Full text of DOI:10.1016/S0022-5347(05)67711-7

0022-5347/00/1634-1144/0
®
THE JOURNAL OF UROLOGY
Vol. 163, 1144–1149, April 2000
Copyright © 2000 by AMERICAN UROLOGICAL ASSOCIATION, INC.®
Printed in U.S.A.
OPTIMAL PREDICTORS OF PROSTATE CANCER ON REPEAT PROSTATE
BIOPSY: A PROSPECTIVE STUDY OF 1,051 MEN
BOB DJAVAN, ALEXANDRE ZLOTTA, MESUT REMZI, KEYWAN GHAWIDEL, ALI BASHARKHAH,
CLAUDE C. SCHULMAN* AND MICHAEL MARBERGER
From the Department of Urology, University of Vienna, Vienna, Austria, and Department of Urology, Erasme Hospital, Brussels, Belgium
ABSTRACT
Purpose: We compare the ability of total prostate specific antigen (PSA), percent free PSA, PSA
density and transition zone PSA density to predict the outcome of repeat prostatic biopsy in men
with serum total PSA 4 to 10 ng./ml. who were diagnosed with benign prostatic hyperplasia after
initial biopsy.
Materials and Methods: In this prospective study 1,051 men with total PSA 4 to 10 ng./ml.
underwent transrectal ultrasound guided sextant biopsy with 2 additional transition zone biop-
sies. In 254 subjects biopsy specimens were also obtained from suspicious areas identified during
transrectal ultrasound and digital rectal examination. All subjects with biopsy specimens neg-
ative for prostate cancer underwent repeat biopsy 6 weeks after initial biopsy. The ability of total
PSA, percent free PSA, PSA density and transition zone PSA density to improve the diagnostic
power of PSA testing was assessed with univariate and multivariate analyses as well as receiver
operating characteristics (ROC) curves.
Results: Initial biopsy was positive (prostate cancer) in 231 and negative (benign prostatic
hyperplasia) in 820 of the 1,051 subjects. Prostate cancer was detected on repeat biopsy in 10%
of subjects (83 of 820) with negative initial biopsy. Percent free PSA and transition zone PSA
density were the most accurate predictors of prostate cancer in these subjects. At a cutoff of 30%
percent free PSA would have detected 90% of cancers (sensitivity) and eliminated 50% of
unnecessary repeat biopsies (specificity). Sensitivity and specificity of transition zone PSA
density at a cutoff of 0.26 ng./ml./cc was 78% and 52%, respectively. ROC curve analysis also
showed that percent free PSA was a significantly better predictor of repeat biopsy results than
total PSA, PSA density and transition zone PSA density. The area under the ROC curve was
74.5% for percent free PSA, 69.1% for transition zone PSA density, 61.8% for PSA density and
60.3% for total PSA.
Conclusions: At least 10% of patients with negative initial prostatic biopsy results will be
diagnosed with prostate cancer on repeat biopsy. Percent free PSA and transition zone PSA
density enhance the specificity of PSA testing compared to total PSA or PSA density when
determining which patients should undergo repeat biopsy. Repeat biopsy should be performed in
patients with percent free PSA less than 30% or transition zone PSA density 0.26 ng./ml./cc or
greater. In our study percent free PSA was the most accurate predictor of prostate cancer in
repeat biopsy specimens.
KEY WORDS: prostatic neoplasms, prostate-specific antigen, prostate, biopsy, prostatic hyperplasia
Except for skin cancer, prostate cancer is the most common ranges from 4.0 to 10.0 ng./ml.3–5 Most studies on early
type of cancer and second leading cause of death due to
cancer among men in the United States. The American Can-
cer Society estimated that 179,300 men were diagnosed with
and 37,000 died of prostate cancer in the United States
during 1999.1 In comparison the incidence of prostate cancer
reached 120 and 98/100,000 per year, respectively, in Austria
and Belgium, reflecting trends similar to those in North
America.
Although prostate specific antigen (PSA) is regarded as the
best biochemical marker for prostate cancer,2 an important
limitation regarding its use in cancer detection is the consid-
erable overlap of patients with prostate cancer and those
with benign prostatic hyperplasia (BPH) when serum PSA
detection of prostate cancer use serum PSA 4.0 ng./ml. as the
cutoff. However, this cutoff does not detect 20% of cancer
cases (false-negative) and incorrectly identifies about 65%
with benign disease as potentially cancerous (false-positive).6
Therefore, the false-positive group will unnecessarily un-
dergo prostatic biopsy, which is invasive and costly.
For subjects within the “gray zone” of serum PSA 4 to 10
ng./ml. several concepts have been introduced to improve
cancer detection (sensitivity) and reduce the number of un-
necessary prostatic biopsies (specificity). These concepts in-
clude PSA density (the ratio of serum PSA-to-prostate vol-
ume), PSA density of the transition zone (the ratio of serum
PSA-to-transition zone volume), PSA velocity (the rate of
change of serum PSA with time), age adjusted PSA reference
ranges and measurement of different molecular forms of PSA
in serum (unbound [free] versus protein bound). During the
last decade percent free PSA (the ratio of free-to-total PSA)
has emerged as the most clinically useful method of improv-
Accepted for publication November 12, 1999.
Presented at annual meeting of American Urological Association,
Dallas, Texas, May 1–6, 1999.
* Financial interest and/or other relationship with Yamanouchi,
Boehringer Ingelheim, Schering-Plough and Zeneca.
1144

OPTIMAL PREDICTORS OF PROSTATE CANCER ON REPEAT BIOPSY
1145
ing prostate cancer detection,7–10 whereas the value of other normal or abnormal (suspicious for prostate cancer). An en-
methods is still the subject of considerable debate.3, 11–19
larged prostate gland was not classified as abnormal. His-
While percent free PSA has been used by clinicians to topathological findings of biopsy specimens were classified as
determine the need for repeat biopsy, clinical studies re- BPH or positive for prostate cancer. PSA density and trans-
ported to date have focused on the performance of percent ition zone PSA density, expressed as ng./ml./cc, were calcu-
free PSA in predicting first biopsy results.20–22 We prospec- lated by dividing serum total PSA by total prostate or trans-
tively evaluate the ability of PSA density, transition zone ition zone volume, respectively.
PSA density and percent free PSA to improve sensitivity and
Statistical analysis. The Wilcoxon rank sum test was used
specificity of PSA testing compared to the use of total PSA to compare continuous data groups. Univariate and multi-
alone in patients who underwent repeat prostatic biopsy variate logistic regression analyses were used to assess con-
after initial results were negative for prostate cancer.23–27
tinuous variables. Sensitivity, specificity, positive predictive
value and negative predictive value of each PSA parameter
were calculated for the overall study population and accord-
ing to repeat biopsy.
MATERIALS AND METHODS
A receiver operating characteristics (ROC) curve was pre-
pared for each PSA parameter by plotting sensitivity versus
the false-positive rate (defined as 100% specificity). The area
under the ROC curve is an indication of the accuracy of a
diagnostic test. The closer the ROC curve is to the upper left
of the graph, the greater the area under the curve and the
better the test. Values for areas under the curve ranged from
50% (no discriminatory and statistical power) to 100% (a
perfect test). Areas under the ROC curve were compared
using the Mann-Whitney U test modulated by Bonferonni-
Holm, with p Ͻ0.05 considered statistically significant.
Subjects. A total of 1,051 subjects 48 to 77 years old were
enrolled in this prospective study from January 1997 through
March 1999 at our institutions. The study population con-
sisted of consecutive referrals for early detection of prostate
cancer or lower urinary tract symptoms. All subjects with
total serum PSA 4 to 10 ng./ml. were enrolled in the study.
Exclusion criteria consisted of history of prostate cancer (193
subjects), prior prostatic biopsy (144), acute or chronic pros-
tatitis and histological evidence of prostatic intraepithelial
neoplasm of any grade, urinary retention, indwelling urinary
catheter or confirmed urinary tract infection. Only 39 pa-
tients were excluded from study because they refused repeat
biopsy.
Study design. Blood specimens for PSA determination
were obtained before any prostatic manipulation, including
digital rectal examination, transrectal ultrasound, biopsy or
any endoscopic procedure. Serum total PSA and free PSA
were measured from deep frozen serum (-70C) with the
AxSYM† total PSA and free PSA assays (intra-observer test
coefficient of variation range 1.2% to 2.4% and 0.8% to 1.5%,
respectively). Both assays are in vitro microparticle enzyme
immunoassays based on a dual monoclonal antibody design.
The AxSYM free PSA assay is intended to be used with the
AxSYM total PSA assay to determine percent free PSA. All
specimens were assayed according to manufacturer recom-
mended procedures.
Digital rectal examination and transrectal ultrasound
guided sextant biopsy with 2 additional transition zone biop-
sies were performed in all subjects. In 254 subjects biopsy
specimens were also obtained from suspicious areas identi-
fied during transrectal ultrasound or digital rectal examina-
tion. All subjects with biopsy specimens negative for prostate
cancer underwent repeat biopsy 6 weeks after initial biopsy.
A biplanar 7.5 MHz. ultrasound probe was used to scan the
prostate gland in the transverse and sagittal planes with the
subject in the left decubitus position. Triaxial distances at
the maximal length, height and width of the total prostate,
and the transition zone were measured as described previ-
ously.15 Volumes of the total prostate and transition zone
were calculated according to the prolate ellipsoid formula
(volume ϭ ␲/6 ϫ length ϫ width ϫ height). Digital rectal
examination was classified by the examining urologist as
RESULTS
Initial biopsy was positive for prostate cancer in 22% of the
subjects (231 of 1,051) with total PSA 4 to 10 ng./ml. Initial
diagnosis was BPH in the remaining 820 subjects. Digital
rectal examination was normal in 64.9% of the subjects (682)
and abnormal in 35.1% (369), and not a significant predictor
of prostate cancer (chi-square 4.5, Pearson correlation p ϭ
0.2121). Additional cores were taken in 254 of 369 patients
with suspicious areas found during transrectal ultrasound
and/or digital rectal examination. Of the cancers 84% were
detected in the peripheral zone compared to 16% in the
transition zone. In general the complication rate of the repeat
biopsy procedure was low. Hematuria was noted in 8% of
patients (no treatment necessary), fever in 1.7% and urinary
tract infection in 3.8%.
Patient age, total prostate volume, transition zone volume
and PSA parameters in subjects diagnosed with BPH or
prostate cancer after initial and repeat biopsy are shown in
table 1. Total prostate volume, transition zone volume and
percent free PSA were significantly higher in men with BPH
compared to those diagnosed with prostate cancer (Wilcoxon
test p Ͻ0.001). At repeat biopsy total prostate volume was
not significantly different between BPH and prostate cancer
cases. Total PSA, PSA density and transition zone PSA den-
sity were significantly higher in subjects diagnosed with
prostate cancer on initial and repeat biopsies (Wilcoxon test
p Ͻ0.01).
All 820 subjects diagnosed with BPH underwent repeat
prostatic biopsy 6 weeks after initial biopsy, and prostate
cancer was detected in 10% (83 of 820). Total prostate volume
† Abbott Laboratories, Abbott Park, Illinois.
TABLE 1. Age, prostate volume and PSA results in 1,051 subjects with prostate cancer or BPH after initial and repeat biopsy
Initial Biopsy
Repeat Biopsy
BPH
820
68 Ϯ 8.5
46.2 Ϯ 17.9
25.2 Ϯ 11.7
6.4 Ϯ 1.8
32.6 Ϯ 17.3
0.156 Ϯ 0.007
0.307 Ϯ 0.181
Prostate Ca
p Value
Prostate Ca
p Value
No. pts.
231
67 Ϯ 7.6
83
68 Ϯ 9
Mean age Ϯ SD
0.32
0.517
0.1981
0.0018
0.0026
Ͻ0.001
Ͻ0.001
Ͻ0.001
Mean total prostate vol. Ϯ SD (ml.)
Mean transition zone vol. Ϯ SD (ml.)
Mean total PSA Ϯ SD (ng./ml.)
Mean % free PSA Ϯ SD
Mean PSA density Ϯ SD (ng./ml./cc)
Mean transition zone PSA density Ϯ SD (ng./ml./cc)
34.3 Ϯ 11.9
14.2 Ϯ 7.1
7.2 Ϯ 1.9
15.7 Ϯ 10.7
0.232 Ϯ 0.109
0.638 Ϯ 0.387
Ͻ0.001
Ͻ0.001
Ͻ0.001
Ͻ0.001
Ͻ0.001
Ͻ0.001
42.5 Ϯ 14.3
20.0 Ϯ 8.7
7.1 Ϯ 1.9
15.4 Ϯ 10.9
0.19 Ϯ 0.08
0.46 Ϯ 0.33

1146
OPTIMAL PREDICTORS OF PROSTATE CANCER ON REPEAT BIOPSY
and transition zone volume were significantly higher (Wil-
coxon test p Ͻ0.001) in the 83 subjects with prostate cancer
detected on repeat biopsy compared to 231 with prostate
cancer on initial biopsy (table 1).
Sensitivity, specificity, and positive and negative values for
predicting repeat biopsy results are shown in table 2 for each
PSA parameter. Cutoffs were selected that produced good
specificity with a sufficiently high level of cancer detection
(sensitivity).15 At a cutoff of 30% and 0.26 ng./ml./cc, respec-
tively, percent free PSA and transition zone PSA density
were the most accurate predictors of positive repeat biopsy.
Although a similar number of unnecessary repeat biopsies
would have been eliminated by percent free PSA or transition
zone PSA density (approximately 50%), prostate cancer de-
tection was much lower for transition zone PSA density (78%)
compared to percent free PSA (90%). Table 2 illustrates the
decreased specificity that would result by selecting cutoffs
that would produce 95% sensitivity for each parameter. At
95% sensitivity percent free PSA would have eliminated the
highest percentage of unnecessary repeat biopsies (33.5%).
The ability of each PSA parameter to predict the outcome
of biopsy results was analyzed further with ROC curves. In
the overall (initial plus repeat biopsy) and repeat biopsy
groups percent free PSA and transition zone PSA density
were the best predictors of prostate cancer (figs. 1 and 2). The
area under the curve for percent free PSA and transition zone
PSA density was 74.2% and 82.7%, respectively, overall, and
74.5% and 69.1% in the repeat biopsy group (p Ͻ0.001, table
3). With the combination of initial and repeat biopsy the area
under the curve was 82.7% for transition zone PSA density
and 74.2% for percent free PSA. Thus, ROC curve analysis
indicated that percent free PSA was a significantly better
predictor of repeat biopsy results than transition zone PSA
density, PSA density and total PSA. In multivariate logistic
regression analysis transition zone PSA density, percent free
PSA and total PSA were significant predictors for cancer
detection in the overall biopsy group (r ϭ 0.4332, p Ͻ0.001).
In the repeat biopsy group only percent free PSA and trans-
ition zone PSA density were significant predictors (r ϭ
0.2150, p Ͻ0.001, table 4).
FIG. 1. Overall ROC curve for prostate cancer detection. AUC,
area under curve. PSA-TZ, transition zone PSA density overall (ϩ).
f/t, free-to-total PSA overall (X). PSAD, PSA density overall (e).
tPSA, total PSA overall (᭛).
though it was previously reported that measurement of trans-
ition zone volume is accurate and reproducible,19 all trans-
rectal ultrasound measurements are operator dependent
and, therefore, subject to possible variability. Furthermore,
measurement of the transition zone in small or large pros-
tates sometimes might be difficult. We reported previously
that transition zone PSA density was not a significant pre-
dictor of biopsy results when total prostate volume was less
than 30 cc.15
Comparison of percent free PSA and transition zone PSA
density revealed that the use of percent free PSA can elimi-
nate costly and invasive procedures. Transition zone PSA
density is much more expensive than percent free PSA and
more subjective because it requires the use of transrectal
DISCUSSION
We prospectively evaluated the ability of percent free PSA, ultrasound. In our repeat biopsy and overall (initial plus
PSA density and transition zone PSA density to increase repeat biopsies) groups percent free PSA resulted in a high
sensitivity and specificity of PSA screening. Whereas most rate of cancer detection without a large number of unneces-
previous studies have focused on the ability of these PSA sary biopsies.
parameters to predict initial biopsy outcome, we focused on
Our results with percent free PSA in predicting repeat
the outcome of repeat biopsy in men with total PSA 4 to 10 biopsy findings were similar to those of others reporting
ng./ml. and initial prostatic biopsy negative for prostate can- initial biopsy results. In a multicenter study of 379 men with
cer. When initial and repeat biopsy results were combined, prostate cancer and 394 with BPH and total PSA 4 to 10
the area under the curve was 82.7% for transition zone PSA ng./ml. a percent free PSA cutoff of 25% had a sensitivity of
density and 74.2% for percent free PSA. It is unclear why 95% and a specificity of 20%.20 Another multicenter study of
transition zone PSA density was superior to percent free PSA 317 men with total PSA 4.0 to 10.0 ng./ml. revealed that a
when initial and repeat biopsy results were combined but percent free PSA cutoff of 26% detected 95% with prostate
not when only repeat biopsy results were considered. Al- cancer and eliminated 29% of negative biopsies.21 In another
TABLE 2. PSA results at specific cutoffs and 95% sensitivity for 820 subjects after repeat biopsy
Pos. Predictive
Value
Neg. Predictive
Value
Cutoff
% Sensitivity
% Specificity
Specific cutoffs:
Total PSA
5.5 ng./ml.
30%
0.13 ng./ml./cc
0.26 ng./ml./cc
80
90
74
78
38
50
44
52
0.12
0.16
0.12
0.15
0.95
0.98
0.93
0.96
% Free PSA
PSA density
Transition zone PSA density
95% Sensitivity:
Total PSA
4.0 ng./ml.
38%
0.09 ng./ml./cc
0.19 ng./ml./cc
95
95
95
95
8.30
33.5
14.7
21.4
0.10
0.14
0.10
0.12
0.93
0.98
0.96
0.98
% Free PSA
PSA density
Transition zone PSA density

OPTIMAL PREDICTORS OF PROSTATE CANCER ON REPEAT BIOPSY
1147
spective study of 51 men with total serum PSA 2 to 15 ng./ml.
demonstrated lower median percent free PSA in those with
positive repeat biopsy compared to those with negative bi-
opsy (15% versus 19%, p ϭ 0.05).24 A percent free PSA cutoff
of 22% yielded a sensitivity of 95% and a specificity of 44% for
predicting repeat biopsy results. In a prospective study of 67
men with persistent total PSA elevations and normal digital
rectal examination low percent free PSA (less than 10%) was
a powerful predictor of prostate cancer, even after 2 negative
biopsies. Disease was detected in 11 patients, and the area
under the ROC curve was 0.93 for percent free PSA compared
to 0.69, 0.66 and 0.51 for free PSA density, PSA density and
total PSA, respectively.25 In another study that used percent
free PSA to predict repeat biopsy results prostate cancer was
detected in 20 of 99 men (20%) with total PSA 4.1 to 10.0
ng./ml. and negative initial biopsy.27 Percent free PSA cutoffs
of 28% and 30% had a sensitivity of 90% and 95%, respec-
tively, and a specificity of 13% and 12%, respectively. For
PSA density cutoffs of 0.10 and 0.08 ng./ml./cc, the respective
sensitivities were 90% and 95%, and the respective specific-
ities were 31% and 12%.
Our prostate cancer detection rate was 10% on repeat
biopsy, which is lower than that observed in previous studies
(range 12% to 30%).23–33 This difference may be related to
the biopsy technique or the fact that our repeat biopsies were
performed only 6 weeks after initial biopsy. The most accu-
rate biopsy technique for repeat biopsies is still a matter of
debate. It is not clear at which prostate volume to start
increasing the number of cores and by how much the total
number of cores should be increased per unit of additional
prostate volume (for example, steps of 10 cc or more). It was
our intent to analyze the value of PSA based parameters for
prostate cancer prediction on initial and repeat biopsy.
Changing the repeat biopsy technique would not have al-
lowed an accurate comparison since an increased or de-
creased detection rate may have been due to an altered
technique and not to the predictive value of the parameters.
In 3 of the previous studies the positive repeat biopsy rate
was 17% after 14.7,23 29% after 19.124 and 30% after 12.826
months between initial and repeat biopsies. Other factors
that must be considered are the range of total PSA in the
study population since men with total PSA greater than 10
ng./ml. may be more likely to have positive repeat biopsy;
whether the results were derived from only 1 repeat bi-
opsy or whether some men underwent additional repeat
biopsies, and whether all men with negative initial biopsy
underwent repeat biopsy or were selected based on factors,
such as persistently elevated total PSA, suspicious digital
rectal examination or symptoms suggestive of prostatic ab-
normality.
FIG. 2. Repeat biopsy ROC curve for prostate cancer detection.
AUC, area under curve. PSA-TZ, transition zone PSA density biopsy
2 (ϩ). f/t, free-to-total biopsy 2 (X). PSAD, PSA density biopsy 2 (e).
tPSA, total PSA biopsy 2 (᭛).
TABLE 3. Area under the ROC curves overall (initial and repeat
biopsy) and for repeat biopsy
Repeat
Overall
Biopsy
No. pts.
1,051
61.2
74.2
55.1
82.7
820
60.3
74.5
61.8
69.1
% Total PSA
% Free PSA
% PSA density
% Transition zone PSA density
TABLE 4. Multivariate logistic regression overall (initial plus repeat
biopsy) and for repeat biopsy
Parameter
Estimate Chi-Square p Value
Overall:
Intercept
Ϫ0.80
Ϫ0.121
6.380
3.72
3.71
0.05
0.05
Ͻ0.001
Ͻ0.001
Ͻ0.001
Total PSA
Transition zone PSA density
Free-to-total PSA
Whole model test (r ϭ 0.4332)
Repeat biopsy:
Intercept
Transition zone PSA density
Free-to-total PSA
Whole model test (r ϭ 0.2150)
128.5
Ϫ0.085
103.2
463.0
Ϫ1.093
2.277
Ϫ0.083
10.3
15.6
41.8
90.5
0.0013
0.001
Ͻ0.001
Ͻ0.001
CONCLUSIONS
study of 308 volunteers with elevated total PSA (2.5 to 10.0
ng./ml.) a percent free PSA cutoff of less than 20% would
Percent free PSA and/or transition zone PSA density
have eliminated 45.5% of negative biopsies. When percent should be determined for patients with total serum PSA 4 to
free PSA was combined with a transition zone PSA density 10 ng./ml. At least 10% of patients with total PSA 4 to 10
cutoff of greater than 0.22 ng./ml./cc, 54.2% of negative biop- ng./ml. and negative prostatic biopsy will be diagnosed with
sies could have been avoided.22
prostate cancer on repeat biopsy. Percent free PSA and trans-
Although some previous studies have focused on repeat ition zone PSA density enhance the specificity of PSA testing
biopsy results, most have been retrospective trials performed compared to total PSA or PSA density in determining which
on relatively small patient populations.23–33 Systematic re- patients should undergo repeat biopsy. Repeat biopsy should
peat biopsies were performed in all of our patients with be performed in men with percent free PSA less than 30% or
negative initial biopsy, whereas this protocol was not usually transition zone PSA density 0.26 ng./ml./cc or greater. The
followed in other studies. Our study also compared several optimal interval between initial and repeat biopsy is not
PSA parameters to predict the outcome of repeat biopsies. In clear but we currently recommend between 6 and 12 months.
a study of 193 men with negative initial biopsy 51 (26%) had Transition zone PSA density is subject to operator dependent
prostate cancer on repeat biopsy, and of the PSA parameters variability, should only be used in patients with total pros-
tested (total PSA, PSA density, age referenced PSA, volume tate volume greater than 30 cc and is associated with rela-
referenced PSA and PSA velocity) total PSA and volume ref- tively high costs. Although transition zone PSA density can
erenced PSA had the highest sensitivity.23 Another retro- be a valuable tool for experienced urologists, for most urolo-

1148
OPTIMAL PREDICTORS OF PROSTATE CANCER ON REPEAT BIOPSY
gists and other physicians percent free PSA is the most
clinically useful tool for determining which patients initially
diagnosed with BPH should undergo repeat biopsy.
zone density and percent free PSA in addition to total PSA
levels. Prostate, 37: 133, 1998
23. Ukimura, O., Durrani, O. and Babaian, R. J.: Role of PSA and its
indices in determining the need for repeat prostate biopsies.
Urology, 50: 66, 1997
REFERENCES
24. Letran, J. L., Blase, A. B., Loberiza, F. R. et al: Repeat ultra-
sound guided prostate needle biopsy: use of free-to-total pros-
tate specific antigen ratio in predicting prostatic carcinoma.
J Urol, 160: 426, 1998
25. Morgan, T. O., McLeod, D. G., Leifer, E. S. et al: Prospective use
of free prostate-specific antigen to avoid repeat prostate biop-
sies in men with elevated total prostate-specific antigen. Urol-
ogy, 48: 76, 1996
26. Fleshner, N. E., O’Sullivan, M. and Fair, W. R.: Prevalence and
predictors of a positive repeat transrectal ultrasound guided
needle biopsy of the prostate. J Urol, 158: 505, 1997
27. Catalona, W. J., Beiser, J. A. and Smith, D. S.: Serum free
prostate specific antigen and prostate specific antigen density
measurements for predicting cancer in men with prior nega-
tive prostatic biopsies. J Urol, 158: 2162, 1997
28. Keetch, D. W., McMurtry, J. M., Smith, D. S. et al: Prostate
specific antigen density versus prostate specific antigen slope
as predictors of prostate cancer in men with initially negative
prostatic biopsies. J Urol, 156: 428, 1996
29. Rietbergen, J. B. W., Boeken Kruger, A. E., Hoedemaeker, R. F.
et al: Repeat screening for prostate cancer after 1-year fol-
lowup in 984 biopsied men: clinical and pathological features
of detected cancer. J Urol, 160: 2121, 1998
30. Durkan, G. C. and Greene, D. R.: Elevated serum prostate spe-
cific antigen levels in conjunction with an initial prostatic
biopsy negative for carcinoma: who should undergo a repeat
biopsy? BJU Int, 83: 34, 1999
31. Noguchi, M., Yahara, J., Koga, H. et al: Necessity of repeat
biopsies in men for suspected prostate cancer. Int J Urol, 6: 7,
1999
32. Keetch, D. W. and Catalona, W. J.: Prostatic transition zone
biopsies in men with previous negative biopsies and persis-
tently elevated serum prostate specific antigen values. J Urol,
154: 1795, 1995
1. Cancer Facts and Figures, 1998. Atlanta: American Cancer So-
ciety, 1998
2. Partin, A. W. and Oesterling, J. E.: The clinical usefulness of
prostate specific antigen: update 1994. J Urol, 152: 1358, 1994
3. Brawer, M. K.: Prostate-specific antigen: critical issues. Urology,
44: 9, 1994
4. Hudson, M. A., Bahnson, R. R. and Catalona, W. J.: Clinical use
of prostate specific antigen in patients with prostate cancer.
J Urol, 142: 1011, 1989
5. Stamey, T. A., Yang, N., Hay, A. R. et al: Prostate-specific anti-
gen as a serum marker for adenocarcinoma of the prostate.
N Engl J Med, 317: 909, 1987
6. Beduschi, M. C. and Oesterling, J. E.: Percent free prostate-
specific antigen: the next frontier in prostate-specific antigen
testing. Urology, suppl., 51: 98, 1998
7. Catalona, W. J., Smith, D. S., Wolfert, R. L. et al: Evaluation of
percentage of free serum prostate-specific antigen to improve
specificity of prostate cancer screening. JAMA, 274: 1214,
1995
8. Morote, J., Ravento´s, C. X., Lorente, J. A. et al: Comparison of
percent free prostate specific antigen and prostate specific
antigen density as methods to enhance prostate specific anti-
gen specificity in early prostate cancer detection in men with
normal rectal examination and prostate specific antigen be-
tween 4.1 and 10 ng/mL. J Urol, 158: 502, 1997
9. Vashi, A. R., Wojno, K. J., Henricks, W. et al: Determination of
the “reflex range” and appropriate cutpoints for percent free
prostate-specific antigen in 413 men referred for prostatic
evaluation using the AxSYM system. Urology, 49: 19, 1997
10. Bangma, C. H., Rietbergen, J. B. W., Kranse, R. et al: The
free-to-total prostate specific antigen ratio improves the spec-
ificity of prostate specific antigen in screening for prostate
cancer in the general population. J Urol, 157: 2191, 1997
11. Bedushi, M. C. and Oesterling, J. E.: Prostate-specific antigen
density. Urol Clin North Am, 24: 323, 1997
33. Keetch, D. W., Catalona, W. J. and Smith, D. S.: Serial prostatic
biopsies in men with persistently elevated serum prostate
specific antigen values. J Urol, 151: 1571, 1994
12. Bazinet, M., Meshref, A. W., Trudel, C. et al: Prospective evalu-
ation of prostate specific antigen density and systematic biop-
sies for early detection of prostatic carcinoma. Urology, 43: 44,
1994
EDITORIAL COMMENT
13. Carter, H. B. and Pearson, J. D.: Prostate-specific antigen veloc-
ity and repeated measures of prostate-specific antigen. Urol
Clin North Am, 24: 333, 1997
14. Richardson, T. D. and Oesterling, J. E.: Age-specific reference
ranges for serum prostate-specific antigen. Urol Clin North
Am, 24: 339, 1997
15. Djavan, B., Zlotta, A. R., Byttebier, G. et al: Prostate specific
antigen density of the transition zone for early detection of
prostate cancer. J Urol, 160: 411, 1998
16. Djavan, B., Zlotta, A., Kratzik, C. et al: PSA, PSA density, PSA
density of transition zone, free/total PSA ratio and PSA veloc-
ity for early detection of prostate cancer in men with serum
PSA 2.5 to 4.0 ng/mL. Urology, 54: 517, 1999
17. Djavan, B., Zlotta, A. R., Remzi, M. et al: Total and transition
zone prostate volume and age: how do they affect the utility of
PSA-based diagnostic parameters for early prostate cancer
detection? Urology, 54: 846, 1999
18. Djavan, B., Seitz, C., Remzi, M. et al: Combination and multi-
variate analysis of PSA-based parameters for prostate cancer
prediction. Tech Urol, 5: 71, 1999
19. Zlotta, A. R., Djavan, B., Roumegue`re, T. et al: Transition zone volume
on transrectal ultrasonography is more accurate and reproducible
than the total prostate volume. BJU Int, 84: 661, 1999
20. Catalona, W. J., Partin, A. W., Slawin, K. M. et al: Use of the
percentage of free prostate-specific antigen to enhance differ-
entiation of prostate cancer from benign prostatic disease: a
prospective multicenter clinical trial. JAMA, 279: 1542, 1998
21. Chan, D. W., Sokoll, L. J., Partin, A. W. et al: The use of percent
free PSA to predict prostate cancer probabilities: an eleven
center prospective study using an automated immunoassay
system in a population with non-suspicious DRE. J Urol,
suppl., 161: 95, abstract 353, 1999
The authors indicate which PSA related factors might indicate
which men at risk of prostate cancer should undergo repeat sextant
biopsies after initial negative results. Their experience with 820
repeat biopsies in men 44 to 77 years old at risk for prostate cancer
with PSA 4 to 10 ng./ml. is the largest study on prospectively per-
forming repeat biopsy in men published so far to my knowledge, and
justifies the use of the PSA ratio or percent free PSA as an indicator.
At a sensitivity of 95% the specificity or true negative value of the
use of percent free PSA was 33.5%, which was significantly more
than that of PSA density (14.7%) or transition zone PSA density
(21.4%) at the given cutoffs. Although 14% of men with percent free
PSA less than 38% had a positive repeat biopsy, no cancer could be
detected in 98% of those with percent free PSA greater than this
cutoff. Therefore, the number of men in whom re-biopsy could be
avoided relatively safely, although PSA was “suspect” was larger
with percent free PSA compared to volume adjusted PSA. As the
number of men included in this study is considerable, the confidence
intervals around these percentages are probably low. I could not
determine easily how these findings related to the results of digital
rectal examination. In this PSA range digital rectal examination
certainly has value for first visit screening,1 although its low inter-
observer agreement is well accepted. Digital rectal examination was
not demonstrated to be a predictor of cancer but did it reinforce the
predicted negative results of percent free PSA?
Remarkably, prostate volume in men with cancer on repeat biopsy
appeared to be larger than that of men with cancer detected initially
(442 participants, p ϭ 0.19). The findings of Reitbergen et al were
similar but significant (p ϭ 0.003) during repeat screening after 1
year, as about 10% of men had positive biopsies (reference 29 in
article). Therefore, the suggestion of Brawer (reference 3 in article)
and others to take more biopsies in men with larger glands, seems to
22. Horninger, W., Reissigl, A., Klocker, H. et al: Improvement of be valid and is undoubtedly applied by many urologists. However,
specificity in PSA-based screening by using PSA-transition the findings of missing up to 25% in clinical series of detectable

OPTIMAL PREDICTORS OF PROSTATE CANCER ON REPEAT BIOPSY
1149
cancers by a set of sextant biopsies probably also have motivated cancers. However, in due time more of that information will become
urologists to repeat biopsies independent of percent free PSA in available.
referred men with initial negative results. While the volumetric
Chris H. Bangma
argument is clear, it remains frustrating not to know in what way
Department of Urology
PSA isoforms relate metabolically to cancer.
The present study supports the use of percent free PSA to provide
additional information for decision making about extra biopsies but
more importantly it also illustrates possibilities to avoid them. That
Erasmus University and Academic Hospital
Rotterdam, The Netherlands
information is time and cost-effective, and will please the patient if
5% uncertainty is accepted. Still, we do not know much about the
histological or biological characteristics of the cancers missed when
restricting the number of biopsies by percent free PSA. This study
was only related to cancer detection and not to the quality of detected
1. Schro¨der, F. H., van der Maas, P., Beemsterboer, P. et al: Eval-
uation of the digital rectal examination as a screening test for
prostate cancer. Rotterdam section of the European Random-
ized Study of Screening for Prostate Cancer. J Natl Cancer
Inst, 90: 1817, 1998