- Email: [email protected]
Postoperative Radiotherapy for Stage pT3 Carcinoma of the Prostate: Improved Local Control
0022-5347/96/1556-1983$03.00/0 TI.E JoL‘KS:ALOF UROLOCY Copyright 0 1996 by AMERICANUROLOGICAL ASSOCIATION, INC,
Vol. 155,1983-1986, June 1996 Printed in U . S A
POSTOPERATIVE RADIOTHERAPY FOR STAGE pT3 CARCINOMA OF THE PROSTATE: IMPROVED LOCAL CONTROL ISABEL SYNDIKUS,* TOM PICKLES, EDMUND KOSTASHUK AND LORNE D. SULLIVAN From the British Columbia Cancer Agency and Department of Surgery, University of British Columbia, Vancouver, British Columbia, Canada
ABSTRACT
Purpose: We determined whether radiotherapy after radical prostatedomy leads to improved results in patients with stage pT3 carcinoma of the prostate. Materials and Methods: In a prospective nonrandomized study of 203 patients with clinical stage T2 prostate cancer treated with radical prostatectomy 88 underwent surgery alone, 89 received early postoperative radiotherapy generally because of pathological stage T3 disease and 26 received delayed radiotherapy for local recurrence. The disease was stage pT3NOK in 135 patients. Results: For patients with pathological stage T3 cancer actuarial local recurrence rates were significantly decreased in the early postoperative radiotherapy group compared to the surgery only group (p = 0.005),while actuarial metastatic rates (p = 0.6) and cause specific survival rates (p = 0.04) were not sigmficantly Merent. Multivariate analysis for all patients in both groups identified adverse features of increased postoperative prostate specific antigen levels, seminal vesicle involvement, lack of postoperative radiotherapy and positive lymph nodes. Late toxicity was severe (Radiation Therapy Oncology Group grade 3 or 4) in 13 surgery only and 17 early postoperative radiotherapy group patients. Of those who were potent postoperatively the incidence of impotence in the early postoperative radiotherapy group was 89% compared to 59% in the surgery only group (p = 0.003).For patients treated with delayed radiation for clinical local recurrence the actuarial local control rate was 54% after 10 years. Conclusions: Local radiotherapy appears to improve local control of stage pT3 cancer but has no impact on overall survival. KEY WORDS: prostatic neoplasms, prostatectomy, radiotherapy, prostate-s&c
After radical prostatectomy for clinically localized prostate cancer approximately 30 to 50% of patients will have extracapsular extension of tumor, positive surgical margins, invasion of seminal vesicles or persistently elevated prostate specific antigen (PSA) indicating residual disease.1.2 Although these patients are at increased risk for local recurrence and metastases, the selection of patients and benefits of adjuvant radiation therapy are not well deiined.”ll In a prospective nonrandomized study we report the outcome of 203 patients with clinical stage T2 cancer treated with radical prostatectomy alone (88),surgery followed by early postoperative radiotherapy (89)or delayed radiotherapy for local recurrence (26).End points were local and distant control, PSA relapse, cause specific survival and complications of treatment. MATERIALS AND METHODS
Patient characteristics. From 1984 to 1989, 203 patients treated with radical prostatectomy for clinically localized prostate cancer without distant metastases were prospectively assessed by 1 of US (E.K.). At the discretion of the surgeons and radiation oncologist, 88 patients underwent surgery alone and 89 received early postoperative radiotherapy. The latter patients were generally selected because of positive surgical margins or stage pT3 disease. Postopera-
antigen
tively, radiation therapy was delayed for an average of 4 months (range 1 to 12) to allow patients to regain urinary continence. No patient had evidence of distant metastases or palpable tumor locally. Preoperative hormonal therapy was given to only 3 patients in each group and was discontinued after 3 to 6 months or at the start of radiotherapy. A total of 26 patients received delayed radiotherapy for a local, biopsy confirmed recurrence after radical prostatectomy. No patient had evidence of metastatic disease and all 26 were treated with radical radiation therapy to the prostate bed at a mean of 3.1years postoperatively (range 0.5to 6.5).These patients were referred at recurrence and are not part of the initial surgery only group. Pathology. All pathological specimens and reports were reviewed independently at referral. Histological grade, evidence of seminal vesicle involvement and disease extending to the surgical resection margin were recorded. Tumor extending through the capsule was classified as extracapsular extension. The 1992 staging system of the International Union Against Cancer was used for analysis. Details of patient characteristics, pathological findings and surgery are listed in table 1. Radiotherapy. External beam radiotherapy was given with a 6, 10 or 25 MeV. linear accelerator to 107 patients or a 60cobaltmachine to 8.All patients received treatment to the prostatic bed only with rotating arcs except for 2 who received radiation to the whole pelvis with a boost to the prostatic bed for a locoregional recurrence. Otherwise, the dose prescription and fractionation plans were the same in the early and late radiation groups. A typical treatment volume (90%isodose) measured 10.5 X 9.6 x 9.3 cm. Daily
Accepted for publication December 1, 1995. Current address: Clatt,t?rbridgeCenter for Oncology, Clatterbridge Rd., Rebington, Wirrai, Merseyside, L63 4JY, United Kingdom. Editor’s Note: This article is the fourth of 5 publishedin this issue for which category 1 CME credits can be earn&. Inf3tructionsfor obcredit#a m given with the qU&lOnS 0x1Pages 2084 and 2065. 1983
1984
POSTOPERATIVE RADIOTHERAPY FOR PROSTATE CANCER
TABLE1. Demographic characteristics, and tumor grade and stage in 203 patients with prostate cancer Surgery Only 88 Total No. pts. 63 (39-731 Median age (range) Median mos. followup (range) 80 (70-130) 40 No. margins pos. 6 No. seminal vesicles pos. 26 No. extracapsular extension No. stage41 PT-2 47 PT3 82 PNO 5 PN 1 1 Pm pT3NOm 43 pT2NONX 40 31/46/11 No. welVmoderateIylpwrly differentiated Ca Median PSA (range): 2.4 (0.2-581 Preop., 65 pts. 0.2 (0.03-17) Postop.. 134 pts. 6 No. given hormones before clinical-radiological fail ure
Early Radiotherapy for Radiotherapy Recurrence 26 63 (53-73) 87 (59-121) 111 146146) 80 18 35 3 53 15
89
63 (49-761
0 89 79 9 1 80 0 18/53/18
3 23 21 2 3 21 2 7/16/3
12 (2-70) Not available 2 (0.1-21) Not available 1 17
fraction size was 2 to 3.1 Gy. (mean 2.76) with treatment administered 4 or 5 days a week, for a total dose of 50 to 55 Gy. (mean 52) in 16 to 20 fractions given in 24 to 33 days. Followup. Patients were seen at yearly intervals after the acute radiation side effects had resolved. At each visit the general progress and side effects of treatment were assessed, and physical examination and a PSA test were performed. PSA levels were measured with the Hybritech assay between 1986 and 1989, the DPI Intermedical assay between 1989 and 1991, and the Abbott IMx assay thereafter. The lower level of assay sensitivity was 2 in 1986 and 1987, 0.2 until 1991 and 0.1 thereafter. Patients were censored for the PSA analysis at the first date of hormonal therapy. Complications were assessed with questionnaires completed yearly by the attending physician for 99% of all patients and with a standardized telephone interview conducted by the radiation oncologist (I. s.) for 84% of all patients. Complications were initially assessed 3 months after completion of all primary therapy and were recorded using the Radiation Therapy Oncology Group scale for late radiation side effects. Any symptom clearly related to a local recurrence was not recorded as a side effect. Additional investigations, such as bone scans or computerized tomography, were done either because of symptoms suggestive of metastatic disease or because of a n elevated PSA at the discretion of the attending physician. Hormonal therapy was instituted only at the time of local, distant or PSA relapse. Statistical analysis. Metastatic disease was diagnosed by a n abnormal test result other than PSA. PSA relapse was defined as doubling of the nadir PSA value and a value of greater than 0.5. PSA relapse alone was not used as a failure criterion. For direct comparison the chi-square and MannWhitney U tests were used. Interval to failure was calculated from the day of surgery until the date of the abnormal examination or test results. Patients were censored for local, distant or initial failure when hormonal manipulation was begun. Curves representing nonparametric estimates of interval to initial evidence of failure were generated by the Kaplan-Meier method. 12 Patients without evidence of cancer were censored at last followup or a t death. Differences between subgroups were tested with the log rank statistic. RESULTS
Patient charactenstics a n d choice of treatment. Patients in the early postoperative radiotherapy group more often had
extracapsular involvement, positive margins and seminal vesicle involvement (table 1, chi-square p <0.00001). In addition, increasing grade ( p = 0.059)and higher postoperative PSA levels (Mann-Whitney u test p = 0.003) were noted more often, and were the reasons why these patients were chosen for early radiation therapy. However, 43 patients in the surgery only group and 80 in the early postoperative radiotherapy group had stage pT3 disease with negative nodes or unknown nodal stage, and they form the basis for most comparisons of outcome in our report. A total of 23 patients received hormonal therapy or underwent orchiectomy before clinical or radiological relapse, and all but 4 of them had increased or increasing PSA postoperatively, Interval to a n d type of initial relapse. For the 123 patients with pathological stage T3NO/X disease the interval to initial failure was significantly delayed in the early postoperative radiotherapy group compared to the surgery only group (log rank p = 0.035, fig. 1). In addition, the pattern of relapse differed with most relapses in the surgery only group being local (log rank p = 0.005, fig. 21, while the actuarial interval to metastatic disease was not different between the 2 groups (log rank p = 0.6). At initial failure 77% of patients in the surgery only group had local recurrence, compared to only
0
8
.t
' .2I-
__
0
2
0
4
6
8
10
years &om surgery
FIG.1. Actuarial interval to initial failure (local or metastatic) in 123 patients with stage pT3NOiX disease treated with early radiation therapy or surgery only. Patients who underwent hormonal manipulation before initial failure were censored on date hormonal therapy began. Median interval to initial failure was 9 years in surgery only group and has not yet been reached in early radiotherapy group. Five-year freedom from initial failure rates were 74% in surgery only group and 93% in early radiotherapy group (log rank p = 0.035).
---
Surgery Only
H
p B
$
5.
---I
I---.
.4
.3. .2.
I ~- .
0 0
2
4
6
8
I0
years from surgery
FIG.2. Actuarial interval to local relapse in 123 patients with stage pT3NOiX disease after radical prostatectomy treated with or without radiotherapy. Patients who underwent hormonal manipulation before initial failure were censored on date hormonal therapy began. Median interval to local relapse was 9 years in surgery only group and has not yet been reached in early radiotherapy group. Five-year freedom from local relapse rates were 79% in surgery only group and loo'% in early radiotherapy group (log rank p = 0.005).
POSTOPERATIVE RADIOTHERAPY FOR PROSTATE CANCER
TABLE2.
Cox m ultiuariate
analysis of prognostic factors for initial
relnpse in 133 paiients in early radiotherapy and surgery only groups for whom complete data were available
-P Value Postop. PSA ing./ml.l Seminal vesicle Early postop. radiotherapy
0.0000 0.0000 0.0057
Relative Adverse Direction of Risk Effect for 1.13 3.13 0.43
Increasing PSA Pos. seminal vesicle ~. Lack of radiation therapy Pos. nodes Pos. margins ~
Lymph nodes 0.024 1.85 0.06 1.7 Pos. margms Extracapsular extension Not significant Pathological grade Not significant __Data were censored at the time of hormonal manipulation.
~
-
-L
1985
Postoperative values were available for 134 patients. The medians and ranges for the groups are shown in table 1. Again, the early postoperative radiotherapy group had significantly higher values (Mann-Whitney U test p = 0.003). The PSA relapse rate was not significantly different between the 2 groups (p = 0.5 for all patients and p = 0.4 for stage pT3NOK cases). When a favorable subgroup of 40 surgery only patients with stage pT2NOK disease was selected, freedom from PSA relapse was observed in 90 and 58% at 5 and 10 years, respectively (fig. 3). Toricity. Toxicity was reported a s t h e long-term morbidity after treatment (table 3). There was no significant difference in either late genitourinary or gastrointestinal toxicity between the groups (chi-square p = 0.16 and p = 0.06), although there were 5 cases of severe late urinary complications in t h e early postoperative radiotherapy group and none i n the surgery only group. Of patients who were potent after surgery 88% in the early postoperative radiotherapy group versus 59% in t h e surgery only group became impotent (chisquare p = 0.003). DISCUSSION
2
4
6
8
10
years from surgery
FIG.3. Freedom from PSA relapse in 40 patients treated with surgery only who had negative surgical margins, no extracapsular extension and no lymph node or seminal vesicle involvement (stage pT2NO). At 5 and 10 years freedom from PSA relapse was observed in 90 and 58% of patients, respectively.
25% in t h e early postoperative radiotherapy group. If t h e remaining 54 patients in these 2 groups are included in the analysis, t h a t is those with stage pT2 or pN1 disease, then a multivariate analysis of factors affecting the risk of initial failure can be performed. Factors identified a s increasing the risk of initial failure were increased postoperative PSA, seminal vesicle involvement, lack of postoperative radiotherapy and positive lymph nodes. Positive surgical margins were of borderline significance (table 2). Overall outcome. Cause specific survival did not differ between t h e 2 groups. The 10-year actuarial survival rate was 89 and 91% for the surgery only and early postoperative radiotherapy groups, respectively (log rank p = 0.4). When considering all causes of death in this elderly population the actuarial 10-year survival rate decreases to 66 and 77Q, respectively (log rank p = 0.5). R a d i o f h w a p yfor recurrence. A total of 26 patients treated with delayed radiotherapy for local recurrence achieved a n actuarial local control rate of 69% a t 5 years and 54% after 10 Years. Lead time bias makes comparisons with other patients problematic. However, when compared to patients with stage PT3NOK disease in the surgery only and early postoperative radiotherapy groups there was no significant difference in interval to metastases ( p = 0.3), although cause specific survival appeared to be borderline significantly worse ( p = 0.06). PSA profile and relapse. During t h e first 3 years of t h e study preoperative PSA tests were not done routinely and. therefore, results were only available for 65 of 177 patients. Values were significantly lower in t h e surgery only group than in t h e early postoperative radiotherapy group (MannWhitney U test, p :0.006, table 1). Because of t h e limited number of studies no further analysis of preoperative PSA results w a s done.
Patients with pathological stage T 3 prostate cancer after radical prostatectomy have a higher risk of local relapse, PSA relapse a n d d e a t h from prostate cancer.'.* If delayed until t h e patient presents with locally recurrent disease, local radiotherapy achieves inferior results for all end points, such as response rate, local control and overall survival compared to early treatment.fi These results could be explained by selection of patients with biologically more aggressive tumors, which not only recur locally but also have a higher metastatic potential. A randomized study (INT 0086, NCIC PR2) evaluating t h e role of early postoperative radiotherapy i s i n progress. Unfortunately, accrual has been slow for t h e first 5 years of t h e study and results other t h a n early toxicity13 have not been reported. I t remains to be seen whether early hormonal therapy14 or postoperative radiotherapy to t h e prostate can improve t h e outcome. Regardless of t h e clinical situation, t h e incidences of urinary incontinence, strictures, radiation induced cystitis and proctitis a r e greater i n treated patients.'j, 10.15-17 Therefore, side effects, cost and time must be weighed against t h e possible benefits. This question was evaluated by several retrospective studies showing a n improvement in local control and normalization of increased PSA values b u t no survival benefit.*-7.9.10.Ifi. l 7 However, small patient numbers and selection bias make general recommendations difficult.'' In these studies interval to local or distant relapse is assumed to be a statistically independent event when assessed by conventional statistical techniques. The computed results may be misleading, since patients who have distant (or local) relapse will generally receive hormonal therapy, which may influence or delay the subsequent local (or distant) relapse. This problem may be overcome by following the suggestions of Gelman e t all2 to evaluate the interval to initial relapse (regardless of site), and then to investigate the proportion of patients by treatment a r m who have local versus distant relapse. Cox multivariate analysis may also be applied to determine which factors predict for initial failure. Our patients with clinical stage T 2 carcinoma of t h e prostate were studied prospectively before surgery. Selection bias could not be avoided, since radiation treatment was not assigned randomly and patients with unfavorable pathological features were selected for treatment. Using t h e aforementioned statistical techniques outlined, radiation therapy given postoperatively on t h e basis of adverse pathological and biochemical reports appears to be associated with a lower local recurrence rate. Although radiotherapy led to a statistically significant decrease i n local
POSTOPERATIVE RADIOTHERAPY FOR PROSTATE CANCER
1986
TABLE 3. Complications after radical prostatectomy with or without radiotheram ~
~~
REFERENCES
1. Partin, A. W., Borland, R. N., Epstein, J . I. and Brendler, C. B.:
~
Influence of wide excision of the neurovascular bundle(s) on prognosis in men with clinically localized prostate cancer with Only established capsular penetration. J. Urol., 150 142, 1993. 2. Zietman, A. L., Shipley, W. U. and Willett, C. G.: Residual disease after radical surgery or radiation therapy for prostate 88 89 26 Total No. pts. cancer. Clinical significance and therapeutic implication. Can0.009 No. complications: 50 48 11 0 cer, suppl., 71: 959, 1993. 31 20 11 1 3. Anscher, M. S. and Prosnitz, L. R.: Postoperative radiotherapy 7 21 4 2 for patients with carcinoma of the prostate undergoing radical 0.16 Urinary complications prostatectomy with positive surgical margins, seminal vesicle (Radiation Therinvolvement and/or penetration through the capsule. J . Urol., apy Oncology 138 1407, 1987. Group grades): 4. Eisbruch, A,, Perez, C. A,, Roessler, E. H. and Lockett. M. A,: 24 22 2 7 Adjuvant irradiation after prostatectomy for carcinoma of the 13 12 4 3 0 5 3 4 prostate with positive surgical margins. Cancer, 73:384, 1994. 0.06 Gastrointestinal com5. Freeman, J. A,, Lieskovsky, G., Cook, D. W., Petrovich, Z., Chen, plications (RadiaS.-C., Groshen, S.and Skinner, D. G.: Radical retropubic prostion Therapy Ontatectomy and postoperative adjuvant radiation for pathologcology Group ical state C (PCNO) prostate cancer from 1976 to 1989: intergrades): mediate findings. J. Urol., 149 1029, 1993. 2 2 9 1 6. Gibbons, R. P., Cole, B. S., Richardson, R. G., Correa, R. J., Jr., 0 0 0 3 4 1 0 0 Brannen, G. E., Mason, J . T., Taylor, W. J . and Hafermann, Potency: M. D.: Adjuvant radiotherapy following radical prostatectomy: No. evaluable 44 7 0.08 32 results and complications. J . Urol., 135 65, 1986. No. impotent at fol39 (89) 19 159) 7(100) 0.003 7. Meier, R., Mark, R., St. Royal, L., Tran, L., Colburn, G. and lowun ($6) Parker, R.: Postoperative radiation therapy after radical pros* Chi-square tests were used to compare the surgery only and early radiotatectomy for prostate carcinoma. Cancer, 7 0 1960, 1992. therapy groups only. 8. Paulson, D. F., M o d , J. W. and Walther, P. J.: Radical prostatectomy for clinical stage T1-2NOMO prostatic adenocarcinoma: long-term results. J. Urol., 144: 1180, 1990. 9. Shevlin, B. E., Mittal, B. B., Brand, W. N. and Shetty, R. M.: The role of adjuvant irradiation following primary prostatectomy, recurrence, it is much less clear whether this was a clinibased on histopathologic extent of tumor. Int. J. Rad. Oncol. cally significant outcome since 25% of the patients in the Biol. Phys., 1 6 1425, 1989. early radiation group had distant metastases after 10 10. Stein, A,, deKernion, J. B., Dorey, F. and Smith, R. B.: Adjuvant years. Our study cannot answer the question whether the radiotherapy in patients post-radical prostatectomy with tumetastatic rate would have been higher if local radiation mor extending through capsule or positive seminal vesicles. had not been given. However, i t is of interest that the rate Urology, 39 59, 1992. of any failure was lower in t h e early radiation group de- 11. Thompson, I. M., Paradelo, J . C., Crawford, E. D., Coltman, C. A. and Blumenstein, B.: An opportunity to determine optimal spite the adverse features that would be expected to portreatment of pT3 prostate cancer: the window may be closing. tend a clearly worse outcome. Because of the identifiability Urology, 44.804, 1994. of local and distance recurrence it is not possible to perform multivariate analyses of prognostic factors separately 12. Gelman, R., Gelber, R., Henderson, I. C., Coleman, C. N. and Harris, J. R.: Improved methodology for analyzing local and for local and distant relapse. distant recurrence. J. Clin. Oncol., 8 548, 1990. The higher rate of severe urinary complications in the 13. Thompson, I. M., Crawford, E. D., Miller, G., Paradelo, J., group receiving early radiotherapy must be balanced against Blumenstein, B., Wolfe, M., Monipour, C., Hayden, K. and the possible gain in terms of local control. The observed Messing, E.: Adjuvant radiotherapy following radical prostacomplication rate was worse than that for the surgery only tectomy for pathologic stage C adenocarcinoma of the prostate: group or results reported after radiation treatment alone,l5 initial evaluation of toxicity. Proc. Ann. Meeting SOC.Clin. Oncol., 11: 212, 1992. possibly related partly to the large individual fraction size, and the rate could perhaps be decreased with use of smaller 14. Narayan, P., h w e , B. A,, Carroll, P. R. and Thompson, I. M.: Neoadjuvant hormonal therapy and radical prostatectomy for fractions. clinical stage C carcinoma of the prostate. Brit. J . Urol., 73 544, 1994. CONCLUSIONS 15. Duncan, W., Warde, P., Catton, C. N., Munro, A. J., Lakier, R., Gadalla, T. and Gospodarowicz, M. K.: Carcinoma of the prosOur study confirmed the efficacy of local radiotherapy for tate: results of radical radiotherapy (1970-1985). Int. J . Rad. the prevention of local recurrence in a high risk group of Oncol. Biol. Phys., 2 6 203, 1993. patients with pathological stage T3 prostate cancer after radical prostatectomy. Because of treatment related toxicity 16. Jacobson, G. M., Smith, J . A,, J r . and Stewart, J. R.: Postoperative radiation therapy for pathologic stage C prostate cancer. we believe that postoperative radiation therapy should be Int. J. Rad. Oncol. Biol. Phys., 13 1021, 1987. reserved for those who are at high risk for local recurrence 17. Kaplan, I. D. and Bagshow, M. A.: Serum prostate-specific antibut at low risk for systemic relapse. This subgroup could not gen after post-prostatectomy radiotherapy. Urology, 39: 401, be identified precisely in our study. 1992. No. Surgery
No. No. Early Radiotherapy p for Value* Radiotherapy Recurrence
Vol. 155,1983-1986, June 1996 Printed in U . S A
POSTOPERATIVE RADIOTHERAPY FOR STAGE pT3 CARCINOMA OF THE PROSTATE: IMPROVED LOCAL CONTROL ISABEL SYNDIKUS,* TOM PICKLES, EDMUND KOSTASHUK AND LORNE D. SULLIVAN From the British Columbia Cancer Agency and Department of Surgery, University of British Columbia, Vancouver, British Columbia, Canada
ABSTRACT
Purpose: We determined whether radiotherapy after radical prostatedomy leads to improved results in patients with stage pT3 carcinoma of the prostate. Materials and Methods: In a prospective nonrandomized study of 203 patients with clinical stage T2 prostate cancer treated with radical prostatectomy 88 underwent surgery alone, 89 received early postoperative radiotherapy generally because of pathological stage T3 disease and 26 received delayed radiotherapy for local recurrence. The disease was stage pT3NOK in 135 patients. Results: For patients with pathological stage T3 cancer actuarial local recurrence rates were significantly decreased in the early postoperative radiotherapy group compared to the surgery only group (p = 0.005),while actuarial metastatic rates (p = 0.6) and cause specific survival rates (p = 0.04) were not sigmficantly Merent. Multivariate analysis for all patients in both groups identified adverse features of increased postoperative prostate specific antigen levels, seminal vesicle involvement, lack of postoperative radiotherapy and positive lymph nodes. Late toxicity was severe (Radiation Therapy Oncology Group grade 3 or 4) in 13 surgery only and 17 early postoperative radiotherapy group patients. Of those who were potent postoperatively the incidence of impotence in the early postoperative radiotherapy group was 89% compared to 59% in the surgery only group (p = 0.003).For patients treated with delayed radiation for clinical local recurrence the actuarial local control rate was 54% after 10 years. Conclusions: Local radiotherapy appears to improve local control of stage pT3 cancer but has no impact on overall survival. KEY WORDS: prostatic neoplasms, prostatectomy, radiotherapy, prostate-s&c
After radical prostatectomy for clinically localized prostate cancer approximately 30 to 50% of patients will have extracapsular extension of tumor, positive surgical margins, invasion of seminal vesicles or persistently elevated prostate specific antigen (PSA) indicating residual disease.1.2 Although these patients are at increased risk for local recurrence and metastases, the selection of patients and benefits of adjuvant radiation therapy are not well deiined.”ll In a prospective nonrandomized study we report the outcome of 203 patients with clinical stage T2 cancer treated with radical prostatectomy alone (88),surgery followed by early postoperative radiotherapy (89)or delayed radiotherapy for local recurrence (26).End points were local and distant control, PSA relapse, cause specific survival and complications of treatment. MATERIALS AND METHODS
Patient characteristics. From 1984 to 1989, 203 patients treated with radical prostatectomy for clinically localized prostate cancer without distant metastases were prospectively assessed by 1 of US (E.K.). At the discretion of the surgeons and radiation oncologist, 88 patients underwent surgery alone and 89 received early postoperative radiotherapy. The latter patients were generally selected because of positive surgical margins or stage pT3 disease. Postopera-
antigen
tively, radiation therapy was delayed for an average of 4 months (range 1 to 12) to allow patients to regain urinary continence. No patient had evidence of distant metastases or palpable tumor locally. Preoperative hormonal therapy was given to only 3 patients in each group and was discontinued after 3 to 6 months or at the start of radiotherapy. A total of 26 patients received delayed radiotherapy for a local, biopsy confirmed recurrence after radical prostatectomy. No patient had evidence of metastatic disease and all 26 were treated with radical radiation therapy to the prostate bed at a mean of 3.1years postoperatively (range 0.5to 6.5).These patients were referred at recurrence and are not part of the initial surgery only group. Pathology. All pathological specimens and reports were reviewed independently at referral. Histological grade, evidence of seminal vesicle involvement and disease extending to the surgical resection margin were recorded. Tumor extending through the capsule was classified as extracapsular extension. The 1992 staging system of the International Union Against Cancer was used for analysis. Details of patient characteristics, pathological findings and surgery are listed in table 1. Radiotherapy. External beam radiotherapy was given with a 6, 10 or 25 MeV. linear accelerator to 107 patients or a 60cobaltmachine to 8.All patients received treatment to the prostatic bed only with rotating arcs except for 2 who received radiation to the whole pelvis with a boost to the prostatic bed for a locoregional recurrence. Otherwise, the dose prescription and fractionation plans were the same in the early and late radiation groups. A typical treatment volume (90%isodose) measured 10.5 X 9.6 x 9.3 cm. Daily
Accepted for publication December 1, 1995. Current address: Clatt,t?rbridgeCenter for Oncology, Clatterbridge Rd., Rebington, Wirrai, Merseyside, L63 4JY, United Kingdom. Editor’s Note: This article is the fourth of 5 publishedin this issue for which category 1 CME credits can be earn&. Inf3tructionsfor obcredit#a m given with the qU&lOnS 0x1Pages 2084 and 2065. 1983
1984
POSTOPERATIVE RADIOTHERAPY FOR PROSTATE CANCER
TABLE1. Demographic characteristics, and tumor grade and stage in 203 patients with prostate cancer Surgery Only 88 Total No. pts. 63 (39-731 Median age (range) Median mos. followup (range) 80 (70-130) 40 No. margins pos. 6 No. seminal vesicles pos. 26 No. extracapsular extension No. stage41 PT-2 47 PT3 82 PNO 5 PN 1 1 Pm pT3NOm 43 pT2NONX 40 31/46/11 No. welVmoderateIylpwrly differentiated Ca Median PSA (range): 2.4 (0.2-581 Preop., 65 pts. 0.2 (0.03-17) Postop.. 134 pts. 6 No. given hormones before clinical-radiological fail ure
Early Radiotherapy for Radiotherapy Recurrence 26 63 (53-73) 87 (59-121) 111 146146) 80 18 35 3 53 15
89
63 (49-761
0 89 79 9 1 80 0 18/53/18
3 23 21 2 3 21 2 7/16/3
12 (2-70) Not available 2 (0.1-21) Not available 1 17
fraction size was 2 to 3.1 Gy. (mean 2.76) with treatment administered 4 or 5 days a week, for a total dose of 50 to 55 Gy. (mean 52) in 16 to 20 fractions given in 24 to 33 days. Followup. Patients were seen at yearly intervals after the acute radiation side effects had resolved. At each visit the general progress and side effects of treatment were assessed, and physical examination and a PSA test were performed. PSA levels were measured with the Hybritech assay between 1986 and 1989, the DPI Intermedical assay between 1989 and 1991, and the Abbott IMx assay thereafter. The lower level of assay sensitivity was 2 in 1986 and 1987, 0.2 until 1991 and 0.1 thereafter. Patients were censored for the PSA analysis at the first date of hormonal therapy. Complications were assessed with questionnaires completed yearly by the attending physician for 99% of all patients and with a standardized telephone interview conducted by the radiation oncologist (I. s.) for 84% of all patients. Complications were initially assessed 3 months after completion of all primary therapy and were recorded using the Radiation Therapy Oncology Group scale for late radiation side effects. Any symptom clearly related to a local recurrence was not recorded as a side effect. Additional investigations, such as bone scans or computerized tomography, were done either because of symptoms suggestive of metastatic disease or because of a n elevated PSA at the discretion of the attending physician. Hormonal therapy was instituted only at the time of local, distant or PSA relapse. Statistical analysis. Metastatic disease was diagnosed by a n abnormal test result other than PSA. PSA relapse was defined as doubling of the nadir PSA value and a value of greater than 0.5. PSA relapse alone was not used as a failure criterion. For direct comparison the chi-square and MannWhitney U tests were used. Interval to failure was calculated from the day of surgery until the date of the abnormal examination or test results. Patients were censored for local, distant or initial failure when hormonal manipulation was begun. Curves representing nonparametric estimates of interval to initial evidence of failure were generated by the Kaplan-Meier method. 12 Patients without evidence of cancer were censored at last followup or a t death. Differences between subgroups were tested with the log rank statistic. RESULTS
Patient charactenstics a n d choice of treatment. Patients in the early postoperative radiotherapy group more often had
extracapsular involvement, positive margins and seminal vesicle involvement (table 1, chi-square p <0.00001). In addition, increasing grade ( p = 0.059)and higher postoperative PSA levels (Mann-Whitney u test p = 0.003) were noted more often, and were the reasons why these patients were chosen for early radiation therapy. However, 43 patients in the surgery only group and 80 in the early postoperative radiotherapy group had stage pT3 disease with negative nodes or unknown nodal stage, and they form the basis for most comparisons of outcome in our report. A total of 23 patients received hormonal therapy or underwent orchiectomy before clinical or radiological relapse, and all but 4 of them had increased or increasing PSA postoperatively, Interval to a n d type of initial relapse. For the 123 patients with pathological stage T3NO/X disease the interval to initial failure was significantly delayed in the early postoperative radiotherapy group compared to the surgery only group (log rank p = 0.035, fig. 1). In addition, the pattern of relapse differed with most relapses in the surgery only group being local (log rank p = 0.005, fig. 21, while the actuarial interval to metastatic disease was not different between the 2 groups (log rank p = 0.6). At initial failure 77% of patients in the surgery only group had local recurrence, compared to only
0
8
.t
' .2I-
__
0
2
0
4
6
8
10
years &om surgery
FIG.1. Actuarial interval to initial failure (local or metastatic) in 123 patients with stage pT3NOiX disease treated with early radiation therapy or surgery only. Patients who underwent hormonal manipulation before initial failure were censored on date hormonal therapy began. Median interval to initial failure was 9 years in surgery only group and has not yet been reached in early radiotherapy group. Five-year freedom from initial failure rates were 74% in surgery only group and 93% in early radiotherapy group (log rank p = 0.035).
---
Surgery Only
H
p B
$
5.
---I
I---.
.4
.3. .2.
I ~- .
0 0
2
4
6
8
I0
years from surgery
FIG.2. Actuarial interval to local relapse in 123 patients with stage pT3NOiX disease after radical prostatectomy treated with or without radiotherapy. Patients who underwent hormonal manipulation before initial failure were censored on date hormonal therapy began. Median interval to local relapse was 9 years in surgery only group and has not yet been reached in early radiotherapy group. Five-year freedom from local relapse rates were 79% in surgery only group and loo'% in early radiotherapy group (log rank p = 0.005).
POSTOPERATIVE RADIOTHERAPY FOR PROSTATE CANCER
TABLE2.
Cox m ultiuariate
analysis of prognostic factors for initial
relnpse in 133 paiients in early radiotherapy and surgery only groups for whom complete data were available
-P Value Postop. PSA ing./ml.l Seminal vesicle Early postop. radiotherapy
0.0000 0.0000 0.0057
Relative Adverse Direction of Risk Effect for 1.13 3.13 0.43
Increasing PSA Pos. seminal vesicle ~. Lack of radiation therapy Pos. nodes Pos. margins ~
Lymph nodes 0.024 1.85 0.06 1.7 Pos. margms Extracapsular extension Not significant Pathological grade Not significant __Data were censored at the time of hormonal manipulation.
~
-
-L
1985
Postoperative values were available for 134 patients. The medians and ranges for the groups are shown in table 1. Again, the early postoperative radiotherapy group had significantly higher values (Mann-Whitney U test p = 0.003). The PSA relapse rate was not significantly different between the 2 groups (p = 0.5 for all patients and p = 0.4 for stage pT3NOK cases). When a favorable subgroup of 40 surgery only patients with stage pT2NOK disease was selected, freedom from PSA relapse was observed in 90 and 58% at 5 and 10 years, respectively (fig. 3). Toricity. Toxicity was reported a s t h e long-term morbidity after treatment (table 3). There was no significant difference in either late genitourinary or gastrointestinal toxicity between the groups (chi-square p = 0.16 and p = 0.06), although there were 5 cases of severe late urinary complications in t h e early postoperative radiotherapy group and none i n the surgery only group. Of patients who were potent after surgery 88% in the early postoperative radiotherapy group versus 59% in t h e surgery only group became impotent (chisquare p = 0.003). DISCUSSION
2
4
6
8
10
years from surgery
FIG.3. Freedom from PSA relapse in 40 patients treated with surgery only who had negative surgical margins, no extracapsular extension and no lymph node or seminal vesicle involvement (stage pT2NO). At 5 and 10 years freedom from PSA relapse was observed in 90 and 58% of patients, respectively.
25% in t h e early postoperative radiotherapy group. If t h e remaining 54 patients in these 2 groups are included in the analysis, t h a t is those with stage pT2 or pN1 disease, then a multivariate analysis of factors affecting the risk of initial failure can be performed. Factors identified a s increasing the risk of initial failure were increased postoperative PSA, seminal vesicle involvement, lack of postoperative radiotherapy and positive lymph nodes. Positive surgical margins were of borderline significance (table 2). Overall outcome. Cause specific survival did not differ between t h e 2 groups. The 10-year actuarial survival rate was 89 and 91% for the surgery only and early postoperative radiotherapy groups, respectively (log rank p = 0.4). When considering all causes of death in this elderly population the actuarial 10-year survival rate decreases to 66 and 77Q, respectively (log rank p = 0.5). R a d i o f h w a p yfor recurrence. A total of 26 patients treated with delayed radiotherapy for local recurrence achieved a n actuarial local control rate of 69% a t 5 years and 54% after 10 Years. Lead time bias makes comparisons with other patients problematic. However, when compared to patients with stage PT3NOK disease in the surgery only and early postoperative radiotherapy groups there was no significant difference in interval to metastases ( p = 0.3), although cause specific survival appeared to be borderline significantly worse ( p = 0.06). PSA profile and relapse. During t h e first 3 years of t h e study preoperative PSA tests were not done routinely and. therefore, results were only available for 65 of 177 patients. Values were significantly lower in t h e surgery only group than in t h e early postoperative radiotherapy group (MannWhitney U test, p :0.006, table 1). Because of t h e limited number of studies no further analysis of preoperative PSA results w a s done.
Patients with pathological stage T 3 prostate cancer after radical prostatectomy have a higher risk of local relapse, PSA relapse a n d d e a t h from prostate cancer.'.* If delayed until t h e patient presents with locally recurrent disease, local radiotherapy achieves inferior results for all end points, such as response rate, local control and overall survival compared to early treatment.fi These results could be explained by selection of patients with biologically more aggressive tumors, which not only recur locally but also have a higher metastatic potential. A randomized study (INT 0086, NCIC PR2) evaluating t h e role of early postoperative radiotherapy i s i n progress. Unfortunately, accrual has been slow for t h e first 5 years of t h e study and results other t h a n early toxicity13 have not been reported. I t remains to be seen whether early hormonal therapy14 or postoperative radiotherapy to t h e prostate can improve t h e outcome. Regardless of t h e clinical situation, t h e incidences of urinary incontinence, strictures, radiation induced cystitis and proctitis a r e greater i n treated patients.'j, 10.15-17 Therefore, side effects, cost and time must be weighed against t h e possible benefits. This question was evaluated by several retrospective studies showing a n improvement in local control and normalization of increased PSA values b u t no survival benefit.*-7.9.10.Ifi. l 7 However, small patient numbers and selection bias make general recommendations difficult.'' In these studies interval to local or distant relapse is assumed to be a statistically independent event when assessed by conventional statistical techniques. The computed results may be misleading, since patients who have distant (or local) relapse will generally receive hormonal therapy, which may influence or delay the subsequent local (or distant) relapse. This problem may be overcome by following the suggestions of Gelman e t all2 to evaluate the interval to initial relapse (regardless of site), and then to investigate the proportion of patients by treatment a r m who have local versus distant relapse. Cox multivariate analysis may also be applied to determine which factors predict for initial failure. Our patients with clinical stage T 2 carcinoma of t h e prostate were studied prospectively before surgery. Selection bias could not be avoided, since radiation treatment was not assigned randomly and patients with unfavorable pathological features were selected for treatment. Using t h e aforementioned statistical techniques outlined, radiation therapy given postoperatively on t h e basis of adverse pathological and biochemical reports appears to be associated with a lower local recurrence rate. Although radiotherapy led to a statistically significant decrease i n local
POSTOPERATIVE RADIOTHERAPY FOR PROSTATE CANCER
1986
TABLE 3. Complications after radical prostatectomy with or without radiotheram ~
~~
REFERENCES
1. Partin, A. W., Borland, R. N., Epstein, J . I. and Brendler, C. B.:
~
Influence of wide excision of the neurovascular bundle(s) on prognosis in men with clinically localized prostate cancer with Only established capsular penetration. J. Urol., 150 142, 1993. 2. Zietman, A. L., Shipley, W. U. and Willett, C. G.: Residual disease after radical surgery or radiation therapy for prostate 88 89 26 Total No. pts. cancer. Clinical significance and therapeutic implication. Can0.009 No. complications: 50 48 11 0 cer, suppl., 71: 959, 1993. 31 20 11 1 3. Anscher, M. S. and Prosnitz, L. R.: Postoperative radiotherapy 7 21 4 2 for patients with carcinoma of the prostate undergoing radical 0.16 Urinary complications prostatectomy with positive surgical margins, seminal vesicle (Radiation Therinvolvement and/or penetration through the capsule. J . Urol., apy Oncology 138 1407, 1987. Group grades): 4. Eisbruch, A,, Perez, C. A,, Roessler, E. H. and Lockett. M. A,: 24 22 2 7 Adjuvant irradiation after prostatectomy for carcinoma of the 13 12 4 3 0 5 3 4 prostate with positive surgical margins. Cancer, 73:384, 1994. 0.06 Gastrointestinal com5. Freeman, J. A,, Lieskovsky, G., Cook, D. W., Petrovich, Z., Chen, plications (RadiaS.-C., Groshen, S.and Skinner, D. G.: Radical retropubic prostion Therapy Ontatectomy and postoperative adjuvant radiation for pathologcology Group ical state C (PCNO) prostate cancer from 1976 to 1989: intergrades): mediate findings. J. Urol., 149 1029, 1993. 2 2 9 1 6. Gibbons, R. P., Cole, B. S., Richardson, R. G., Correa, R. J., Jr., 0 0 0 3 4 1 0 0 Brannen, G. E., Mason, J . T., Taylor, W. J . and Hafermann, Potency: M. D.: Adjuvant radiotherapy following radical prostatectomy: No. evaluable 44 7 0.08 32 results and complications. J . Urol., 135 65, 1986. No. impotent at fol39 (89) 19 159) 7(100) 0.003 7. Meier, R., Mark, R., St. Royal, L., Tran, L., Colburn, G. and lowun ($6) Parker, R.: Postoperative radiation therapy after radical pros* Chi-square tests were used to compare the surgery only and early radiotatectomy for prostate carcinoma. Cancer, 7 0 1960, 1992. therapy groups only. 8. Paulson, D. F., M o d , J. W. and Walther, P. J.: Radical prostatectomy for clinical stage T1-2NOMO prostatic adenocarcinoma: long-term results. J. Urol., 144: 1180, 1990. 9. Shevlin, B. E., Mittal, B. B., Brand, W. N. and Shetty, R. M.: The role of adjuvant irradiation following primary prostatectomy, recurrence, it is much less clear whether this was a clinibased on histopathologic extent of tumor. Int. J. Rad. Oncol. cally significant outcome since 25% of the patients in the Biol. Phys., 1 6 1425, 1989. early radiation group had distant metastases after 10 10. Stein, A,, deKernion, J. B., Dorey, F. and Smith, R. B.: Adjuvant years. Our study cannot answer the question whether the radiotherapy in patients post-radical prostatectomy with tumetastatic rate would have been higher if local radiation mor extending through capsule or positive seminal vesicles. had not been given. However, i t is of interest that the rate Urology, 39 59, 1992. of any failure was lower in t h e early radiation group de- 11. Thompson, I. M., Paradelo, J . C., Crawford, E. D., Coltman, C. A. and Blumenstein, B.: An opportunity to determine optimal spite the adverse features that would be expected to portreatment of pT3 prostate cancer: the window may be closing. tend a clearly worse outcome. Because of the identifiability Urology, 44.804, 1994. of local and distance recurrence it is not possible to perform multivariate analyses of prognostic factors separately 12. Gelman, R., Gelber, R., Henderson, I. C., Coleman, C. N. and Harris, J. R.: Improved methodology for analyzing local and for local and distant relapse. distant recurrence. J. Clin. Oncol., 8 548, 1990. The higher rate of severe urinary complications in the 13. Thompson, I. M., Crawford, E. D., Miller, G., Paradelo, J., group receiving early radiotherapy must be balanced against Blumenstein, B., Wolfe, M., Monipour, C., Hayden, K. and the possible gain in terms of local control. The observed Messing, E.: Adjuvant radiotherapy following radical prostacomplication rate was worse than that for the surgery only tectomy for pathologic stage C adenocarcinoma of the prostate: group or results reported after radiation treatment alone,l5 initial evaluation of toxicity. Proc. Ann. Meeting SOC.Clin. Oncol., 11: 212, 1992. possibly related partly to the large individual fraction size, and the rate could perhaps be decreased with use of smaller 14. Narayan, P., h w e , B. A,, Carroll, P. R. and Thompson, I. M.: Neoadjuvant hormonal therapy and radical prostatectomy for fractions. clinical stage C carcinoma of the prostate. Brit. J . Urol., 73 544, 1994. CONCLUSIONS 15. Duncan, W., Warde, P., Catton, C. N., Munro, A. J., Lakier, R., Gadalla, T. and Gospodarowicz, M. K.: Carcinoma of the prosOur study confirmed the efficacy of local radiotherapy for tate: results of radical radiotherapy (1970-1985). Int. J . Rad. the prevention of local recurrence in a high risk group of Oncol. Biol. Phys., 2 6 203, 1993. patients with pathological stage T3 prostate cancer after radical prostatectomy. Because of treatment related toxicity 16. Jacobson, G. M., Smith, J . A,, J r . and Stewart, J. R.: Postoperative radiation therapy for pathologic stage C prostate cancer. we believe that postoperative radiation therapy should be Int. J. Rad. Oncol. Biol. Phys., 13 1021, 1987. reserved for those who are at high risk for local recurrence 17. Kaplan, I. D. and Bagshow, M. A.: Serum prostate-specific antibut at low risk for systemic relapse. This subgroup could not gen after post-prostatectomy radiotherapy. Urology, 39: 401, be identified precisely in our study. 1992. No. Surgery
No. No. Early Radiotherapy p for Value* Radiotherapy Recurrence