Japanese Journal of Clinical Oncology 32:536-542 (2002)
© 2002 Foundation for Promotion of Cancer Research
Predicting Prostate Specific Antigen Failure after Radical Retropubic Prostatectomy for T1c Prostate Cancer
1 Division of Urology, Department of Surgery, 3 Department of Pathology, Taichung Veterans General Hospital, National Yang-Ming University School of Medicine, Taiwan, 2 Institute of Medicine, 4 Institute of Toxicology and 5 Institute of Biochemistry, Chung Shan Medical University, Taiwan
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Purpose: Clinicopathological data were reviewed to find a predictor of prostate specific antigen (PSA) failure in Taiwanese patients who had received radical retropubic prostatectomy (RRP) for stage T1c prostate cancer (PC).
Methods: Fifty-five consecutive men who underwent RRP for stage T1c PC were included. The clinical end point was PSA failure (PSA >0.2 ng/ml). Preoperative PSA, free-to-total PSA ratio, prostate volume, PSA density, transrectal sextant biopsy and whole mount of RRP parameters were analyzed for their ability to predict postoperative PSA failure.
Results: Fifteen of the 55 patients developed PSA failure during the follow-up period. Those with PSA failure had higher PSA, higher percentage of cancer in biopsies and higher biopsy Gleason score than the freedom from PSA failure group (all P < 0.05). The PSA failure group had higher pathology Gleason score and a higher incidence of extracapsular tumor extension than the freedom from PSA failure group (all P < 0.05). The PSA failure group had a larger tumor volume and higher incidence of combined peripheral lobe with transitional lobe involvement than the freedom from PSA failure group (all P < 0.05). Multivariate analysis revealed that the predictors for PSA failure after RRP were biopsy Gleason score 
6, tumor volume 2.5 ml and PSA 10 ng/ml.
Conclusion: The single most significant predictor for PSA failure in T1c PC patients after RRP was tumor volume 2.5 ml.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Since the use of prostate-specific antigen (PSA) has become widespread for the detection of prostate cancer, impalpable PSA-detected tumors (clinical T1c) have become the most commonly diagnosed prostate cancers (1). The incidence of T1c prostate cancer (PC) has increased from 2.7% of all prostate cancers in 1989 to 47.2% in 1994 (2). In one screening study, 78% of PC detected were stage T1c (3). There is a considerable difference in the age-specific prevalence of clinical prostate cancers between Asians and Caucasians (4). However, in our previous study, the incidence of latent PC in patients with transitional cell carcinoma of the urinary bladder receiving radical cystoprostatectomy was 33%, similar to those in Japan and the USA (5–7). The clinicopathological features of T1c PC in Japanese patients were similar to those in Caucasians reported elsewhere (8). In this study, clinicopathological features of stage T1c PC Taiwanese patients were evaluated.
Results of radical retropubic prostatectomy (RRP) for clinically localized PC vary depending on pathological status, with 5 year rates of biochemical disease control from 15% for lymph node involvement to 97% for organ confined disease (9). PSA failure after curative treatment is used as a primary treatment end point. These patients may have detectable PSA only, local recurrence or distant metastasis. To predict the likelihood of PSA failure after RRP, outcomes of preoperative transrectal ultrasound-guided sextant biopsy of the prostate along with clinicopathological data obtained from these patients were reviewed.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
From January 1996 to December 2000, 385 patients with prostate cancer received treatment in this hospital. Among them, 58 patients (15.1%) had stage T1c PC and received RRP. Three patients with T1c prostate cancer who received neoadjuvant hormonal therapy were excluded owing to the difficulty of pathological interpretation. Fifty-five patients were enrolled in this study. Clinical data, including patient age, initial PSA and free-to-total PSA ratio were collected. The serum total PSA and free PSA concentrations were measured with TPSA-RIACT and FPSA-RIACT kits (Cis Bio International, France), using a solid-phase two-sided immunoradiometric assay. The inter- and intra-assay variation in our laboratory was less than 10%. Free PSA was measured in only 41 of 55 patients.
Sextant biopsies of the prostate using the Stamey modified technique (10) were performed using a spring-loaded 18-gauze biopsy gun (Bard, USA) at an angle of ~30° under transrectal urological ultrasound (TRUS) guidance (Bruel and Kjer 3535 urological ultrasound coupled with a biplane transrectal probe, Denmark). The PSA density (PSAD) was calculated by dividing the serum PSA by the prostate volume measured by transrectal ultrasound. In each needle biopsy certain variables were assessed, including the Gleason score, number of cancer positive cores and percentage of cancers in all sextant biopsies.
RRP and bilateral pelvic lymphadenectomy were done. Prostatectomy specimens were fixed, coated with Indian ink and cut into systemic stepwise sections at 3 mm intervals. The tumors were evaluated for Gleason score and tumor location. Tumor volume was calculated by multiplying the resulting tumor areas by section thickness (3 mm) and correcting for tissue shrinkage factor (1.5) following fixation (11). Radical prostatectomy stage was classified as either extracapsular tumor extension (ECE) or organ-confined disease (OCD). ECE was defined as tumor-involved capsular perforation extending entirely through the prostate capsule. OCD was defined as tumor limited within the prostate capsule, with no seminal vesicle or lymph node involvement. We defined PSA or biochemical failure as two serial serum PSA >0.2 ng/ml. Patients were observed every month until PSA was undetectable (PSA <0.01 ng/ml), then every 3 months for 3 years, followed by an evaluation every 6 months thereafter. The mean follow-up periods were 2.19 ± 1.45 years. The time of PSA failure was taken to be the time of the first of two detectable PSA measurements postoperatively.
All data were expressed as mean ± standard deviation. The SPSS 10.0 for Windows program package was used for basic statistical calculation. Receiver-operating characteristics (ROC) curves were generated for various factors, plotting sensitivity versus 1 – specificity. Areas under the ROC curves (AUC) were calculated for various factors. The cutoffs of various factors were determined from the point that would provide the greatest sensitivity and specificity. Statistical analysis was performed using the non-parametric Mann–Whitney U test, Fisher’s exact test, Yate’s correction of contingency, Pearson chi-squared test or linear by linear association (trend) as appropriate. To find the most significant variables for clinicopathological parameters using the Cox proportional hazardous model with forward regression multivariate survival analysis. Multivariate survival analyses were carried out with the SPSS-Cox program and the P values were assessed by the Wald statistic. In the first step, the most significant variable is accepted into this model. Then the second significant variable is chosen from other variables when P < 0.05. Actuarial freedom from PSA failure was calculated using the Kaplan–Meier method and comparisons were made using the log rank test. A P value <0.05 was considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
During the follow-up after RRP, PSA failure (PSA +) was noted in 15 patients. Freedom from PSA failure (PSA –) was noted in 40 patients. Clinical characteristics of the 55 patients with stage T1c PC and comparisons between PSA failure and freedom from PSA failure are shown in Table 1. The PSA failure group had a significantly higher PSA level than the freedom from PSA failure group. From TRUS-guided sextant needle and pathology results, the PSA failure group had more cancer-positive biopsy cores than the freedom from PSA failure group, but the difference was not statistically significant. The PSA failure group had a higher percentage of cancer in biopsies and a higher Gleason score than the freedom from PSA failure group, all statistically significant.
|
Pathological findings of RRP specimens from patients with stage T1c PC are shown in Table 2. The whole mount of RRP specimens revealed 37 cases of OCD and 18 cases of ECE. Among the 18 cases of ECE, 14 cases were margin positive, three had seminal vesicle invasion and two had combined lymph node metastasis and seminal vesicle invasion. The PSA failure group had higher Gleason scores than the non-failure group (P = 0.037). The PSA failure group had a higher incidence of extracapsular extension than the freedom from PSA failure group (P = 0.0002). There was a higher frequency of bilateral lobe cancer and more than one focus of cancer in the PSA failure group than the freedom from PSA failure group, but these were not statistically significant. There was a significantly higher frequency of combined peripheral lobe and transitional lobe involvement than the freedom from PSA failure group. PSA failure group had a larger tumor volume than the freedom from PSA group (P = 0.0005). The incidence of PSA failure was 0% (0/11), 28.6% (8/28) and 43.8% (7/16) in tumor volume <0.5, 0.5–3.9 and
4ml, respectively (Table 2).
|
Univariate analyses for predicting the various parameters are shown in Tables 3 and Table 4. The cutoffs of clinical parameters for predicting PSA failure in stage T1c PC from the receiver-operating characteristic (ROC) curve analysis are shown in Table 3. Biopsy Gleason score
6 is the best predictor. The second best predictor of PSA failure in stage T1c PC is PSA density 0.35. Free-to-total PSA ratio 15% increased the relative risk of PSA failure 2.9 times over free-to-total PSA ratio <15%. However, because free PSA was only measured in 41 cases, this did not quite reach statistical significance (P = 0.0645). The cutoffs of pathology parameters for predicting PSA failure in stage T1c PC are shown in Table 4. The three best predictors of PSA failure in stage T1c PC were tumor volume 2.5 ml, extracapsular tumor extension and pathology Gleason score 6. Multivariate survival analysis for the Cox proportional hazardous model in clinical parameter predictors revealed biopsy Gleason score 6 as the best predictor (Table 5). Multivariate survival analysis for Cox proportional hazardous model in pathological parameter predictors revealed only tumor volume 2.5 ml as the best predictor (Table 6). In this model, in the first step extracapsular tumor extension (ECE) was accepted as the most significant variable. However, in the second step tumor volume was chosen as the most significant variable because ECE and tumor volume have multicollinearity and the validity of ECE and tumor volume was mixed up (Table 6). Multivariate survival analysis for the Cox proportional hazardous model in clinical and pathological parameter predictors revealed tumor volume 2.5 ml still as the most significant predictor and PSA 10 ng/ml as the second predictor (Table 7). The 4 year freedom from PSA disease survival for tumor volume <2.5 ml was 95% and for tumor volume 2.5ml was 39%, which was statistically significant (Fig. 1).
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Serum PSA has been adopted to monitor disease status after patients have received RRP (9). An undetectable serum PSA level after RRP is synonymous with tumor-free status. The advantage of utilizing PSA after RRP is early detection of persistent or locally recurrent microscopic disease, which sometimes precedes clinical symptoms by years. Results of RRP for clinically localized PC vary depending on pathological status (9). Using PSA failure as our treatment end point, we evaluated clinicopathological results to identify predictive factors for patients with stage T1c PC receiving RRP.
Stage T1c PC is a unique group which may include insignificant cancer, localized PC, locally advanced PC or metastatic PC. Epstein and co-workers (12,13) defined four categories with T1c PC: 1, insignificant (confined tumor smaller than 0.2 ml with a Gleason score <7); 2, minimal (confined tumor size 0.2 to less than 0.5 ml with a Gleason score <7); 3, moderate (tumor 0.5 ml or larger or capsular penetration with a Gleason score <7); and 4, advanced (capsular penetration with a Gleason score of 7 or positive margins, seminal vesicle or lymph node involvement). In our series, the incidences of insignificant, minimal, moderate and advanced disease were 3.6, 12.7, 40 and 43.6%, respectively. The same group reported, in a T1c PC series from 1988 to 1992, that the incidences of insignificant, minimal, moderate and advanced disease were 16, 10, 37 and 37%, respectively (14). In the series from 1994 to 1996, the incidences were 17, 12, 52 and 19%, respectively (13). The authors suggested that an increase in the detection of early cancers was manifested by an increase in the prevalence of organ-confined disease and a decrease in the positive margin rate (13). Our series had more cases of advanced disease than the above series, because our patients were from a non-screened population and most men are unaware of prostate cancer. In Ogawa et al.’s series of T1c in Japanese patients, the incidence of organ-confined disease was 59% and the incidence of tumor volume <0.5, 0.5–3.9 and 4.0 ml were 22, 49 and 29%, respectively (8). In our series among Taiwanese patients, the incidence of organ-confined disease was 67.2% and the incidence of tumor volume <0.5, 0.5–3.9 and 4.0 ml were 20% (11/55), 50.9% (28/55) and 29.1% (16/55), respectively. The pathological features of stage T1c PC are very similar between the two Asian groups. Although age-specific prevalence of clinical prostate cancers is higher in Caucasians than Asians (4), the incidence of latent PC is similar in Taiwan, Japan and the USA (5–7). The incidence of ECE in T1c PC in other series has been ~30–45%, which is similar to that in our series (32.8%) (2,8,12–14). Likewise, stage T1c PC in Asians (Taiwan and Japan) share similar pathological features to Caucasians. Lerner et al. suggested that disease-free survival in the stage T1c group was similar to that in the clinical stages T1a to T2a group, but significantly better than in the T2b/c group (15). Cookson et al. also reported that stage T1c PC is pathologically similar to stage T2 PC (16).
Risk factors for PSA failure in patients with clinically localized or pathologically organ-confined PC have been studied before (9,17–20). Olumi et al. reported that the calculated volume of PC is a good predictor of biochemical recurrence (PSA failure) in patients with clinical stage T1c PC (17). However, we did not find a good correlation between calculated volume of PC and actual tumor volume measured from whole-mount step section of RRP (data not shown). Epstein et al. reported that the progression rate was 49% for tumors >4 ml and 26% for tumors <4 ml in patients with clinical stage B PC who received RRP (21) or a relative risk of progression of 1.9-fold. In our series, we classified tumor volume 4 ml or <4 ml with relative risk of PSA failure of only 2.1-fold. From the ROC curve, we defined tumor volume 2.5 ml as cutoffs, with a relative risk of PSA failure of 15.6-fold (Table 4), 12.5-fold (Table 6) and 27.5-fold (Table 7) in univariate and multivariate survival analysis. The 4 year freedom from PSA disease survival for tumor volume <2.5 ml was 95% and for tumor volume 2.5 ml was 39%, which was also statistically significant (P < 0.0001).
The pathological status of extracapsular tumor extension was a good predictor of PSA failure in our series of limited cases of T1c PC; however, in multivariate analysis extracapsular tumor extension was not a significant predictor due to a large case overlap between tumor 2.5 ml and presence of ECE. Among 18 cases of ECE, three had seminal vesicle invasion and two had combined lymph node metastasis and seminal vesicle invasion and they were not excluded from this series. All five of these patients developed PSA failure from 3 to 16 months after RRP. Freedom from PSA failure is rare in patients with PC with lymph node metastasis or seminal vesicle invasion following RRP (13).
Partin et al. collected data on 894 men with clinically localized PC receiving RRP and found that the 5 year freedom from PSA failure was 80% for low-grade capsular penetration tumor and 68% for high-grade capsular penetration tumor (9). D’Amico et al. (18) reported that PC patients with pathological organ-confined and margin-negative diseases but preoperative PSA level >10 ng/ml or a pathological Gleason score 7 have significant decrements in short-term freedom from PSA failure. Similar results reported by Kupelian et al. showed that the 5 year freedom from PSA failure survival was 81% for low risk (PSA <10 ng/ml, Gleason score <7) and 40% for high risk (PSA >10 ng/ml, Gleason score 7) (19). Lerner et al. reported that the risk factors of progression (including PSA failure and local recurrence) in pathologically confined PC after RRP included preoperative PSA level, clinical stage, grade and DNA ploidy (20). Epstein et al. also reported that Gleason score was the best predictor of progression (21). Our results revealed that pathology Gleason score is good predictor of PSA failure. The cutoff of Gleason score was 6 according to the ROC curves. The relative risk of PSA failure was 7.4-fold higher for pathology Gleason score 6 versus <6 (P = 0.019) and the relative risk of PSA failure was 2.6-fold higher for pathology Gleason score 7 versus <7 (P = 0.0829). Obviously, from this result, we divided the Gleason score into 6 versus <6 better than Gleason score 7 versus <7. Preoperative biopsy Gleason score is also a good predictor, with the relative risk of PSA failure being 4.8-fold higher for biopsy Gleason score 6 versus <6. We also find preoperative PSA to be a good predictor of PSA failure in multivariate analysis.
From our data, PSA density and free-to-total PSA ratio were borderline predictors. Zentner et al. defined PSA density 0.3 as a prognostic indicator for PC receiving definitive conformal radiation therapy. They demonstrated that patients with a PSA density <0.3 had 100% actuarial disease-free survival at 30 months compared with 62% for patients with PSA density >0.3 (22). Only 41 of 55 patients received free PSA determinations in our series, hence the role of free-to-total PSA ratio in PSA failure needs the study of more cases.
Although the number of cases was limited and the follow-up period was not long in this series, this is a report relating to patients with stage T1c prostate cancer in an area with low age-specific prevalence of clinical prostate cancers (Taiwan). We evaluated clinicopathological results to identify predictive factors of PSA failure in patients receiving RRP for stage T1c PC. The single most significant predictor was tumor volume 2.5 ml. Other significant predictors included PSA 10 ng/ml, biopsy Gleason score 6, PSAD 0.35, pathology Gleason score 6 and the presence of ECE.
|
Acknowledgement |
|---|
The authors thank Miss Hui-Chin Ho for her assistance with the statistical analyses.
|
FOOTNOTES |
|---|
+ For reprints and all correspondence: Chi-Rei Yang, 160, Sect. 3, Taichung-Kang Road, Division of Urology, Department of Surgery, Taichung Veterans General Hospital, Taichung 40705, Taiwan. E-mail: ycou@vghtc.vghtc.gov.tw
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 Brawer MK. Prostate-specific antigen. Current status. CA Cancer J Clin 1999;49:264–81.[Abstract]
2 Pound CR, Walsh PC, Epstein JI, Chan DW, Partin AW. Radical prostatectomy as treatment for prostate-specific antigen-detected stage T1c prostate cancer. World J Urol 1997;15:373–7.[ISI][Medline]
3 Humphrey PA, Keetch DW, Smith DS, Shepherd DL, Catalona WJ. Prospective characterization of pathological features of prostatic carcinomas detected via serum prostate specific antigen based screening. J Urol 1996;155:816–20.[ISI][Medline]
4 Boring CC, Squires TS, Tong T. Cancer statistics, 1992. Cancer 1992;42:19–39.
5 Yang CR, Ou YC, Ho HC, Kao YL, Cheng CL, Chen JT, et al. Unsuspected prostate carcinoma and prostatic intraepithelial neoplasm in Taiwanese patients undergoing cystoprostatectomy. Mol Urol 1999;3:33–9.[ISI][Medline]
6 Yatani R, Shiraishi T, Nakakuki K, Kusano I, Takanari H, Hayashi T, et al. Trends in frequency of latent prostate carcinoma in Japan from 1965–1979 to 1982–1986. J Natl Cancer Inst 1988;80:683–7.
7 Kabalin JN, Mcneal JE, Price HM, Freiha FS, Stamey TA. Unsuspected adenocarcinoma of prostate in patients undergoing cystoprostatectomy for other cause: incidence, histology and morphometric observations. J Urol 1988;141:1091–4.
8 Ogawa O, Egawa S, Arai Y, Tobisu KI, Yoshida O, Kato T. Preoperative predictors for organ-confined disease in Japanese patients with stage T1c prostate cancer. Int J Urol 1998;5:454–8.[Medline]
9 Partin AW, Pound CR, Clemens JQ, Epstein JI, Walsh. PC. Serum PSA after anatomic radical prostatectomy. Urol Clin N Am 1993;20:713–25.[ISI][Medline]
10 Stamey TA. Making the most out of six systematic sextant biopsies. Urology 1995;45:2–12.[ISI][Medline]
11 Arnauld V, John EM, Fuad SF, Stamey TA. Multiple cancers in the prostate. Morphological features of clinically recognized versus incidental tumors. Cancer 1992;70:2313–8.[ISI][Medline]
12 Carter HB, Sauvageot J, Walsh PC, Epstein JI. Prospective evaluation of men with stage T1c adenocarcinoma of prostate. J Urol 1997;157:2206–9.[ISI][Medline]
13 Epstein JI, Pizov G, Walsh PC. Correlation of pathological findings with progression after radical retropubic prostatectomy. Cancer 1993;71:3582–93.[ISI][Medline]
14 Epstein JI, Walsh PC, Carmichael MJ, Brendler CB. Pathological and clinical findings to predict tumor extent of nonpalpable (stage T1c) prostate cancer. J Am Med Assoc 1994;270:368–74.
15 Lerner SE, Seay TM, Bluye ML, Bergstralh EJ, Barrett D, Zincke H. Prostate specific antigen detected prostate cancer (clinical stage T1c): an interim analysis. J Urol 1996;155:821–6.[ISI][Medline]
16 Cookson MS, Fleshner NE, Soloway SM, Fair WR. Prognostic significance of prostate-specific antigen in stage T1c prostate cancer treated by radical prostatectomy. Urology 1997;49:887–93.[ISI][Medline]
17 Olumi AF, Richie JP, Schultz DJ, D’Amico AV. Calculated volume of prostate cancer identifies patients with clinical stage T1c disease at high risk of biochemical recurrence after radical prostatectomy: a preliminary study. Urology 2000;56:273–7.[ISI][Medline]
18 D’Amico AV, Whittington R, Malkowicz SB, Schultz D, Tomaszewski JE, Wein A. Prostate-specific antigen failure despite pathologically organ-confined and margin-negative prostate cancer: the basis for an adjuvant therapy trial. J Clin Oncol 1997;15:1465–9.[Abstract]
19 Kupelian PA, Katcher J, Levin HS, Klein EA. Stage T1-2 prostate cancer: a multivariate analysis of factors affecting biochemical and clinical failure after radical prostatectomy. Int J Radiat Oncol Biol Phys 1997;37:1043–52.[ISI][Medline]
20 Lerner SE, Blute ML, Bergstralh EJ, Bostwick DG, Eickholt JT, Zincke H. Analysis of risk factors for progression in patients with pathologically confined prostate cancers after radical retropubic prostatectomy. J Urol 1996;156:137–43.[ISI][Medline]
21 Epstein JI, Carmichael M, Partin AW, Walsh PC. Is tumor volume an independent predictor of progression following radical prostatectomy? A multivariate analysis of 185 clinical stage B adenocarcinoma of the prostate with 5 years of followup. J Urol 1993;149:1478–81.[ISI][Medline]
22 Zentner PG, Pao LK, Benson MC, McMahon DJ, Schiff PB. Prostate-specific antigen density: a new prognostic indicator for prostate cancer. Int J Radiat Oncol Biol Phys 1993;27:47–55.[ISI][Medline]
Received May 27, 2002; accepted September 4, 2002
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||