Patterns of recurrence and outcomes following induction bacillus Calmette-Guerin for high risk Ta, T1 bladder cancer.

PURPOSE

The standard approach to treatment for patients with high risk Ta, Tis, or T1 bladder cancer that persists or recurs after bacillus Calmette-Guerin is radical cystectomy in medically fit patients. Maintenance bacillus Calmette-Guerin has been shown in both SWOG (Southwest Oncology Group) and EORTC (European Organization for Research and Treatment of Cancer) studies to reduce the probability of disease worsening events. As new drugs come on line and experience with maintenance and combination immunotherapy increases, there may be a tendency to delay definitive local therapy and thereby expose patients to a higher risk of progression to invasive and potentially metastatic disease. We explored a large prospective data set from the SWOG 8507 randomized trial of maintenance bacillus Calmette-Guerin to better understand this risk and specifically to assess the impact of timing of recurrence on survival.

MATERIALS AND METHODS

The database includes 501 evaluable patients who were treated with induction bacillus Calmette-Guerin and then were randomized to maintenance bacillus Calmette-Guerin or observation. Recurrence patterns were defined as early (less than 12 months following randomization) or late (12 or more months after randomization). Patients were identified who underwent cystectomy at any time after induction bacillus Calmette-Guerin. All patients were followed for life for determination of vital status. Outcome measure of overall survival was assessed using Kaplan-Meier analysis and adjustment for covariates was done with proportional hazards models. Survival was defined from date of randomization to death from any cause.

RESULTS

A total of 501 patients were randomized after induction bacillus Calmette-Guerin, of whom 251 had recurrence and 229 died. Of the patients who died 59% had recurrence following randomization. Early recurrence was not associated with a higher risk of death compared to late recurrence (p=0.68). There was no evidence that bacillus Calmette-Guerin affected the relationship of timing of relapse and survival. There was no difference in progression to T2 or greater between early and late recurrence (38 of 117, 32% vs 34 of 134, 25%; p=0.21). Cystectomy was performed infrequently as 56 of 251 patients who had recurrence underwent the operation. Patients who had early recurrence had a slightly higher cystectomy rate than those with late recurrence (32 of 117, 27% vs 24 of 134, 18%; p=0.07). Among 394 patients with no evidence of disease at randomization those who underwent cystectomy for T2 or greater disease had a higher risk of death compared to patients who underwent cystectomy for Tis or T1 disease (HR 1.76; 95% CI 0.77, 4.00; p=0.18).

CONCLUSIONS

There was no association of the timing of recurrence after induction bacillus Calmette-Guerin on long-term survival probability. When patients had early recurrence there was a slightly higher probability of cystectomy but not progression to muscle invasive cancer. When cystectomy was performed the 5-year postoperative survival probability was lower than that reported in contemporary series.