the multiple scattered radiation has dimin-
ished with distance down the maze and can, in any case,
be estimated from existing charts. It is here, beyond the
maze entrance, that the oblique-scatter plots should
prove useful in ensuring that scatter from the inner maze
wall to the outside is kept to acceptably low levels.
Instead of building barriers inside the linear accele-
rator bunker, another way to test the plots is to measure
the dose rate beyond the maze entrance where the
detector is protected from head leakage by the outside
secondary wall. The detector should be positioned
slightly to one side of the maze barrier, so that it can
‘‘see’’ the area where the primary beam exits the inner
maze wall, but not the far end of the maze, which is a
source of multiple scatter.
There is only one linear accelerator bunker in our
radiotherapy centre where the primary beam can be
916
D C Benito and A L McKenzie
directed at the inner maze wall. This wall consists of a
0.6 m concrete barrier to which has been added 0.25 m of
steel on the inside of the room. A Southern Scientific RO-
10 survey meter using a 400 cm3 ionization chamber was
placed outside the end of the maze, and slightly to one
side of it, in order to minimize the level of multiple
scatter entering the detector. The instrument had
been calibrated less than 2 months previously in
accordance with the laboratory accreditation require-
ments of the UK Accreditation Service. This detector was
4.43 m from the centre of the linear accelerator beam at
the point where it exited the inner maze wall. The scatter
angle from the centre of the beam to the detector was
62.5 ˚.
The primary beam of the linear accelerator was aimed at
normal incidence onto the inner maze wall, and set to
deliver a 40 cm620 cm beam as defined at a target
distance of 1 m. The beam was narrower in the horizontal
direction so that the oblique scatter entering the detector
would be confined within a smaller range of scatter angles.
The measured dose rate was 0.7 mGy h21. When the
detector was moved out of the line of sight of the exit
beam, the measured dose rate fell to between 0.1 mGy h21
and 0.2 mGy h21. This is representative of the dose rate
from multiple scatter reaching the detector even although
it was within the ‘‘shadow’’ of the main component of
multiple scatter from the far end of the maze. Hence the
dose rate attributable to oblique scatter from the inner
maze wall was in the order of 0.5 mGy h21.
This dose rate has to be compared with that predicted
by the plots. We accounted for the two-component
nature of our maze wall by multiplying by the ratio of
the densities of steel and concrete to estimate the
thickness of an equivalent wall built either entirely of
concrete or entirely of steel. The calculated thicknesses
are 1.435 m and 0.43 m, respectively. For an oblique
angle of 62.5 ˚, the charts in Figures 2 and 3 predict a
relative scattered dose rate of 2.161028 and 1.261028,
respectively. Using the linear accelerator output of
240 Gy h21 at 1 m from the source, and taking care to
estimate the incident beam size at the same distance from
the source as that at which the output is calculated, the
predicted dose rates at the detector are 1.1 mGy h21 or
1.9 mGy h21 depending upon whether an all-concrete or
an all-steel wall is assumed.
Hence, in comparison with the measured rate of
0.5 mGy h21, the plots predict between two and four
times the actual dose rate. Several uncertainties attend
this conclusion, however. The measured dose rates were
at the very limits of detection of the instrument. In
addition to the concrete and steel composition of the
wall, there is a layer of plaster of indeterminate
thickness, but which will reduce the measured dose
rate. The predicted dose rates had to be made using a
calculated effective thickness, which turned out to be so
large that the plots even had to be extrapolated slightly
to determine the relative scattered dose.
Conclusion
The charts presented here are for calculating the
intensity of radiation emerging from a wall and scattered
in different directions, and are particularly helpful in
The British Journal of Radiology, November 2006
Short communication: Aiding bunker design
designing bunkers where a primary beam is directed at a
maze wall.
A ratio of between two and four between the predicted
and the measured dose rate is equivalent to an
uncertainty of 100–200 mm concrete or 30–60 mm steel.
The lower values in these ranges may be acceptable,
particularly since the plots appear to be conservative,
that is, a design based on the plots might tend to
overestimate the protection required. The discrepancy
may arise because of the very low dose rates that are
available and because of the uncertainty of the effect of
the additional plaster layer to the wall which would
reduce the amount of the discrepancy.
In the light of the remaining uncertainties, we should
welcome experimental verification by other centres who
may happen to have thinner maze walls, ideally of
uniform composition. Such independent corroboration
would lend weight to these plots as a primary tool in
room design.
The British Journal of Radiology, November 2006
Acknowledgments
We are grateful for the assistance of Paul Stevens and
Henry Lawrence in implementing the Geant4 code.
References
1. Handbook of Radiological Protection. Part 1: Data. SBN 11
360079 8. London, UK: HMSO, 1971.
2. Recommendation for Data on Shielding from Ionizing
Radiation, Part 2 Shielding from X-radiation. British
Standards: BS4094. ISBN 058006522-7. London, UK: British
Standards Institution, 1971.
3. GEANT4 (Agostinelli S, Allison J, Amako K, Apostolakis J,
Araujo H, Arce P, et al), http://cern.ch/geant4 [Accessed 18
September 2006].
4. Sheikh-Bagheri D, Rogers DW. Monte Carlo calculations of
nine megavoltage photon beam spectra using the BEAM
code. Med Phys 2002;29:391–402.
917
 The British Journal of Radiology, 79 (2006), 918–921

SHORT COMMUNICATION

The use of megavoltage cone-beam CT to complement CT for

target definition in pelvic radiotherapy in the presence of hip
replacement
M AUBIN, MSc, O MORIN, BSc, J CHEN, PhD, A GILLIS, MD, B PICKETT, MSc, J F AUBRY, MSc,
C AKAZAWA, CMD, J SPEIGHT, MD, M ROACH III, MD and J POULIOT, PhD

Department of Radiation Oncology, University of California San Francisco, Comprehensive Cancer

Center, San Francisco, CA 94143, USA

ABSTRACT. In Europe and the USA combined, over half a million people had a hip joint
replaced in 2005, contributing to the increasing number of radiotherapy patients with
metallic hip prostheses. The treatment plan for external beam radiation therapy is
based on the delineation of the anatomy in the planning CT scan. When implanted Received 20 January 2006
objects of high atomic number (Z) material are present, however, severe image Revised 22 June 2006
artefacts are generated in conventional CT, strongly hindering the ability to delineate Accepted 4 July 2006
some organs. This is particularly the case for the planning of prostate patients with hip
DOI: 10.1259/bjr/19559792
prostheses. This short communication presents the use of a new imaging modality,
megavoltage cone-beam CT, to complement the regular CT for target definition of ’ 2006 The British Institute of
prostate cancer treatment of patients with hip replacements. Radiology

Treatment planning for external beam radiation
therapy is based on the delineation of the anatomy
visualized in the planning CT scan. However, image
artefacts caused by the presence of a hip replacement
often render CT images useless for prostate delineation
(Figure 1, left) and preclude precise dose calculation.
It was recently suggested that CT-MR image registra-
tion could facilitate target definition for a prostate patient
with hip replacements [1]. Effectively, MR images
provide an accurate definition of the clinical target
volume (CTV) and better visualization of normal
structures. A number of factors, however, were found
to affect image quality and/or accuracy of target
definition. The standard MR couch, different from a CT
or linac treatment couch, might result in different patient
position, and the presence of the metallic implants may
create significant distortion. Finally, in the presence of a
hip replacement, neither the regular CT nor the MR can
provide accurate electron density information for hetero-
geneous dose calculation. Promising artefact reduction
techniques on regular CT are being developed to
minimize the impact of streaking artefacts [2]. At that
time, they may require manual image post-processing
and most CT scanners available in radiation oncology
departments are not equipped with these features.
We report on the use of megavoltage cone-beam CT
(MV CBCT) to complement the conventional CT for
Address correspondence to: Dr Jean Pouliot , Department of
Radiation Oncology, University of California San Francisco, 1600
Divisadero Street, Suite H1031, San Francisco, CA 94941-1708,
USA. E-mail: pouliot@radonc17.ucsf.edu.
This research was supported by Siemens Oncology Care Systems.
target organ definition in seven patients treated for
prostate cancer with external beam radiation therapy. In
this study, we exploited the predominantly Compton
scattering of high-energy photons delivered in MV CBCT
systems to obtain 3D images of the anatomy in the
presence of unilateral or bilateral hip replacements and
complement the planning CT during the target delinea-
tion process.
Method
A cone-beam CT image is reconstructed from a large
set of projection images acquired at various angles
around a patient in a process similar to that of
conventional CT. In cone-beam CT, a two-dimensional
(2D) array of detectors, in our case a portal imager
attached to the linear accelerator (linac), is used to
reconstruct a three-dimensional (3D) image. For MV
CBCT, the 6 MV treatment beam of the accelerator,
containing photons primarily in the MeV range, is used
for the imaging.
A MV CBCT system [3, 4] integrated onto an ONCOR
clinical Linac (Siemens Oncology Care Systems,
Concord, CA) was used to acquire 3D images in
treatment position for seven prostate patients with
unilateral or bilateral hip prostheses. MV CBCT acquisi-
tions were performed by rotating the linac in a
continuous 200 ˚ arc (270 ˚ to 110 ˚, clockwise) acquiring
one portal image for each degree. Because MV CBCT
uses the treatment beam, the treatment planning system
(Pinnacle, Phillips, Best, The Netherlands) was used to
evaluate the dose delivered during an MV CBCT

918 The British Journal of Radiology, November 2006

Short communication: Target definition in pelvic radiotherapy with MV CBCT
acquisition. For a typical acquisition procedure, the dose
at the isocentre was 0.05 Gy and the acquisition lasted
45 s. The reconstructed image, a typical 25662566274
volume (27.4 cm627.4 cm627.4 cm), was available less
than 2 min after the acquisition of the first portal image.
A first order scatter correction was applied to facilitate
the adjustment of window/levels. For each patient, the
MV CBCT images were imported into the treatment
planning system and registered with the original plan-
ning CT using bony anatomy contoured on each image
set. The target volumes and organs at risk for prostate
treatment were contoured using both the CT and the MV
CBCT for single hip replacement, and using only the MV
CBCT for bi-lateral hip prostheses.
Results
The MV CBCT images could be used to visualize
clearly the hip prosthesis and bony anatomy and provide
sufficient soft-tissue contrast to help delineate the
The British Journal of Radiology, November 2006
Figure 1. Comparison of a conven-
tional CT (left) and megavoltage
cone beam CT (MV CBCT) (right).
(A) Axial and (B) coronal views are
shown for a unilateral hip replace-
ment. (C) Axial view for a bilateral
hip replacement.
prostate, bladder and rectum. The artefacts on the
regular CT obscure the border between the prostate
and anterior wall of the rectum (Figure 1A, left) and the
interface between the prostate base and the bladder neck
(Figure 1B, left). The MV CBCT images were particularly
useful to help delineate these structures as well as the
lateral extension of the prostate in the axial plane, the
seminal vesicles and the lymph nodes. Also, normal
anatomy such as pelvic bones, penile bulb, bladder,
femoral heads, rectum and small bowel can be deli-
neated with higher accuracy as well.
An example of organ segmentation is presented in
Figure 2. The change of shape of the prostate (red),
bladder (yellow) and rectum (blue) between the CT (left)
and the MV CBCT (right) can be easily observed. In this
study, the time intervals between the CT and the MV
CBCT ranged from a few hours to 1 week. By itself, this
can explain the change in rectum and bladder volumes
due to different fillings. For the six patients with a single
hip replacement, the posterior limit of the prostate was
found to be more anterior, and therefore the prostate
919

volumes contoured with the help of MV CBCT were
generally smaller than the volumes that would have been
estimated using only the regular CT containing severe
artefacts. These smaller prostate volumes may prevent
overdosage of the rectum. Target delineation for the
patient with bilateral hip prostheses was entirely
performed using the MV CBCT, since the relevant
organs were totally obscured due to the severity of the
artefacts on the conventional CT.
Discussion
Image artefacts (Figures 1 and 2, left) caused by the
presence of hip replacements render regular CT images
difficult to use for prostate delineation. In contrast, the
presence of high-Z material has relatively little impact on
the image quality of MV CBCT.
Tests performed on phantoms [5] showed that the
presence of a metallic object strongly impacts on
Hounsfield numbers (up to 70% error) of a conventional
CT image and therefore disturbs the electron density
even far away from the object, making CT inaccurate for
dose calculation. For this reason, CT treatment plans of
patients with hip replacements are generally produced
without density correction. Similar tests performed with
MV CBCT on a phantom with and without a metallic
object demonstrated that Hounsfield numbers remain
unchanged (within 3%) in the presence of metallic
objects, allowing for significantly more accurate dose
calculation. Thus, the next step after using MV CBCT for
image segmentation will be to use MV CBCT for dose
calculation. Research to develop calibration procedures
920
M Aubin, O Morin, J Chen et al
Figure 2. Segmentation of the
bladder (yellow), prostate (red) and
rectum (bule) using the conven-
tional CT (left) and the megavoltage
cone beam CT (MV CBCT) (right)
shown on the (A) axial, (B) sagittal
and (C) coronal views.
to use MV CBCT for dose calculations is being
performed. There is also ongoing technical development
to increase the field of view of the current version of MV
CBCT (27 cm627 cm) to encompass the entire pelvic
region.
The possibility of considering dose escalation proto-
cols depends on the ability to identify the prostate
volume and critical structures for treatment planning,
and the accuracy of the dose calculation. Because of
degradation of image quality in the presence of high-Z
material, treatment planning and dose calculations are
limited in these settings. Consequently, patients with hip
prostheses may not be candidates for advanced treat-
ment planning like intensity-modulated radiotherapy
(IMRT). With the advance capability of using MV CBCT
in the treatment planning software, it is now possible to
deliver accurately higher doses of radiation to the
prostate in patients with hip prostheses. While MV
CBCT acquisition dose is two to three times higher than a
conventional CT, it represents only a very small fraction
(,0.1%) of the treatment dose. With dose escalation these
patients may benefit from a decreased risk of treatment
failure.
MV CBCT in the presence of high-Z material may
improve treatment planning, allowing patients with
pelvic malignancies and hip prostheses the most
advanced form of radiation therapy.
Conclusion
MV CBCT provides 3D anatomical information of the
patient in the treatment position, even in the presence of
The British Journal of Radiology, November 2006
Short communication: Target definition in pelvic radiotherapy with MV CBCT
‘‘CT non-compatible’’ objects. MV CBCT registered with
the planning CT can complement missing information
and facilitate segmentation for planning purposes when
hip prostheses are present.
References
1. Charnley N, Morgan A, Thomas E, Wilson S, Bacon S, Wilson
D, et al. The use of CT-MR image registration to define target
volumes in pelvic radiotherapy in the presence of bilateral
hip replacements. Br J Radiol 2005;78:634–6.
2. Yazdia M, Gingras L, Beaulieu L. An adaptive approach
to metal artifact reduction in helical computed tomography
The British Journal of Radiology, November 2006
for radiation therapy treatment planning: experimental
and clinical studies. Int J Radiat Oncol Biol Phys 2005;62:
1224–31.
3. Pouliot J, Bani-Hashemi A, Chen J, et al. Low-dose
megavoltage cone-beam CT for radiation therapy. Int J
Radiat Oncol Biol Phys 2005;61:238–46.
4. Morin O, Gillis A, Chen J, Aubin M, Bucci MK, Pouliot J.
Megavoltage cone-beam CT: system description and IGRT
clinical applications. Special issue on image-guided radiation
therapy (IGRT). Med Dosim 2006;31:51–61.
5. Aubin M, Morin O, Bucci K, Chan A, Chen J, Ghelmansarai
F, et al. Megavoltage conebeam CT to complement prostate
planning CT in presence of hip prosthesis. ESTRO Annual
Meeting (abstract), Lisbon, 2005.
921
 The British Journal of Radiology, 79 (2006), 922–928

PICTORIAL REVIEW

CT of thoracic lymph nodes. Part I: anatomy and drainage

1
T SUWATANAPONGCHED, MD and 2D S GIERADA, MD

1
Department of Radiology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, 270,
Rama VI Road, Rajthevi, Bangkok 10400, Thailand and 2Mallinckrodt Institute of Radiology,
Washington University School of Medicine, 510 South Kingshighway Blvd, Campus Box 8131, Saint
Louis, Missouri 63110, USA

ABSTRACT. CT is the primary non-invasive technique for the diagnostic evaluation of

thoracic lymph nodes. The CT patterns and anatomic location of thoracic lymph node Received 11 April 2005
involvement can provide important clues in the diagnosis of many diseases. Part I of the Revised 23 June 2005
pictorial review illustrates the anatomic location and drainage of thoracic lymph nodes Accepted 11 July 2005
in the chest wall, mediastinum and lungs through examples of pathologic involvement.
DOI: 10.1259/bjr/26411607
Part II of the pictorial review focuses on CT patterns of lymph node involvement in
various pulmonary and extrapulmonary diseases, differential diagnoses based on CT ’ 2006 The British Institute of
findings and pitfalls. Radiology

CT is the primary non-invasive technique for the
diagnostic evaluation of thoracic lymph nodes. Lymph
node abnormalities are depicted by CT as an increase in
nodal size and/or number or change in attenuation.
Although these findings are non-specific, patterns of
thoracic lymph node involvement can provide important
clues in the diagnosis of many pulmonary and extra-
pulmonary diseases. Part I of this pictorial review
illustrates the anatomic location and drainage of thoracic
lymph nodes in the chest wall, mediastinum and lungs
through examples of pathologic involvement. Part II
focuses on CT patterns of lymph node involvement in
various pulmonary and extrapulmonary diseases.
Classification of thoracic lymph nodes
As in other parts of the body, thoracic lymph nodes are
named using descriptive terminology according to the
blood vessels or visceral structures to which they are
most closely related, or by their general anatomic
location. Although there are slight differences in the
classification of the thoracic nodes [1–5], they can be
divided into those of the chest wall and those of the
intrathoracic contents. To facilitate accurate pathologic
staging and analysis of treatment outcomes in lung
cancer, a classification scheme for mediastinal and
pulmonary lymph nodes (Figure 1) has been devised
by the American Joint Committee on Cancer (AJCC) and
the Union Internationale Contre le Cancer (UICC) [5],
based on surgically recognizable anatomic landmarks.
Chest wall nodes
The axillary nodes (Figures 2 and 3) receive superficial
lymphatic drainage from the upper limbs, breasts and
Address correspondence to: D S Gierada.
pectoral muscles excluding their medial portions, pari-
etal pleura, and skin and muscles of the trunk above the
umbilicus and iliac crest [1, 2]. The lymph flow is
directed toward the terminal nodal group in the axillary
apices. The efferent vessels from this group unite as the
subclavian trunk, which finally drains directly or indi-
rectly into the jugulo-subclavian venous confluence [1, 2,
6]. A few efferents usually reach the supraclavicular
nodes, a well-recognized route for the spread of breast
cancer [1, 2, 6].
The internal mammary (internal thoracic or paraster-
nal) nodes (Figure 4) lie at the anterior ends of the
intercostal spaces, along the internal mammary (internal
thoracic) vessels. They receive lymphatic drainage from
the anterior diaphragmatic nodes, anterosuperior por-
tion of the liver, medial part of the breasts, and deeper
structures of the anterior chest and upper anterior
abdominal wall [2]. Their efferent channels may empty
into the right lymphatic duct, the thoracic duct, or the
inferior deep cervical nodes [3, 6].
The posterior intercostal nodes (Figures 5 and 6),
located near the heads and necks of the posterior ribs,
receive lymphatic drainage from the posterolateral
intercostal spaces, posterolateral breasts, parietal pleura,
vertebrae and spinal muscles [2–4]. The efferent vessels
from the upper intercostal spaces end in the thoracic
duct on the left, and in one of the lymphatic trunks on
the right [2–4]. Those from the lower four to seven
intercostal spaces unite to form a common trunk, which
empties into the thoracic duct or cisterna chyli [2–4]. The
juxtavertebral (pre-vertebral or paravertebral) nodes lie
along the anterior and lateral aspects of the vertebral
bodies, most numerous from T8 to T12 (Figures 5 and 6)
[3, 4]. They communicate with posterior mediastinal
lymph nodes [3] and the posterior intercostal nodes, and
similarly drain to the right lymphatic duct or thoracic
duct [3, 4].

922 The British Journal of Radiology, November 2006

Pictorial review: CT of thoracic lymph nodes

(a) (b)

Figure 1. Revised American Joint Committee on Cancer (AJCC) and the Union Internationale Contre le Cancer (UICC) regional
nodal stations for lung cancer staging. (From Mountain CF, Dresler CM. Regional lymph node classification for lung cancer
staging. Chest 1997;111:1718–23 [5]. Reprinted with permission). (a) Drawing illustrates mediastinum lymph node stations in the
frontal projection. Ao 5 aortic arch, PA 5 main pulmonary artery, 1 (red) 5 highest mediastinal nodes, 2R and 2L (dark blue) 5
right and left upper paratracheal nodes, 4R and 4L (orange) 5 right and left lower paratracheal nodes, 7 (blue) 5 subcarinal
nodes, 8 (grey) 5 para-oesophageal nodes, 9 (brown) 5 pulmonary ligament nodes, 10R and 10L (yellow) 5 right and left hilar
nodes, 11R and 11L (green) 5 right and left interlobar nodes, 12R and 12L (pink) 5 right and left lobar nodes, 13R and 13L (pink)
5 right and left segmental nodes, 14R and 14L (pink) 5 right and left subsegmental nodes. (b) Illustration of mediastinum lymph
node stations in the left anterior oblique projection. Ao 5 aortic arch, PA 5 main pulmonary artery, 3 (pink) 5 pre-vascular and
retrotracheal nodes, 5 (black) 5 subaortic nodes, 6 (red) 5 para-aortic nodes.

The diaphragmatic nodes are located on or just above Mediastinal lymph nodes
the thoracic surface of the diaphragm and are divided
into three groups [2–4, 7]. The anterior (pre-pericardial or Anterior mediastinal group
cardiophrenic) group (Figure 6) is located anterior to the This group includes the highest mediastinal (station 1,
pericardium, posterior to the xiphoid process, and in the Figures 1 and 3a), pre-vascular (station 3A, Figures 1
right and left cardiophrenic fat. This node group receives and 3b), and para-aortic (station 6, Figures 1 and 9)
afferent drainage from the anterior part of the diaphragm
and its pleura, and the anterosuperior portion of the
liver. They drain to the internal mammary nodes
alongside the xiphoid and can provide a route for
retrograde spread of breast cancer to the liver, via
lymphatics of the rectus abdominis muscle when the
upper internal thoracic trunks are blocked [4]. The
middle (juxtaphrenic or lateral) (Figure 7) group receives
lymph from the central diaphragm and from the convex
surface of the liver on the right [2]. The posterior
(retrocrural) group (Figure 8), lying behind the dia-
phragmatic crura and anterior to the spine, receives
lymph from the posterior part of the diaphragm and
communicates with the posterior mediastinal nodes and
para-aortic nodes in the upper abdomen [2, 4]. When
diaphragmatic nodes are enlarged, widespread disease Figure 2. Enhanced CT scan in a 66-year-old woman with
in other locations is usually present, so biopsy of these lymphoma showing multiple enlarged bilateral axillary
sites is uncommon [7]. lymph nodes (arrows).

The British Journal of Radiology, November 2006 923

 T Suwatanapongched and D S Gierada

(a) (b)

Figure 3. A 65-year-old man with chronic lymphocytic leukaemia. (a) Enhanced CT scan demonstrates enlarged right axillary
nodes (arrowheads) and right interpectoral (Rotter) node (black arrow) lying between pectoralis major (M) and minor (m)
muscles. Nodes in the subpectoral and interpectoral regions are included in the axillary nodal group. Also seen are enlarged
highest mediastinal nodes (station 1; white arrows) defined by their location cranial to the superior margin of the left
brachiocephalic vein, behind and to the right and left sides of the trachea. (b) Enhanced CT scan at the lower level shows
bilaterally enlarged axillary nodes (arrowheads), including left subpectoral nodes (open arrow) underneath the left pectoralis
minor muscle (m). There are enlarged pre-vascular nodes (station 3A; white arrows), which lie between the superior margin of
the left brachiocephalic vein (V) and the superior margin of the aortic arch, and anterior to its large arterial branches; enlarged
retrotracheal node (station 3P; black arrow), which lies behind the trachea and above the inferior aspect of azygos vein arch;
and enlarged right upper paratracheal nodes (station 2R; wavy arrow), which are located above the superior margin of the
aortic arch.

Figure 5. Enhanced CT scan of a 31-year-old man with

Figure 4. Enhanced CT scan at the level of the main
pulmonary artery in a 55-year-old woman with left breast
cancer demonstrating enlarged left internal mammary node
(arrow). Note normal right internal mammary vessels (wavy
arrow) and a portion of primary cancer in the left breast
(asterisk).
lymphoma showing enlarged, necrotic right and left inter-
costal nodes (white arrows) as well as enlarged left
paravertebral (arrowheads) and retrocrural (black arrows)
nodes. Note a left pleural effusion (E) with pleural nodules
(small white arrows), splenectomy clips and coeliac adeno-
pathy (N). A 5 aorta.

924 The British Journal of Radiology, November 2006

Pictorial review: CT of thoracic lymph nodes
Figure 6. Enhanced CT scan in a 69-year-old woman with
lymphoma showing enlarged bilateral paravertebral nodes
(white arrows), left intercostal node (open arrow) and
anterior diaphragmatic nodes (black arrows). Note bilateral
pleural effusions (E).
nodes [2, 3, 5, 8]. They receive afferent vessels from the
thymus, thyroid, heart and pericardium, diaphragmatic
and mediastinal pleura, and middle diaphragmatic
nodes [2, 3]. Their efferent channels join those from the
paratracheal, tracheobronchial and internal mammary
nodes, to form the right and left bronchomediastinal
trunks, which may empty to the right lymphatic duct, the
thoracic duct, or open independently into the jugulo-
subclavian venous confluence [2, 3].
Paratracheal and tracheobronchial groups
These groups receive drainage from most parts of
the lungs and bronchi, thoracic trachea, heart and some
efferents from the upper para-oesophageal nodes of the
posterior mediastinal group [2, 4]. The nodes comprising
Figure 7. Non-enhanced CT scan in a 28-year-old woman
with metastatic papillary serous adenocarcinoma of the
ovary revealing enlarged, densely calcified right middle
diaphragmatic nodes (arrow), located lateral to the
intrathoracic end of the inferior vena cava (V) and near the
insertion of the right phrenic nerve.
The British Journal of Radiology, November 2006
Figure 8. CT scan through the upper abdomen in a 45-year-
old man with distal oesophageal carcinoma (not shown)
revealing enlarged retrocrural lymph nodes (large arrows)
and liver metastases (small arrows).
these groups include the upper (station 2R, 2L, Figures 1
and 3b) and lower (station 4R, 4L, Figures 1, 9 and 10)
paratracheal, subaortic (aortopulmonary window, sta-
tion 5, Figures 1, 11, and 12), retrotracheal (station 3P,
Figures 1 and 3b), and subcarinal (station 7, Figures 1
and 12) nodes [2, 3, 5]. The azygos node, located medial
to the azygos arch, is included in station 4R [5]. The
upper paratracheal nodes link the lower paratracheal
and inferior deep cervical nodes [10]. The subcarinal
nodes are contiguous with the hilar nodes and drain to
the paratracheal nodes, preferentially to the right [11].
Figure 9. Non-enhanced CT scan in the same patient as in
Figure 7 revealing enlarged, calcified para-aortic nodes
(station 6; arrows), lying anterior and lateral to the aortic
arch (A) below its superior margin. Also seen is right lower
paratracheal lymphadenopathy (station 4R; open arrow). V
5 superior vena cava. (From Glazer HS, Molina PL, Siegel MJ,
Sagel SS. High-attenuation mediastinal masses on unen-
hanced CT. AJR Am J Roentgenol 1991;156:45–50 [8].
Reprinted with permission).
925

Figure 10. Enhanced CT scan in a 73-year-old man with left
lower lobe lung cancer (not shown) showing enlarged right
lower paratracheal nodes (large arrow) lying medial to the
azygos vein (V) and enlarged left lower paratracheal nodes
(station 4L; open arrow) lying medial to ligamentum
arteriosum (small arrows). Lower paratracheal nodes lie
caudal to the top of the aortic arch. (From Sagel SS, Slone
RM. Lung. In: Lee JKT, Sagel SS, Stanley RJ, Heiken JP, editors.
Computed body tomography with MRI correlation, 3rd edn.
Philadelphia, PA: Lippincott-Raven Publishers, 1998:351–454
[9]. Reprinted with permission).
Figure 11. Enhanced CT scan in a 58-year-old woman with
carcinoid tumour showing enhancing subaortic lymphade-
nopathy (station 5; arrows) within the aortopulmonary
window region. This group is located lateral to the
ligamentum arteriosum (not seen). Note primary tumour in
the left upper lobe (open arrow).
926
T Suwatanapongched and D S Gierada
Figure 12. Enhanced CT scan in a 65-year-old man with
diffuse pulmonary lymphangitic carcinomatosis secondary to
non-small cell lung cancer (not shown) demonstrating
enlarged subcarinal (station 7; curved arrow), para-oesopha-
geal (black arrow), right hilar (station 10R; large white
arrows) and left hilar (station 10L; open arrow) nodes. Hilar
nodes are outside the mediastinal pleura, below the top of
the upper lobe bronchi. Note enlarged subaortic (arrow-
head) and para-aortic (small white arrow) nodes. Oe 5
oesophagus.
Figure 13. Enhanced CT scan in a 65-year-old man with non-
small cell lung cancer demonstrating metastasis to left
pulmonary ligament node (station 9; curved arrow) from
left lower lobe lung cancer (straight arrow). Oe 5 oesopha-
gus, A 5 aorta. (From Sagel SS, Slone RM. Lung. In: Lee JKT,
Sagel SS, Stanley RJ, Heiken JP, editors. Computed body
tomography with MRI correlation, 3rd edn. Philadelphia,
USA: Lippincott-Raven Publishers, 1998:351–454 [9].
Reprinted with permission).
The British Journal of Radiology, November 2006
Pictorial review: CT of thoracic lymph nodes
Figure 14. Enhanced CT scan in a 29-year-old woman with
sarcoidosis demonstrating enlarged right lobar node (station
12R; arrowhead) at the bifurcation of the bronchus inter-
medius, right segmental node (open arrow) adjacent to the
right middle lobe lateral segmental bronchus, and left
interlobar nodes (station 11R and 11L; white arrows)
between the lingular and left lower lobe superior segmental
bronchus. Note enlarged subcarinal nodes (black arrows) and
bilateral pulmonary involvement.
Thus, the left lower lobe is the most common primary
site for contralateral mediastinal lymph node metastasis
in lung cancer.
Posterior mediastinal group
The posterior mediastinal nodes are comprised of the
para-oesophageal (station 8, Figures 1 and 12) and
pulmonary ligament (station 9, Figures 1 and 13) nodes
[4, 5]. The para-oesophageal nodes receive afferent
vessels from the thoracic oesophagus, posterior pericar-
dium, diaphragm, posterior diaphragmatic nodes and
Figure 15. Enhanced CT scan in the same patient as in
Figure 12 showing enlarged right and left segmental nodes
(station 13R and 13L; large white arrows) lying adjacent to
the segmental bronchi (small white arrows) and enlarged
para-oesophageal nodes (black arrows).
The British Journal of Radiology, November 2006
Figure 16. Axial CT scan with lung-window setting in a 59-
year-old man with myocardial infarction showing a 1 cm,
indeterminate, solitary pulmonary nodule containing an
eccentric calcific focus in the right middle lobe (arrow).
Wedge resection revealed a subsegmental lymph node
(station 14R) with calcified granuloma.
the left hepatic lobe, and are more numerous on the left
[2]. The pulmonary ligament nodes receive drainage
from the basilar segments of the lower lobes and lower
half of the oesophagus [4]. The efferents from the
posterior mediastinal nodes communicate with the
tracheobronchial group, particularly subcarinal nodes,
and drain chiefly into the thoracic duct, but also drain to
the subdiaphragmatic para-aortic or coeliac nodes [3, 4].
Lymph nodes of the lungs
Lymph nodes are located along the bronchi and can be
divided into hilar (station 10R, 10L, Figures 1 and 12)
and intrapulmonary nodes [5, 10]. The latter consist of
interlobar (station 11R, 11L, Figures 1 and 14), lobar
(station 12R, 12L, Figures 1 and 14), segmental (station
13R, 13L, Figures 1, 14 and 15), subsegmental (station
14R, 14L, Figures 1 and 16) and intraparenchymal
intrapulmonary (Figure 17) nodes [5, 9, 12]. Most of
the lymphatic flow of the lungs is directed toward the
interlobar and hilar nodes, which drain into the
subcarinal nodes or directly into the lower paratracheal
nodes [3, 4, 10, 11].
The normal hilar and interlobar nodes are frequently
visible, particularly with thinner (1–3 mm) collimation and
intravenous contrast administration [13]. Recognition of
these nodes is important to avoid misdiagnosis of
pulmonary embolism. Intraparenchymal intrapulmonary
nodes may present as indeterminate subpleural pulmon-
ary nodules in the lower parts of the lungs [12].
927

(a)
Figure 17. A 58-year-old man with bronchioloalveolar carcinoma of the left upper lobe (not shown). (a) CT scan with lung-
window setting demonstrates a tiny, subpleural nodule in the lingular segment (arrow). (b) Histological examination reveals a
normal lymph node (arrows), surrounded by alveolar tissue. It had capsule with visible germinal centres and contains histiocytes
and carbon pigment (haematoxylin and eosin 640).
References
1. Johnson D, Ellis H. Pectoral girdle and upper limb. In:
Standring S, editor. Gray’s anatomy, 39th edn. Edinburgh,
Scotland: Churchill Livingstone, 2005:817–49.
2. Johnson D, Shah P. Thorax. In: Standring S, editor. Gray’s
anatomy, 39th edn. Edinburgh, Scotland: Churchill
Livingstone, 2005:951–1079.
3. Fraser RS, Müller NL, Colman N, Paré PD. The lymphatic
system of the lungs, pleura, and mediastinum. In: Fraser RS,
Müller NL, Colman N, Paré PD, editors. Fraser and Paré’s
diagnosis of diseases of the chest, 4th edn. Philadelphia, PA:
W.B. Saunders Company, 1999:172–95.
4. Weinberg JA. The intrathoracic lymphatics. In: Haagensen
CD, editor. The lymphatics in cancer. Philadelphia, PA:
W.B. Saunders Company, 1972:231–99.
5. Mountain CF, Dresler CM. Regional lymph node classifica-
tion for lung cancer staging. Chest 1997;111:1718–23.
6. Haagensen CD. Lymphatics of the breasts. In: Haagensen
CD, editor. The lymphatics in cancer. Philadelphia, PA: W.B
Saunders Company, 1972:300–98.
7. Aronberg DJ, Peterson RR, Glazer HS, Sagel SS. Superior
diaphragmatic lymph nodes: CT assessment. J Comput
Assist Tomogr 1986;10:937–41.
928
T Suwatanapongched and D S Gierada
(b)
8. Glazer HS, Molina PL, Siegel MJ, Sagel SS. High-attenuation
mediastinal masses on unenhanced CT. AJR Am J
Roentgenol 1991;156:45–50.
9. Sagel SS, Slone RM. Lung. In: Lee JKT, Sagel SS, Stanley RJ,
Heiken JP, editors. Computed body tomography with MRI
correlation, 3rd edn. Philadelphia, PA: Lippincott-Raven
Publishers, 1998:351–454.
10. Nohl-Oser HC. Surgery of the lung, general management
and operative technique. In: Nohl-Oser HC, editor. Surgery
of the lung. New York, NY: Thieme-Stratton Inc., 1981:38–
184.
11. Shields TW. Mediastinal lymph nodes. In: Shields TW,
editor. Mediastinal surgery. Philadelphia, PA: Lea &
Febiger, 1991:14–8.
12. Bankoff MS, McEniff NJ, Bhadelia RA, Garcia-Moliner M,
Daly BDT. Prevalence of pathologically proven intrapul-
monary lymph nodes and their appearance on CT. AJR Am
J Roentgenol 1996;167:629–30.
13. Shimoyama K, Murata K, Takahashi M, Morita R.
Pulmonary hilar lymph node metastases from lung
cancer: evaluation based on morphology at thin-
section, incremental, dynamic CT. Radiology 1997;203:
187–95.
The British Journal of Radiology, November 2006
The British Journal of Radiology, 79 (2006), 929–930
Book reviews
Clark’s Positioning in Radiography, 12th edn. By
A Stewart Whitley, Charles Sloane, Graham Hoadley,
Adrian D Moore and Chrissie W Alsop. pp. 544, 2005
(Hodder Arnold, London, UK) £75.00
ISBN 0-340-76390-6
By any standards, Kathleen Clara Clark (1896–1968) was
a remarkable woman. She entered radiography as the
profession was developing. The Society of
Radiographers had been set up in 1920 and letters had
been written from the new society to the assistants in the
various X-ray departments inviting applications for
membership. Those who had been in active practice for
over 10 years were given membership without examina-
tion; all other applicants had to take a new examination.
The first group comprising 45 students was entered for
examination in January 1922, 20 of whom passed and
were duly awarded the certificate of the new Society (the
MSR). Miss ‘Katie’ Clark completed her training course
at Guy’s Hospital in 1921 and passed the first qualifying
examination ever held by the Society of Radiographers.
She then worked (1922–1927) at the Princess Mary’s
Hospital (for Surgical Tuberculosis) and Margate
(General) Hospital before moving to the Royal
Northern Hospital in London (1927–1935).
Kathleen Clark was aware of the lack of adequate
training for radiographers, and so founded a School of
Radiography at the Royal Northern Hospital, which
became a model for schools elsewhere. She was also
tutor to the X-ray Training School at the Royal Northern
from 1930-1935.
In 1935 she became co-founder and Radiographer-in-
charge of the Ilford Radiographic Technical and
Demonstration Department based at Tavistock House,
where she was involved in instruction and research into
radiography and medical photography. Under her
guidance, the department developed a worldwide
reputation.
Kathleen Clark was President of the Society of
Radiographers from 1935 to 1937 (and the first woman
President). She was also the first President to wear the
President’s chain of office, having been presented with it
by her predecessor Dr Leo Rowden.
The first edition of her classic book Positioning in
Radiography (Heinemann Medical Books) was published
in 1939. The book became the standard work of reference
for radiographers and has been through many editions,
most of which are held in the Information Centre at the
British Institute of Radiology. Generations of radio-
graphers will have used the textbook. Her slide collec-
tion has been preserved and is also available in the
Information Centre at the British Institute of Radiology
as the K C Clark Slide Library. Although primarily of
historical interest, the slide collection remains a useful
teaching resource and may be copied.
The original Positioning in Radiography is a very
remarkable book for several reasons. First, Katie Clark
was keen to standardize both positioning and exposure,
and in her book she standardized the radiographic
The British Journal of Radiology, November 2006
projections so that similar projections could be made in
all hospitals. Second, the book is very artistic. The
illustrations do not come across as cold and entirely
objective scientific images. It is therefore not surprising
that the artist Francis Bacon acknowledged Positioning in
Radiography as a crucial source of inspiration and that it
was his favorite medical textbook. Lawrence Gowling
indicated that Bacon repeatedly borrowed from the
photographs in the book for his work. The images of
the body that Francis Bacon made have an almost
radiographic quality and there is the impression that
multiple layers of the body are seen at the same time and
that one is not just looking at the skin’s surface. This was
well illustrated by Professor Robert Clark from the
University of South Florida in his inaugural lecture ‘‘The
Art of Radiography: how Positioning in Radiography by
radiographer Kathleen Clara Clark influenced the art
of Francis Bacon’’, for the British Society for the History
of Radiology held recently at the British Institute of
Radiology.
Kathleen Clark was awarded the MBE in 1945 for her
services to radiography, particularly for mass miniature
radiography of the chest. Her book Mass Miniature
Radiography of Civilians (MRC Special Report, series
No. 251) was written jointly with P D’Arcy Hart, Peter
Kerley and Brian Thompson, appeared in 1945 and was
the definitive guide to mass radiography of the chest. She
was committed to fostering co-operation and contact
between radiographers throughout the world, and was a
driving spirit behind the formation of the International
Society of Radiographers and Radiological Technologists
(ISRRT). She remained as principal at Ilford until 1958
and acted as Consultant in Radiography until 1964. The
story of her life remains an inspiration to radiographers
everywhere.
In front of me is the first edition of Positioning in
Radiography (1939) and the new Clark’s Positioning in
Radiography’ (2005). The new book certainly feels very
different from the first edition; however, comparisons
are always difficult. It is rather like comparing an ‘‘Old
Master’’ such as the Flemish baroque painter Peter Paul
Rubens (1577–1640), who was probably the most
renowned northern European artist of his day, to
modern artists such as the American pop artist Andy
Warhol (1928–1987) or to the contemporary British
conceptual artist Tracey Emin (born 1963). A comparison
can be made, but is any comparison really meaningful?
Certainly, the new work by Stewart Whitley and his co-
authors feels less artistic that the first edition. However,
the world has changed and a different approach is
needed.
The new book is an excellent account of current plain
radiographic techniques in a single volume. Techniques
that are no longer used or that are associated with a high
radiation dose to the patient have been removed from
the text and a modern alternative is suggested. Many of
the radiographic positions in the older editions are no
longer used. There are excellent sections on mammo-
graphy and dental radiography, and there is a new
929

section on forensic radiography. The trauma section is
excellent and there is a reference to Advanced Trauma
Life Support (ATLS).
The introduction of digital radiography has pro-
foundly affected our departments and this is acknowl-
edged in the book. The expression ‘‘film’’ has been
replaced by ‘‘cassette’’ or ‘‘‘image receptor’’. I would
have liked to see a little more detail on post-processing
and image handling, and marking in a digital environ-
ment. The radiographic images in the modern text have
less contrast than the images in the first edition;
however, the earlier images were obtained using old
high silver film. The older images look very different
from our modern digital images.
I am pleased that the book also reflects the extended
role of the technologist/radiographer and there is a
useful ‘‘Radiological Considerations’’ sub-heading to all
of the chapters, giving the clinical context and an
understanding of image interpretation requirements.
The anatomical diagrams are good and the labelling is
clear.
I showed the book to the radiographers in my
department and the book was universally greeted with
enthusiasm. Everyone thought that the department
needed a new book on radiographic technique to reflect
modern practice. It is easy for a department to feel that as
they already have a radiographic positioning book, they
do not need a new one since positioning does not change.
This is not really the case as techniques develop and are
refined as time passes. Whilst we should keep our old
volumes of Positioning in Radiography (and many with
missing pages), I would recommend that all departments
buy a copy of Clark’s Positioning in Radiography – my
hospital Trust has ordered four copies for our various
departments.
A THOMAS
MRI of the Musculoskeletal System, 5th edn. By
TH Berquist. pp. 1008, 2005 (Lippincott Williams &
Wilkins, Philadelphia, PA) $199.00
ISBN 0-7817-5502-6
This is an updated edition of a well-established text of
MRI of the musculoskeletal system. This is essentially an
in-house text with sixteen of the eighteen contributors
coming from the Mayo clinic. The text and most of the
images have been updated from the fourth edition,
which was published in 2001.
The book is an excellent authoritative text on MR of the
appendicular system. It has a traditional textbook
format, with the chapters being divided mainly along
anatomical lines. The book is quite comprehensive,
covering most conditions that one is likely to see in a
musculoskeletal practice. Loose ends are dealt with in
chapters on musculoskeletal neoplasms, infections, mar-
row disease and a final chapter entitled ‘‘Miscellaneous
conditions’’. The book has many strengths, not least the
comprehensive and readable initial chapters on the basic
principles of MRI. Chapters are introduced with good
descriptions of the basic anatomy and are illustrated
with good quality images.
930
Book reviews
If you are looking for an excellent reference book on
MR of the appendicular musculoskeletal system, you
will not be disappointed. However, do not expect
anything more than a tome dedicated to MRI: what
you see on the title is what you get. Only lip service is
paid to the varying merits of MRI with other imaging
techniques or the clinical application of MRI in everyday
practice. The chapter on the spine is written by
neuroradiologists and there are large gaps related to
musculoskeletal aspects of spinal imaging. There is no
mention of some conditions that are frequently encoun-
tered in a musculoskeletal practice, such as ankylosing
spondylitis and pars defects. Primary bone tumours of
the spine such as osteoblastoma, giant cell tumour or
chordoma do not feature. There is a disproportionate
emphasis on intradural disease, reflecting the interest of
the authors.
I judge this to be a first class reference book for MRI of
the musculoskeletal system, although the spine chapter
would benefit from some input by a musculoskeletal
radiologist to fill some significant gaps. The book
represents excellent value, being priced at just under
$200.00.
S OSTLERE
NeuroPET: Positron emission tomography in neuroscience
and clinical neurology. By Karl Herholz, Peter
Herscovitch and Wolf-Dieter Heiss. pp. xv+297, 2004
(Springer-Verlag, Heidelberg, Germany)
ISBN 3-540-00691-5
This is a very well prepared and concise book at just over
250 pages. Very heavily referenced, it provides the
reader with an excellent and well-illustrated introduc-
tion to the concepts associated with the use of positron
emission tomography (PET) in the human brain. It has a
number of sections; in the first part of the book, these are
related to individual disease states, with chapters on
dementia, movement disorders, brain tumours, etc. The
next section is related to the imaging of brain function,
with specific chapters on blood flow, glucose consump-
tion, oxygen consumption, amino acid transport, etc. The
third and final portion of the book is technical, dealing
with the data acquisition, reconstruction, modelling and
statistics associated with the collection of PET data.
A colleague once told me that PET is the technology of
the future and always will be. At least with this book you
can have some understanding as to the basic concepts,
applications and potential values of this increasingly
important clinical methodology. The book is probably
too detailed for the clinicians starting to do a little bit of
PET reporting, but I think those are few and far between
with the majority of people involved in this technology
wanting to understand it at a significant level, which this
book is ideal for. The authors are to be congratulated on
putting together a very concise and well presented book,
which I would recommend to anyone involved in PET
imaging.
A JACKSON
The British Journal of Radiology, November 2006
The British Journal of Radiology, 79 (2006), 931–932

CASE OF THE MONTH

Various holes and lesions
1 2
S PUNEKAR, MBBS, MRCP, D J A BUTTERISS, BSc, MBBS (Hons), FRCR and 2D BIRCHALL, MB BChir, FRCR

1
Freeman Hospital, Newcastle upon Tyne and 2Newcastle General Hospital, Newcastle upon Tyne,
UK
Received 21 July 2005
Accepted 11 October 2005

DOI: 10.1259/bjr/62312165

’ 2006 The British Institute of

Radiology

A 14-year girl presented with progressive painless visual
loss in her right eye. She was otherwise well and on
examination abnormal findings were localized to the right
globe. Further questioning revealed that her mother had
undergone surgery for a brain tumour several years
previously. MRI of the brain and orbits pre- and post-
administration of gadolinium was performed (Figure 1a–d).
What are the MRI findings? What is the most likely
diagnosis?
As a result of the cranial MRI findings, imaging of the
abdomen was also performed (Figure 2).
What abnormality is shown? Does this help to confirm
or disprove the diagnosis? What other abnormalities
would you wish to exclude in this case?
Address correspondence to: Dr Samad Punekar, Radiology,
Freeman Hospital, Newcastle Upon Tyne, High Heaton NE7
7DN, UK. E-mail: samad@doctors.org.uk.

(a) (b) (c)

Figure 1. (a) Cranial axial proton density-weighted image. (b) Axial post-
gadolinium T1 weighted image. (c) Axial T2 weighted image. (d) Coronal post-
(d) gadolinium T1 weighted image.

The British Journal of Radiology, November 2006 931


Figure 2. Abdominal axial T2 weighted image.
Discussion
Figure 1 shows the presence of two synchronous
central nervous system (CNS) lesions.
First, there is a proliferative process affecting the right
retina, which fills most of the vitreous (Figure 1a). This
demonstrates enhancement following gadolinium
administration (Figure 1b). Second, there is a cystic
abnormality in the right cerebellar hemisphere
(Figure 1c), with an associated enhancing mural nodule
laterally (Figures 1b,d). MRI of the abdomen showed
several pancreatic cysts (Figure 2). These appearances
are in keeping with the presence of retinal and cerebellar
haemangioblastomas, features that in combination with a
positive family history allow the diagnosis of von
Hippel-Lindau disease (vHL). Pancreatic cysts are found
in approximately 75% of patients with vHL.
vHL is an inherited multisystem disorder character-
ized by a variety of highly vascularized tumours of the
CNS and the viscera, and is an eponym that carries the
names of two eminent European physicians [1, 2]. Eugen
von Hippel, a late 19th century German ophthalmologist,
described familial retinal capillary angiomatosis in 1895,
a lesion that has subsequently been coined ‘‘the von
Hippel tumour’’ [3]. In 1926, Arvid Lindau, a Swedish
neurologist, reported the occurrence of hereditary
cerebellar haemangioblastoma (later termed ‘‘the
Lindau tumour’’) and, having noted an association with
retinal angiomatosis and renal cell carcinoma, linked the
cerebellar, retinal and visceral components into a single
coherent entity [4]. He corresponded with von Hippel,
and their joint observations were subsequently widely
disseminated, largely through the work of Harvey
Cushing. The term ‘‘von Hippel-Lindau disease’’ was
first used in 1936 by Davison et al [5].
The most characteristic lesion of vHL is the CNS
haemangioblastoma, and these are the cause of the
majority of morbidity and mortality associated with the
condition. They are most common in the cerebellum, but
may be found anywhere in the neuraxis from the
cerebral hemispheres and optic nerves to the spinal
cord. The typical appearance is of a cystic lesion with an
avidly enhancing mural nodule, but there is a spectrum
of appearances to a completely solid lesion. The cyst
932
S Punekar, D J A Butteriss and D Birchall
fluid is usually slightly hyperdense compared with
cerebrospinal fluid (CSF). Simple cysts of multiple
intra-abdominal organs including liver, pancreas, kidney
and adrenal are common, with renal cell carcinoma
(RCC) and phaeochromocytoma the most common non-
CNS neoplastic abnormalities. vHL-associated RCCs act
differently from sporadic RCCs and may be multiple,
bilateral and cystic. Phaeochromocytomas in vHL occur
in certain lineages only, and this has been used to classify
the condition:
Diagnostic criteria are summarized below:
1. CNS and retinal haemangioblastoma.
2. Haemangioblastoma and at least one of the following:
renal, pancreatic, hepatic, epididymal cyst.
Phaeochromocytoma, renal carcinoma.
3. Family history and at least one of the following:
haemangioblastoma, visceral changes, phaeochromo-
cytoma, renal cancer.
This disorder has an incidence of approximately 1 in
36 000 live births. It is a highly penetrant autosomal
dominant trait, and has been attributed to a germline
mutation of a tumour suppressor gene on the short arm
of chromosome 3. Active research continues into the
molecular consequences of the vHL gene.
Screening of affected patients and at-risk relatives is
necessary because of the high incidence of RCC
(occurring in up to 45% of cases) and because of the
morbidity associated with CNS lesions. The Cambridge
protocol was devised by Maher et al for screening
patients with vHL disease or at-risk relatives [6].
Affected asymptomatic patients and at-risk relatives
should have annual physical examination, ophthalmo-
scopy and urine testing, brain imaging every 3 years, and
abdominal CT scanning every 3 years.
Treatment involves resection of the offending tumour,
aspiration of the cysts causing pressure-related symp-
toms and cryotherapy or phototherapy of retinal lesions.
Radiology has a central role in managing vHL. Because a
conservative approach to the treatment of some vHL
lesions is now more widely accepted, ongoing follow-up
with careful cross-sectional imaging plays a central role
in evaluating the progression of disease.
References
1. Richard S, Graff J, Lindau J, Resche F. Von Hippel-Lindau
disease. Lancet 2004;363:1231–4.
2. Lonser RR, Glenn G, Walther M, et al. Von Hippel-Lindau
disease. Lancet 2003;361:2059–67.
3. Von Hippel E. Vorstellung eines Patienten mit einer sehr
ungewohnlichen Netzhaut. XXIV Verstellung der ophthal-
molgischen Gesellschaft (Heidelberg, 1895). Wisebaden,
Germany: Bergmann Verlag, 1896:269.
4. Lindau A. Studien eber Kleinhirnzysten. Bau, Pathogenese
und Beziehungen zur Angiomatosis retinae. Acta Pathol
Microbiol Scand 1926;S1:1–128.
5. Davison C, Brock S, Dyke CG. Retinal and central
nervous hemangioblastomatosis with visceral changes (von
Hippel-Lindau’s disease). Bull Neurol Instit NY 1936;5:
72–93.
6. Maher ER, et al. Clinical features and natural history of von
Hippel-Lindau disease. Quart J Med 1990;77:1151–63.
The British Journal of Radiology, November 2006
The British Journal of Radiology, 79 (2006), 933
Book reviews
Basics of PET imaging: Physics, Chemistry and Regulations.
By Gopal B Saha. pp. xv+206 (Springer, NY) £35.00
ISBN 0-387-21307-4
Gopal Saha is a World expert in radiopharmacy. His text
book on Fundamentals of Nuclear Pharmacy is now in its
fifth edition. Here, he has produced a very readable text
covering all aspects of the basics of PET imaging.
The book is divided into 11 chapters covering radio-
active decay and interaction of radiation with matter;
PET scanning systems; data acquisition and corrections;
image reconstruction, storage and display; performance
characteristics of PET scanners; cyclotron and production
of PET radionuclides; synthesis of PET radiopharmaceu-
ticals; regulations governing PET radiopharmaceuticals;
reimbursement for PET procedures; design and cost of
PET centre; and procedures for PET studies;. There are
several appendices, including a useful glossary of terms
and answers to the numerical and yes/no questions at
the end of each chapter. There is a also good index.
The lay-out of the book is logical and, for anyone with
some relevant background knowledge either in radiation
physics or nuclear medicine, the text is easy to follow. Some
features I particularly liked were a good, brief description
of factors affecting acquired data –normalisation, photon
attenuation, random coincidences, scatter coincidences,
dead time and radial elongation; a good discussion of the
performance characteristics of PET scanners – spatial
resolution, sensitivity, noise equivalent count rate, scatter
fraction and contrast. There is a very clear description of the
various factors that contribute to limiting spatial resolution,
including positron range and non-collinearity of the
coincidence gammas; also, I enjoyed the useful list, with
brief details, of the physicochemical and biological tests
that are necessary as part of a quality assurance pro-
gramme for PET pharmaceuticals.
The text is generally accurate and has been carefully
proofread, although one or two numerical errors seem to
have crept in, especially in Table 1.3 – half value layers for
511 keV photons. Also, the quoted effective dose (assumed
to be in mSv MBq21) for an 18F-FDG examination cannot be
correct if the stated organ doses are right (p. 193).
UK readers may have a few problems with this book,
written for the US market, including use of units – although
the SI equivalent of the units still currently used in the US
(curies, rads etc.) is frequently given, a reader who cannot
readily switch from one system to the other may have
difficulty. Also, radiation protection is a major problem.
In a recent review of PET: molecular imaging and its
biological applications, edited by ME Phelps; see Br J Radiol
78:2005;871, Lesley Malone commented that radiation safety
had been neglected. It is covered at length by Saha in the
longest chapter in the book. However, organisational
details, with the involvement of both Federal and State
Regulatory Bodies, are completely different in the US from
those in the UK. Also, even the regulations themselves show
variations, e.g. some differences in dose limits and
descriptors of designated areas. Since the section on
principles of radiation protection is not specific to PET and
The British Journal of Radiology, November 2006
the same information can be found in a number of other
texts, there is a case for skipping this chapter completely.
The subject of patient doses appears to receive rather less
attention in the US than in the UK and Europe. By
combining the information in Appendix D with that in
Chapter 11, the effective dose for a few clinical PET
procedures can be worked out. However, there is no
attempt to put these into context with other nuclear
medicine and radiological examinations.
In conclusion, who should read this book? The
preface suggests that it is suitable for nuclear medicine
technologists, radiographers, radiation therapy technolo-
gists and physicians studying PET courses. This is unlikely
to be the situation in the UK because PET technology has
not diffused so rapidly and most persons in the above-
mentioned groups would find that there were just too
many unfamiliar concepts.
However, the book can be recommended to experienced
nuclear medicine staff and more junior medical physics
staff who wish to get involved in PET or simply to find out
more about it.
P DENDY
Imaging of the head and neck, 2nd edn. By Mahmood
F Mafee, Galdino E Valvassori and Minerva Becker.
pp. 866, 2004 (Thieme Medical Publishers Stuttgart,
Germany) J229.95
ISBN 3 13 100942 X
This is the second edition of this superb reference text which
is, against my expectations, improved over the original first
edition. The book is quite capable of causing significant
damage if you try and read it on your knee, as I did, with
over 800 pages of heavily and well-illustrated clinical
material, presented in excellent and logical order. The
information in the book is presented in six anatomically
defined sections, starting with the temporal bone, the eye,
orbit and base of skull, nasal cavity and paranasal sinuses,
the masticatory system, the suprahyoid neck, and the
infrahyoid neck. In each of these there is excellent illustrated,
descriptive anatomy of both the pure anatomy and the
radiological important components. No time or effort has
been spared to ensure that the reader is provided with any
audio visual aid in terms of diagrams, 3D renderings,
histological sections or images that could possibly help their
comprehension of the underlying anatomical components.
The anatomical descriptions are then followed by detailed
discussion of the relevant pathologies affecting the area,
lavishly illustrated with an absolutely excellent series of
cases, the majority of which are collected on recent
generation imaging equipment and are of very high quality.
The discussion and presentation is of more than adequate
detail for the specialist radiologist interested in these areas,
whilst at the same time, on the whole, being easily so well
presented that it could be perused by the non-specialist
radiologist or the radiologist in training without them
feeling particularly threatened. Overall, I have to admit that
I am fairly jealous that I did not write this book and it
certainly has a strong place on my desk top from now on.
A JACKSON
933
BJR
The British Journal
of Radiology

September
2006
Volume 79
Special
Issue
The British Journal of Radiology, 79 (2006), S1
PREFACE
Imaging in radiotherapy treatment planning and delivery
R M HARRISON
Deputy Editor (Physics & Technology)
Recent years have seen major developments in radio-
therapy techniques and equipment, focusing on con-
formal external beam techniques with the emphasis on
intensity-modulated radiotherapy (IMRT). The progress
of IMRT implementation has been summarized in a
series of review articles in the British Journal of Radiology
during 2003 and 2004. One of the fundamental pre-
requisites for conformal radiotherapy is the localization
of the target, starting with the gross tumour volume
(GTV) and moving outwards to the planning target
volume (PTV). The judgement of how to define these
volumes is crucial. The future value of IMRT and related
techniques may depend on the underpinning images of
these volumes and how they are interpreted.
Fortunately, in parallel with the developments in radio-
therapy delivery techniques has come an equally
impressive and timely development in associated
imaging.
Although CT remains the gold standard for radio-
therapy planning, MRI has much to offer. Superior soft
tissue characterization and new developments in
dynamic contrast enhanced, diffusion-weighted and
diffusion tensor imaging are making their mark in target
delineation. Magnetic resonance spectroscopy (MRS)
promises insights into tissue metabolism and, although
currently limited by large voxel sizes, may provide
information on tumour response and recurrence to
complement the structural information of MRI and CT.
The traditional role of the simulator is now comple-
mented by the use of the CT scanner as a virtual
simulator, thus making 3D image datasets available.
The British Journal of Radiology, Special Issue 2006
DOI: 10.1259/bjr/14035252
’ 2006 The British Institute of
Radiology
Verification on a fraction-by-fraction basis is now
possible by several means. CT scanners within the
treatment room have been used to compare images at
treatment time with planning CT scans and also to
measure and correct for interfraction variations in
anatomy or positioning. Portal imaging has seen equally
impressive developments as a result of the introduction
of flat panel imaging systems. As well as their improved
image quality for traditional radiographic projection
imaging, when used in conjunction with cone beam
reconstruction algorithms, they have made megavoltage
CT a reality. Alternatively, if used with a kilovoltage
X-ray source and cone beam CT algorithm, the CT
reconstructions can harness the improved tissue contrast
associated with the lower X-ray energies.
The development of PET/CT imaging for radiotherapy
planning provides, like MRS, another example of the
imaging of metabolic function, but in this case the
functional information is displayed within the same
coordinate system as the familiar structure provided by
CT. This is an unusual luxury, however, and the
problems of extracting and overlaying data from
different imaging studies and modalities has prompted
several advances in image registration and fusion
algorithms.
So these are exciting times in radiotherapy imaging
and this special issue reflects both the current practice
and future opportunities for exploiting advances in
several imaging modalities in order to increase the
confidence with which modern conformal radiotherapy
treatments may be planned and delivered.
S1
 The British Journal of Radiology, 79 (2006), S2–S15

New developments in MRI for target volume delineation in

radiotherapy
1,2
V S KHOO, FRACR, FRCR, MD and 3D L JOON, FRACR

1
Royal Marsden Hospital, Institute of Cancer Research, Fulham Road, London SW3 6JJ, 2University
of Manchester, Manchester, UK and 3Austin Health Radiation Oncology Centre, Heidelberg
Repatriation Hospital, Victoria, Australia

ABSTRACT. MRI is being increasingly used in oncology for staging, assessing tumour
response and also for treatment planning in radiotherapy. Both conformal and
intensity-modulated radiotherapy requires improved means of defining target volumes
for treatment planning in order to achieve its intended benefits. MRI can add to the
radiotherapy treatment planning (RTP) process by providing excellent and improved
characterization of soft tissues compared with CT. Together with its multiplanar
capability and increased imaging functionality, these advantages for target volume
delineation outweigh its drawbacks of lacking electron density information and
potential image distortion. Efficient MR distortion assessment and correction
algorithms together with image co-registration and fusion programs can overcome
these limitations and permit its use for RTP. MRI developments using new contrast
media, such as ultrasmall superparamagnetic iron oxide particles for abnormal lymph
node identification, techniques such as dynamic contrast enhanced MRI and diffusion
MRI to better characterize tissue and tumour regions as well as ultrafast volumetric or Received 9 August 2005
cine MR sequences to define temporal patterns of target and organ at risk deformity Revised 6 January 2006
and variations in spatial location have all increased the scope and utility of MRI for RTP. Accepted 10 March 2006
Information from these MR developments may permit treatment individualization,
DOI: 10.1259/bjr/41321492
strategies of dose escalation and image-guided radiotherapy. These developments will
be reviewed to assess their current and potential use for RTP and precision high dose ’ 2006 The British Institute of
radiotherapy. Radiology

The increased sophistication of modern radiotherapy
planning techniques such as conformal (CFRT) and
intensity-modulated radiotherapy (IMRT) necessitates
improved means of defining target volumes for treat-
ment. This is needed to achieve the intended benefits of
using CFRT and IMRT. This step remains the most
crucial and difficult part of the radiotherapy planning
process, otherwise a geographical miss of the tumour or
a systematic error will be perpetuated throughout
therapy. MRI is being increasingly used in oncology for
staging, assessing tumour response and evaluating
disease recurrence. As a result of the enhanced imaging
properties of MR, it has been estimated to be a more cost
effective diagnostic tool in the management of some
diseases [1]. Similarly, the improved characterization of
soft tissues and visualization of tumour extent using MRI
can be used to benefit the radiotherapy treatment
planning (RTP) process from delineation of target
volumes to determining planning margins and treatment
response [2].
There are many current areas of development in MRI.
These include developments in hardware technology,
such as 3 Tesla machines, and the use of new MR
contrast media, such as ultrasmall superparamagnetic
iron oxide particles for lymph node evaluation [3]. MR
techniques and sequences previously used for research
are now becoming available for general use. MR
techniques such as dynamic contrast enhanced and
diffusion weighted MRI may provide further character-
ization of tissue and tumour regions [4, 5]. MR sequences
such as ultrafast volumetric and 3D cine sequences can
offer the opportunity to assess target/organ motion and
deformity [6, 7]. Temporal-spatial information gleaned
from MRI can then be used for image-guided strategies
in radiotherapy delivery. All these features have the
potential to increase the scope and utility of MRI for RTP.
It is worthwhile briefly reviewing the background to
the use of MRI for RTP in order to understand the
rationale and issues with its use. Some methods of
utilizing MRI in RTP will be outlined. This article will
then discuss the new MRI developments in terms of their
current and potential impact in target volume definition
for treatment planning with examples of applications at
some cancer subsites.
MRI rationale for RTP
Any additional procedures used for RTP must add
value to the planning process. Standard RTP uses CT
data. CT images are good at distinguishing between
structures that have substantially different X-ray
attenuation properties or Hounsfield units, such as
between air, tissue and bone. It is more difficult to
discriminate between adjoining soft tissue structures
using CT if these soft tissue structures possess similar

S2 The British Journal of Radiology, Special Issue 2006

MRI for target volume delineation in radiotherapy
Hounsfield units unless there is a fat, air or bone
interface between these structures. The imaging para-
meters for CT scanning are much more limited compared
with the range available with MRI.
In the case of MRI, the contrast from soft tissue
structures can be widely varied by extensively mani-
pulating the imaging parameters, which include proton
densities and tissue relaxation times (spin-lattice or T1
and spin–spin or T2). This increased flexibility in varying
tissue contrast or signal intensities offers much better
characterization of soft tissues even when these struc-
tures possess very similar X-ray attenuation properties or
electron densities. Tumours often have similar electron
densities to their neighbouring soft tissues. By using
different MRI sequences, better tissue discrimination can
be obtained between the extent of tumour with its
boundaries of infiltration and the adjacent normal
structures. In this manner, MRI provides improved
target delineation for RTP. This utility of MRI applies
not only to the initial radiotherapy treatment of tumours
but also potentially for re-treatments by being able to
differentiate between changes due to recurrent cancer or
that secondary to post-treatment fibrosis. It can also
provide better delineation of organs at risk (OARs) for
dose avoidance in RTP.
An obvious benefit of enhancing the visualization of
volumes of interest (VOI) is the increased reliability and
consistency of target definition. This will improve both
interobserver and intraobserver variability for outlining.
This has value for institutional and multicentre trials in
radiotherapy where it is important to maintain consistent
and accurate target and OAR volumes. Substantial and
inappropriate variations in target volumes can impact on
trial outcomes, with geographical misses for poorer local
control rates or unnecessary inclusion of normal tissue
for higher toxicity rates.
MRI can avoid bony and metal artefacts seen with CT.
Large thick bony sections attenuate X-rays and reduce
the adjacent soft tissue image quality. This can obscure
identification of nearby tumours and internal anatomy.
(a)
Figure 1. A comparison of sagittal views of the pelvis for prostate radiotherapy with (a) CT reconstructed from 2.5 mm slices
and (b) MR image obtained in-plane in the same patient. Some of the relevant structures of interest for radiotherapy are
labelled on the MR image. These structures are not visualized well enough on CT to provide confident
MR images are not affected by this. MRI can thus further
improve the delineation of both tumour and OAR
volumes for RTP in these regions.
Another feature of the increased functionality of MRI
is its true multiplanar capability. This ability to image in
any oblique plane can reduce the ‘‘partial volume’’
imaging effect that often results from conventional
transaxial CT imaging, particularly where the 3D shape
of the target is extreme or changes substantially between
conventional transaxial CT slices. Para-sagittal or para-
coronal views can also permit better understanding of
the boundaries of target volumes with the surrounding
normal tissues leading to better target volume deline-
ation (Figures 1 and 2).
Furthermore, MRI can provide functional and biologi-
cal information for tumour regions that may improve
target definition and permit new opportunities for novel
radiotherapy strategies. Some of the salient features of
using MRI for RTP are summarized in Table 1. However,
it is important to be aware that implicit in the use of MRI
for RTP is that oncology clinicians should have the
necessary training to comprehend MR images and
understand how to use them appropriately for defining
VOIs. Thus it is important to undertake suitable
supervised training [8]. Even if there is relevant
experience, it is still beneficial to liaise closely with local
diagnostic colleagues who have MR expertise for the
cancer subsite(s) in question. Ideally there should be an
oncology team collaboration for the definition of target
volumes in radiotherapy similar to the multidisciplinary
team arrangement that exists for general cancer manage-
ment. This approach has been endorsed by national
bodies, such as the Royal College of Radiologists in a
recent publication [9].
All of these advantages add to the RTP process and
outweigh its current drawbacks, which include the lack
of electron density information, potential image distor-
tion and specific patient considerations with MR
scanners. Some examples of the current impact of using
MRI for target volume delineation and some of these
(b)