See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/12690614

A comparison of image reject rates when using film, hard copy computed
radiography and soft copy images on picture archiving and communication
systems (PACS) workstations

Article  in  British Journal of Radiology · August 1999

DOI: 10.1259/bjr.72.859.10624322 · Source: PubMed

CITATIONS READS

40 421

3 authors, including:

Stirling Bryan
University of British Columbia - Vancouver
283 PUBLICATIONS   9,872 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Clinical decision making and patient preferences View project

The Use of Economic Evaluation in Health Care Decision Making View project

All content following this page was uploaded by Stirling Bryan on 30 May 2014.

The user has requested enhancement of the downloaded file.

The British Journal of Radiology, 72 (1999), 653±660 E 1999 The British Institute of Radiology

A comparison of image reject rates when using ®lm,

hard copy computed radiography and soft copy images
on picture archiving and communication systems
(PACS) workstations
1
G C WEATHERBURN, MSc, TDCR(R), 2S BRYAN, BSc, MSc and 3M WEST, MSc
1
Health Economics Research Group, Brunel University, Uxbridge, Middlesex UB8 3PH, 2Health Services
Management Centre, University of Birmingham, 40 Edgbaston Park Road, Birmingham B15 2RT, and
3
Radiological Sciences Unit, Hammersmith Hospital, Du Cane Road, London W12 0HS, UK

Abstract. A comparison has been made of the reject rates of plain images for three separate
periods when ®lm, computed radiography (CR) and PACS systems were in operation throughout
the Hammersmith Hospital, London. There was a statistically signi®cant reduction in the overall
percentage reject rate across all examinations from 9.9% to 8.1% when the hospital changed from
using a conventional ®lm based system to a CR system. There was a further reduction in the
reject rate to 7.3% when the hospital moved to a hospital-wide PACS system, but this change was
not statistically signi®cant. Using estimations of the total number of images used, the percentage
reject rates were 6.6% for ®lm, 5.5% for CR and 5.5% for PACS. Thus, if the radiation dose for
each image is unchanged, and the same types of images are used for the examination of each body
area, a move from conventional ®lm imaging to phosphor plate imaging provides the potential to
reduce the patient population dose.

``Reject analysis'' is described as ``the critical
evaluation of radiographs which are used as part
of the imaging service but do not play a useful
part in the diagnostic process'' [1]. The analysis of
rejected ®lms gives an indication of the sources of
radiographic errors and highlights areas where
improvements can be made. Reject rates vary
between 2% and 13% of all ®lms taken [2±9]. In
1990, the average reject rate in the UK was 10%
of all ®lms [10]. This implies that, nationally,
rejected ®lms are responsible for an unnecessary
increase in the radiation dose to the patient
population [11] and for additional staff, ®lm and
processing costs.
The introduction of computed radiography
(CR) systems is expected to reduce reject rates
because the phosphor plates used for acquiring
the images have a wider latitude than conven-
tional ®lm and repeats resulting from incorrect
exposure factors are therefore largely eliminated
Received 26 November 1998 and in revised form 17
March 1999, accepted 24 March 1999.
This study was one component of the evaluation of a
hospital-wide picture archiving and communication
system which was undertaken by the Health
Economics Research Group and funded by a grant
from the Policy Research Programme of the
Department of Health.
[12]. For images of body areas where there are
large differences in density and thickness within
the patient, or there is unexpected pathology
which causes the area to be either particularly
radiolucent or dense, manipulation of soft copy
images should allow most areas of the images to
be visualized [13]. Thus, when CR systems are
incorporated into a picture archiving and com-
munication system (PACS), which includes
manipulation facilities such as grey scale and
density variation, the reject rates should be lower
than those for both ®lm and CR. A General
Electric (GE) hospital-wide PACS is now in
operation at the Hammersmith Hospital,
London and has been the subject of an indepen-
dent evaluation by the Health Economics
Research Group (HERG) at Brunel University
[14]. As part of this evaluation it was possible to
compare the results of reject analyses when each
of the three types of radiographic images were
used: ®rst with Film, second with CR, and third
with PACS soft copy images.
Two hypotheses were tested in this study:
(1) the reject rate of images would be reduced
after the introduction of phosphor plate
computed radiography; and
(2) the reject rate would be reduced further after
the introduction of PACS workstations with
manipulation facilities for soft copy images.

The British Journal of Radiology, July 1999 653


Methods
Three detailed reject analyses have been under-
taken as a joint project involving the Departments
of Medical Physics and Radiology at
Hammersmith Hospital and HERG. Data were
collected in 1992 when conventional ®lm was still
being used, in 1994 after radiology started using
CR hard copy images, and 1995 after PACS was
in wide use throughout the hospital. The focus of
the study was on plain images which involved
irradiation of the patient.
Details of the numbers of each body area
examined during each period were obtained
from the radiology information system (RIS).
Throughout the period of the study there was no
formal policy in the department for de®ning
images as rejects. If the radiographer was
responsible for producing the image, the decision
to reject was made by that radiographer, based on
professional judgment. In the few instances where
the radiographer was uncertain, the decision was
made by a radiologist who would subsequently
produce the clinical report. Thus, all decisions
were to some degree subjective judgments follow-
ing professional guidelines.
Reject analysis of ®lms
This part of the study was undertaken during
August 1992. All examinations of all patients
being radiographed were included in the study.
Each ®lm was viewed and a decision was made as
to whether it was acceptable for diagnosis or
should be rejected.
Reasons for rejection of ®lms were as follows
(the codes de®ned here are used later in the
paper).
A Incorrect patient position. The patient has not
been positioned to show the whole of the body
area or to demonstrate pathology adequately.
This is less subjective than some other aspects
since there are set protocols which should be
followed. However, some variations in these
protocols occur and tend to be dependent upon
where the radiographer trained and has had
previous experience.
B Unsharpness or blurring of the image. This
results from movement of the patient or
equipment and may be voluntary, such as
breathing, or involuntary, such as heart beat.
This category of rejects is less subjective than
some others but there will be differing degrees
of movement and a decision has to be made on
whether the diagnosis can be made from the
®lm and whether the movement can be
eliminated with a repeat exposure.
654
G C Weatherburn, S Bryan and M West
C Incorrect exposure factors. If incorrect exposure
factors have been used, the ®lm density
(blackening) may be too high or too low to
demonstrate the body area of interest. This
overexposure and underexposure may be
caused by incorrect selection of exposure
factors for the size and density of the patient
or by the presence of unexpected pathology,
such as emphysema, resulting in overexposed
lung ®elds. There is no absolute value de®ning
``correct density''. Decisions about rejections
are subjective with quite considerable varia-
tions in density between ®lms of the same body
area being accepted.
D X-ray equipment faults. Rejects can result from
inaccurate exposures delivered by faulty X-ray
equipment, such as timer errors, automatic
exposure device errors, or a fall in X-ray tube
output.
E Processing equipment faults. These are wet
processing errors such as static build-up
during processing, physical damage resulting
in the ®lm emulsion bearing the image being
removed. This category of rejection is less
subjective than some. In addition a ®lm will be
rejected if light enters the cassette and produces
an area of high density fogging on the ®lm.
There is some degree of subjectivity in this
category; some staff will reject all ®lms with
any areas of fogging, whereas others only reject
the ®lm if the fogging obscures part of the area
of interest.
F Films which are not required. Films will also be
rejected when they are considered to be of no
value in aiding diagnosis. This decision is
normally made by the radiologist. It should
be emphasized that some rejected images
printed on laser ®lms do not re¯ect additional
exposure to the patient and, since they have a
cost implication only, have not been included
in the study.
G Miscellaneous reasons. This category includes
double exposure of the ®lm resulting in two
superimposed images and the presence of
extraneous opacities such as jewellery.
When a decision was made to classify a ®lm as
a ``reject'' the radiographer involved in the
examination attached a label to the ®lm which
indicated details of the exposure factors, patient
size (small, medium, large), identity of the X-ray
room and the reason for rejection. The rejected
®lms with labels attached were then placed in a
The British Journal of Radiology, July 1999
A comparison of image reject rates when using ®lm, hard copy CR and PACS
box in the ®lm viewing area. At the end of the
study all rejected ®lms were viewed in order to
identify the body areas examined and the reason
for rejection of the image. On the few occasions
when a label was missing, one of the authors
(MW), in consultation with a radiographer from
the department, coded the ®lm for the reason for
rejection.
Hard copy CR images
This part of the study was undertaken in
September 1994 at a time when the CR system
had been in use in the hospital for 18 months and
radiology staff were generally satis®ed with the
system. The method used was essentially the same
as for ®lm but the rejected images were sorted,
and on the rare occasions when a code was
missing, coded by another author (GW) in
consultation with departmental radiographers.
In addition to the rejection codes indicated
above, the following CR-speci®c codes were used:
H CR technique/scatter/position of image on
cassette/irradiated area too small. A pattern
recognizer for iris of exposure ®eld (PRIEF) is
an algorithm used by the image plate reader to
detect the irradiated area on the plate. Each
PRIEF has speci®c exposure precautions
relating to irradiated ®eld positioning, size
and shape. There are ®ve PRIEFs, one of
which is pre-selected by the radiographer for
each individual examination. Data obtained
outside the irradiated ®eld, for example,
scattered radiation, can adversely affect the
data analysis and the resultant image. The
positioning of the irradiated ®eld within the
imaging plate area and the limitation of scatter
is therefore very important.
J Incorrect organ code. The radiographer has to
select an organ code for each plate before it is
processed. The plate reader uses the organ code
to identify the image processing parameters for
the plate. Thus, if the incorrect organ code is
selected, the plate may not be processed under
optimum conditions for the body area under
examination and cannot be reprocessed.
K Digiscan fault. The Digiscan refers to the plate
reader in which the phosphor plate is scanned
by a laser beam to convert the latent image
into a digital image which, if required, will be
transferred to a ®lm for hard copy CR
production. Faults in the Digiscan include
artefacts on plates, plates jammed and plates
incorrectly read.
L Hard copy processor faults. When a hard copy
The British Journal of Radiology, July 1999
CR image is required, the ®lm has to be
transported via a roller system to a standard
chemical processor. Faults in this system are
similar to processing faults of ®lm as described
previously in section E.
Soft copy PACS images
This study took place during November and
December 1995, 18 months after the department
began to use PACS and at a time when the staff
had adapted to using the new system. The method
used here was different from the previous two
rounds because of changes which had occurred
with the use of PACS relating to the rejected
images and the examination request forms. After
PACS was fully operational, no hard copy images
were produced routinely. It was not possible to
delete from the system an image which was
identi®ed on the workstation as unsatisfactory,
but it was transferred to the electronic reject
folder by the radiographer. The reject folder
contained a permanent work list which retained
details of all patients and their examinations from
which images had been rejected. It did not
provide details of how many images or which
views were rejected or the reason for their
rejection. However, the reject process included a
compulsory ®eld in which the reason for rejecting
the image had to be given. The images were then
annotated with the reason for rejection for the
purpose of audit, using the codes listed above.
Images could only be viewed in the reject folder
for 8 days and, owing to logistic dif®culties, data
on the reason for rejection were not collected on
36 (7%) rejected images, although the body areas
were known.
Units of measurement
The majority of previous studies of image reject
rates have typically reported the number of
rejected images as a percentage of the total
number of images taken [3±10]. In this study,
data were not available on the total number of
images taken for all three periods. The total
number of examinations, provided by the RIS,
was therefore used as the denominator through-
out. An examination may include more than one
image and so the rejects rates calculated by this
method, and expressed as a percentage of all
examinations, will be higher than if the more
common baseline of rejects as a percentage of all
images were used. However, the comparison of
the reject rates for the three periods when ®lm,
CR and PACS were used, was made consistently.
Given the unconventional, but necessary, use of
655

Table 1. Comparison of reject rates when the calculations are based on the total number of examinations
Modality Total of plain
radiography
examinations
Film 3904
CR 4502
PACS 6617
Film±CR Difference in proportions50.017.
99% con®dence interval for the difference in proportions is 0.00136 to 0.033, p,0.01.
CR±PACS Difference in proportions50.008.
95% con®dence interval for the difference in proportions is 20.00206 to 0.0182.
examinations as the denominator, care should be
taken in attempting to relate these results to local
situations elsewhere. In an attempt to help readers
extrapolate the results to local settings, the PACS
database was interrogated to obtain annual mean
values of the number of images used for each
body area. Since these values and the examination
protocols remained virtually unchanged for the 3
years after this study, the values provided for
1995 were applied to all three phases of the study
to provide a broad estimate of the reject rates
based on the total number of images used.
Results
Rejects expressed as rates across all
examinations undertaken
There was a decrease in reject rates across all
examinations from 9.9% when ®lm was used to
8.1% when hard copy CR was used and 7.3%
when PACS was used (Table 1). The reductions in
the reject rates per examination were statistically
signi®cant for both CR (p,0.01) and PACS
(p,0.01) compared with ®lm but there was no
statistically signi®cant reduction when comparing
PACS with CR.
Rejects expressed as rates across all images
undertaken
The reject rates across all images undertaken
followed a similar trend. There was a statistically
signi®cant decrease in the reject rates for CR hard
Table 2. Comparison of reject rates when the calculations are based on the total number of images used
Modality Total of plain
radiography
images
Film 5791
CR 6597
PACS 9593
Film±CR Difference in proportions50.011.
99% con®dence interval for the difference in proportions is 0.00269 to 0.0196, p,0.01.
CR±PACS Difference in proportions50.001.
95% con®dence interval for the difference in proportions is 20.00596 to 0.00829.
656
G C Weatherburn, S Bryan and M West
No. of rejects % Reject
involving irradiation rate/examination
385 9.9
365 8.1
483 7.3
copy compared with ®lm (6.6% for ®lm and 5.5%
for hard copy CR), but no further reduction when
PACS soft copy images were used (Table 2).
Reject rates by body area
The reject rates by body areas are shown in
Table 3. Examinations of the chest accounted for
the majority of rejects during all periods. This was
expected since this is the body area most
frequently examined. When ®lm was used, the
body area with the highest reject rate over all
examinations was the thoracic spine and this rate
dropped noticeably in the CR and PACS periods.
For examinations of the hip and the skull, the
reject rates in the CR and PACS periods were
higher than in the ®lm period, and there was an
increase in rejects owing to errors in patient
positioning and radiographic technique (Tables 4±
6). For all other body areas the reject rates were
very similar for all three periods.
Reject rates by reason for rejection
The reasons for rejection of images are shown
in Table 7. Positioning and radiographic techni-
que errors were the major reasons for rejects
throughout all periods and contributed to a
greater proportion of all rejects when CR and
PACS were used than when ®lm was used. The
rejects which were expected to be completely
eliminated when ®lm was no longer used were
rejects associated with ®lm processing and fogging
of ®lms, which accounted for 3.4% of all rejects.
No. of rejects % Reject rate/all
involving irradiation images
385 6.6
365 5.5
483 5.5
The British Journal of Radiology, July 1999
A comparison of image reject rates when using ®lm, hard copy CR and PACS

Table 3. Rejects by body area for plain radiography images

Body area Film CR PACS

No. Rejects No. Rejects No. Rejects
of exams (% exams of of exams (% exams of of exams (% exams of
body area) body area) body area)

Chest 2148 161 (7.5) 2413 105 (4.4) 3653 235 (8.9)
Abdomen 231 20 (8.7) 307 26 (8.5) 386 40 (10.4)
Skull 132 34 (25.8) 146 53 (36.3) 203 47 (23.2)
Cervical spine 126 30 (23.8) 134 32 (23.9) 209 41 (19.6)
Thoracic spine 57 29 (50.9) 43 3 (7.0) 79 9 (11.4)
Lumbar spine 161 26 (16.1) 223 22 (9.9) 284 38 (13.4)
Pelvis 122 17 (13.9) 178 23 (12.9) 219 10 (4.6)
Hip 71 8 (11.3) 78 29 (37.2) 137 29 (21.2)
Upper limbs & 403 33 (8.2) 485 45 (9.3) 677 45 (6.7)
shoulder girdle
Lower limbs 453 27 (6.0) 495 27 (5.5) 706 31 (4.4)
excluding hips
Total 3904 385 4502 365 6617 525a
a
This total includes 36 rejects which did not involve irradiation of the patient.

These rejects were eliminated but new CR-speci®c
reasons for rejection of images emerged which
were responsible for 6.3% of all CR rejects and
7.4% of all PACS rejects.
Discussion
The reject rates over all examinations under-
taken for ®lm (9.9%), CR (8.1%) and PACS
(7.3%) can be compared with a recent study from
Nottingham [15] which reports its ®lm reject rates
over all examinations as 8%. The rates calculated
across all images undertaken can be directly
compared with other published results which are
based on reject rates calculated as a percentage of
all images taken. The published reject rates [2±10]
vary between departments and range from 2% to
13%, the rates found in this study (6.6% for ®lm
and 5.5% for CR and PACS) falling within this
range.
The reject rates reported here are from the same
hospital and calculated by the same method and it
is therefore the change in reject rate rather than
the actual value of the reject rate which is of
central importance. This shows that the reject rate
was reduced signi®cantly when the CR system was
introduced in the hospital, but that there was no
further signi®cant reduction when the PACS
system was used. If the radiation dose for each
image is unchanged, and the same types of images
are used for the examination of each body area,
there is therefore the potential to reduce the
patient population dose by moving from ®lm
imaging to phosphor plate imaging [16].
These results must be interpreted with caution
as the process of rejection is subjective and the
changes following the introduction of CR and

Table 4. Reason for rejecting image (number (%) of rejects): ®lm

Body Area A B C D E F G Total rejects

for body area

Chest 64 (39.8) 11 (6.8) 54 (33.5) 3 (1.9) 6 (3.7) 5 (3.1) 18 (11.2) 161

Abdomen 4 (20) 1 (5) 9 (45) 1 (5) 0 (0) 1 (5) 4 (20) 20
Skull 22 (64.7) 3 (8.8) 4 (11.8) 0 (0) 2 (5.9) 0 (0) 3 (8.8) 34
Cervical spine 13 (43.3) 3 (10) 11 (36.7) 2 (6.7) 0 (0) 1 (3.3) 0 (0) 30
Thoracic spine 13 (44.8) 0 (0) 13 (44.8) 2 (6.9) 1 (3.4) 0 (0) 0 (0) 29
Lumbar spine 18 (69.2) 1 (3.8) 4 (15.4) 1 (3.8) 1 (3.8) 1 (3.8) 0 (0) 26
Pelvis 5 (29.4) 0 (0) 6 (35.3) 3 (17.6) 1 (5.9) 2 (11.8) 0 (0) 17
Hip 6 (75) 0 (0) 2 (25) 0 (0) 0 (0) 0 (0) 0 (0) 8
Upper limbs & 13 (39.4) 1 (3.0) 12 (36.4) 1 (3.0) 1 (3.0) 2 (6.1) 4 (12.1) 33
shoulder girdle
Lower limbs 13 (48.1) 0 (0) 10 (37.0) 1 (3.7) 1 (3.7) 0 (0) 3 (11.1) 27
excluding hips
Total5385

A, patient position, radiographic technique; B, patient movement; C, incorrect exposure; D, X-ray equipment fault; E, processing,
fogging; F, not required; G, miscellaneous.

The British Journal of Radiology, July 1999 657

 G C Weatherburn, S Bryan and M West

Table 5. Reason for rejecting image (number (%) of rejects for body area): CR

Body Area A B C D H J K L F G Total rejects

for body area

Chest 88 (83.8) 0 (0) 2 (1.9) 0 (0) 2 (1.9) 2 (1.9) 4 (3.8) 2 (1.9) 0 (0) 5 (4.8) 105
Abdomen 21 (80.8) 1 (3.8) 0 (0) 0 (0) 3 (11.5) 0 (0) 0 (0) 0 (0) 0 (0) 1 (3.8) 26
Skull 48 (90.1) 0 (0) 0 (0) 0 (0) 0 (0) 1 (1.9) 0 (0) 0 (0) 0 (0) 4 (7.5) 53
Cervical spine 23 (71.9) 0 (0) 1 (3.1) 0 (0) 1 (3.1) 0 (0) 0 (0) 0 (0) 0 (0) 7 (21.9) 32
Thoracic spine 2 (66.7) 0 (0) 1 (33.3) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 3
Lumbar spine 17 (77.3) 0 (0) 2 (9.0) 0 (0) 0 (0) 1 (4.5) 0 (0) 0 (0) 0 (0) 2 (9.0) 22
Pelvis 19 (82.6) 0 (0) 0 (0) 0 (0) 1 (4.3) 0 (0) 0 (0) 0 (0) 0 (0) 3 (13.0) 23
Hip 27 (93.1) 0 (0) 0 (0) 0 (0) 1 (3.5) 0 (0) 0 (0) 0 (0) 0 (0) 1 (3.5) 29
Upper limbs & 34 (75.5) 2 (4.4) 0 (0) 0 (0) 4 (8.9 1 (2.2) 0 (0) 0 (0) 0 (0) 4 (8.9) 45
shoulder girdle
Lower limbs 25 (92.6) 0 (0) 0 (0) 0 (0) 1 (3.7) 1 (3.7) 0 (0) 0 (0) 0 (0) 0 (0) 27
excluding hips
Total5365

A, patient position, radiographic technique; B, patient movement; C, incorrect exposure; D, X-ray equipment fault; H, CR technique;
J, incorrect organ code; K, Digiscan fault; L, hard copy processor fault; F, not required; G, miscellaneous.
Code E not applicable.

PACS could be a result of changes in the
threshold of acceptance of images by the current
staff or indeed by changes in staff over the period
of the study.
As was anticipated, the introduction of CR
reduced the high percentage of thoracic spine
rejects which were caused by incorrect exposure
factors. It was expected that the use of PACS
workstations might reduce some of the rejects of
cervical spine images resulting from errors in
positioning and radiographic technique. This was
because many of the rejects were caused by
inadequate demonstration of the cervicothoracic
junction in the lateral view and it was anticipated
that manipulation of the soft copy image would
allow this area to be visualized through the mass
of the shoulders. Overall, the reject rates for the
cervical spine remained a reasonably constant
proportion of all rejects throughout all three
studies (®lm 23.8%, CR 23.9% and PACS 19.6%).
However, whereas the reject rates caused by
incorrect exposure factors decreased, the reject
rates resulting from incorrect positioning and
radiographic technique increased. The expected
overall reduction in cervical spine rejects was
therefore not achieved.
It was expected that when PACS was fully
operational, the rejects caused by wet processing

Table 6. Reason for rejecting image (number (%) of rejects for body area): PACS

Body area A B C D H K Fa G Rejects Total

with rejects
unknown for body
reasonb area

Chest 178 (75.7) 4 (1.7) 1 (0.4) 5 (2.1) 7 (3.0) 2 (0.8) 12 (5.6) 5 (2.1) 21 235
Abdomen 29 (72.5) 1 (2.5) 1 (2.5) 4 (10.0) 0 (0) 1 (2.5) 4 (10.0) 0 (0) 0 40
Skull 41 (87.2) 1 (2.1) 1 (2.1) 0 (0) 0 (0) 0 (0) 0 (0) 2 (4.2) 2 47
Cervical spine 28 (68.3) 0 (0) 1 (2.4) 0 (0) 4 (9.8) 0 (0) 1 (2.4) 1 (2.4) 5 40
Thoracic spine 6 (66.7) 0 (0) 1 (11.1) 0 (0) 0 (0) 0 (0) 2 (22.2) 0 (0) 0 9
Lumbar spine 25 (65.8) 0 (0) 0 (0) 0 (0) 3 (7.9) 1 (2.6) 8 (21.1) 0 (0) 1 38
Pelvis 7 (80.0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 1 (12.5) 0 (0) 2 10
Hip 20 (69.0) 0 (0) 3 (10.3) 1 (3.5) 4 (13.8) 0 (0) 0 (0) 0 (0) 2 30
Upper limbs & 22 (48.9) 0 (0) 10 (23.2) 0 (0) 8 (17.8) 1 (2.2) 2 (4.4) 0 (0) 2 45
shoulder girdle
Lower limbs 23 (76.7) 0 (0) 0 (0) 1 (3.2) 1 (3.2) 1 (3.2) 1 (3.2) 0 (0) 4 31
excluding hips
Total5525
a
These rejects do not involve irradiation of the patient.
b
Images could not be fetched and viewed, so reason for rejection unknown.
A, patient position, radiographic technique; B, patient movement; C, incorrect exposure; D, X-ray equipment fault; H, CR technique;
K, Digiscan fault; F, not required; G, miscellaneous.
Codes E, J and L not applicable.

658 The British Journal of Radiology, July 1999

A comparison of image reject rates when using ®lm, hard copy CR and PACS
Table 7. Reasons for rejection of plain radiography images
Reason for rejection No. of rejects (% rejects)
Film
(total rejects5385)
Patient position, radiographic 171 (44.7)
technique (A)
Patient movement (B) 20 (5.2)
Incorrect exposure (C) 125 (32.5)
X-ray equipment fault (D) 12 (3.1)
Processing, fogging (E) 15 (3.4)
CR technique (H) n/a
Incorrect organ code (J) n/a
Digiscan fault (K) n/a
Hard copy processor fault (L) n/a
Not requiredb (F) 12 (3.1)
Miscellaneous (G) 32 (8.3)
a
Some images could not be fetched and viewed, so reasons for rejection could not be identi®ed.
b
Some of these rejects do not involve irradiation of the patient.
faults would be eliminated but that PACS might
produce new types of rejects, such as the incorrect
choice of algorithm. This expectation proved to
be true and the number of rejects caused by the
CR-speci®c reasons accounted for more rejects
than those which were eliminated.
The number of rejects of images of the skull
and hip increased during the CR period. This was
probably because of problems associated with
learning CR techniques, although rejects of skull
images could have been the result of the
installation of a new type of skull unit with
which the radiographers were unfamiliar.
In conclusion, two hypotheses were tested in
this study. The ®rst hypothesis that the reject rate
of images at the Hammersmith would be reduced
after the introduction of phosphor plate technol-
ogy (CR) was accepted since a statistically
signi®cant difference between the reject rates of
Film and CR was found. The second hypothesis
that the reject rate would be further reduced after
the introduction of PACS and soft copy images
with manipulation facilities was not accepted
because no statistically signi®cant difference was
found between the reject rates for CR and PACS.
Acknowledgments
The authors wish to thank Professor Martin
Buxton, Director of HERG for his advice and
continued support throughout this study, Alan
McBride in the Radiological Sciences Unit and all
radiographers at the Hammersmith Hospital for
their assistance with the reject analysis studies and
Karen Arnold for her secretarial expertise. This
study was funded by the Policy Research
Programme of the Department of Health but
the views expressed are those of the authors.
The British Journal of Radiology, July 1999
CR PACSa
(total rejects5365) (total rejects5498)
304 (83.3) 391 (78.5)
2 (0.55) 7 (1.4)
6 (1.6) 24 (4.8)
0 15 (3.0)
n/a n/a
14 (3.8) 30 (6.0)
5 (1.4) n/a
4 (1.1) 7 (1.4)
0 n/a
3 (0.82) 16 (3.2)
27 (7.4) 8 (1.6)
References
1. British Institute of Radiology. Assurance of Quality
in Diagnostic X-ray Departments. London: BIR,
1988.
2. Mazzafero RJ, Balter S, Janower ML. The
incidence and causes of repeated radiographic
examinations in a community hospital. Radiology
1974;112:112±71.
3. McKinlay A, McCanley B. Spoilt ®lms in X-ray
departments (correspondence). Br J Radiol 1977;
50:233±4.
4. Bowne D. Repeats: an aspect of departmental
management. Radiography 1969;35:257±61.
5. Mustafa AA, Vasisht CM, Sumanasekara SJ.
Analysis of wasted X-ray ®lms: experience in two
Kuwait hospitals. Br J Radiol 1987;60:513±5.
6. Arvantis TN, Parizel PM, Degryse HR, DeSchepper
AMA. Reject analysis: a pilot programme for
image quality management. Eur J Radiol 1991;
12:171±6.
7. Gadeholt G, Geitung J, Gothlin J. Continuing
reject-repeat ®lm analysis program. Eur J Radiol
1989;9:137±41.
8. Nixon PP, Thorogood J, Holloway J, Smith NJ.
An audit of ®lm reject and repeat rates in a
department of dental radiology. Br J Radiol 1995;
68:1304±7.
9. Lewentat G, Bohndorf K. Analysis of reject X-ray
®lms as a quality assurance element in diagnostic
radiology. Rofo Fortschr Geb Rontegenstr Neuen
Bildgeb Verfahr 1997;166:376±81.
10. Francis R, editor. Quality assurance guidelines for
radiographers in general radiology. North East
Thames Regional Health Authority, 1990.
11. Berry RJ, Oliver R. Spoilt ®lms in X-ray depart-
ments and radiation exposure to the public from
medical radiology (correspondence). Br J Radiol
1976;49:475±6.
12. Sagel SS, Jost G, Glazer HS, Molina PL, Anderson
DJ, Solomon SL, et al. Digital mobile radiography.
J Thorac Imaging 1990;5:36±48.
13. Weatherburn GC, Watkins J, Bryan S, Cocks R.
The effect of PACS on the visualisation of the
lateral cervical spine and the management of
patients presenting with trauma. Medical
Informatics 1997;22:359±68.
659

14. Bryan S, Weatherburn G, Buxton M, Watkins J,
Keen J, Muris N. Evaluation of a hospital picture
archiving and communication system. Journal of
Health Services Research and Policy (forthcoming).
15. Dunn MA, Rogers AT. X-Ray ®lm reject analysis
as a quality indicator. Radiography 1998;4:29±31.
Book review
Modelling in Clinical Radiobiology Based on an
International Workshop in WuÈrzburg. Ed. by
K Baier, D Baltas, pp. 345, 1997 (Albert-
Ludwigs-University Freiburg Publishing,
Germany)
ISBN 3980413901
Radiobiology continues to be an area of much
activity and interest in radiotherapy treatment. It
combines the efforts of clinicians, physicists and
biologists in deriving accurate parameters for use
in modelling the biological effects of radiation on
both individual patients and on populations in
order to fully optimize radiotherapy treatment. As
more reliable patient data is collected, greater
accuracy in predicting responses will be achiev-
able and increasingly sophisticated models will
become accessible.
This monograph reports on the papers pre-
sented at an International Workshop in
WuÈrzburg, Germany in 1993, updated in 1995.
A total of 31 papers are presented in the form of
chapters, covering a broad range of topics within
clinical radiobiology from both theoretical and
clinical perspectives. Approximately two-thirds of
the contributions are from the host country,
additional contributions being from researchers in
the USA, UK, Sweden, Canada, France and
Greece.
The monograph begins with a review by J F
Fowler on the history of modelling radiobiology,
its successes and failures. The papers that follow
consider the latest in current thinking and future
developments in a broad range of areas.
Contributions from physicists, clinicians and
biologists are presented side-by-side. In this way
the monograph provides the reader with a balance
of new theoretical models and results of clinical
studies. This combination of disciplines also
serves to remind the reader that any applications
of the models considered must be tempered by
clinical judgement.
Amongst the many interesting chapters are the
following notable contributions: Modelling the
660
View publication stats
G C Weatherburn, S Bryan and M West
16. Weatherburn GC, Bryan S. The effect of a picture
archiving and communication system (PACS) on
patient radiation doses for examination of the
lateral lumbar spine. Br J Radiol 1999;72:534±545.
Practical Requirements of Clinical Radiotherapy
(R G Dale); this includes a discussion on where
the linear-quadratic (LQ) model should be applied
in the clinic, factors involved in moving from low
dose rate (LDR) to high dose rate (HDR)
brachytherapy, combining treatment modalities
and hypofractionation for palliative treatment.
Pulsed Brachytherapy Radiobiological Aspects
(J F Fowler); dependence on dose-per-pulse and
frequency of pulse delivery and the reliability of
radiobiological parameters.
Determination of the Relative Seriality of a
Tissue from its Response to Non-uniform Dose
Delivery (A A Ê gren, P KaÈllmann and A Brahme);
a mathematical treatment of tissues behaving
either as a serial (e.g. lung, liver) or parallel (e.g.
spinal cord) structure in terms of NTCP.
Volume Effect and Normal Tissue Responses
(D Baltas); discussion and comparison of various
dose-response models for NTCP.
Extraction of Average Values for Tumour
Parameters from Clinical Local Tumour Control
Probability Data (J van de Geijn, H Xie, J Chen);
application of a 6-parameter TCP model using
clinical data.
Radiobiologically Based Treatment Plan
Optimisation (A Brahme); current and emerging
technology for diagnosis and treatment, inverse
planning, molecular biology and methods for
delivering non-uniform dose distributions.
As is common with conference proceedings, a
small number of misprints or errors have
remained in some contributions. However, on
the positive side, a useful index is provided and
each chapter is supported by a large number of
references.
In summary, these proceedings provide a
valuable source of reference for anyone beginning
or currently undertaking research in radiobiology,
whether from a medical or scienti®c perspective,
covering as they do a full review and the status of
research in clinical radiobiology up to 1995.
C R BAKER
The British Journal of Radiology, July 1999