BACKGROUND. Determining the ratio of free to total prostate specific antigen (fPSA to t-PSA, calculated as the percentage of f-PSA [f-PSA%]) in serum allows for
a clearer distinction between patients with prostate carcinoma (PCa) and patients
with benign prostate hyperplasia (BPH) than determining the level of t-PSA alone.
To find influencing factors on f-PSA%, the authors investigated prostate volume,
TNM classification, and tumor stage.
METHODS. The authors measured f-PSA and t-PSA in 36 men with untreated PCa
(tumor classification: T1, 2, 3pN0, M0), 44 patients with BPH, and 54 healthy
controls. Prostate volume was determined by transrectal ultrasound.
RESULTS. The median values of t-PSA and f-PSA% were 7.8 mg/L and 10.5% in PCa
patients, 4.3 mg/L and 20.8% in patients with BPH, and 1.4 mg/L and 23.6% in the
control group. Patients with PCa had a significantly lower proportion of f-PSA than
BPH patients and healthy men. There was no correlation of f-PSA% to TNM stage
or tumor grade. In PCa patients a significant positive correlation (correlation coefficient [r] 0.51, P 0.001) was found between f-PSA% and prostate volume,
whereas there was no significant correlation in BPH patients (r 00.27, P 0.05).
There was a significant difference in f-PSA% between PCa and BPH patients with
prostate volumes smaller than 40 cm3 (9.0% vs. 21.6%, P 0.01) but not between
patients in these 2 groups with prostate volumes exceeding 40 cm3 (15.1% vs.
18.2%, P 0.11).
CONCLUSIONS. Determining the ratio of f-PSA to t-PSA to discriminate between
PCa and BPH patients yields significant results only in men with a prostate volume
of less than 40 cm3. Cancer 1997; 79:1049. q 1997 American Cancer Society.
Supported in part by a grant from the Foundation of the Klee family (Michael Lein) and from
the Fonds der Chemischen Industrie (Klaus
Jung, project no. 400700).
The authors thank Silke Klotzek for her valuable
technical assistance.
This work includes parts of the doctoral thesis
of Carsten Stephan.
Address for reprints: Klaus Jung, Department of
Urology, University Hospital Charite, Humboldt
University Berlin, Schumannstrae 20/21,
10098 Berlin, Germany.
Received May 1, 1996; revision received August
15, 1996; accepted August 30, 1996.
/ 7b48$$0817
12-10-96 11:23:14
cana
W: Cancer
/ 7b48$$0817
12-10-96 11:23:14
cana
105
TABLE 1
Total PSA, Free PSA, and Free PSA% in Healthy Persons, Patients
with Benign Prostate Hyperplasia, and Patients
with Prostate Carcinoma
Measurement
Controls
(n 54)
Benign prostate
hyperplasia
(n 44)
Prostate
carcinoma
(n 36)
1.4 (0.493.15)
0.29 (0.170.53)
23.6 (1345.8)
4.3b (1.313.6)
0.81b (0.381.85)
20.8 (11.438.3)
7.8b,c (3.9516.4)
0.8b (0.411.7)
10.5b,c (6.817.6)
PSA: prostate specific antigen; PSA%: percentage of free prostate specific antigen.
a
Values are given as median (and 1090 percentiles).
b
Significantly different from controls by at least P 0.01 (MannWhitney U test).
c
Significantly different from patients with benign prostate hyperplasia by at least P 0.01 (Mann
Whitney U test).
were incubated together with alkaline phosphataseconjugated monoclonal anti-PSA antibody and formed
a PSA-antibody sandwich complex. After removing the
unbound conjugate by centrifugal wash, the chemiluminescent substrate, a phosphate ester of adamantyl
dioxetane, was added. In the presence of alkaline
phosphatase, that substrate was cleaved into unstable
intermediate emitting photons that were measured by
the luminometer of the analyzer. For determining fPSA, f-PSA was bound to the polystyrene bead during
the first incubation period and an alkaline phosphatase-labeled polyclonal goat anti-PSA antibody was
bound to that formed complex of f-PSA/anti-f-PSA antibody in a second incubation cycle. The following
steps corresponded to those described for the determination of t-PSA.
Statistical Analysis
The percentage of free PSA (f-PSA%) was calculated as
the ratio of f-PSA to t-PSA multiplied by 100. The authors used the Mann-Whitney U test, variance analysis
(ANOVA), and the rank correlation coefficient according to Spearman with the statistical package Statgraphics, version 5.01 (Statistical Graphics Corp.,
Rockville, MD). The diagnostic validity was evaluated
by the receiver operation characteristic (ROC) curve
analysis.15 The GraphROC for Windows software was
used for calculations of areas under the curves.16
RESULTS
Table 1 shows the results of t-PSA and f-PSA measurements and for f-PSA% calculation. Patients with PCa
had a lower proportion of f-PSA than BPH patients
and healthy men. The PCa-free groups showed no differences in f-PSA% values.
TNM stage and grade had no influence on the
W: Cancer
106
/ 7b48$$0817
12-10-96 11:23:14
cana
and BPH patients, as the authors suggested in a previous paper.17 The analysis of these two volume groups
resulted in two important findings (Table 2): 1) BPH
patients with prostate volumes 40 cm3 and 40 cm3
did not differ in f-PSA% values (P 0.46). Prostate
volume had no influence on f-PSA% in BPH patients;
and 2) the f-PSA% values in PCa patients differed significantly in both volume groups (9% vs. 15.1%; P
0.002). This confirms the increase of f-PSA% with increasing prostate volume. Hence it follows that there
is a significant difference in f-PSA% (P 0.0001) in
BPH and PCa patients with volumes 40 cm3. If the
prostate volume exceeds 40 cm3, both groups do not
differ significantly in f-PSA% (18.2% vs. 15.1%; P
0.11). It should be noted that the evaluation of the
subset of patients only with PSA values within the
range of 4 10 mg/L resulted in nearly identical median
f-PSA% values as given for the total group in Table 2.
Whereas BPH patients with a prostate volume 40
cm3 (n 5) and 40 cm3 (n 13) had f-PSA% values
of 20.9% and 18.1%, respectively, PCa patients (n
14; n 8) had values of 9.1% (n 14) and 15.4%
(n 8), respectively. Thus, the different effect of the
prostate volume in BPH and PCa patients on the fPSA% values as described above does not depend on
the total PSA value within the range of 0 23 mg/L.
Only one PCa patient with a prostate volume of
40 cm3 exceeded the value of 15% f-PSA% (Fig. 1).
Of the BPH patients, only one patient was below this
cutoff limit. Thus, values for sensitivity and specificity
were high (93.8% and 96.2%, respectively). Compared
with f-PSA% calculation alone, this leads to an increase
of efficiency to 95%.
Figure 2 includes the ROC curves of t-PSA and fPSA% for all PCa and BPH patients (n 80) and the
f-PSA% curves for all PCa and BPH patients with prostate volumes 40 cm3 (n 42). The t-PSA curve for
the patients with prostate volumes 40 cm3 is not
indicated but used in calculations. The f-PSA% curve
of all 80 patients and the t-PSA curve of patients with
volumes 40 cm3 runs significantly above the t-PSA
curve of all patients (P 0.05). The f-PSA% curve of
patients with volumes 40 cm3 runs again significantly above the f-PSA% curve of all BPH and PCa
patients (P 0.05). It follows that the diagnostic validity of f-PSA% as a diagnostic tool to discriminate between PCa and BPH patients would be essentially improved if it is used in combination with prostate volume (40 cm3 and 40 cm3).
DISCUSSION
PSA is the most important serum parameter for early
detection of PCa.18 Increasing PSA concentrations are
also found in BPH, inflammatory prostatic disease, or
W: Cancer
107
TABLE 2
Total PSA and Free PSA% in Dependence on Prostate Volume in Patients with Benign Prostate Hyperplasia
and Prostate Carcinoma
Benign prostate hyperplasia
Prostate carcinoma
Measurement
Vol 40 cm3
(n 16)
Vol 40 cm3
(n 28)
Vol 40 cm3
(n 26)
Vol 40 cm3
(n 10)
2.4 (0.965.1)
21.6 (15.235.7)
6.2b (2.519.6)
18.2 (9.241.3)
7.6b (3.916.4)
9b (6.314.2)
7.9 (4.117.6)
15.1c (8.422.5)
PSA: prostate specific antigen; PSA%: percentage of free prostate specific antigen.
a
Values are given as median (and 1090 percentiles).
b
Significantly different from patients with benign prostate hyperplasia and prostate volume 40 cm3 by at least P 0.01 (MannWhitney U test).
c
Significantly different from patients with prostate carcinoma and prostate volume 40 cm3 by at least P 0.01 (MannWhitney U test).
FIGURE 2.
/ 7b48$$0817
12-10-96 11:23:14
cana
W: Cancer
108
/ 7b48$$0817
12-10-96 11:23:14
cana
creasing the sensitivity and specificity to approximately 95% may lead to a more effective detection of
PCa.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
W: Cancer
/ 7b48$$0817
12-10-96 11:23:14
cana
22.
23.
24.
25.
26.
27.
109
cancer volume, location and therapeutic selection of resistant clones. J Urol 1994;152:15104.
Villers A, McNeal JE, Freiha FS, Stamey TA. Multiple cancers
in the prostate. Morphologic features of clinically recognized
versus incidental tumors. Cancer 1992;70:23138.
Aihara M, Wheeler TM, Ohori M, Scardino PT. Heterogeneity
of prostate cancer in radical prostatectomy specimens. Urology 1994;43:606.
Grignon DJ, Sakr WA. Pathologic staging of prostate carcinoma. What are the issues? Cancer 1996;78:33740.
Catalona WJ, Smith DS. Free serum prostate-specific antigen
and screening for prostate cancer. JAMA 1996;275:838.
Bangma CH, Kranse R, Schroder FH. Free serum prostatespecific antigen and screening for prostate cancer. JAMA
1996;275:8378.
Spitz J, Benz P, Jauch M, Eiffler E, Kollermann M. Critical
aspects related to the interpretation of the ratio free/total
PSA, estimated by different assay methods. Anticancer Res
1995;15:24001.
W: Cancer