Differences in time to disease progression do not predict for cancer-specific survival in patients receiving immediate or deferred androgen-deprivation therapy for prostate cancer: final results of EORTC randomized trial 30891 with 12 years of follow-up.

BACKGROUND: Trials assessing the benefit of immediate androgen-deprivation therapy (ADT) for treating prostate cancer (PCa) have often done so based on differences in detectable prostate-specific antigen (PSA) relapse or metastatic disease rates at a specific time after randomization.

OBJECTIVE: Based on the long-term results of European Organization for Research and Treatment of Cancer (EORTC) trial 30891, we questioned if differences in time to progression predict for survival differences.

DESIGN, SETTING, AND PARTICIPANTS: EORTC trial 30891 compared immediate ADT (n=492) with orchiectomy or luteinizing hormone-releasing hormone analog with deferred ADT (n=493) initiated upon symptomatic disease progression or life-threatening complications in randomly assigned T0-4 N0-2 M0 PCa patients.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Time to first objective progression (documented metastases, ureteric obstruction, not PSA rise) and time to objective castration-resistant progressive disease were compared as well as PCa mortality and overall survival.

RESULTS AND LIMITATIONS: After a median of 12.8 yr, 769 of the 985 patients had died (78%), 269 of PCa (27%). For patients receiving deferred ADT, the overall treatment time was 31% of that for patients on immediate ADT. Deferred ADT was significantly worse than immediate ADT for time to first objective disease progression (p<0.0001; 10-yr progression rates 42% vs 30%). However, time to objective castration-resistant disease after deferred ADT did not differ significantly (p=0.42) from that after immediate ADT. In addition, PCa mortality did not differ significantly, except in patients with aggressive PCa resulting in death within 3-5 yr after diagnosis. Deferred ADT was inferior to immediate ADT in terms of overall survival (hazard ratio: 1.21; 95% confidence interval, 1.05-1.39; p [noninferiority]=0.72, p [difference] = 0.0085).

CONCLUSIONS: This study shows that if hormonal manipulation is used at different times during the disease course, differences in time to first disease progression cannot predict differences in disease-specific survival. A deferred ADT policy may substantially reduce the time on treatment, but it is not suitable for patients with rapidly progressing disease.