Fax: (714) 456-7189
Seminal vesicle involvement in patients with D1 disease predicts early prostate specific antigen recurrence and metastasis after radical prostatectomy and early androgen ablation
Article first published online: 15 MAR 2002
Copyright © 2002 American Cancer Society
Volume 94, Issue 6, pages 1648–1653, 15 March 2002
How to Cite
Ahlering, T. E., Skarecky, D. W., McLaren, C. E. and Weinberg, A. C. (2002), Seminal vesicle involvement in patients with D1 disease predicts early prostate specific antigen recurrence and metastasis after radical prostatectomy and early androgen ablation. Cancer, 94: 1648–1653. doi: 10.1002/cncr.10378
- Issue published online: 15 MAR 2002
- Article first published online: 15 MAR 2002
- Manuscript Accepted: 5 NOV 2001
- Manuscript Received: 15 OCT 2001
- lymph nodes;
- seminal vesicle involvement (SVI)
Controversy persists regarding the management of patients who present with locally advanced metastatic prostate carcinoma. Although radical prostatectomy is not curative, there is growing evidence that survival may be prolonged when the surgery is combined with early androgen ablation. In the current study, the authors present data with which to evaluate and define factors for disease progression in patients undergoing radical prostatectomy with lymph node positive disease who are treated with early endocrine ablation.
Data from 40 patients undergoing radical prostatectomy and early androgen ablation between 1987–1998, all of whom had lymph node positive disease, were analyzed. Age, preoperative prostate specific antigen (PSA) level, clinical and pathologic Gleason score, surgical margin, seminal vesicle involvement (SVI), and the number and percentage of involved positive lymph nodes were analyzed to predict PSA progression, metastasis, and death using univariate and multivariate statistical techniques.
Univariate analysis identified only SVI as a statistically significant predictor of PSA progression and metastasis. Twenty-seven patients (67.5%) were found to have SVI. Multivariate analysis failed to identify other factors that added significantly to the predictive ability of SVI. Kaplan–Meier estimates of time to PSA recurrence and metastasis demonstrated that SVI was highly predictive of disease progression. The median time to PSA progression for the 27 patients with SVI was 7.5 years compared with no progression reported in the 13 patients without SVI (P = 0.011).
VI is a very powerful predictor of disease progression in patients with lymph node positive disease who undergo radical prostatectomy and early androgen ablation. In the current study, preoperative PSA, clinical or pathologic Gleason scores, and other clinical factors were not found to be predictive of disease outcome. Cancer 2002;94:1648–53. © 2002 American Cancer Society.
It has long been realized that the presence of lymph node metastasis is an ominous sign in patients selected to undergo radical prostatectomy. In the early 1980s, nearly 20–30% of men who underwent radical prostatectomy were found to have pelvic lymph node involvement. With the advent of prostate specific antigen (PSA) monitoring and its widespread use for diagnosis, the incidence of positive lymph nodes at the time of surgery has decreased to 6–12%.1–5 Previously, after positive lymph nodes were identified at lymphadenectomy, the majority of surgeons aborted the procedure and offered palliative radiation therapy or hormonal therapy because of the belief the disease was incurable. Over the last decade, radical prostatectomy in lymph node positive patients has been proffered therapeutically to control local disease and possibly prolong survival.
In 1994, Sgrignoli et al.6 examined prognostic factors that would predict early metastasis and shortened survival in patients with lymph node positive disease. The study group was comprised of 113 men, all of whom had positive lymph nodes at the time of radical prostatectomy; however none of these patients received androgen ablation until symptomatic metastasis occurred. Sgrignoli et al.6 reviewed multiple factors and identified only preoperative Gleason score as an independent predictor of distant metastasis. In contrast, Cheng et al.7 published results of a study in which the majority of patients with lymph node positive disease were managed with early androgen ablation after undergoing a radical prostatectomy. The authors found that lymph node metastasis volume was the only significant predictor of disease progression. Factors limiting this study include the fact that not all the patients were managed with early androgen ablation (6%), some of the patients (13%) received adjuvant radiation therapy, seminal vesicle involvement (SVI) was not analyzed independently, and volumetric evaluation of lymph node involvement was not performed readily by community or university pathologists.
Much controversy exists with regard to the timing of androgen ablation in these patients. Many articles have suggested a survival advantage in these patients when early androgen ablation is instituted.4, 8–11 Ultimately, treatment depends on the ideology and preferences of the individual surgeon and patient. In the current study, a number of prognostic factors were evaluated to identify clinical and pathologic factors with regard to their ability to predict PSA progression, metastasis, and death in lymph node positive patients undergoing radical prostatectomy and early androgen ablation.
MATERIALS AND METHODS
Patients and Procedures
Between 1987–1995, 73 patients with pathologic confirmation of lymph node metastasis were identified retrospectively from three institutions (the University of California at Irvine Medical Center, the Long Beach Veterans Administration Medical Center, and the St. Jude Hospital in Fullerton, CA). The surgical technique varied according to the surgical preference but included a pelvic and obturator lymph node resection and a radical retropubic prostatectomy. Adjunctive interventions varied according to the preferences of the patient and surgeon and included radiation therapy and early or delayed hormonal ablation. We identified 40 patients, all of whom were treated with early androgen ablation only, either medical or surgical. Early androgen ablation was defined as that initiated immediately after surgery or at the first indication of an elevation in the PSA level. Thirteen men (33%) underwent orchiectomy at the time the prostatectomy was performed. The remaining 27 men (66%) received hormonal therapy, either bicalutamide or flutamide in combination with leuprolide acetate. Eleven of these men subsequently chose to undergo orchiectomy. All patients who received radiation therapy were excluded from the study.
Using the database we evaluated patient age, preoperative and postoperative PSA level and Gleason score, pathologic stage (particularly surgical margin status and SVI), and the number and percentage of lymph nodes involved. The prostate and seminal vesicle specimens were examined according to standard clinical protocols. None of the institutions performed whole block step sections and, accordingly, the pathologic specimens were not adequate for precise tumor volume estimations or other in-depth pathologic detail (i.e., mode of SVI). Follow-up of current disease status was obtained by a review of medical records and/or telephone interview. Institutional Review Board approval was obtained at all sites.
Patient data are listed in Table 1. After undergoing radical prostatectomy and hormonal ablation, all patients were followed for disease progression with PSA and radiographic evaluation when accompanied by patient-reported symptoms. PSA progression was defined as a reproducible detectable value (> 0.5 ng/mL). Time to metastasis was defined as the interval from surgery to the date of radiographic or biopsy confirmation of metastasis. Gleason scores (clinical or pathologic) were categorized as ≤ 6, 7, or 8–10. The cause of death was attributed to prostate carcinoma in all patients who died with the presence of symptomatic metastasis. One unrelated death was censored.
|All patients (n = 40)||No progression (n = 30)||PSA progression (n = 10)||P value||Metastasis (n = 9)||Death (n = 6)|
|Mean age (yrs)||64.4||64.4||64.4||0.72||64.4||62.7|
|Mean preop PSA (ng/mL)||30.2||27.5||39.3||0.45||43.5||21.0|
|Preop Gleason score < 7||11 (28%)||10||1||1.00||1||1|
|Postop Gleason < 7a||2 (5%)||2||0||0.72||0||0|
|Surgical margin positivea||24 (62%)||17||7||0.71||7||5|
|Surgical margin negative||15 (38%)||12||3||2||1|
|Seminal vesicle positive||27 (68%)||17||10||0.02||9||6|
|Seminal vesicle negative||13 (32%)||13||0||0||0|
|% positive lymph nodes||1.00|
|> 20%||16 (40%)||11||5||4||4|
The following independent variables were subject to univariate analysis (using the Fisher exact tests for association) for PSA progression, metastasis, and death. The variables included patient age, preoperative PSA level, preoperative Gleason score (< 7, 7, or 8–10), postoperative Gleason score (< 7, 7, or 8–10), surgical margin status, SVI, and percentage positive lymph nodes (log of positive lymph nodes). These variables also were analyzed using the log-rank and Wilcoxon tests for homogeneity of the time to PSA recurrence and time to metastasis using the Kaplan–Meier life table analyses. Multivariate analyses utilizing logistic regression and Cox proportional hazards regression for PSA progression and time to PSA progression were performed. Based on the results of univariate analysis, predictors analyzed were narrowed to preoperative Gleason score, postoperative Gleason score, surgical margins, and SVI. Analysis was performed using the SAS statistical package software.12
The average length of follow-up for the 40 patients was 6.8 years (range, 2.2–12.7 years; median, 6.3 years). Thirty of the patients had follow-up > 5 years. At the time of last follow-up, 10 patients had developed a PSA progression, 9 patients had developed metastases, and 6 patients had died of prostate carcinoma. One patient died of other causes 5 years after surgery without a detectable PSA level. In the current study, PSA progression was found to be associated strongly with metastasis (P ≤ 0.0001) and death (P = 0.0001) using the Fisher two-tailed exact test. With regard to the 6 patients who died, the average length of time between PSA progression and death was 1.9 years (range, 0.1–5.0 years). Table 1 lists pertinent patient data.
Of the variables examined using univariate analysis to determine risk of PSA progression, only SVI was found to have a statistically significant association (P = 0.016, Fisher exact test) (Table 1). A total of 27 patients (67%) were found to have SVI; 7 patients had just 1 involved seminal vesicle and 20 patients had bilateral involvement. PSA progression based on preoperative PSA levels, clinical and pathologic Gleason score, and surgical margin status all were found to be not significant according to both log-rank (all P ≥ 0.52) and Wilcoxon (all P ≥ 0.30) statistics. Figure 1 illustrates Kaplan–Meier event curves for SVI and PSA progression. Kaplan–Meier event curves also were generated for metastasis and SVI (Fig. 2). For patients with SVI, the median time to PSA progression was 7.5 years and the median time to metastasis was 8.7 years. Statistical significance for death was not attained; however, with sufficient time significance appears to be inevitable. Multivariate analysis utilizing logistic regression and Cox proportional hazards models failed to identify factors that added a significant additional effect to the predictive value of SVI for PSA progression and metastasis.
The optimal management of patients with lymph node positive disease is continuing to develop. Historically, patients with lymph node positive disease were deemed incurable and were managed palliatively. Since the 1980s some authors have advocated radical prostatectomy as a means to control local disease and possibly prolong survival. Steinberg et al.13 and others14 have reported good control of local disease in patients with lymph node positive disease who are undergoing radical prostatectomy. Some authors suggest a survival advantage in lymph node positive patients who are treated with radical prostatectomy with or without hormonal ablation.7–9, 15–17 Two case–control studies supporting removal of the prostate have been reported by Cadeddu et al.15 and Ghavamian et al.18 Ghavamian et al. demonstrated nearly double the 10-year disease specific survival in 79 patients with pTxN+ disease who were undergoing radical prostatectomy and immediate orchiectomy compared with 79 case-matched men with pTxN+ disease who were treated only with immediate orchiectomy (79% ± 5% vs. 28% ± 6%). Other authors have reported similar retrospective data, suggesting a benefit for prostatectomy in lymph node positive patients.3, 4 Ultimately, the decision to proceed with radical prostatectomy in the presence of lymph node positive disease rests with the patient and surgeon. If radical prostatectomy is performed, one must then decide on the timing of hormonal ablation because there is strong evidence (87% progression at 5 years) that essentially all patients will develop disease progression in time.5
Recently, evidence strongly supporting the role of early androgen ablation was reported. Messing et al.11 published the results of a randomized prospective study of lymph node positive patients who were undergoing radical prostatectomy and compared immediate androgen ablation with androgen ablation performed at the time of symptomatic disease progression. With a mean follow-up of 7.2 years, 6.4% of the early ablation patients had died of prostate carcinoma compared with 30.8% in the delayed androgen ablation arm. This finding strongly suggests that, with time, patients with prostate carcinoma spontaneously will develop clones of androgen-independent cells. Factors that help predict disease progression and death need to be defined to identify at-risk groups and to understand the biology of prostate carcinoma further.
The extent of lymph node involvement has been reported to be a possible predictor of disease progression. A variety of authors have examined the prognostic significance of the number, percentage, largest dimension, and volume of involved lymph nodes.7, 11, 13, 14, 18–24 In general, the results of previous studies have been divided between positive findings7, 13, 14, 19, 20, 23 and negative findings.5, 6, 11, 21, 23, 24 A number of reasons have been offered for this variability including treatment differences, technical issues, patient numbers, and follow-up. In 1998, Cheng et al.7 reported on lymph node positive patients who were treated with radical prostatectomy and immediate hormonal ablation and found that tumor volume in the lymph nodes was the best predictor of disease progression. In a review of 269 patients, they evaluated multiple factors but focused on lymph node tumor volume. The authors did not examine SVI as an independent variable and the follow-up period was only 5 years, which may account for the difference between their results and ours.
We focused on pathologic factors performed commonly in the community. The largest dimension or estimates of tumor volume either in the prostate or lymph nodes is expensive, technically involved, and operator-dependent. The pathologic finding of SVI represents a simple and reproducible tool for essentially all pathologists. Similarly, the percent of lymph nodes involved was used because of its ease and reproducibility technically. It also represents a reasonable approximation of the extent of tumor, although not as accurately as other models such as tumor greatest dimension or volume.
We evaluated factors commonly employed by urologists for their ability to predict disease progression in lymph node positive patients who undergo radical prostatectomy and EET. The results of the current study demonstrate that PSA progression is a very strong predictor of metastasis and death. On average, the patients in the current study with PSA progression developed bone metastasis and died within 1.9 years. Pound et al.1 also reported PSA progression to be a strong predictor of metastasis but with significantly longer intervals, most likely due to hormonal naiveté. Patient age, preoperative PSA and Gleason score, pathologic Gleason score, surgical margin status, and percent of positive lymph nodes were not found to be predictive of PSA progression or metastasis (Table 1) (Fig. 1). Utilizing univariate analysis, SVI was the only factor found to be predictive of PSA progression and metastasis. When analyzed with Kaplan–Meier event curves, SVI was found to be a strong predictor of early PSA progression and metastasis. The median time to PSA progression with SVI was 7.5 years compared with no events reported to have occurred in 13 patients without SVI.
The finding that SVI is a strong predictor of disease progression is not new. Increasing volumes of prostate tumors has been found to be correlated strongly with SVI and the risk of disease progression.2, 3 However, to our knowledge, the significance of concomitant SVI has not been reported in the presence of lymph node metastasis. It is our opinion that concomitant SVI is a bad prognostic finding because it represents a greater volume of disease. Furthermore, we suspect that increasing volumes of disease could result in more androgen-independent clones at the time of initiation of hormonal ablation and hence more rapid disease progression. This retrospective review lacks the methodology to assess accurately the influence of tumor volume and the risk of disease progression.
Radical prostatectomy patients with lymph node positive disease were found to have concomitant involvement of the seminal vesicles 67.5% of the time. Patients with lymph node positive disease treated with early androgen ablation and SVI have been reported to have a significantly greater risk for PSA progression and more ominous outcomes. In the future, delaying hormonal intervention until the PSA becomes detectable may be considered in patients without SVI.
The authors thank John Davis, M.D., Charles Streit, M.D., Ms. Pamela Randall, and the St. Jude Hospital Tumor Registry for their generous help in this project. The authors also thank Chao Cancer Center biostatisticians Khanh Nguyen, Kuo-Tung Li, and Sandhya Upasani for their technical assistance.
- 12SAS Institute, Inc. SAS® STAT Users's Guide. Version 6. Cary, NC: SAS Institute Inc., 1990.