NEOADJUVANT CISPLATIN CHEMOTHERAPY BEFORE RADICAL CYSTECTOMY IN INVASIVE TRANSITIONAL-CELL CARCINOMA OF THE BLADDER - A PROSPECTIVE RANDOMIZED PHASE-III STUDY

From November 1984 to April 1989, 122 patients with clinical T2-4a Nx-2 M0 transitional cell carcinoma of the bladder were entered in a prospective randomized trial to compare survival between a control group of 60 patients treated only with radical cystectomy (arm A) and a group of 62 patients treated with 3 cycles of 100 mg./m.(2) neoadjuvant cisplatin before radical cystectomy (arm B). Secondary objectives of the trial were comparison of the disease-free interval and time to death, significance of response of the primary tumor to cisplatin, pattern of relapse and toxicity. As of April 1993 after a median followup of 78.2 months (range 48 to 101) no difference in survival between the control patients and those who received neoadjuvant cisplatin has been observed. The overall direct survival is 37.3% for arm A and 35.5% for arm B. The survival rate of the 109 patients who complied with the protocol is 38.2% for 55 patients of the control group and 40.7% for 54 patients of the cisplatin group. Survival rates of patients theoretically rendered free of disease by radical cystectomy (complete response pT0-4a, pN0-2, M0) is 43.7% for 40 control patients and 47.8% for 41 cisplatin treated patients. The time to relapse in complete response patients was significantly longer (p = 0.0298) for those who received cisplatin (arm A 13.1 months versus arm B 30.3 months). The time to death (cause specific) did not differ significantly between both groups overall (p = 0.1349) but it was significantly different between controls and responders (p = 0.0501). Preoperative cisplatin downstaged the primary tumor in 19 patients (33.9%), of whom 11 (19.6%) had no tumor in the cystectomy specimen (pT0) and 8 (14.3%) had superficial tumor (pTis pTa pT1). In 6 patients (9.7%) disease progressed during chemotherapy. The survival of the responders was significantly better than that of nonresponders (p = 0.0142), with specific death rate of 26.3% and 62.5%, respectively, and a median time to death of 43 months for responders and 30.5 months for nonresponders. Patients without nodal involvement (pN0) or with only 1 micrometastasis (pN1) fared significantly better (p 0.0001) than those with major node invasion (pN2-4), irrespective of the treatment received. The survival rate is 48.6% for patients with pN0 disease, 37.5% for pN1 and 5% for pN2-4. Toxicity of cisplatin was minimal and there were no differences in perioperative morbidity between the arms. In conclusion, 3 cycles of neoadjuvant cisplatin did not improve the overall survival of patients undergoing cystectomy for muscle infiltrating transitional cell bladder cancer, although they did prolong significantly the disease-free interval (p = 0.0298). Moreover a subgroup of patients whose primary tumor was downstaged by cisplatin showed a significant survival advantage over nonresponders (p = 0.0142).