Predictive factors for outcome in invasive bladder cancer treated with alternating chemoradiotherapy.

PURPOSE: In order to select patients properly for a bladder preservation program, this retrospective study aimed to evaluate the predictive role of pretreatment- and treatment-related factors in a group of patients with invasive bladder cancer treated with alternating chemoradiotherapy at a single institution. METHODS AND MATERIALS: From 1986 to 1994, 72 patients with invasive bladder cancer, stages T1 poorly differentiated or T2-4M0 refusing surgery or not eligible for surgery, were treated with alternating chemoradiotherapy. Each patient had a pretreatment cystoscopy with an attempted complete transurethral resection of the bladder tumor (TURB). The treatment schedule consisted of chemotherapy (cisplatin, 5-fluorouracil, or methotrexate) alternated with radiotherapy. Over the years, the treatment schedule was modified with respect to the total number of chemotherapy cycles, the type of chemotherapy drugs, the dose per fraction and total dose of radiation therapy, and the presence of a planned treatment gap at midtreatment. Treatments were aligned in order of their received average relative dose intensities of both chemotherapy (ARDICT) and radiotherapy (RDIRT). RESULTS: Twenty-two patients (76%) developed infiltrative bladder recurrences for an estimated 5-year pelvic control rate of 68% +/- 6%; 5-year actuarial survival with intact bladder is 40% +/- 6%. Obstructive uropathy at diagnosis, residual disease after TURB, and ARDICT value equal or below the median were independent predictive factors for pelvic failure, with hazard ratios of 2.87 (95% confidence interval [CI], 1.16-7.04), 8.13 (95% CI, 2.74-24.1), and 3.36 (95% CI, 1.29-8.74), respectively. A more detailed model including interactions among these factors showed that the negative prognostic effect of obstructive uropathy at diagnosis was not modified by ARDICT or TURB resection; on the contrary, the risk of local failure for patients with incomplete TURB was markedly affected by different levels of ARDICT. Also, a trend toward a better local outcome was observed for patients with RDIRT above the median. Hydronephrosis and incomplete TURB were also independent predictors of distant metastases and overall survival, but no effect was found for ARDICT on these endpoints. DISCUSSION: As a result of this analysis we believe that (1) patients with obstructive uropathy should not be offered a bladder-sparing approach, (2) gross total TURB of the primary tumor should be maximized, (3) prompt surgery should be considered for patients with incomplete TURB who are not compliant with the combined-modality treatment, and (4) the intrinsic value of dose intensity of both chemotherapy and radiotherapy should be confirmed in a prospective, controlled study.