T he British Journal of Radiology, 71 (1998), 1130–1135 © 1998 The British Institute of Radiology

Prostate specific antigen level and Gleason score

in predicting the stage of newly diagnosed
prostate cancer
1J A SPENCER, MD, FRCR, 2W J CHNG, MB, ChB, 3E HUDSON, RGN, 4A P BOON, MD, MRCPath
and 3P WHELAN, MS, FRCS

Departments of 1Radiology, 3Urology and 4Pathology, St James’s University Hospital, Leeds LS9 7TF, and
2University of Leeds School of Medicine, Leeds, UK

Abstract. The purpose of this study was to determine the utility of prostate specific antigen (PSA)
level and Gleason score in the prediction of disease stage in men with newly diagnosed prostate
cancer. 102 consecutive men, newly diagnosed with prostate cancer and candidates for radical
therapy, underwent contrast enhanced pelvic CT and skeletal scintigraphy. Staging examinations
used the TNM classification and were reported prospectively with the radiologist blinded to the
patient’s Gleason score and level of PSA. Lymph node metastasis was confirmed by CT guided
biopsy, lymphadenectomy or response to therapy in some cases of massive disease. There were
significant differences between the mean PSA values of 18 men with and 84 men without skeletal
metastases ( p=0.01) and between men with locally confined and non-confined disease ( p=0.02).
There was no difference between PSA values of 13 men with and 89 men without lymph node
metastasis ( p=0.9). Only one man with CT evidence of nodal metastasis (N+ve) had a PSA value
below 20 ng ml−1. Two men with Gleason scores below 6 were N+ve and both had PSA values
over 20 ng ml−1. One man with skeletal metastasis had a PSA value below 20 ng ml−1 but had
bone pain. For this study group if only those men with PSA values over 20 ng ml−1 had been
examined, sensitivity for lymphatic and skeletal metastasis would have been 92%. Using this
threshold about one-third would have been spared imaging investigation. In conclusion, pelvic CT
and skeletal scintigraphy are unlikely to show metastatic disease in a man newly diagnosed with
prostate cancer who has no suggestive clinical features, a PSA level below 20 ng ml−1 and a
Gleason score below 6.

The goal in staging men with newly diagnosed
prostate cancer is to distinguish those with disease
confined to the prostate from those who have local
extraprostatic extension, locoregional or distant
metastatic disease. Options for radical local ther-
apy in men without skeletal or lymphatic meta-
stasis include surgery and radiotherapy. The
operation of radical prostatectomy is preceded
by staging lymphadenectomy which determines
the presence of lymph node metastasis (N stage
assessment). Imaging has an important role in
determining N stage with candidates for radical
radiotherapy. CT has shortcomings as a pre-
surgical technique [1] but is a useful examination
in assessment of candidates for radiotherapy [2,
3]. CT demonstrates unsuspected lymphatic met-
astasis in about one in 10 cases [3] and may also
show more extensive local disease than is clinically
suspected.
Prostate specific antigen (PSA) values are pro-
portional to the pathological stage at initial
Received 9 February 1998 and in revised form 12 May
1998, accepted 13 July 1998.
diagnosis of prostate cancer [4, 5]. However,
prediction of local disease extent (T stage) for
individual patients is unreliable. PSA values predict
those men unlikely to have skeletal metastasis (M
stage assessment) at initial diagnosis [6]. There
have been suggestions that PSA estimation might
thus be used to reduce the number of scintigrams
performed in newly diagnosed cases of prostate
cancer. The predictive value of PSA assay for
assessment of lymphatic metastasis (N stage assess-
ment) is uncertain. The differentiation of the pri-
mary tumour (Gleason score) also influences the
likelihood of distant metastasis [7]. The Gleason
score summates a score for the predominant cellu-
lar differentiation and architecture (ranged from 1
to 5) with that for any secondary pattern on the
specimen. The total score thus ranges from 2
(indolent tumour) to 10 (highly malignant).
Gleason score estimation is widely available at
diagnosis of prostate cancer from material obtained
by transurethral resection or transrectal biopsy.
The aim of this study was to evaluate whether
PSA level and Gleason score could act as predic-
tors of the staging findings of pelvic CT and

1130 T he British Journal of Radiology, November 1998

PSA for staging newly diagnosed prostate cancer
skeletal scintigraphy in newly diagnosed patients,
and thus to determine their potential to reduce the
need for imaging studies in these men.
Patients and methods
102 consecutive men referred for staging exam-
inations were examined with CT and skeletal scin-
tigraphy at initial diagnosis of prostate cancer. All
had PSA estimation at the time of staging examin-
ation and remote from any prostate manipulation.
The men ranged in age from 55 to 90 years (mean
72.6 years). 69 men presented with symptoms of
chronic bladder outflow obstruction, 11 presented
in acute retention, 15 were referred primarily on
the basis of raised PSA or abnormal digital rectal
examination and seven were found to have prostate
cancer incidentally during evaluation of other
urological symptoms, principally impotence.
Examinations were performed between 1993 and
1996.
CT staging examinations of the 102 men with
known prostate cancer were performed using either
a Somatom Plus S or AR.T machine (Siemens,
Erlangen, Germany) with a standardized CT tech-
nique. 200 ml of oral contrast medium (2.5%
Gastrografin, Schering, Burgess Hill, UK) was
administered at bedtime on the day prior to the
CT examination and 1000 ml of 3% Gastrografin
in the hour prior to the examination. Intravenous
contrast medium was routinely administered,
100 ml of Omnipaque 300 (Nycomed, Birmingham,
UK), at a rate of 2 ml s−1 using an injection pump
or by hand injection via an 18 G cannula as a
bolus. The dynamic CT examination was per-
formed upwards from the pelvic floor with 5 mm
contiguous sections obtained through the prostate
and seminal vesicles followed by 10 mm contiguous
sections up to the aortic bifurcation. The examin-
ation was complete at this point if no lymphadeno-
pathy was found, but the examination continued
craniad to the upper limit of disease if nodal
enlargement was found in the pelvis.
Assessment of local disease spread (T stage)
employed standard CT criteria [8]. An upper limit
of normal of 10 mm for short axis measurement of
pelvic lymph nodes was used. CT T stage assess-
ment of the prostate used the following criteria:
disease was staged as T2 or less if there was no
evidence of local extension beyond the capsule of
the prostate, and included prostates with smooth
nodular deformity; staged as T3 when there was
evidence of local periprostatic extension of disease
beyond the capsule or into the seminal vesicles but
not to other adjacent organs or to the pelvic
sidewall; a CT staging of T4 indicated extension
of disease either to the pelvic sidewall or into
adjacent organs other than the seminal vesicles,
for example the rectum. N stage assessment
T he British Journal of Radiology, November 1998
followed the prevailing TNM classification [9]
and was as follows: N1, single node abnormal and
less than 2 cm in size; N2, single node abnormal
and 2–5 cm in size or multiple abnormal nodes;
N3, lymph nodes more than 5 cm in greatest
dimension.
Estimation of Gleason score for the primary
prostate cancer was performed by the same pathol-
ogist on specimens obtained from either trans-
urethral resection or needle biopsy. This examin-
ation was performed retrospectively, blinded to
clinical and imaging details. 426 specimens were
examined. Estimation of prostate specific antigen
levels in the peripheral blood was performed
remote from any prostate manipulation using the
monoclonal assay, with a normal range of
0–4 ng ml−1.
Either laparoscopic or CT guided biopsy was
performed to confirm the nature of equivocal pelvic
lymph node enlargement. In patients with bulky
lymphadenopathy, regression of nodes following
anti-androgen therapy was used as proof of meta-
static involvement. CT guided biopsy used an
anterior transpelvic approach or a trans-sciatic
approach. Biopsy was performed following
informed consent and as an elective in-patient
procedure having excluded any bleeding diathesis
by measurement of blood platelets and INR
(International Normalized Ratio). Biopsy was
contraindicated if the INR was greater than 1.5 or
if the platelet count was below 100 000 ml−1.
Aftercare included bedrest for a minimum of 6 h
with blood pressure and pulse rate checked at half-
hourly intervals for 2 h and then hourly for 4 h.
For skeletal scintigraphic examination, 600 MBq
of 99Tcm Osteoscan HDP (Mallinckrodt Medical,
Petten, The Netherlands) was administered by
intravenous injection with images obtained at
2–4 h after injection. Oral intake of 2 l of fluid was
encouraged to promote clearance of radionuclide
from the soft tissues.
Results
Of the 102 consecutive men staged with CT at
initial diagnosis, 13 men (13%) had evidence of
lymphadenopathy (N+ve). 10 men had pelvic
lymphadenopathy alone and three had additional
retroperitoneal disease. Five of these 13 men had
skeletal metastasis. Eight of the 13 men considered
N+ve by CT had T3 or T4 disease as assessed by
CT. Mean PSA values for men with lymphadeno-
pathy of 105.9 ng ml−1 and without lymphadeno-
pathy (N−ve) of 101.6 ng ml−1 were not
significantly different. The lowest PSA value for a
man with CT evidence of lymphadenopathy was
18.6 ng ml−1. These data are shown in Table 1.
Overall, 18% of men (18 of 102) had positive
skeletal scintigraphy. The mean PSA values of men
1131
 J A Spencer, W J Chng, E Hudson et al

Table 1. Comparison of PSA values for node negative Table 4. Comparison of Gleason scores with CT assess-
and node positive men (102 cases) ment of nodal status (100 of 102 newly diagnosed men)

Node status Number of Mean PSA PSA range Gleason score Number of men Node positive by
men (ng ml−1) (ng ml−1) CT assessment

N−ve (N0) 89 101.6a 1.9–3526.0 2–4 28 2a

5–7 57 10
N+ve 13 105.9a 18.6–360.0
8–10 15 0
N1 8 66.3 20.9–242.0
N2 4 224.7 18.6–360.0
a Both men Gleason score 4; PSA 28.7, 55.5 ng ml−1
N3 1 288 respectively.

a Not significant.
PSA values of 28.7 and 55.5 ng ml−1, respectively.
Table 2. Comparison of PSA values with results of skel- A Gleason score was not available for one of the
etal scintigraphy (102 cases) 13 men with nodal metastasis. This man’s prostate
needle biopsy was consistent with prostate cancer,
Scintigraphy Number of Mean PSA PSA range
confirmation of metastatic disease resulting from
men (ng ml−1) (ng ml−1)
CT guided lymph node biopsy.
Positive 18 371a 2.9–3526.0 Overall, 30 of the 102 men had PSA values
Negative 84 46.6a 1.2–273.0 below 20 ng ml−1. If imaging had been restricted
to men with PSA values above 20 ng ml−1 there
a p=0.014.
would have been a sensitivity of 92% (24 of 26
men) and a specificity of 67% (48 of 72 men) for
Table 3. Comparison of PSA values with CT T stage
(102 cases) metastatic disease. The negative predictive value
of this threshold was 93% (28 of 30 men).
CT T stage Number of Mean PSA PSA range Conversely, if imaging had been denied only to
men (ng ml−1) (ng ml−1) those men with both a PSA level below 20 ng ml−1
T2 or less 72 48.7a 1.2–150.0
and a Gleason score below 6, imaging sensitivity
would have increased to 96% but 12 further men
T3 and T4 30 236a 1.9–360.0 would have required imaging.
T3 24 76.2 1.9–360.0
T4 6 799 11.9–3526.0
Discussion
a p=0.016.
The incidence of metastatic lymph node disease
who had positive skeletal scintigraphy of in the present group of men studied by CT is
371 ng ml−1 and those with negative skeletal broadly similar to other series using either CT or
scintigraphy of 46.6 ng ml−1 were significantly lymphography as methods of staging assessment
different ( p=0.01). These data are shown in [3, 10, 11]. In this group of 102 men staged with
Table 2. Five men with CT evidence of lymph- CT at initial diagnosis of prostate cancer, the
adenopathy had positive skeletal scintigraphy, with lowest PSA level for a man with nodal enlargement
PSA values in the range 55–360 ng ml−1, and was 18.6 ng ml−1. 11 of the 13 men with nodal
Gleason scores of 5–8. Only one of the men with enlargement had Gleason scores of 6 or more; the
positive skeletal scintigraphy had a PSA value of two with lower Gleason scores had PSA levels
less than 20 ng ml−1. This man with a PSA value above 20 ng ml−1 (Table 4). There is thus a case
of 2.9 ng ml−1 had a Gleason score of 4 and CT for a selective use of CT for staging newly diag-
stage of T2N0, and had metastatic (pelvic and nosed cases of prostate cancer. Our data lend
spinal) bone pain. support to the contention that digital rectal exam-
CT T stage assessments were also compared ination, PSA estimation and Gleason grade are
with PSA values. There was a clear difference able to predict nodal status [12].
between the mean PSA values of men with CT However, our data show a marked overlap
confined disease (T2 or less), 48.7 ng ml−1, and CT between PSA values for men with different T and
non-confined disease (T3/4), 236 ng ml−1 ( p= N stages (Tables 1 and 3). Whilst PSA prediction
0.02). However, there was a marked overlap in the of disease stage may not be possible on an individ-
range of values and no PSA threshold could be ual basis, it is possible to suggest a PSA threshold
suggested to predict T status (Table 3). below which imaging tests are unlikely to reveal
It was possible to estimate formal Gleason scores metastases. Only two of the 102 men had N+ve
from the prostate biopsies of 100 of the 102 men. or M+ve disease with PSA levels less than
Gleason scores are compared with CT assessment 20 ng ml−1. A threshold of 20 ng ml−1 for per-
of nodal status in Table 4. The two men with nodal forming imaging would have resulted in a sensi-
metastases had Gleason scores less than 6 but had tivity of 92% (24 of 26 men) for metastasis and

1132 T he British Journal of Radiology, November 1998

PSA for staging newly diagnosed prostate cancer
would have reduced the number of men examined
by about one-third. Important savings would arise
from a more selective use of imaging in this
common cancer.
PSA estimation has already been suggested as a
means of reducing the number of skeletal scintig-
rams performed [6]. The negative predictive value
for skeletal metastasis for men with PSA values
below 20 ng ml−1 was in excess of 99%. Other
workers have found similar results but a single
small UK study demonstrated a significant min-
ority of men with low PSA values and positive
skeletal scintigrams [13]. Caution must therefore
be exercised before changing local practice based
upon findings from patient populations with
different referral patterns and primary treatment
intent.
A number of studies have evaluated the value of
PSA for N stage assessment. In a group of 300
newly diagnosed cases examined with CT/MRI,
the likelihood of nodal disease below 20 ng ml−1
was very low [14]. This study was retrospective,
not all patients had complete radiological staging
and no imaging criteria for metastatic disease were
stated. The study by Levran et al [15] also con-
cluded that the use of pelvic CT and bone scans
was not cost effective for staging in men with PSA
levels below 20 ng ml−1. In this study of 861 men,
only eight (0.9%) had positive bone scans and 13
(1.5%) had evidence of nodal disease on CT. These
incidences are some of the lowest ever reported
but follow a trend in studies reported from the
USA of falling incidence of metastatic disease in
newly diagnosed cases of prostate cancer [16]. A
study of 173 candidates for surgery with a mean
PSA value of 42 ng ml−1 found only nine CT scans
(5%) positive for lymphadenopathy and concluded
that CT was not indicated for men with PSA
values below 25 ng ml−1 [17].
The vast majority of men in our series presented
with symptoms of prostatic disease and none was
from screening programmes. Furthermore, the men
imaged in this study represent a selected group of
newly diagnosed cases at our institution, i.e. those
considered by consultant urologists as being likely
to require active treatment. The prior probability
of abnormality was high and is reflected in the
incidence of metastatic disease. 72 of our 102 men
had PSA values above 20 ng ml−1 and the mean
PSA value was in excess of 100 ng ml−1.
The use of imaging in the assessment of men
with newly diagnosed prostate cancer has received
scrutiny based on economic considerations [18,
19]. Forman et al [18] examined the use of CT of
the abdomen and pelvis to detect co-morbid dis-
ease in prostate cancer patients and found its use
in this context unjustifiable. Our use of CT was
restricted to assessment of the pelvis for staging
purposes. Wolf et al argued that to justify the use
T he British Journal of Radiology, November 1998
of CT for staging prior to radical prostatectomy
there would need to be a likelihood of nodal
metastasis of 50% and a sensitivity for metastasis
of 36%, a figure averaged from reported series.
This figure is, however, much lower than a recently
reported large series [20] and heavily biased by
small studies from the early 1980s. Wolf et al
propose that CT would be reserved for men with
a PSA of greater than 20 ng ml−1 and a Gleason
score of 7 or more [19]. Other authorities
reviewing available data have suggested that pre-
operative CT be reserved for men with PSA greater
than 25 ng ml−1, Gleason score above 6 and a
positive digital rectal examination [21]. Our pro-
spective data lend further support for a policy of
selective use of CT based upon available clinico-
pathological findings. A PSA threshold appears
more useful than their Gleason score threshold
based on examination of our data. Five of our
men with metastatic disease had Gleason scores
below 7.
The size criterion chosen for pelvic lymph node
assessment by CT influences the sensitivity and
specificity of the examination. It is also a reflection
of the staging philosophy. The larger the size
threshold chosen, the lower the sensitivity to nodal
metastasis but the fewer cases denied radical ther-
apy. The lower the size threshold chosen, the
greater the sensitivity but with increased require-
ment for biopsy to achieve acceptable specificity.
For many years the accepted upper limit of normal
size was 15 mm [8]. The size criterion has reduced
in recent years. There has been a renewed interest
in cytological aspiration of ‘‘enlarged’’ pelvic lymph
nodes [20]. All nodes above 6 mm size were
biopsied in this impressive study of 285 men. 43
men had nodes sampled; 10 were cytologically
negative. A further 10 men who underwent lym-
phadenectomy had micrometastases in nodes less
than 6 mm. This 6 mm threshold gave a sensitivity
for metastasis of 78% which is higher than any
previously reported for CT. The specificity of 94%
is also most impressive and increased to 100%
with aspiration cytology. In Oyen et al’s study
[20] the actual nodes sampled ranged from 6 to
25 mm, mean 13 mm. Our study used an upper
limit of normal size of 10 mm which reflected
prevailing opinion at the inception of the study.
The amount of undiagnosed metastatic disease in
pelvic nodes smaller than this is unknown.
The value of CT in the staging evaluation of
newly diagnosed patients entirely depends upon
the referral population and the type of treatment
planned. The technique fell into disrepute in the
USA with the resurgence of interest in radical
prostatectomy [1]. CT was shown to be inferior
to MRI and transrectal ultrasound (TRUS)
in local disease assessment as a pre-operative
procedure in patients planned for radical
1133

prostatectomy [22, 23]. This calls into question
the validity of T stage assessments based on CT
findings when the findings are not validated by
surgical pathology. Using CT to distinguish T3a
from T2 disease is particularly difficult when one
simply relies upon minor periprostatic fat changes
which might result from recent instrumentation or
low grade infection. It is interesting therefore that
such a clear difference is apparent between PSA
values for CT confined and unconfined cases, a
finding that is not due to lumping of T4 and T3
cases (Table 3).
There are strong arguments for the continued
use of CT to stage candidates for radical radio-
therapy. It has a central role in planning of this
treatment and influences management [2, 3]. In
the largest published comparative trial of CT and
MRI involving 46 men with surgical pathology
proof, all understaging failures of CT involved lack
of detection of microscopic capsular breach or
invasion of normal sized seminal vesicles [22].
Whilst surgical margins are often only a few milli-
metres, the boosted radiation field usually includes
a margin of 1–2 cm around the prostate gland—
according to the likelihood of capsular invasion.
In centres which use conformal radiotherapy, the
treatment volume may not totally encompass the
seminal vesicles (in order to reduce the radiation
dose to the rectum). A margin of 10 mm around
the identifiable prostate is used, and only 6 mm at
its interface with the rectum [24]. The greater
ability of MRI to identify tumour in the seminal
vesicles is a clear theoretical advantage. More work
is required to assess the role of MRI in the staging
of candidates for radiotherapy, particularly with
trends towards smaller conformal fields.
Further powerful evidence for the use of CT is
its ability to provide prognostic information. Data
from an outcome study of 483 men suggest that
CT evidence of seminal vesicle invasion has import-
ant prognostic significance [25]. No other CT
findings of local disease status had any importance.
However, all men with CT evidence of nodal
disease were dead at 5 years. Similar outcome data
are not yet available for men staged with MRI.
Finally, PSA estimation is now widely used in
follow-up of men with prostate cancer. PSA values
should fall to zero after radical surgery. There is a
rise after radical radiotherapy due to prostate
‘‘injury’’ but then a progressive decline over the
first year with a further fall in a minority of men
in the second year [5, 26]. Stable levels after 2
years suggest cure, but 40% of those with rising
levels will have metastatic disease or local recur-
rence/persistence. CT is inadequate to assess local
recurrence and TRUS biopsy is required [27]. CT
has a limited role in the minority of men whose
symptomatic relapse is within the soft tissues [28].
Retroperitoneal and upper abdominal disease
1134
J A Spencer, W J Chng, E Hudson et al
may predominate and thus CT examination of the
pelvis alone may underestimate the extent of
relapse. PSA is not absolutely reliable in monitor-
ing treatment of men with metastatic bone disease
[29]. Nonetheless, the requirement for imaging is
likely to be increasingly dictated by PSA estimation
in follow-up of men with prostate cancer.
In conclusion, clinical assessment, PSA esti-
mation and Gleason score may be used in combi-
nation to predict the need for imaging of men with
newly diagnosed prostate cancer. Pelvic CT and
skeletal scintigraphy are most unlikely to show
metastatic disease in a man without bone pain, a
PSA level below 20 ng ml−1 and a Gleason score
below 6.
Acknowledgments
We are grateful to Drs A Parkin and D Wilson
for statistical advice. WJC was attached to the
Department of Radiology as a Special Study
Module from the University of Leeds School of
Medicine.
References
1. Platt JF, Bree RL, Schwab RE. The accuracy of CT
in the staging of carcinoma of the prostate. AJR
1987;149:315–8.
2. Dobbs HJ, Husband JE. The role of CT in the
staging and radiotherapy planning of prostatic
tumours. Br J Radiol 1985;58:429–36.
3. Spencer J, Golding S. CT evaluation of lymph node
status at presentation of prostatic carcinoma. Br
J Radiol 1992;65:199–201.
4. Stamey TA, Kabalin JN. Prostate-specific antigen in
the diagnosis and treatment of adenocarcinoma of
the prostate. J Urol 1989;141:1070–4.
5. Stamey TA, Kabalin JN, McNeal JE, et al.
Prostate-specific antigen in the diagnosis and treat-
ment of adenocarcinoma of the prostate. II. Radical-
prostatectomy-treated patients. J Urol 1989a;
141:1076–83.
6. Chybowski FM, Larson Keller JJ, Bergstralh EJ,
Oesterling JE. Predicting radionuclide bone scan
findings in patients with newly diagnosed, untreated
prostate cancer: prostate-specific antigen is superior
to all other clinical parameters. J Urol 1991;
145:313–8.
7. Kramer SA, Spahr J, Brendler CB, Glenn JF,
Paulson DF. Experience with Gleason’s histopathol-
ogic grading in prostatic cancer. J Urol 1980;
124:223–5.
8. Husband JES. Staging of bladder and prostate
cancer. In: Husband J, editor. CT review. Edinburgh:
Churchill Livingstone, 1989:203–15.
9. Spiessl B, Beahrs OH, Hermanek P, et al, editors.
TNM atlas. Berlin: Springer-Verlag, 1992:248–50.
10. Zingg EJ, Fuchs WA, Heritier P, Gothlin J.
Lymphography in carcinoma of the prostate. Br J
Urol 1974;46:549–54.
11. Hayward SJ, McIvor J, Burdge AH, Jewkes RF,
Williams G. Staging of prostatic carcinoma with
radionuclide bone scintigraphy and lymphography.
Br J Radiol 1987;60:79–81.
T he British Journal of Radiology, November 1998
PSA for staging newly diagnosed prostate cancer
12. Amis ES. Role of CT and CT-guided nodal biopsy
in staging of prostatic cancer. Radiology 1994;
190:309–10.
13. Huncharek M, Muscat J. Serum prostate-specific
antigen as a predictor of radiographic staging studies
in newly diagnosed prostate cancer. Cancer Invest
1995;13:31–5.
14. Miller PD, Eardley I, Kirby RS. Prostate specific
antigen and bone scan correlation in the staging and
monitoring of patients with prostate cancer. Br
J Urol 1992;70:295–8.
15. Levran Z, Gonzalez JA, Diokno AC, et al. Are pelvic
computed tomography, bone scan and pelvic lym-
phadenectomy necessary in the staging of prostatic
cancer? Br J Urol 1995;75:778–81.
16. Petros JA, Catalona WJ. Lower incidence of unsus-
pected lymph node metastasis in 521 consecutive
patients with clinically localized prostate cancer.
J Urol 1992;147:1574–5.
17. Flanigan RC, McKay TC, Olson M, et al. Limited
efficacy of preoperative computed tomography scan-
ning for the evaluation of lymph node metastasis in
patients before radical prostatectomy. Urology
1996;48:428–32.
18. Forman HP, Heiken J, Brink JA, et al. CT screening
for co-morbid disease in patients with prostatic
carcinoma; is it cost-effective? AJR 1994;162:1125–8.
19. Wolf JS, Cher M, Dall’era M, et al. The use and
accuracy of cross-sectional imaging and fine needle
aspiration cytology for detection of pelvic lymph
node metastases before radical prostatectomy. J Urol
1995;153:993–9.
20. Oyen RH, van Poppel HP, Ameye FE, van de
Voorde WA, Baert AL, Baert LV. Lymph node
staging of localised prostatic carcinoma with CT
and CT-guided fine-needle aspiration biopsy: pro-
spective study of 285 patients. Radiology 1994;
190:315–22.
T he British Journal of Radiology, November 1998
21. Rees MA, Resnick MI, Oesterling JE. Use of
prostate-specific antigen, Gleason score, and digital
rectal examination in staging patients with newly
diagnosed prostate cancer. Urol Clin North Am
1997;24:379–88.
22. Hricak H, Dooms GC, Jeffrey RB, et al. Prostatic
carcinoma: staging by clinical assessment, CT, and
MR imaging. Radiology 1987;162:331–6.
23. Salo JO, Kisivaari L, Rannikko S, et al.
Computerised tomography and transrectal ultra-
sound in the assessment of local extension of pros-
tatic cancer before radical retropubic prostatectomy.
J Urol 1987;137:435–8.
24. Liebel SA, Heimann R, Kuthcher GJ, et al. Three-
dimensional conformal radiation therapy in locally
advanced carcinoma of the prostate: preliminary
results of a phase I dose-escalation study. Int
J Radiat Oncol Biol Phys 1993;28:55–65.
25. Perez CA. Carcinoma of the prostate: a model for
management under impending health care system
reform. Radiology 1995;196:309–22.
26. Stamey TA, Kabalin JN, Ferrari M. Prostate-specific
antigen in the diagnosis and treatment of adenocarci-
noma of the prostate. III. Radiation-treated patients.
J Urol 1989;141:1084–7.
27. Kramer S, Gorich J, Gottfried HW, et al. Sensitivity
of computed tomography in detecting local recur-
rence of prostatic carcinoma following radical pros-
tatectomy. Br J Radiol 1997;70:995–9.
28. Spencer JA, Golding SJ. Patterns of lymphatic met-
astases at recurrence of prostate cancer: CT findings.
Clin Radiol 1994;49:404–7.
29. Sissons GRJ, Clements R, Peeling WB, Penney MD.
Can serum prostate-specific antigen replace bone
scintigraphy in the follow-up of metastatic prostate
cancer? Br J Radiol 1992;65:861–4.
1135