Full text of DOI:10.1016/S0022-5347(05)67710-5

0022-5347/00/1634-1138/0
®
THE JOURNAL OF UROLOGY
Vol. 163, 1138–1143, April 2000
Copyright © 2000 by AMERICAN UROLOGICAL ASSOCIATION, INC.®
Printed in U.S.A.
THE POSITIVE YIELD OF IMAGING STUDIES IN THE EVALUATION OF
MEN WITH NEWLY DIAGNOSED PROSTATE CANCER: A POPULATION
BASED ANALYSIS
PETER C. ALBERTSEN, JAMES A. HANLEY, LINDA C. HARLAN, FRANK D. GILLILAND,
ANN HAMILTON, JONATHAN M. LIFF, JANET L. STANFORD AND ROBERT A. STEPHENSON
From the Division of Urology, University of Connecticut Health Center, Farmington, Connecticut, National Cancer Institute, Applied
Research Branch, Bethesda, Maryland, Norris Comprehensive Cancer Center, University of Southern California, Los Angeles, California,
Department of Epidemiology, Rollins School of Public Health of Emory University, Atlanta, Georgia, Division of Public Health Sciences,
Fred Hutchinson Cancer Research Center, Seattle, Washington, Division of Urology, University of Utah, Salt Lake City, Utah, and
Department of Epidemiology and Biostatistics, McGill University, Montreal, Quebec, Canada
ABSTRACT
Purpose: We determine the positive yield of imaging studies performed on men with newly
diagnosed prostate cancer.
Materials and Methods: A prospective, population based survey was conducted on 3,690 men
with prostate cancer diagnosed between October 1, 1994 and October 31, 1995. Cases were
identified by the rapid case ascertainment systems used in 6 geographic regions participating in
the Surveillance, Epidemiology and End Results Program. Based on information captured in
primary medical record reviews we estimated the positive yield of bone scans, computerized
tomography (CT) and magnetic resonance imaging.
Results: The positive yield of bone scan and CT was less than 5% and 12%, respectively, for all
men with prostate specific antigen (PSA) 4 to 20 ng./ml., and less than 2% and 9%, respectively,
for those who also had a Gleason score of 6 or less. Only men with PSA greater than 50 ng./ml.
and those with Gleason scores 8 to 10 and PSA greater than 20 ng./ml. had positive yields greater
than 10% and 20% for bone scan and CT, respectively.
Conclusions: Imaging studies designed to identify metastases and/or extracapsular extension
in men with newly diagnosed prostate cancer frequently have a low positive yield. Wide varia-
tions exist in the use of imaging studies and are associated with tumor factors, such as Gleason
score and serum PSA, and nontumor factors, such as state of residence. More extensive cost-
effectiveness analyses are needed to define appropriate guidelines for ordering imaging studies
to optimize the positive yield among men with newly diagnosed prostate cancer.
KEY WORDS: prostatic neoplasms, neoplasm staging, tomography, magnetic resonance imaging, prostate
In 1999 approximately 179,300 American men were diag- prostate cancer beyond the prostate capsule in the pelvic
nosed with prostate cancer.1 Physicians performed radionu- lymph nodes or bone.
clide bone scans, computerized tomography (CT) and mag-
netic resonance imaging (MRI) on many of these men as part
of the initial staging evaluation to determine whether dis-
MATERIALS AND METHODS
ease extended beyond the prostate capsule to pelvic lymph
nodes or bone. Men with metastatic disease are usually not
advised to undergo definitive radiation therapy or surgery.
Several investigators have suggested that serum prostate
specific antigen (PSA) can be used to predict the results of
imaging examinations. O’Dowd et al suggested that radio-
graphic imaging had a narrow role in the staging of newly
diagnosed prostate cancer.2 A recent report by Kindrick et al
suggests that physicians may not need to perform imaging
studies, such as bone scans, CT and pelvic MRI, on many men
when evaluating newly diagnosed prostate cancer.3 We ana-
lyzed community practice patterns in 6 population based
regions in the mid 1990s to quantify the use and outcome of
imaging studies among men with newly diagnosed prostate
cancer. We assessed the usefulness of imaging studies in the
setting of other available pretreatment factors, including
Prostate Cancer Outcomes Study. Data were obtained from
the Prostate Cancer Outcomes Study, which is a prospective,
population based analysis of men with newly diagnosed dis-
ease.4 A total of 11,137 men were diagnosed with prostate
cancer between October 1, 1994 and October 31, 1995 in 6 of
the 11 participating regions monitored by the Surveillance
Epidemiology and End Results Program (SEER). Cases were
identified using rapid case ascertainment systems. After re-
ceiving institutional review board approval to contact pa-
tients, a random sample was invited to participate in the
study. A total of 5,667 men were asked to complete 6, 12 and
24-month surveys concerning health related quality of life. In
addition, consent was obtained for office and hospital medical
records to be reviewed by trained medical abstractors. The
information obtained from these assessments was recorded
in a database. Black, Hispanic and white men younger than
Gleason score, disease stage, serum PSA, and patient age, 60 years were over sampled to obtain more extensive infor-
race, education, income and geographic region to identify mation among these subgroups.
Subjects. Of the 5,667 men sampled approximately 7%
could not be located and 2.6% were reported to be too ill or
mentally incompetent to complete surveys. Permission to
abstract medical records was not granted by the treating
Accepted for publication November 19, 1999.
Supported by National Cancer Institute Grants N01-PC-
67000,67005, 67006, 67007, 67009 and 67010.
1138

IMAGING STUDIES AND PROSTATE CANCER
1139
physician or patient for 6.7% of men sampled. Medical
TABLE 1. Patient and tumor characteristics
Weighted %
records were abstracted in 3,826 cases and information about
imaging studies was available in 3,690. Data on the outcome
of the imaging study (positive, negative or equivocal for ex-
tracapsular or metastatic disease) were obtained from a re-
view of office and/or hospital records. Information was col-
lected on the use of bone scans, CT and MRI. Information was
not collected concerning the use of other imaging studies,
such as excretory urography or endorectal coil MRI.
No. Pts.
Diagnosed With Ca
(11,137 pts.)
Total No. pts.
3,690
Pt. age:
Younger than 45
13
76
0.2
1.3
45–49
50–54
254
494
625
797
727
436
268
4.2
55–59
8.3
60–64
17.8
21.0
23.3
14.0
9.9
Data analysis. Descriptive statistics were calculated for
the entire group of 3,690 men. Results of imaging studies and
the yield of a positive test were tabulated according to pre-
diagnosis PSA and biopsy Gleason score. The impact of clin-
ical factors on test use was determined by calculating the
proportion of subjects stratified by clinical characteristics
undergoing each imaging test during the initial staging eval-
uation. Patients were categorized into groups based on pa-
tient and tumor characteristics. In this preliminary analysis
the proportions of patients were weighted by the inverses
of the sampling fractions to reflect more closely proportions
in the actual age and race distribution of the entire set of
11,137 patients diagnosed in the 6 different SEER regions.
In a second analysis multiple logistic regression of the
weighted variables was used to determine the independent
contribution of each characteristic to variation in the use of
imaging studies. Multiple regression was used to assess the
difference in the use of staging examinations among patients
with different levels of a particular factor but the same pro-
file on all other factors. For example, we compared the use of
bone scans for men with different levels of serum PSA at
diagnosis while controlling for age, geographic region, race/
ethnicity, education, household income and Gleason score.
This approach allowed us to assess the extent to which phy-
sicians differ in the use of imaging studies among different
subgroups of patients and which information is most critical
in decisions to order imaging studies. Independent variables
tested included patient specific factors (age, geographic re-
gion, race/ethnicity and socioeconomic status) and tumor fac-
tors (PSA at diagnosis and biopsy Gleason score). The stan-
dard error of each coefficient in the logistic regressions was
used to determine whether a factor was significant after
adjusting for the other variables. Analysis of geographic re-
gion and race/ethnicity was limited to situations when there
were adequate numbers of patients.
65–69
70–74
75–80
Older than 80
Region:
Connecticut
670
439
24.1
8.3
New Mexico
Seattle
398
6.3
Utah
704
442
1,037
9.2
11.2
40.9
Atlanta
Los Angeles
Race/ethnicity:
Nonhispanic white
2,502
651
537
77.2
13.4
9.5
Black
Hispanic
Education:
Did not graduate high school
High school graduate
Some college
College graduate
Advanced/graduate training
Annual household income ($1,000):
Less than 10
746
692
802
472
617
21.3
20.1
23.7
15.8
19.1
295
533
531
453
327
409
476
8.4
17.7
17.4
15.4
10.9
13.9
16.3
10–20
20–30
30–40
40–50
50–75
Greater than 75
PSA at diagnosis (ng./ml.):
Less than 4
358
1,652
743
416
317
9.2
48.5
22.3
11.6
8.5
4–10
10–20
20–50
Greater than 50
Gleason score on biopsy/transure-
thral resection:
2–4
557
647
884
791
438
14.5
18.5
28.9
24.7
13.4
5
6
7
8–10
Clinical stage:
T1/T2
3,265
120
180
91.9
3.2
4.9
T3
T4
RESULTS
Bone scan (3,670 pts.)
CT (3,644 pts.)
MRI (3,652 pts.)
2,549
1,099
152
69.5
32.0
5.0
Clinical and demographic characteristics of the 3,690 pa-
tients with available data are summarized in table 1. Be-
cause of the sampling scheme these 3,690 men were slightly
younger than the 11,137 diagnosed in the 6 SEER regions,
and there were greater proportions of black and Hispanic
Frequencies do not always total 3,690 because of missing data. Weighted
proportions are based on the sample size available for each variable.
men. Estimates of the true distributions of the 11,137 men rather than to confirm the presence of suspected metastatic
are provided by the weighted proportions shown in table 1. disease.
The majority of patients were 60 to 74 years old (range less
than 45 to greater than 90). Weighted mean patient age of pelvic CT in 30% and pelvic MRI in 4%. Of the cases 40% had
the study sample was 69 years. bone scan only, 26% bone scan and CT, 1% all 3 studies and
We estimated that bone scan was performed in 69% of men,
Based on the distribution of patients sampled or weighted 27% no imaging study. Results were located for 2,532 of 2,549
distribution nearly 50% had PSA 4 to 10, 10% PSA less than bone scans (99%), 1,066 of 1,099 CTs (97%) and 149 of 155
4 and 40% PSA greater than 10 ng./dl. Most patients had MRIs (96%). Table 2 shows the positive yield of studies strat-
biopsy tumor Gleason score 5, 6 or 7 (19%, 28% and 24%, ified by pre-diagnosis PSA and biopsy Gleason score. Bone
respectively) with the remainder almost equally divided be- scan was positive in 171, equivocal in 208 and negative in
tween high (Gleason 8 to 10) and low (Gleason 2 to 4) scores. 2,153 of 2,532 cases. The positive yield of bone scans was less
Final clinical staging suggested that 92% of the weighted than 5% for all men with PSA 4 to 20 ng./dl., and less than 2%
sample had clinically localized (stage T1 or T2), 3% regional for those with PSA 4 to 20 ng./ml. and Gleason score 6 or less.
(T3) and 5% metastatic (T4) disease. A review of patient Only men with PSA greater than 50 ng./dl., and those with
symptoms suggests that few had clinical evidence Gleason scores 8 to 10 and PSA greater than 20 ng./ml. had
of metastatic disease. Only 63 patients (1.8%) complained of positive yields greater than 10%. The exceptions were men
weight loss or anorexia and only 81 (2.2%) had bone pain. with PSA less than 4 ng./dl. and Gleason scores 8 to 10 but
Therefore, most studies appear to have been ordered to de- there could be random fluctuations because of the small cell
termine the presence of extracapsular or metastatic disease sample sizes on which some percentages were based.

1140
IMAGING STUDIES AND PROSTATE CANCER
TABLE 2. Yield of positive bone scans and CT stratified by pre-diagnosis PSA and Gleason score on biopsy or transurethral resection
% Yield (No. men)/95% CI
Gleason Score
PSA Less Than 4
PSA 4–10
PSA 10–20
PSA 20–50
PSA Greater Than 50
(9)/ 0–50
All pts.
Bone scan:
2–4
4
1
0
2
(41)/ 0–20
(40)/ 0–15
(39)/ 0–12
(24)/ 0–21
0
1
0
1
2
(164)/ 0–4
(226)/ 0–4
(271)/ 0–2
(204)/ 0–4
(70)/ 0–11
0
(66)/ 0–5
0
1
(28)/ 0–20
(51)/ 0–12
0
1
1
4
5
(323)/ 0–3
(450)/ 0–3
(567)/ 2–6
(609)/ 3–8
5
6
7
0 (108)/ 0–5
2 (126)/ 0–7
4 (168)/ 1–9
10 (13)/ 0–46
32 (35)/ 15–53
22 (85)/ 12–34
51 (101)/ 38–63
38 (276)/ 31–45
10 (82)/ 3–22
3 (105)/ 0–9
18 (73)/ 9–32
7 (362)/ 4–11
8–10
All pts.
CT:
2–4
5
11 (20)/ 1–36
3 (180)/ 1–8
5
(75)/ 1–14
21 (355)/ 16–27
6 (2,532)/ 5–8
1 (1,013)/ 0–2
2 (597)/ 1–4
5
0
0
(17)/ 0–32
(13)/ 0–29
(16)/ 0–26
0
3
0
4
1
2
(62)/ 0–8
(84)/ 0–10
(109)/ 0–5
(90)/ 1–11
(41)/ 0–12
(420)/ 1–4
0
9
5
1
(23)/ 0–21
(52)/ 2–23
(46)/ 1–19
(75)/ 0–7
10 (14)/ 0–46
5
15 (4)/ 0–82
3
4
5
7
(126)/ 0–8
(185)/ 1–9
(227)/ 2–10
(269)/ 4–12
(24)/ 0–29
0
(4)/ 0–71
6
19 (36)/ 6–41
14 (44)/ 4–30
17 (36)/ 5–37
14 (162)/ 8–22
20 (15)/ 3–53
21 (41)/ 8–39
62 (34)/ 40–81
36 (111)/ 25–48
7
18 (12)/ 2–52
16 (11)/ 1–55
8–10
All pts.
12 (42)/ 3–30
5 (251)/ 3–10
19 (167)/ 13–27
8 (1,066)/ 7–11
8
(77)/ 3–18
Yield expressed as a percentage projected to the sampled population, and numbers of men refer to those in the dataset, who underwent the examination and
may not add up across rows or columns because data were not always available on PSA and Gleason score.
CT was positive in 88, equivocal in 100 and negative in 878
of 1,066 cases. Men with PSA 4 to 20 ng./dl. had a positive
yield of 12% or less with most combinations of PSA and
Gleason scores. More than 10% of men with PSA greater than
TABLE 3. Use of imaging tests stratified by patient and tumor
20 ng./dl. and Gleason score 6 or greater were likely to have
characteristics
CT positive for metastatic disease. For combinations of high
% Bone
Scan
% Pelvic
MRI
Gleason scores and PSA greater than 50 ng./dl. the positive
yield was as high as 62%. Findings were similar for pel-
vic MRI, although the sample sizes were much smaller.
Pelvic MRI was positive in 27, equivocal in 14 and negative in
108 of 149 cases.
% CT
Pt. age:
Younger than 45
77
69
77
69
68
73
71
69
61
38
38
38
33
29
31
33
26
17
0
5
8
6
6
5
5
3
1
45–49
50–54
Variations in the use of bone scan and CT in relation to
patient and tumor characteristics are shown in table 3 and
figure 1. Percentages were weighted to reflect the actual
age/race distribution of the total number of men with pros-
tate cancer diagnosed in the 6 SEER regions sampled. Figure
2 demonstrates the extent to which individual patient and
tumor factors influence the use of imaging examinations,
even after adjusting for the other factors. Serum PSA was a
major predictor of the decision to order an imaging study.
55–59
60–64
65–69
70–74
75–80
Older than 80
Region:
Connecticut
New Mexico
Seattle
83
64
58
70
78
62
61
23
25
17
41
19
5
2
2
2
7
6
Utah
Atlanta
Los Angeles
DISCUSSION
Race:
Nonhispanic white
71
69
66
32
28
26
5
4
3
Testing for serum PSA during the last decade has resulted
in a dramatic increase in the number of incidental cases of
prostate cancer in the United States. Prostate cancer has
become a significant medical problem and is estimated to
account for $4.75 billion in health care expenditures annual-
ly.5 Routine use of PSA testing has identified men with early
stage disease and produced a significant shift toward clini-
cally localized disease during the last 5 years.6 Historically,
appropriate staging studies included imaging studies, such
as chest x-ray, pelvic CT, bone scan and possibly pelvic MRI.
Others have suggested that many of these studies are no
longer necessary for men with low serum PSA since results
are frequently negative.2, 3, 7, 8
Accurate pretreatment staging of newly diagnosed pros-
tate cancer is critical in determining whether patients will
benefit from surgery or radiation targeted at localized dis-
ease. Under staging may result in ineffective local therapy,
while over staging risks withholding therapy that may con-
trol or even cure local disease. From the perspective of re-
source use physicians must balance the need to document the
presence or absence of metastatic disease against the cost
and morbidity of these studies, and the probability that they
will provide information that will alter clinical decision
making.
Nonhispanic black
Hispanic
Education:
Did not graduate high school
High school graduate
73
69
70
70
69
32
33
28
32
31
4
5
5
4
6
Some college
College graduate
Advanced/graduate training
Annual household income ($1,000):
Less than 10
67
69
72
71
73
67
71
23
31
31
28
34
30
33
3
3
4
7
5
4
8
10–20
20–30
30–40
40–50
50–75
Greater than 75
PSA at diagnosis (ng./ml.):
Less than 4
50
62
81
87
88
24
30
36
41
35
5
5
5
4
7
4–10
10–20
20–50
Greater than 50
Gleason score on biopsy/transurethral
resection:
2–4
59
70
65
78
82
23
34
29
36
41
3
6
4
6
5
5
6
7
8–10
Clinical stage:
In 1995 physicians ordered bone scans for approximately
two-thirds and CT for a third of all new patients. Our find-
ings document that in less than 5% of most of these patients
imaging studies yielded positive results. Bone scan or CT was
positive in 10% to 20% of men with serum PSA 20 to 50 ng./dl.
T1/2
T3
T4
67
92
94
30
60
48
4
5
10
Percentages are weighted to reflect the age-race-region distribution in sam-
pled population.

IMAGING STUDIES AND PROSTATE CANCER
1141
FIG. 1. Variation in use of bone scan (●) and CT (X) in relation to patient and tumor characteristics. Percentages are weighted to reflect
actual age/race distribution of total numbers of patients diagnosed in 6 areas combined. UT, Utah. NM, New Mexico. CT, Connecticut.
or biopsy Gleason scores 8 to 10. Imaging studies were pos- specify an appropriate test yield to justify ordering an imag-
itive in more than 20% of men with serum PSA greater than ing study, the high cost of these examinations will likely
50 ng./dl. and more than 60% with a high probability of prompt closer scrutiny in the future. Well designed cost-
metastatic disease (serum PSA greater than 50 and Gleason benefit analyses can help define optimal practice guidelines.
scores 8 to 10).
In 1995 only 1.0% of bone scans and 2.7% of CTs were
Since physicians sometimes order imaging studies to con- positive for extracapsular or metastatic disease for men with
firm the absence of extracapsular or metastatic disease, no PSA less than 10 ng./dl. and Gleason score of 6 or less.
findings of metastatic disease can also provide valuable in-
Our analysis also documented wide variations in the use of
formation. Patients with a low probability of metastatic dis- imaging studies in a community based sample of prostate
ease (PSA less than 20 ng./ml. and Gleason score less than 8) cancer cases. As expected, our data indicate that in general a
had a 98.6% chance (95% confidence interval [CI] 98 to 99) of greater proportion of patients at high risk for systemic me-
having a negative bone scan. Only 136 of 819 men (17%) with tastases underwent bone scan and CT compared to those at
a higher risk of metastatic disease (PSA greater than 20 low risk. Pelvic MRI was not frequently used in the initial
ng./ml. or Gleason score 8 to 10) had evidence of metastatic evaluation of men with clinically localized disease.
disease on bone scan. Similar values were found for CT.
In 1995 there appears to have been very little unanimity in
Patients with a low probability of metastatic disease had a the use of imaging studies among practicing clinicians. Our
97.3% chance (95% CI 96 to 98) that CT would not demon- study documents variations in ordering patterns by geo-
strate such disease. Only 59 of 370 men (16%) with a higher graphic region, Gleason score, biopsy PSA and age groups. In
risk of metastatic disease had evidence of such disease on CT. general physicians ordered bone scans twice as often as CT.
Although no practice guidelines have been developed to This finding was relatively consistent across all variables

1142
IMAGING STUDIES AND PROSTATE CANCER
FIG. 2. Extent to which individual patient and tumor factors influence use of bone scans (●) and CT (X) after adjusting for all other factors.
Percentages are predicted from multiple logistic regression of weighted data on white men using average value of other variables. UT, Utah.
NM, New Mexico. CT, Connecticut.
tested. A possible explanation includes the perceived need for to 21%). Even after adjusting for these treatment patterns
a baseline bone scan so that future metastatic progression
can be monitored more accurately. Other possible explana-
tions include a belief that bone scans are more sensitive
indicators of metastatic disease compared to CT, that bone
metastases precede lymph node metastases or simply be-
cause bone scans are less costly and less invasive than CT. As
expected, the use of imaging studies was greatest for patients
with high PSA and high Gleason scores.
Imaging studies were used much more frequently among
practitioners in Connecticut and Atlanta compared to those
on the West Coast. This finding may reflect regional practice
standards, the impact of managed care or the fact that radi-
ation therapy was performed more frequently in these 2
regions in the East compared to the 4 regions in the West.
Radiation therapists often order CT to assist in pretreatment
planning. The use of imaging studies positively correlated
with the use of radiation therapy as the primary treatment.
Radiation therapy was used most frequently in Connecticut
(30%) and Seattle (27%) followed by the other 4 regions (16%
regional differences in the use of imaging studies persisted.
Ordering patterns did not vary much by race or age with the
exception of men older than 80 years. Imaging studies for
these men were ordered less frequently possibly because
elderly men were less likely to receive definitive therapy for
disease.
Several factors contribute to the strength of our study. The
population based study design ensures that our findings are
more generalized to community practices than reports origi-
nating from tertiary medical centers. Cases were randomly
selected to be included in this study by rapid case ascertain-
ment systems operating in large geographic areas. Further-
more, patients were sampled prospectively and during a rel-
atively short time, minimizing the chances of reporting
biases and changing practice patterns. Rather than relying
on physician surveys, we used primary data collection meth-
ods that verified the ordering of imaging studies and even-
tual clinical outcomes. Therefore, our estimates for positive

IMAGING STUDIES AND PROSTATE CANCER
1143
yields should be free of response biases that frequently occur
in physician surveys.9–11
Hospital systems participating in the SEER network
in Connecticut, New Mexico and Utah, and in the Atlanta,
A potential limitation of our study stems from the potential Los Angeles and Seattle areas provided valuable research
selection bias introduced by incomplete survey responses. assistance.
Some patients chose not to participate in our study and some
physicians did not permit access to patients. Based on infor-
REFERENCES
mation on file with the SEER system, these patients do not
appear to differ from those sampled.
1. Landis, S. H., Murray, T., Bolden, S. et al: Cancer statistics,
1999. CA Cancer J Clin, 49: 8, 1999
CONCLUSIONS
2. O’Dowd, G. J., Veltri, R. W., Orozco, R. et al: Update on the
appropriate staging evaluation for newly diagnosed prostate
Based on our analysis bone scan and CT should probably
cancer. J Urol, 158: 687, 1997
only be ordered for men with newly diagnosed prostate can-
3. Kindrick, A. V., Grossfeld, G. D., Stier, D. M. et al: Use of
cer with PSA greater than 20 ng./ml. or PSA greater than 10
imaging tests for staging newly diagnosed prostate cancer:
ng./ml. and Gleason scores 8 to 10. Our data are consistent
with previous reports that also document low positive yields
of imaging studies.2, 3 To our knowledge only 1 professional
organization, the American College of Radiology, has devel-
oped guidelines for ordering imaging studies.12 The guide-
lines discourage imaging with bone scan and CT for men with
low grade disease (Gleason scores 2 to 5) and/or serum PSA
10 ng./ml. or less. Our data suggest that bone scans should be
restricted to men with serum PSA greater than 20 ng./dl. or
biopsy Gleason scores 8 to 10. These men have a higher risk
of extracapsular extension and had positive yields greater
than 10% on bone scans. Men in lower risk groups had
positive yields that generally range from 0% to 5%. Positive
yields for CT were similar but somewhat higher than those
for bone scans for all men with PSA greater than 20 ng./dl. or
Gleason score 8 to 10. It is noteworthy that positive yields for
bone scan and CT were higher for men with PSA less than 4
ng./dl. and Gleason score greater than 6 but these findings
may simply reflect small sample sizes in these groups.
Our data also documented wide variations in the use of im-
trends from the CaPSURE database. J Urol, 160: 2102, 1998
4. Potosky, A. L., Harlan, L. C., Stanford, J. L. et al: Prostate
cancer practice patterns and quality of life: the Prostate Can-
cer Outcomes Study. J Natl Cancer Inst, 91: 1719, 1999
5. Brown, M. L. and Fintor, B.: The economic burden of cancer. In:
Cancer Prevention and Control. Edited by P. Greenwald, B. S.
Kramer and D. L. Weed. New York: Marcel Dekker, Inc. pp.
69–81, 1995
6. Stamford, J. L., Stephenson, R. A., Coyle, L. M. et al: Prostate
cancer trends 1973–1995, SEER program, NIH. Bethesda,
MD: National Cancer Institute, 1998
7. Chybowski, F. M., Keller, J. J., Bergstrahl, E. J. et al: Predicting
radionuclide bone scan findings in patients with newly diag-
nosed, untreated prostate cancer: prostate specific antigen is
superior to all other clinical parameters. J Urol, 145: 313, 1991
8. Levran, A., Gonzalez, J. A., Diokno, A. C. et al: Are pelvic
computed tomography, bone scan, and pelvic lymphadenec-
tomy necessary in the staging of prostatic cancer? Br J Urol,
75: 778, 1995
9. Gee, W. F., Holtgrewe, H. L., Albertsen, P. C. et al: Practice
trends in the diagnosis and management of prostate cancer in
the United States. J Urol, 154: 207, 1995
aging studies, some of which can be attributed to clinical fac- 10. Barry, M. J., Fowler, F. J., Jr., Bin, L. et al: A nationwide survey
of practicing urologists: current management of benign pros-
tatic hyperplasia and clinically localized prostate cancer.
J Urol, 158: 488, 1997
tors, such as serum PSA and biopsy Gleason scores. However, a
considerable portion of the variation can be attributed to differ-
ent practice patterns by geographic region. More extensive cost-
effectiveness studies are needed to define the optimal use of
imaging studies in the evaluation of men with newly diagnosed
prostate cancer. Until then clinicians should carefully evaluate
the potential yield of an imaging study based on serum PSA and
Gleason score before recommending testing for men with newly
diagnosed prostate cancer.
11. Plawker, M. W., Fleisher, J. M., Vapnek, E. M. et al: Current
trends in prostate cancer diagnosis and staging among United
States urologists. J Urol, 158: 1853, 1997
12. American College of Radiology, Expert Panel of Urologic Imag-
ing.: Pretreatment staging of clinically localized prostate can-
cer: appropriateness criteria. American College of Radiology,
September, 1995