The study evaluated whether the use of 3 years as compared with 6 months of androgen suppression therapy (AST) combined with external beam radiation therapy (RT) in the treatment of high-risk prostate cancer was associated with prolonged survival in advanced age men.
A pooled analysis of 311 men enrolled in 3 prospective randomized trials between 1987 and 2000 who received 6 months or 3 years of AST and RT for locally advanced or high-grade localized adenocarcinoma of the prostate comprised the study cohort. Cox regression multivariable analysis was performed adjusting for known prognostic factors to determine whether the treatment received was associated with time to death after randomization. The median age and follow-up was 70 and 5.9 years, respectively, during which 82 (26%) deaths occurred.
Treatment received was not significantly associated with survival time after randomization (adjusted hazard ratio [AHR]: 1.1; 95% confidence interval [CI]: 0.7, 1.8; P = .70), whereas age at randomization (AHR: 1.05; 95% CI: 1.01, 1.09; P = .02) was. The presence of Gleason score 8 to 10 cancers approached significance (AHR: 1.6; 95% CI: 0.9, 2.6; P = .09).
Several randomized trials1–3 have documented a prolongation in the time to prostate-specific antigen (PSA) recurrence, prostate cancer-specific mortality (PCSM), and/or all cause mortality (ACM) when as little as 6 months or as much as 3 years of androgen suppression therapy (AST) and external beam radiation therapy (RT) as compared with RT was used to treat men with higher-risk prostate cancer based on the PSA level (>10 ng/mL), Gleason score (7–10), and/or the 2002 AJCC4 clinical T (tumor)-category (T2b to T4). Therefore, for men with localized high-grade or locally advanced prostate cancer the combination of RT and AST has become a standard of care, although the optimal duration of AST remains uncertain.
Based on this evidence,1–3 the use of AST in conjunction with RT has increased markedly.5 Specifically, an increasing number of men in the US, especially over the age of 80, are receiving AST alone or as part of their management for newly diagnosed or recurrent prostate cancer.6 Yet AST can also cause anemia, an increase in body mass index, insulin resistance, increased arterial stiffness, and less favorable lipid profiles.7–11 These changes may explain the association between the adverse impact on health-related quality of life (HRQOL) and the use of neoadjuvant AST before prostate brachytherapy as compared with prostate brachytherapy monotherapy.12 These changes can also persist after discontinuing AST, especially in men of advanced age.13 Some of these symptoms satisfy the criteria for metabolic syndrome,14 a condition associated with an increased risk of death due to a myocardial infarction (MI), despite the absence of cardiovascular disease or diabetes.15 To that end, a meta-analysis of randomized trials2–3 shows that 6 months of AST administration shortens the time to a fatal MI in men of advanced age.16 In addition, a review of the Surveillance, Epidemiology, and End Results (SEER) and Medicare databases shows an association with new diagnoses of diabetes and cardiac disease and AST use.17 Concomitant with these potential adverse findings is the knowledge that the effective duration of androgen suppression (AS) after the discontinuation of AST increases with advancing age18–20 and increasing durations of AST.21 Therefore, within the frame of a public health policy, evidence is needed to individualize the administration of AST in order to give physicians the ability to determine the minimum duration needed to maximize the prolongation in survival while minimizing the impact on HRQOL for an individual man planning to undergo RT and AST for high-risk localized or locally advanced prostate cancer.
The first attempt toward achieving this goal was provided by the Radiation Therapy Oncology Group (RTOG) 9202 study22 that randomized 1554 men with node-negative and clinical category T2b to T4 prostate cancer to RT with either 4 months or 28 months of AST. In that study, no survival benefit was noted. With regard to the question of 6 months vs 3 years of AST the European Organization for Research and Treatment of Cancer (EORTC) completed a 1117-patient randomized Phase III clinical trial (EORTC 22961)23 comparing the use of 2.5 years of luteinizing hormone-releasing hormone (LHRH) agonist monotherapy to no further AST in men with locally advanced node-negative or node-positive prostate cancer who underwent RT and 6 months of combined AST. The results of that trial will help to determine whether a course of AST beyond 6 months is needed in men undergoing RT and AST for high-risk localized or locally advanced prostate cancer to prolong survival, but may not completely answer the question due to the inclusion of men with node-positive disease. While awaiting the results of that randomized study, the current study was designed to assess whether an association exists between length of survival after randomization and duration of AST administered from a pooled analysis of men who received either 6 months of AST and RT2, 3 vs 3 years of AST and RT1 for node-negative localized high-grade or node-negative locally advanced prostate cancer.
MATERIALS AND METHODS
Study Cohort, Staging, and Treatment
The study cohort comprised 311 men of median age 70 (interquartile range [IQR]: 66, 74) enrolled in 3 randomized trials conducted in Australia and New Zealand by the Trans-Tasman Radiation Oncology Group (TROG 9601;2 N = 125) between 1996 and 2000, in Europe by the EORTC1 (N = 175) between 1987 and 1995, and in the US3 (N = 11) between 1995 and 2001. These men included those who received RT and 6months of AST or RT and 3 years of AST on the TROG and US as compared with the EORTC studies, respectively. These men also had either node-negative and locally advanced (clinical T-category 3 or 4) and any grade prostate cancer or node-negative and clinically localized (clinical T-category 1 or 2) and high-grade (Gleason score 8–10 or World Health Organization [WHO] grade 3) prostate cancer and in addition had a serum PSA level obtained within 1 month before randomization. All patients read and signed an approved internal review board consent form before study entry and had a life expectancy of at least 10 years in the US study3 or 5 years in the TROG and EORTC studies.1, 2 A summary and comparison of the prerandomization patient characteristics and follow-up stratified by treatment received are listed in Table 1.
Table 1. Distribution [Number (%)] and Comparison of the Pretreatment Clinical Characteristics at Randomization and Follow-up of the 175 Men Treated With Radiation Therapy and 3 Years of Androgen Suppression Therapy and the 136 Men Treated With Radiation and 6 Months of Androgen Suppression Therapy
RT and 3 y of AST n=175
RT and 6 mo of AST n=136
RT indicates radiation therapy; AST, androgen suppression therapy; IQR, interquartile range; WHO, World Health Organization; PSA, prostate-specific antigen; T, tumor.
Percentages may not sum to 100% due to rounding.
Wilcoxon 2-sample test compares the distributions of these covariates.24
The details of the RT techniques, doses, and fields have been previously described but the main differences included the use of an initial pelvic field in the EORTC patients but not the TROG and US studies and a final total dose to the tumor volume of 70 Gy in the US and EORTC studies and 66 Gy in the TROG study.1–3 AST consisted of a combination of an LHRH agonist (leuprolide acetate) or goserelin and a nonsteroidal antiandrogen in the US study3 and goserelin in the TROG2 and European1 studies coupled with the nonsteroidal antiandrogen flutamide in the TROG study and cyproterone acetate in the EORTC study. Leuprolide acetate was delivered intramuscularly each month at a dose of 7.5 mg or 22.5 mg every 3 months. Goserelin was administered subcutaneously each month at a dose of 3.6 mg in the TROG2 and EORTC1 studies or at the same monthly dose or 10.8 mg every 3 months in the US study.5 Both LHRH agonists were permitted in the US study5 because they have been shown to have equivalent efficacy in the treatment of prostate cancer.2 Flutamide was taken orally at a dose of 250 mg every 8 hours and starting 1 to 3 days before the LHRH agonist to block the transient increase in testosterone caused by the LHRH agonist. Cyproterone acetate was taken orally at a dose of 50 mg 3 times daily for 1 week before the start of goserelin to also block the testosterone flare reaction and continued for 1 month and then stopped. In the US study3 RT was given during Months 3 and 4 of the 6-month course of AST, whereas in the TROG study2 RT was initiated during the last month of AST and conventional RT began concurrent with the administration of goserelin and cyproterone acetate in the EORTC study.1
Follow-up and event size
Follow-up visits were performed after the end of RT every 3 months for 2 years, every 6 months for an additional 3 years, and then annually thereafter in the US study3 and 2 months after the end of RT, then every 4 months for 2 years, every 6 months for an additional 3 years, and annually thereafter in the TROG study.2 The follow-up visits in the EORTC study1 occurred at 1 month after the end of RT and then every 6 months for 5 years and annually thereafter. The median follow-up (IRQ) for all patients was 5.9 years (4.2, 7.2) during which 82 (26%) deaths occurred. Given 82 events this study has 88% power using a 2-sided P-value with a significance level at .05 to detect a doubling of the median survival or a hazard ratio of 2.0, which was the approximate hazard ratio observed in both the prior US3 and the EORTC1 studies.
The data fields required for study from each cooperative group were pooled into a central database held by the study statistician and then analyzed.
Comparison of Baseline Clinical Characteristics and Follow-up
The initial treatments were compared on their distribution of the pretreatment clinical covariates. The chi-square test24 was used for comparisons of the categorical covariates of treatment, clinical T-category, and biopsy Gleason score, and the Wilcoxon 2-sample test25 was used for comparisons of the continuous variables of age, PSA level, and follow-up.
Assessment of Associations With Survival Time
Cox regression univariable and multivariable analyses26 were performed to assess whether the treatment received, the PSA level, age, and biopsy Gleason score at randomization were significantly associated with the survival time after randomization. For the purpose of the Cox regression analyses the PSA level and age were considered continuous variables, whereas clinical T-category, biopsy Gleason score, and initial treatment were analyzed as categorical variables defined as T-category (3, 4 vs 1, 2), high-grade (Gleason score 8–10 or WHO grade 3) vs all others as the baseline (Gleason score 7 or less or WHO grade 1 or 2) and RT and 3 years of AST vs RT and 6 months of AST (baseline), respectively. The PSA value at randomization was log-transformed so that these values followed a normal distribution. For all categorical variables the cutpoints selected were made before examining the data and were based on established strata.27 For all Cox regression analyses residuals26 were examined to check for the proportional hazard (PH) assumption and no violation of the PH assumption was found. Unadjusted and adjusted hazard ratios (HRs) and the associated 95% confidence intervals (CIs) and P-values for probability estimates of time to death were calculated for all covariates using a Cox proportional hazards model.26
Age-adjusted estimates of survival28 after randomization to RT and 3 years of AST or RT and 6 months of AST were computed and graphically displayed. Comparisons of these estimates across treatments were made using the age-adjusted log-rank test.29 SAS v. 9.1.3 (SAS Institute, Cary, NC) was used for all statistical analyses. All statistical tests were 2-sided and P < .05 was considered statistically significant.
Baseline Characteristics Stratified by Treatment
As shown in Table 1, men who were treated using RT and 3 years of AST as compared with RT and 6 months of AST had a significantly higher proportion of men with clinical category T3,4 disease (91% vs 75%; P <.001), median PSA level (32 vs 17 ng/mL; P< .001), and median age at randomization (71 vs 69 years; P <.001), respectively. However, there was no significant difference in the median follow-up (5.6 vs 5.9 years; P = .61) or proportion of men with Gleason score 8 to 10 cancers (34% vs 40%; P = .28) among the 3-year and 6-month AST arms, respectively.
Factors Associated With Survival Time
Treatment received (adjusted hazard ratio [AHR]: 1.1; 95% CI: 0.7, 1.8; P = .70), clinical T-category 3,4 (AHR: 1.01; 95% CI: 0.50, 2.1; P = .98), and the serum PSA level at randomization (AHR: 1.07; 95% CI: 0.9, 1.3; P = .52) were not significantly associated with survival time after randomization, whereas age at randomization (AHR: 1.05; 95% CI: 1.01, 1.09; P = .02) was. Gleason score 8 to 10 cancers (AHR: 1.6; 95% CI: 0.9, 2.7; P = .09) approached significance, as shown in Table 2.
Table 2. Summary of the Treatment and Baseline Prognostic Factor Unadjusted and Adjusted Hazard Ratios for the Time to All-Cause Mortality After Randomization
As shown in Figure 1, age-adjusted estimates of survival were not significantly different (P = .80) across treatment arms. Specifically, age-adjusted estimates of survival 5 years after randomization were 78% (95% CI: 71, 85) as compared with 83% (95% CI: 75, 90) for men who received RT and 3 years of AST as compared with RT and 6 months of AST, respectively.
Based on the results of several prospective randomized clinical trials1–3 that have documented a cancer-specific and/or overall survival benefit when RT and 6 months to 3 years of AST as compared with RT is used to treat higher-risk prostate cancer, the combination of RT and AST has become a standard of care for men with high-risk disease. As evidence accumulates regarding the toxicity of AST,7–12, 16, 17 especially in advanced age men, the question of whether more than 6 months of AST is needed to achieve prolonged survival remains unanswered. Therefore, the current study was performed to glean insight into what that randomized study is likely to show for the node-negative subset.
In the current study, information regarding baseline prognostic factors and survival times after randomization were compiled from the 3 randomized studies1–3 that documented an overall or cancer-specific survival benefit when men with node-negative and locally advanced or high-grade localized prostate cancer were treated using RT and 6 months2–3 or RT and 3 years of AST1 as compared with RT. The main finding of the current study is that a significant association between treatments with 6 months or 3 years of AST and survival time was not found after adjusting for known prognostic factors. This result may reflect the median age of 70 years of the study cohort and the observation from prior reports that the duration of AS after the completion of AST increases with advancing age.18–20 Specifically for men age 70 or more the median duration of AS below the baseline testosterone levels is 18 months after the administration of 6months of AST result in 2 years of AS.30 This result is also consistent with the randomized RTOG 9202 study22 that also showed no difference in overall survival in 1554 men with high-risk prostate cancer randomized to RT and 4 months as compared with RT and 2 years and 4 months of AST after a median follow-up of approximately 10 years. However, this result is in contrast to what has been observed for breast cancer where adjuvant hormonal therapy of 5-year duration as compared with shorter durations lengthens survival.31
Several issues require further discussion. First, this study provides evidence to support a lack of association between survival time after randomization and the addition of 6 months as compared with 3 years of AST to RT after adjusting for known prognostic factors, but does not prove that these 2 treatments will produce similar overall survival. Only large randomized studies such as those performed by the RTOG,22 EORTC,23 and TROG32 can make definitive statements about the impact of long- vs short-term AST on survival because a randomized study can control for both unknown and known prognostic factors. Second, radiotherapeutic treatment of the pelvic lymph nodes was performed in the EORTC study,1 whereas pelvic lymph nodes were not treated in the TROG2 and US3 studies and the dose was slightly lower in the TROG as compared with US and EORTC study: 66 compared with 70 Gy, respectively. It is noteworthy, however, that despite the additional 2.5-year use of AST and the more extensive RT that encompassed the pelvic lymph nodes in the EORTC as compared with the TROG and US studies, an association with prolonged survival did not emerge. Third, a substantial reduction in death (HR ≈2.0) was observed in the prior EORTC1 and US3 studies of 3 years vs 0 and 6 months vs no AST, respectively, making it likely that a further reduction in death, if present, when increasing AST length from 6 months to 3 years, would be smaller and this study was not powered to measure such a small difference.
Finally, whereas the PSA level and proportion of men with clinical T3,4 disease were significantly higher in men treated with 3 years as compared with 6 months of AST, these factors were not significantly associated with length of survival. However, age was associated with length of survival and the results in Figure 1 showing no difference in time to death when short- or long-course AST was delivered were age-adjusted. In addition, whereas the proportion of men with Gleason score 8 to 10 cancers and the length of median follow-up were not significantly different between the 2 treatment arms, there were numerical increases in both of these parameters of 6% and 4 months, respectively, in men who received the 6-month as compared with a 3-year course of AST. These increases could have contributed to a shorter survival time in men receiving the 6-month as compared with a 3-year course of AST, especially given the near-significant association of Gleason score 8 to 10 cancer and a shorter time to death in the adjusted Cox regression analysis; however, such a difference was not observed. Therefore, despite the use of an additional 2.5 years of AST, treatment of the pelvic lymph nodes, 6% less Gleason 8 to 10 cancers, a median follow-up shorter by 4 months, the age-adjusted survival estimates were not significantly longer for men who received long-term as compared with short-term AST. Taken together, these data support the hypothesis that survival is not likely to be significantly different after either 6 months or 3 years of AST with RT delivered for the treatment of men with node-negative and locally advanced or localized and high-grade prostate cancer.
In conclusion, after adjusting for known prognostic factors, the treatment of node-negative high-risk prostate cancer using RT and 3 years as compared with RT and 6 months of AST was not associated with prolonged survival in men of advanced age. The EORTC randomized trial23 will help to answer whether unknown confounding factors will affect the results of this study.