prediction of post-radical prostatectomy pathological outcome for stage t1c prostate cancer with...

0022-5347/!39/1624-1346/0 THE JOVRUL OF UROLOGY Copyright 0 1999 by AMEIUC~\N UKOLOGICM. ASSOCIATION, Isc

Vol. 162, 1346-1351, October 1999 Printed in U.S.A.

PREDICTION OF POST-RADICAL PROSTATECTOMY PATHOLOGICAL OUTCOME FOR STAGE Tlc PROSTATE CANCER WITH PERCENT FREE

PROSTATE SPECIFIC ANTIGEN: A PROSPECTIVE MULTICENTER CLINICAL TRIAL

PAULA C. SOUTHWICK, WILLIAM J. CATALONA,",? ALAN W. PARTIN, KEVIN M. SLAWIN, MICHAEL K. BRAWER, ROBERT C. FLANIGAN, ANUP PATEL, JEROME P. RICHIE,

PATRICK C. WALSH, PETER T. SCARDINO,$ PAUL H. LANGE, GAIL H. GASIOR, ROBERT E. PARSON AND KATHLEEN G. LOVELAND

From the Division o f Urologic Surgery, Washington University School of Medicine, St. Louis, Missouri, Department o f Urology, The Johns Hopkins Hospital, Baltimore, Maryland, Baylor College of Medicine, Houston, Texas, University of Washington, Seattle, Washington,

Loyola University Medical Center, Maywood, Illinois, Department of Research and Development, Hybritech Inc., Sun Diego and University of California Los Angeles, School of Medicine, Los Angeles, California, and Harvard Program in Urology, Brigham and Women's

Hospital, Boston, Massachusetts

ABSTRACT

Purpose: Prostate specific antigen (PSA) exists in bound (complexed) and unbound (free) forms in serum. The percentage of free PSA enhances the specificity of PSA testing for prostate cancer detection. We evaluated the use of percent free PSA preoperatively to predict pathological stage.

Materials and Methods: A total of 379 men with prostate cancer and 394 with benign prostatic disease 50 to 75 years old were enrolled in this prospective study a t 7 medical centers. All subjects had a palpably benign prostate gland, serum PSA 4.0 to 10.0 ng./ml. and a histologically confirmed diagnosis. The Hybritech Tandems PSA and free PSA assays were used. Of the 379 cancer patients 268 (71%) underwent radical prostatectomy.

Results: Higher percent free PSA levels were associated with more favorable histopathological findings in prostatectomy specimens. A value of 15% free PSA provided the greatest discrimination in predicting favorable pathological outcome. Organ confined cancer, Gleason sum less than 7 and small tumors (10% or less involvement of the prostate) were noted in 75% of patients with greater than 15% and only 34% with 15% or less free PSA (p <0.001). Multivariate logistic regression analysis revealed percent free PSA to be the strongest predictor of postoperative pathological outcome (odds ratio 2.25), followed by biopsy Gleason sum (2.06) and patient age (1.35). Total PSA was not predictive in this cohort but has been shown in prior studies to be predictive of outcome when a broader range of PSA values is evaluated.

Conclusions: Percent free PSA may be used for risk assessment of the presence (diagnosis) and stage of prostate cancer in men with PSA between 4 and 10 ng./ml. Percent free PSA may be combined with PSA, digital rectal examination and biopsy findings to help predict postoperative pathological stage and grade, and may assist the patient and physician in making more informed treatment decisions.

KEY WORDS: prostate, prostate-specific antigen, prostatic neoplasms, neoplasm staging

Prostate cancer is the most common cancer in American men and the second leading cause of cancer deaths.' Poten- tially curative surgical therapy is most effective when the tumor is organ confined, and assessment of long-term prognosis and rational treatment decisions depend on accurate prediction of pathological stage and grade before treatment. Unfortunately, current methods of clinical staging do not accurately assess disease status. At least a third of clinically localized cancers are under staged (pathologically advanced

Accepted for publication April 9, 1999. Supported bv a research grant from Hybritech, Inc., San Diego,

All authors have financial interest and/or other relationship with

* Re uests for reprints: Division of Urologic Surgery, 4960 Chil-

1 Financial interest and/or other relationship with Boehringer,

j: Financial interest and/or other relationship with Zeneca. 'j Hybritech, Inc., San Diego, California.

California (subsidiary of Beckman Coulter, Fullerton, California).

Hybritech.

dr:n's%laee, St. Louis, Missouri 63110.

UroMed and Monsanto Searle.

disease).2 Selection of appropriate therapy could be aided by more accurate clinical staging.

Measurement of serum prostate specific antigen (PSA) concentration in combination with digital rectal examination, transrectal ultrasonography and biopsy Gleason grade is commonly used to assess prostate cancer status and predict pathological stage.3 PSA has been shown to exist in multiple forms in serum and is predominantly bound to proteins but 1 form remains uncomplexed and is called free PSA.4.5 Recent studies have demonstrated that the percentage of free PSA in serum may be used to enhance the specificity of total PSA measurements for prostate cancer ~ c r e e n i n g . ~ - ' ~ Others have also suggested that lower percent free PSA may be associated with more aggressive tumors."~'6-'9 In this prospective, multicenter study we examined whether percent free PSA could be useful in the staging of prostate cancer, and added to biopsy findings and other clinical information to predict the postoperative histopathological features of the tumor more accurately. We have previously reported our

1346

PERCENT FREE PROSTATE SPECIFIC ANTIGEN IN PROSTATE CANCER STAGING 1347

results summarizing the use of percent free PSA for cancer screening.'"

METHODS

Study design. Serum samples were obtained from men 50 to 75 years old meeting study entry criteria at 7 university medical centers between July 1994 and December 1996. Of the samples 91% were prospective and 9% were from serum banks of recently evaluated patients. Subjects received no prior treatment for prostatic disease at the time of blood draw, and had serum PSA between 4.0 and 10.0 ng./ml. and no suspicious findings of cancer on digital rectal examination. These cases represent the diagnostic gray zone. Total PSA identifies them as high risk (25% cancer rate compared to 4% for the general population of men older than 50 years) but specificity could be improved since 75% of biopsies are negative.', ' O A histologically confirmed diagnosis based on ultrasound guided 6-sector minimum needle biopsies of the prostate was also required before study entry. Cases were classified as prostate cancer or benign prostatic disease. Free PSA was not used for biopsy or diagnosis decisions but was determined after subjects were enrolled to ascertain which would have been correctly or incorrectly identified by the new marker. In this blinded study pathologists did not have access to the percent free PSA value and laboratory scientists did not have access to the diagnosis. Patients with acute prostatitis, urinary tract infection, prior transurethral resec- tion of the prostate, or recent prostatic manipulation or med- ications (for example finasteride) that might alter serum PSA concentrations were excluded from study. The study protocol was approved by the institutional review board at each site before initiation, and the study was performed in compliance with their respective requirements.

Subjects were enrolled primarily through prostate cancer screening programs, although those screened elsewhere and referred for treatment to the study centers were also enrolled. This population of men with biopsies based solely on an elevated PSA level represents those in whom percent free PSA would be used in clinical practice, and cancer patients were representative of those evaluated for clinically localized prostate cancer a t these institutions.

Histological grade of the biopsy and prostatectomy specimens was determined using the Gleason grading system, and clinical and pathological stage was determined using the TNM system.'l Pelvic lymph node dissections were performed in all radical prostatectomy cases. Percent tumor involvement (tumor volume/prostate volume) was also re- corded if available. Prostatectomy specimens were com- pletely embedded, and intraglandular tumor extent was quantified as percentage of carcinoma by the grid morpho- metric technique of Arcangeli et a1 at Washington Universi- ty.IH To obtain tumor volume we multiplied the sum of all tumor areas by section thickness and a constant to correct for shrinkage of tissue with fixation as previously described." Favorable pathological outcome was defined as organ confined cancer (stages pT1 and pT2), pathological Gleason sum less than 7 and small tumor ( 10% or less involvement of the prostate).

Serum sample collection, storage and assays. Serum samples were processed and refrigerated within 3 hours of blood draw. If the serum sample was to be assayed within 24 hours after collection, the specimen was stored at 2 to 8C. Speci- mens held for longer times were stored at -70C until ana- lyzed,22.23 Concentrations of PSA were determined using the Hybritech Tandem and free PSA monoclonal antibody assays.'" The same serum sample was used to determine total and free PSA concentrations as would be done in clinical practice. Total PSA testing was performed and then the sample was frozen. The free PSA concentration was measured after determining that all study entry criteria were met.

aultiple freeze-thaw cycles do not affect total and free PSA neasurements, and both analytes are stable under the afore- nentioned condition^."'^'^

Statistical analysis. Percent free PSA was calculated as the -atio of free-to-total PSA multiplied by 100. The chi-square test vas used to compare proportions in analyzing categorical vari- 3bles. To evaluate the ability of various preoperative variables XI predict favorable pathological outcome a stepwise (forward selection) logistic regression analysis'" provided variable selection (assessment of percent free PSA, biopsy Gleason sum, age md total PSA) and statistical significance used to derive multivariate logistic models, which were compared using the like- lihood ratio chi-square test. Odds ratios and 959 confidence intervals (CI) using the Wald method were derived from the logistic regression model.

RESULTS

A total of 379 men with prostate cancer and 394 with benign prostatic disease and histologically confirmed diag- noses were enrolled in the study. Of the 379 cancer patients 268 (71%) underwent radical prostatectomy. Pathological tumor stage and grade were available for these patients, and morphometrically determined percent tumor involvement was available for 96. Median age of all cancer patients was 62 years (range 50 to 75), median total PSA was 5.7 (range 4 to 10) and median percent free PSA was 12 (range 2 to 42). Additional characteristics of cancer patients are shown in tables 1 and 2. Biopsy Gleason sum was less than 7 in 844 of patients. Prostatectomy specimen Gleason sums were higher, with only 71% of patients having a sum less than 7. The majority of tumors were organ confined (72%). Negative surgical margins were present in 83% of cases.

Higher percent free PSA levels were associated with more favorable histopathological findings in prostatectomy specimens. Of the 192 patients with organ confined cancer (stages pT1 and pT2) 38% had greater than 15% free PSA. The percentage of patients with greater than 15% free PSA decreased as pathological stage became more advanced, and no patient with advanced stage tumors had greater than 15% free PSA (table 3, fig. 1). This same relationship was ob- served with pathological tumor grade. Greater than 15% free PSA was noted in 67% of subjects with a Gleason sum of 4 and only 19% with a Gleason sum of 7 or higher (table 3).

Chi-square analysis revealed a statistically significant in- crease in the probability of favorable pathological outcome as percent free PSA increased (table 4). Various percent free PSA values were assessed, ranging from 10 to 25%, and 15% provided highly significant predictive values and the greatest discrimination. Organ confined cancer was noted in only 66% of patients with 15% or less compared to 83% with greater than 15% free PSA (p <0.004). Higher percent free PSA (greater than 15%) also predicted more favorable Gleason sums and smaller tumor volumes (table 4). Organ confined cancer, low Gleason sums and small tumors (10% or less involvement of the prostate) were noted in 754 of patients (30 of 40) with higher and only 34% (19 of 56) with lower percent free PSA (fig. 2, p <0.001). Thus, cancer patients with higher percent free PSA tended to have less aggressive disease. Surgical margin status was not associated with per-

TABLE 1. Distribution of biopsy arid prostatectomy specimen Gleason sums

Gleason Sum No. Biopsy ('7) No. Prostatectomy (r; 1

4 17 16.3) 6 (2 .2) 5 6 7 8

39 114.6) 169 (63 01 38 114.2) 3 11.1)

42 115.7) 143 (53.4) 71 (26.5) 4 (1.5)

9 2 (0.7) 2 (0.7)

1348 PERCENT FREE PROSTATE SPECIFIC ANTIGEN IN PROSTATE CANCER STAGING

TABLE 2. Distribution of pathological stage No. Pts. (%I

Organ confined Extracapsular extension Seminal vesicle invasion Invasion of bladder neck, rectum Lvmuh node involvement

192 (71.6) 64 (23.9) 6 (2.2) 4 (1.5) 2 (0.7)

TABLE 3. Percent free PSA, pathological s t a e and grade Free PSA

&15 Greater Than 15

% Pathological stage (No. ptsftotal No.): Organ confined 62 (119/192) 38 (731192) Extracapsular extension 78 (50164) 22 (14/64) Seminal vesicle invasion 83 (516) 17 (Y6) Invasion of adjacent structures, or 100 (6/6) 0 (016)

lymph node metastasis 96 Gleason sum (No. pts./total No.): 4 33 (2/6) 67 (4/6) 5 45 (19142) 55 (23142) 6 68 (971143) 32 (461143) 7-9 81 (62/77) 19 (15/77) As pathological stage and grade became more advanced, the percentage of

patients with higher percent free PSA (greater than 15%) decreased (fig. 1).

8

I

8

seminal Veslde Lymph Node Yzz? IllWhnW Mel&aam M pT4 OrOM confir*d

m s t s o e

FIG. 1. Percent free PSA and pathological stage

cent free PSA. Negative margins were found in 82% of patients with 15% or less and 84% with greater than 15% free PSA (p = 0.65).

Table 5 shows that percent free PSA may be combined with biopsy specimen Gleason sums to predict pathological stage (organ confined cancer) or grade (prostatectomy specimen Gleason sum). The probability of a favorable outcome was highest in patients with greater than 15% free PSA and biopsy Gleason sum less than 7 (87% organ confined disease, 90% pathological Gleason sum less than 7). and lowest in

TABLE 4. Favorable pathological outcome and percent bee PSA % Free PSA

Greater p Than 15 0-15

% Organ coniined disease 66 (1191180) 83 (73188) 0.004

% Pathological Gleason sum less 66 (1191180) 83 (73188) 0.003

40 Tumor vol. 10% or less 71 (40156) 90(36/40) 0.03

(No. ptsJtotal No.)

than 7 (No. pts./total No.)

(No. ptsJtotal No.) There was a higher probability of favorable pathological outcome for patients

with greater than 1540 versus 15% or less free PSA.

greater (63% organ confined disease, 31% pathological Gleason sum less than 7). Subjects with high percent free PSA rarely had high biopsy Gleason sums. The majority of patients with clinically localized cancer have a biopsy Gleason sum of less than 7, and so in most patients percent free PSA would add predictive value.

A stepwise logistic regression analysis was initially performed to evaluate the statistical significance of clinical and demographic variables (percent free PSA, biopsy Gleason sum, age and total PSA). Percent free PSA was the most significant predictor of unfavorable pathological outcome (odds ratio for 10-point decline in percent free PSA 2.25,95% CI 1.44 to 3.601, followed by biopsy Gleason sum (odds ratio 2.06, 95% CI 1.43 to 3.07) and patient age (odds ratio 1.35, 95% CI 1.06 to 1.72). Total PSA was not predictive and did not meet the 0.05 significance level for entry into the model.

DISCUSSION

Results from this prospective multicenter clinical trial show that preoperative serum percent free PSA can be used to help predict pathological outcome of prostate cancer patients with moderately elevated PSA levels. Higher percent free PSA levels were associated with less aggressive disease. A value of 15% free PSA provided the greatest discrimination in predicting favorable pathological outcome. Organ confined cancer, Gleason s u m less than 7 and small tumors were noted in 75% of patients with greater than 15%, and only 34% with lower free PSA levels.

Percent free PSA significantly predicted pathological stage, grade and tumor volume but was not related to surgical margin status. A tumor marker would be expected to correlate with the first 3 variables but not necessarily with margin status, which is most likely to be confounded by surgical technique and pathological interpretation. Others have reported a significant association between margin status and percent free PSA for data from a single center" but this relationship may be masked by site to site variability in a multicenter trial. Percent free PSA may also be combined with biopsy Gleason sum to predict pathological stage and grade. The probability of a favorable pathological outcome was highest in patients with greater than 15% free PSA and biopsy Gleason sum less than 7, and lowest in those with 15% or less free PSA and biopsy Gleason sum 7 or greater.

Partin et a1 presented nomograms for prediction of pathological stage by preoperative PSA, biopsy Gleason sum and clinical stage.3 We demonstrate that percent free PSA can be added as a preoperative variable in similar nomograms (table 5) but addressed only cases with PSA between 4 and 10 ng./ml., and clinical stage Tlc disease, which represent at least a third of detected prostate cancers.' We did not include men with clinical stage T2 because percent free PSA would not typically be used for biopsy decisions in those with suspicious digital rectal examination findings. The study was designed to test the efficacy of percent free PSA for diagnosis and staging in men who would undergo biopsy based solely on an elevated PSA level. However. it is Dossible that Dercent

those-with 15% or less free PSA and biopsy Gleason sum 7 or free PSA might be useful in staging pafpable cancers. Addi-

PERCENT FREE PROSTATE SPECIFIC ANTIGEN IN PROSTATE CANCER STAGING 1349 TABLE 5. Probability of organ confined disease, pathological

Gleason sum less than 7, PSA 4 to 10 ng. Iml. and clinical stage TIC disease

100 - 95 -

90 - 85 -

80 - 75 -

7 0 -

8 6 5 -

5 6 0 - n Q, 5 5 - li 5 5 0 -

h - )1

0

(II

-

P I? 4 5 -

.- 5 4 0 -

Q 3 5 -

rc 0

- .- Q

g 3 0 -

25 - 20

15

10

5 -

- - -

0 1 I 1

0 - 15% > 15%

Percent Free PSA FIG. 2. Probability of favorable pathological outcome (organ con-

fined, pathological Gleason sum less than 7, lymph nodes urlth negative test results for metastasis and small tumor). Cancer patients with greater than 15% tended to have less aggressive disease than those with 15% or less free PSA (p CO.001).

tional studies are needed to complete the nomograms and determine whether percent free PSA is useful for staging in men with a PSA outside of the 4 to 10 ngJml. range and in those with clinical stage other than Tlc disease.

Multivariate logistic regression analysis revealed percent free PSA to be the strongest predictor of postoperative pathological outcome, followed by biopsy Gleason sum and patient age. Total PSA was not predictive in this cohort of men with total PSA between 4.0 and 10.0 ng./ml. Biopsy Gleason sum and total PSA would be expected to be stronger predictors if the PSA range were not truncated. However, a biopsy Gleason sum of 6 is common in this group of men with a palpably benign prostate gland and moderately elevated PSA levels, reducing the ability of biopsy tumor grade to predict

Biopsy Gleason Sum

2-6 7-10

46 Probability organ confined cancer: 0-15% Free PSA 67 63 Greater than 15% free PSA 07 1

0-1546 Free PSA 73 31 Greater than 15% free PSA 90 * * Sample size tw small to calculate probability. Only 11 of the 268 cancer

patients (4%) had elevated percent free PSA and a high biopsy Gleason sum, supporting our findings that elevated percent free PSA is not associated with unfavorable pathological outcome. Few patients would have results in this cell in clinical practice.

46 Probability pathological Gleason sum less than 7:

postoperative pathological outcome. In these patients percent free PSA provides the greatest predictive value.

The 96 tumors with measured volume were generally con- sidered medically important (94% larger than 0.5 and 100% larger than 0.2 CC). '~ Other studies of men with lower PSA and smaller tumors have demonstrated that percent free PSA may help identify possibly harmless t u m ~ r s , ' ~ * ' ~ and support our findings that higher percent free PSA is associated with more favorable pathological outcomes. Percent free PSA has recently been shown to predict insignificant tumor preoperatively better than PSA density, using the same 15% free PSA cutoff and manufacturer a~says . '~

Our results verify earlier studies suggesting that percent free PSA may be associated with tumor pathology. Arcangeli et a1 noted that lower percent free PSA significantly corre- lated with larger tumors, and more advanced pathological stage and grade.16 Carter et a1 reported that percent free PSA may be the earliest marker of tumor behavior, predicting aggressive cancer 10 years before diagnosis, whereas total PSA predicted aggressive cancer 5 years before diagnosis.17 Studies of patients with moderately elevated PSA have revealed a significant relationship between percent free PSA and tumor pathology, with lower percent free PSA associated with features of more aggressive tumors, such as capsular penetration, higher Gleason scores and larger tumor vol-

Others have not reported this relationship but generally included patients with a wide range of PSA values or used different assay^.'^-^' As seen with the use of free PSA testing to distinguish between cancer and benign hyper- plasia, percent free PSA appears to be most effective in staging when PSA levels are in the gray zone, since it is difficult for a new marker to improve the considerable predictive value of total PSA in patients with low or high PSA.

Our results were obtained using the Hybritech Tandem free and total PSA assays. Percent free PSA cutoffs differ when using various combinations of free and total PSA assays from different manufacturer^?^-^^ Mean percent free PSA from identical serum samples may be 2-fold higher using different assay combination^.^' Thus, our results apply only to the assays used in our study.

ume.ll. 18.19

CONCLUSIONS

Percent free PSA may be used for diagnosis and staging of prostate cancer. When used for diagnosis, patients with greater than 25% free PSA need not undergo biopsy unless family o r medical history suggests otherwise. This approach would detect 95% of cancers and spare 20% of men with benign disease from biopsy.15 The missed cancers with high percent free PSA are more likely to be in older men and are primarily organ contined, small tumors with low Gleason sums.15 With annual screening it would be possible to mon- itor patients and perform biopsies only in those with increas-

1350 PERCENT FREE PROSTATE SPECIFIC ANTIGEN IN PROSTATE CANCER STAGING

ing PSA levels or decreasing percent free PSA, to detect what tend to be less advanced tumors.

Percent free PSA also may be used for staging cancer. Patients with greater than 158 free PSA had the most favorable pathological outcomes. However, 15% should not be used as an absolute cutoff to determine appropriate therapy. Rather, percent free PSA may be combined with PSA, digital rectal examination and biopsy findings to predict probable postoperative pathological stage and grade more accurately. This information may assist the patient and physician in making more informed treatment decisions.

REFERENCES

1. Landis, S. H., Murray, T., Bolden. S. and Wingo, P. A,: Cancer Statistics, 1998. CA Cancer J. Clin., 4 8 6, 1998.

2. Catalona, W. J., Richie, J . P., Ahmann, F. R., Hudson, M. A., Scardino, P. T., Flanigan, R. C., deKernion, J. B., Ratliff, T. L., Kavoussi, L. R., Dalkin, B. L., Waters, W. B., MacFarlane, M. T. and Southwick P. C.: Comparison of digital rectal examination and serum prostate specific antigen in the early detection of prostate cancer: Results of a multicenter clinical trial of 6,630 men. J . Urol., 151: 1283, 1994.

3. Partin, A. W., Kattan, M. W., Subong, E. N. P., Walsh, P. C., Wojno, K. J., Oesterling, J. E., Scardino, P. T. and Pearson, J . D.: Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update. J.A.M.A., 277: 1445, 1997.

4. Christensson, A., Laurell, C. B. and Lilja, H.: Enzymatic activity of prostate-specific antigen and its reactions with extracellular serine proteinase inhibitors. Eur. J. Biochem., 194: 755, 1990.

5. Lilja, H., Christensson, A,, Dahlen, U., Matikainen, M. T., Nilsson, O., Pettersson, K. and Lovgren, T.: Prostate-specific antigen in human serum occurs predominantly in complex with alpha 1-antichymotrypsin. Clin. Chem., 37: 1618, 1991.

6. Stenman, U. H., Leinonen, J., Alfthan, H., Rannikko, S., Tuhkanen. K. and AlRhan. 0.: A comdex between Drostate- specific antigen and alpha-1-antichymotrypsin is t6e major form of prostate-specific antigen in serum of patients with prostatic cancer: assay of the complex improves clinical sensi- tivity for cancer. Cancer. Res., 51: 222, 1991.

Christensson, A., Bjork, T., Nilsson, O., Dahlen, U., Matikainen, M. T., Cockett, A. T. K., Abrahamsson, P.-A. and Lilja, H.: Serum prostate specific antigen complexed to alpha-l- antichymotrypsin as an indicator of prostate cancer. J . Urol., 150 100, 1993.

Catalona, W. J., Smith, D. S., Wolfert, R. L., Wang, T. J., Rittenhouse, H. G., Ratliff, T. L. and Nadler, R. B.: Evaluation of percentage of free serum prostate-specific antigen to improve specificity of prostate cancer screening. J.A.M.A., 274 1214, 1995.

Partin, A. W., Catalona, W. J., Southwick, P. C., Subong, E. N., Gasior, G. H. and Chan, D. W.: Analysis of percent free prostate-specific antigen (PSA) for prostate cancer detection: Influence of total PSA, prostate volume, and age. Urology, SUPDI., 4 8 55, 1996.

10. Luderer, A. A,, Chen, Y., Soriano, T. F., Kramp, W. J., Carlson, B., Cuny, C., Sharp, T., Smith, W., Petteway, J., Brawer, M. K. and Thiel, R.: Measurement of the proportion of free to total prostate-specific antigen improves diagnostic performance of prostate-specific antigen in the diagnostic gray zone of total prostate-specific antigen. Urology, 4 6 187, 1995.

11. Elgamal, A.-A. A,, Cornillie, F. J . , Van Poppel, H. P., Van de Voorde, W. M., McCabe, R. and Baert, L. V.: Free-to-total prostate specific antigen ratio as a single test for detection of significant stage T lc prostate cancer. J . Urol., 156 1042, 1996.

12. Van Cangh, P. J., De Nayer, P., Sauvage, P., Tombal, B., Elsen, M., Lorge, F., Opsomer, R. and Wese. F. X.: Free to total prostate-specific antigen (PSA) ratio is superior to total PSA in differentiating benign prostate hypertrophy from prostate cancer. Prostate, suppl., 7: 30, 1996.

13. Chen. Y. T., Luderer, A. A., Thiel, R. P., Carlson, G., Cuny, C. L. and Soriano, T. F.: Using proportions of free to total prostate- specific antigen, age, and total prostate-specific antigen to predict the probability of prostate cancer. Urology, 47: 518, 1996.

14. Woodrum, D. L., Brawer, M. K., Partin, A. W., Catalona, W. J. and Southwick, P. C.: Interpretation of free prostate specific antigen clinical research studies for the detection of prostate cancer. J. Urol., 1 5 9 5, 1997.

15. Catalona, W. J., Partin, A. W., Slawin, K. M., Brawer, M. K.. Flanigan, R. C., Patel, A,, Richie, J. P., deKernion, J. B., Walsh, P. C., Scardino, P. T., Lange, P. H., Subong, E. N. P.. Parson, R. E., Gasior, G. H., Loveland, K. G. and Southwick, P. C.: Use of the percentage of free prostate-specific antigen to enhance differentiation of prostate cancer from benign prostatic disease: a prospective multicenter clinical trial. J.A.M.A., 279 1542, 1998.

16. Arcangeli, C. G., Shepherd, D. L., Smith, D. S., Humphrey, P. A,, Keetch, D. W. and Catalona, W. J. : Correlation of percent free PSA and pathologic features of prostatic carcinomas. J. Urol., part 2, 155 415A, abstract 420, 1996.

17. Carter, H. B., Partin, A. W., Luderer, A. A., Metter, E. J . , Landis, P., Chan, D. W., Fozard, J. L. and Pearson, J. D.: Percentage of free prostate-specific antigen in sera predicts aggressiveness of prostate cancer a decade before diagnosis. Urology, 4 9 379, 1997.

18. Arcangeli, C. G., Humphrey, P. A,, Smith, D. S., Harmon, T. J., Shepherd, D. L., Keetch, D. W. and Catalona, W. J.: Percent- age of free serum prostate-specific antigen as a predictor of pathologic features of prostate cancer in a screening population. Urology, 51: 558, 1998.

19. Pannek, J., Rittenhouse, H. G., Chan, D. W., Epstein, J. I., Walsh, P. C. and Partin, A. W.: The use of percent free prostate specific antigen for staging clinically localized prostate cancer. J. Urol., 159 1238, 1998.

20. 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.

21. Schroder, F. H., Hermanek, P., Denis, L., Fair, W. R., Gospodarowicz, M. K. and Pavone-Macaluso, M.: The TNM classification of prostate cancer. Prostate, suppl., 4 129, 1992.

22. Woodrum, D. L., French, C. and Shamel, L. B.: Stability of free P-SA in serum samples under a variety of sample collection and sample storage conditions. Urology, suppl., 48: 33, 1996.

23. Woodrum, D. L. and York, L.: Two-year stability of free and total PSA in frozen serum samples. Urology, 5 2 247, 1998.

24. Woodrum, D. L., French, C. M., Hill, T. M., Roman, S. J., Slatore, H. L., Shaffer, J. L., York, L. G., Eure, K. L., Loveland, K. G., Gasior, G. H., Southwick, P. C. and Shamel, L. B.: Analytical performance of the Tandem@-R free PSA immunoassay meas- uring free prostate specific antigen. Clin. Chem., 43: 1203, 1997.

25. Hosmer, D. W. and Lemeshow, S.: Applied Logistic Regression. New York John Wiley & Sons, 1989.

26. Epstein, J . I., Walsh, P. C., Carmichael, M. and Brendler, C. B.: Pathologic and clinical findings to predict tumor extent of nonpalpable (stage Tlc) prostate cancer. J.A.M.A., 271: 368, 1994.

27. Epstein, J. I., Chan, D. W., Sokoll, L. J., Walsh, P. C., Cox, J. L., Rittenhouse, H., Wolfert, R. and Carter, H. B.: Nonpalpahle (stage Tlc) prostate cancer: Prediction of insignificant disease using freeltotal (F/T prostate specific antigen levels and needle biopsy findings. J. Urol., part 2, 1 5 9 2407, 1998.

28. Pannek, J., Subong, E. N., Jones, K. A,, Marschke, P. L. S., Epstein, J. I., Chan, D. W., Carter, H. B., Luderer, A. A. and Partin, A. W.: The role of freeltotal prostate-specific antigen ratio in the prediction of final pathologic stage for men with clinically localized prostate cancer. Urology, suppl., 4 8 51, 1996.

29. Lerner, S. E., Jacohsen, S . J., Lilja, H., Bergstralh, E. J. , Ransom, J. , Klee, G. G., Piironen, T., Blute, M. L., Lieber. M. M., Zincke, H., Pettersson, K., Peterson, D. and Oesterling, J . E.: Free, complexed, and total serum prostate-specific antigen concentrations and their proportions in predicting stage, grade, and deoxyrihonucleic acid ploidy in patients with ade- nocarcinoma of the prostate. Urology, 4 8 240, 1996.

30. Bangma, C. H., Kranse, R., Blijenberg, B. G. and Schroder, F. H.: The free-to-total serum prostate specific antigen ratio for staging prostate carcinoma. J. Urol., 157: 544, 1997.

31. Chan, D. W., Sokoll, L. J., Jones, K. A., Kelley, C. A,, Subong, E. N. P. and Partin, A. W.: Clinical evaluation of two free PSA assays in combination with different total PSA assays. J. Urol., suppl., 157: 112. abstract 435, 1997.

PERCENT FREE PROSTATE SPECIFIC ANTIGEN IN PROSTATE CANCER STAGING 1351

32. Fortunato, A., Dorizzi, R. M., Marchi, G., Cocco, C. and Davini, A,: PSNtPSA ratios obtained using nine commercial assays. Clin. Chem., suppl., 4 2 268, 1996.

33. Jung, K., Stephan, C., Lein, M., Henke, W., Schnorr, D., Brw, B., Schurenkamper, P. and Loening, S. A,: Analytical performance and clinical validity of two free prostate-specific antigen assays compared. Clin. Chem., 42: 1026, 1996.

EDITORIAL COMMENT

A major challenge in the treatment of prostate cancer is the accurate preoperative staging of an apparent localized carcinoma. Men who benefit from local therapies tend to have cancer localized within the gland. However, some who exhibit local extension (pT3) will benefit if they possess favorable factors, such as low Gleason grade and minimal capsular penetration.

The diagnostic imaging techniques currently available to detect intraprostatic cancer are not accurate enough to be useful in most patients. Transrectal ultrasound, computerized tomography and magnetic resonance imaging have provided little assistance in dis- criminating locally confined from extracapsular carcinoma. Other parameters promoted to aid in this distinction, including ploidy status, p53, bcl-2 and reverse transcriptase polymerase chain reac- tion PSA, also have not provided consistent and meaningful information. Therefore, most clinicians rely on PSA, Gleason score and clinical stage to decide whether local therapy such as radical prostatectomy will extend the life of the patient. Therefore, we seek an ideal marker that is simple and reproducible, and has a high positive and negative predictive value.

The authors review an extensive multicenter experience evaluat- ing free PSA in predicting pathological outcome following radical prostatectomy. This study of 773 men confirms several prior obser- vations concerning the salutary benefit of this marker in prostatic cancer (references 11, 18 and 19 in article). They determined that a cutoff of 15% for free PSA seems to provide the greatest discrimination in predicting favorable pathological outcome. This simple and reproducible test outperformed Gleason score and total PSA. Since all men included in this prospective trial had Tlc disease, clinical stage could not be evaluated. However, the majority of patients whom I treat for localized cancer have the parameters evaluated by these investigators.

What are the potential weaknesses of this clinical study? One already mentioned is the select population. The findings may not apply to men with an abnormal digital rectal examination and/or a PSA of less than 4 or above 10 ng./ml. Further scrutiny of these cohorts is indicated. Gleason score was reported but would the results vary if the individual patterns were reported? A pattern of 4,3 carries a worse prognosis than a pattern of 3,4. Additionally, only 16% of the patients had Gleason score sum 7 or greater. Would the inclusion of more patients in the high risk group change the results? In my practice the percentage of patients with a Gleason score of 7

seems to be increasing. Widespread screening for prostate cancer has decreased the number of low Gleason score cancers and has resulted in an increasing number of Gleason score 7 cancers. The vast majority (71%) of these men underwent radical prostatectomy. Could a selection bias exist which excluded favorable or unfavorable cases from radical prostatectomy? In many series surgical margin status is predictive of ultimate failure yet in this series it was not associated with the free PSA level. My final concern is biochemical failure, stage D 1.5 disease.' Clinicians and patients are interested in the chances of failure. We are provided with no information regarding failure rates for patients with less than 15% free PSA compared to those with higher levels. What the authors describe as favorable and unfavorable pathological outcomes are based on their extensive experience with pathological features associated with success or failure. I t would seem reasonable to assume that this will be their experience when these men are followed for an extended time but we cannot be sure. These surrogate end points, although pathological, need to be confirmed.

What do we do with this provocative information? Do we deny a potentially curative radical prostatectomy in a 50-year-old man who has a PSA of 5 ng./ml., a biopsy Gleason score of 6 and a free PSA of 26? I do not t h n k so but predicated on what these investigators have presented, I might reconsider radical prostatectomy in someone with a Gleason score of 7, a total PSA of 9 and a free PSA of 2%.

The authors identify what appears to be a powerful tool to predict the final pathological outcome following radical prostatectomy. I believe the power in ths marker is going to be combining it with other reproducible and simple markers as well as pre-prostatectomy pathological findings. We have recently reported on the use of artificial neural networks to enhance clinician ability to determine pathological findings and failure rates following radical prostatectomy! Artificial neural networks seem to be more specific to the individual patient than probability tables. Additionally, they can accommodate multiple variables in determining outcome. I believe the information provided in this report will further enhance these and other promising tools.

E. David Crawford Division of Urology University of Colorado Denver, Colorado

1. Crawford, E. D. and Blumenstein, B. A,: Proposed substages for metastatic prostate cancer. Urology, 5 0 1027, 1998.

2. Snow, P. B., Levine, R. F., Ziada, A. M., Price, B. C. P., Lisle, T. C. and Crawford, E. D.: Impact of different variables on outcome of treatments for clinically confined prostate cancer: prediction of response and pathological stage of the disease using a neural network. J. Urol. part 2, 159 295, abstract 1136, 1998.

prediction of post-radical prostatectomy pathological outcome for stage t1c prostate cancer with...

Documents