Salvage radiotherapy after radical prostatectomy for prostate adenocarcinoma: analysis of efficacy and prognostic factors

ADULT UROLOGY

SALVAGE RADIOTHERAPY AFTER RADICAL PROSTATECTOMY FOR PROSTATE ADENOCARCINOMA: ANALYSIS OF EFFICACY AND PROGNOSTIC FACTORS ASHISH K. CHAWLA, HARJOT K. THAKRAL, ANTHONY L. ZIETMAN,

AND

WILLIAM U. SHIPLEY

ABSTRACT Objectives. To determine the probability of biochemical control for patients treated with salvage irradiation and identify prognostic factors associated with successful salvage. The optimal management of prostate cancer in patients with an elevated prostate-specific antigen (PSA) level after radical prostatectomy remains unclear. Methods. We reviewed the records of 54 patients with node-negative prostate cancer treated with radiotherapy alone between 1991 and 1998 for isolated biochemical relapse after prostatectomy. The median preoperative PSA level was 15 ng/mL, and the median salvage PSA level was 1.3 ng/mL. Complete pathologic information was recorded, as was the interval to postoperative PSA failure. Radiotherapy was delivered to the prostatic fossa using appropriate techniques. The primary endpoint was biochemical failure, measured from radiotherapy initiation to the first detectable PSA level. Biochemical control rates were determined using Kaplan-Meier methods. The median follow-up was 45 months. Results. The initial complete response rate was 76%. Only seminal vesicle status demonstrated borderline significance for the rate of the initial complete response. The 5-year actuarial biochemical control rate was 35%. The presence of seminal vesicle invasion, Gleason score greater than 6, and an immediately detectable postoperative PSA level all predicted for decreased 5-year biochemical control. Gleason score and detectable postoperative PSA retained significance on multivariate analysis. Those with a salvage PSA level of 1.2 ng/mL or less had a trend toward a decreased 5-year biochemical control rate (P ⫽ 0.07). Conclusions. Salvage radiotherapy yields a 76% complete response rate, with 35% of treated patients free of a detectable PSA at 5 years. Those with favorable biochemical and pathologic tumor features are most likely to remain disease free. UROLOGY 59: 726–731, 2002. © 2002, Elsevier Science Inc.

M

ore than 40,000 radical prostatectomies are performed yearly for the attempted cure of localized prostate adenocarcinoma.1,2 Surgical resection alone is effective for pathologically confirmed organ-confined disease.3 A significant proportion of patients, however, are surgically upstaged because of extracapsular extension or seminal vesicle invasion (pT3 disease).4 – 6 These patients are at a substantial risk of persistent or recurrent disease after prostatectomy. From the Department of Radiation Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts Reprint requests: Ashish K. Chawla, M.D., Department of Radiation Oncology, Massachusetts General Hospital, Blossom Street, Cox 302, Boston, MA 02114 Submitted: November 15, 2001, accepted (with revisions): January 16, 2002

726

© 2002, ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED

Historical studies have demonstrated significant rates of local and distant clinical failure for patients with surgically resected pT3N0 prostate cancer.7–10 With the advent of prostate-specific antigen (PSA) testing, residual cancer is often detected after prostatectomy without any clinical findings. The treatment of such patients is complex, because the evaluation rarely can differentiate between local and distant disease. Local salvage probably benefits only those in whom the disease has not metastasized. Although the current published studies report high initial response rates to salvage irradiation, longer follow-up has demonstrated modest durability of biochemical control.11–14 In addition, the tumor characteristics predicting the likelihood of cure with irradiation have varied somewhat among different institutional experiences. Factors including Gleason score, seminal vesicle status, 0090-4295/02/$22.00 PII S0090-4295(02)01540-6

salvage PSA level, and surgical margin status have been shown in varying degrees to influence biochemical control in these analyses.11–17 To contribute further to the body of data regarding these issues, we report an update of our experience with salvage radiotherapy after prostatectomy. MATERIAL AND METHODS We reviewed the records of 76 patients with prostate cancer treated with salvage irradiation at the Massachusetts General Hospital (68 patients) or Boston Medical Center (8 patients) between 1991 and 1998. One patient was excluded because of inadequate follow-up. Fifteen patients had clinical or radiographic evidence of a recurrent nodule in the prostatic fossa. Twelve patients received androgen suppression therapy before or during radiotherapy, of whom 6 had clinical local recurrence. The exclusion of these subgroups left 54 patients with only biochemical evidence of disease treated with salvage irradiation alone. All patients initially underwent radical prostatectomy and were treated for a persistent or rising PSA level (0.5 ng/mL or greater) after surgery. This PSA value had been the detectability level at our institutions through 1999 and was therefore used for this analysis. This level would not be considered appropriate for rigorous screening in contemporary practice, in which values of 0.1 ng/mL or less are used. Before irradiation, patients were evaluated with physical examination and appropriate imaging studies. No patient had evidence of metastases before irradiation. Of 54 patients, 51 had negative pelvic lymph nodes; the other 3 did not undergo sampling. The initial preoperative PSA level was available in 49 patients (91%). The PSA level at radiotherapy initiation was recorded for all patients. Complete pathologic information for all prostatectomy specimens, including T stage, margin status, Gleason score, perineural invasion status, and seminal vesicle status was recorded. For 47 patients with adequate PSA follow-up between radical prostatectomy and irradiation (either persistent postoperative detectable PSA or a documented undetectable PSA within 12 months of recurrence), the time to PSA failure after surgery was also recorded. Radiotherapy was delivered to the prostatic fossa to a median dose of 64.8 Gy in 1.8-Gy daily fractions. A four-field contoured technique was used with high-energy beams. Rectal barium, cystography, and urethrography were used to define the target and nontarget tissues. After irradiation, patients were followed up every 3 to 6 months by a urologist and/or radiation oncologist with physical examination and PSA analysis. Additional testing to exclude distant disease was performed as warranted. For patients with biochemical or clinical failure after irradiation, hormonal therapy was given at the clinician’s discretion. The interval from irradiation to subsequent androgen suppression was recorded. The median follow-up from radiotherapy initiation was 45 months (mean 53; standard deviation 28). Forty-eight patients were followed up for more than 2 years; 28 of these had a sustained response of at least 2 years’ duration. Our primary endpoint was biochemical failure, defined as the failure to reach an undetectable PSA level after irradiation or two consecutive detectable PSA levels after initial control. The initial complete response rate (percentage of patients achieving an undetectable PSA after irradiation) was recorded. Differences in the complete response rates were compared among the patient subgroups using chi-square analysis. To assess response durability, 5-year biochemical diseasefree (bNED) rates were determined actuarially using the Kaplan-Meier method, with time to failure measured from radiotherapy initiation to the first detectable PSA. The differUROLOGY 59 (5), 2002

TABLE I. Patient and tumor characteristics Age (yr) Median Range Preoperative PSA (ng/mL) Median Range Salvage PSA (ng/mL) Median Range Total dose (cGy) Median Range T stage pT2 pT3 Gleason score ⱕ6 7 ⱖ8 Seminal vesicles Positive Negative Perineural invasion Positive Negative Margins Positive Negative

67 48–78 15 1.2–148 1.3 0.5–30.0 6480 6040–6480 5 (9) 49 (91) 13 (24) 22 (41) 19 (35) 14 (26) 40 (74) 25 (46) 29 (54) 43 (80) 11 (20)

KEY: PSA ⫽ prostate-specific antigen. Numbers in parentheses are percentages.

ences in the bNED rates among patient subgroups were tested using univariate (log-rank) and multivariate (Cox proportional hazard) analyses. Prognostic factors were then separately analyzed for patients with persistent postoperative disease and those with failure at any later interval. Additionally, the actuarial 5-year rate of freedom from androgen therapy was determined. Chronic urinary and rectal complications (persisting or appearing 90 days or more after treatment) were scored using the Radiation Therapy Oncology Group/Eastern Cooperative Oncology Group late effects scale. Late urinary toxicity was scored with respect to the baseline postoperative symptoms.

RESULTS INITIAL RESPONSE The patient and tumor characteristics are listed in Table I. Of 54 patients, 41 (76%) achieved an undetectable PSA level after radiotherapy. Seminal vesicle status had a borderline significant impact on the initial complete response rate (seminal vesicles positive 57% versus negative 83%, P ⫽ 0.06). The initial complete response rates for all other subgroups were greater than 64%, with no significant differences among these groups (data not shown). 727

FIGURE 1. Actuarial biochemical control rate for all salvage irradiation patients (n ⫽ 54).

SALVAGE THERAPY DURABILITY The overall 5-year actuarial bNED rate was 35% (Fig. 1). On univariate analysis, significant pathologic prognostic factors for biochemical control included seminal vesicle status and Gleason score (Table II). Patients with seminal vesicle invasion had a 5-year bNED rate of 11% compared with 45% for those without seminal vesicle invasion (P ⫽ 0.03). Patients with Gleason score 7 or higher tumors had a 5-year bNED rate of 31% compared with 45% for those with lower grade tumors (P ⫽ 0.04). Within the high-grade subset, no significant difference was found between Gleason 7 and Gleason 8 to 10 tumors (data not shown). A range of PSA cutpoints (from 0.5 to 2.0 ng/mL) was evaluated in a stepwise fashion to determine whether a distinction could be drawn between those with a low and high PSA level at salvage. Patients with a salvage PSA of 1.2 ng/mL or less had a 5-year bNED rate of 43%, compared with 26% for those with a higher salvage PSA level. This was not statistically significant (P ⫽ 0.07). The time to postoperative failure was analyzed as a predictor of control for the 47 patients with adequate postoperative PSA follow-up. No significant difference was found in the biochemical control between patients with failure before or after 12 months (P ⫽ 0.82). A cutoff of 24 months also demonstrated no significant difference (P ⫽ 0.90). Patients with immediately detectable postoperative PSA levels, however, fared worse than those with failure at any later interval (P ⫽ 0.05). A multivariate analysis of the pathologic, biochemical, and clinical factors was performed using 728

the Cox proportional hazards model. Gleason score (P ⫽ 0.01) and postoperative PSA status (undetectable versus detectable) (P ⫽ 0.01) remained statistically significant predictors of the 5-year bNED rate. The salvage PSA level and seminal vesicle status demonstrated a trend toward significance. When the 21 patients with immediate postoperative failure were analyzed separately, no significant prognostic variables were identified. A trend toward improved biochemical control was found for those with a Gleason score of less than 7 (P ⫽ 0.13). Among those with recurrent rather than persistent disease, a significant decrease in biochemical control was noted for patients with initial negative margin status (P ⫽ 0.02). No other factors were found to be significant within this subgroup (data not shown). FREEDOM FROM SUPPRESSIVE HORMONAL THERAPY We also measured the ability of salvage irradiation to delay hormonal therapy for postirradiation biochemical or clinical failure. Hormonal therapy was never administered without at least biochemical failure. Twenty patients received hormonal therapy after irradiation. Actuarially, the 5-year freedom from hormonal therapy was 61%. TOXICITY Late rectal sequelae, seen in 9 patients (17%), were limited to grade 1-2 rectal urgency or bleeding. No grade 3 or 4 bowel complications were seen. Seven patients (13%) had grade 1-2 urinary complications. Five had self-limited hematuria requiring cystoscopy with or without fulguration, and two reported mild worsening of stress incontinence that did not require pads. Two additional patients required surgery for anastomotic site bladder neck obstruction, probably resulting from the initial surgery and irradiation; both did well after surgical intervention. No other urinary complications were demonstrated. Among patients with full urinary continence before irradiation, none developed incontinence after salvage therapy. COMMENT This study describes the postoperative salvage radiotherapy results for patients with biochemical relapse of prostate cancer. By excluding patients with clinical disease and those receiving initial systemic salvage, we demonstrated the efficacy of radiotherapy alone in controlling potential microscopic disease in the prostatic fossa. About three quarters of patients experienced an initial complete response, in concordance with similar studies12–14 and suggesting a component of local failure in most patients. Of those initially responding, UROLOGY 59 (5), 2002

TABLE II. Prognostic factors for biochemical control Prognostic Factor Gleason score ⱕ6 ⬎6 Seminal vesicles Negative Positive Perineural invasion Negative Positive Margins Negative Positive Salvage PSA (ng/mL) ⱕ1.2 ⬎1.2 Time to detectable postoperative PSA (mo) 0 ⬎0 ⱕ12 ⬎12 ⱕ24 ⬎24

5-yr bNED Univariate Multivariate (%) P Value P Value 46 31

0.04

0.01

45 11

0.03

0.10

36 29

0.45

0.73

27 41

0.31

0.76

43 26

0.07

0.08

0.05

0.01

0.82

0.39

0.90

0.12

27 52 38 42 39 34

KEY: bNED ⫽ biochemical disease-free; PSA ⫽ prostate-specific antigen.

however, only about one half were disease free at 5 years. The actuarial curve for biochemical control did not demonstrate a plateau 5 years after radiotherapy, indicating the need for longer follow-up. Our 5-year bNED rate of 35% compares with rates of 20% to 55% in series with a similar followup.12,13,15–17 This variation may be accounted for by differences in the tumor characteristics among the patients treated. Almost three quarters of our patients had Gleason 7 or higher disease, and most had treatment failure within 1 year of prostatectomy. In addition, 12 of 14 patients with seminal vesicle invasion had failure. A significant percentage of our patients with salvage failure were likely those with subclinical metastases, for whom local treatment cannot achieve durable remission. The endpoint used to define failure also influences the apparent biochemical control. We measured the time to failure from irradiation to the first detectable PSA level; a different endpoint (ie, halfway between the last undetectable PSA and the first detectable level) would result in lower apparent control rates. These issues must be considered when comparing contemporary series. Because relapse is common after salvage irradiation, it is important to identify the subgroups likely to remain disease free. In this study, seminal vesicle status, Gleason score, and time to postoperative UROLOGY 59 (5), 2002

failure were important prognostic factors for biochemical control. Seminal vesicle involvement is an established marker for metastatic disease,18 and only 11% of patients with seminal vesicle invasion were disease free actuarially at 5 years. Seminal vesicle status did not retain significance on multivariate analysis, probably because of small patient numbers. The Gleason score also predicted a decrease in 5-year biochemical control, retaining significance on multivariate analysis. This suggests that higher rates of occult metastases for high-grade tumors reduces the efficacy of local salvage. With a 5-year bNED rate of more than 30%, salvage irradiation appears reasonable for patients with Gleason 7 or higher tumors. When analyzing various salvage PSA cutpoints for prognostic value, none were statistically significant. Patients with salvage PSA of 1.2 ng/mL or less, however, were almost twice as likely to be disease free at 5 years as those with higher PSA levels (P ⫽ 0.07). The American Society of Therapeutic Radiology and Oncology consensus panel acknowledged the importance of this factor, recommending radiotherapy before the PSA level reaches 1.5 ng/mL.15 Higher salvage PSA levels likely indicate an increased risk of micrometastases, limiting the efficacy of local salvage. It there729

fore seems prudent to administer irradiation at low PSA levels for the best chance of cure. Patients with failure immediately after prostatectomy are thought most likely to harbor occult distant disease. Partin et al.19 have shown that patients with rising PSA levels in the first year after surgery are less likely to develop clinical local failure than those failing at later intervals. We analyzed the time to postoperative failure and found significantly higher durable salvage rates for patients with recurrent rather than persistent biochemical disease. Patients with immediate postoperative failure were then analyzed separately from those with later recurrence, because these subgroups may be approached differently in the clinical setting. Although the small number of patients within each subgroup precludes definitive conclusions, a significant decrease in biochemical control for patients with negative surgical margins was detected within the recurrent disease subgroup. This suggests that in the setting of a late-rising postoperative PSA, negative margin status helps predict for disease outside the prostatic fossa. Larger studies are required to confirm these findings. Recently published data suggest the salvage radiation dose may also have an impact on biochemical control.12 Because almost all our patients received 64.8 Gy, we could not analyze dose as a prognostic factor. However, we have observed only one clinical local failure using this dose, and published studies on salvage irradiation lack the follow-up required to assess the late effects of higher doses to bladder and rectal tissues. Prospective data will help determine the optimal dose in this setting. The benefit of androgen deprivation after biochemical recurrence remains unproven. To evaluate radiotherapy as a single modality, patients treated with initial androgen deprivation were omitted from our analysis. Currently, the Radiation Therapy Oncology Group is conducting a randomized trial to determine the benefit of concurrent and adjuvant bicalutamide antiandrogen therapy for patients with prostate cancer with postoperative PSA elevation. For patients with salvage irradiation failure, options include observation or androgen ablation to delay disease progression. The common side effects of antiandrogens include weight gain, hot flashes, decreased libido, and fatigue. We therefore analyzed the freedom from long-term hormonal therapy as an additional endpoint. Almost two thirds of patients were predicted to be free of hormonal treatment 5 years after radiotherapy. This study may act as a yardstick against which other series, in which salvage irradiation was not used, can be measured. 730

CONCLUSIONS Salvage irradiation is a well-tolerated, reasonable therapeutic option for patients with biochemical failure after radical prostatectomy. This treatment is more effective for patients with tumor factors suggesting a low metastatic disease risk. Longer follow-up is required to assess the efficacy of this treatment definitively. Our results currently indicate that no more than one third of these patients will ultimately be cured. REFERENCES 1. Zietman AL: Salvage and adjuvant radiation after radical prostatectomy, in Vogelzang NJ, Scardino PT, Shipley WU, et al (Eds): Comprehensive Textbook of Genitourinary Oncology, 2nd ed. Philadelphia, Lippincott Williams & Wilkins, 2000, pp 804 – 813. 2. Jones GW, Mettlin C, Murphy GP, et al: Patterns of care for carcinoma of the prostate gland: results of a national survey of 1984 and 1990. J Am Coll Surg 180: 545–554, 1995. 3. Eastham JA, and Scardino PT: Radical prostatectomy for clinical stage T1 and T2 prostate cancer, in Vogelzang NJ, Scardino PT, Shipley WU, et al (Eds): Comprehensive Textbook of Genitourinary Oncology, 2nd ed. Philadelphia, Lippincott Williams & Wilkins, 2000, pp 723–738. 4. Catalona WJ, and Stein AJ: Staging errors in clinically localized prostatic cancer. J Urol 127: 452– 456, 1982. 5. Lu-Yao GL, Potosky AL, and Albertsen PC, et al: Follow-up prostate cancer treatments after radical prostatectomy: a population-based study. J Natl Cancer Inst 88: 166 –173, 1996. 6. Zincke H, Blute ML, Fallen MJ, et al: Radical prostatectomy for stage A adenocarcinoma of the prostate: staging errors and their implication for treatment recommendations and disease outcome. J Urol 146: 1053–1058, 1991. 7. Gibbons RP, Cole BS, Richardson G, et al: Adjuvant radiotherapy following radical prostatectomy: results and complications. J Urol 135: 65– 68, 1986. 8. Shevlin BE, Mittal BB, Brand WN, et al: The role of adjuvant irradiation following primary prostatectomy, based on histopathologic extent of tumor. Int J Radiat Oncol Biol Phys 16: 1425–1430, 1989. 9. Schellhammer PF: Radical prostatectomy: patterns of local failure and survival in 67 patients. Urology 31: 191–197, 1988. 10. Myers RP, and Fleming TR: Course of localised adenocarcinoma of the prostate treated by radical prostatectomy. Prostate 4: 461– 472, 1983. 11. Morris MM, Dallow KC, Zietman AL, et al: Adjuvant and salvage irradiation following radical prostatectomy for prostate cancer. Int J Radiat Oncol Biol Phys 38: 731–736, 1997. 12. Anscher MA, Clough R, and Dodge R: Radiotherapy for a rising prostate-specific antigen after radical prostatectomy: the first 10 years. Int J Radiat Oncol Biol Phys 48: 369 –375, 2000. 13. Pisansky TM, Kozelsky TF, Myers RP, et al: Radiotherapy for isolated serum prostate specific antigen elevation after prostatectomy for prostate cancer. J Urol 163: 845– 850, 2000. 14. American Society for Therapeutic Radiology and Oncology Consensus Panel: Consensus statements on radiation therapy of prostate cancer: guidelines for prostate rebiopsy after radiation and for radiation therapy with rising prostateUROLOGY 59 (5), 2002

specific antigen levels after radical prostatectomy. J Clin Oncol 17: 1155–1163, 1999. 15. Do T, Parker RG, Do C, et al: Salvage radiotherapy for biochemical and clinical failures following radical prostatectomy. Cancer J Sci Am 4: 324 –330, 1998. 16. Cadeddu J, Partin A, DeWeese T, et al: Long-term results of radiation therapy for prostate cancer recurrence following radical prostatectomy. J Urol 159: 173–178, 1998. 17. Crane C, Rich T, Read P, et al: Preirradiation PSA predicts biochemical disease-free survival in patients treated with

UROLOGY 59 (5), 2002

postprostatectomy external beam irradiation. Int J Radiat Oncol Biol Phys 39: 681– 686, 1997. 18. Anscher MS, and Prosnitz LP: Multivariate analysis of factors predicting local relapse after radical prostatectomy— possible indications for postoperative radiotherapy. Int J Radiat Oncol Biol Phys 21: 941–947, 1991. 19. Partin AW, Pound CR, Pearson JD, et al: Evaluation of serum PSA velocity after radical prostatectomy to distinguish local recurrence from distant metastases. Urology 43: 649 – 659, 1994.

731