Prognostic factors in patients treated with intravesical bacillus Calmette-Guerin for superficial bladder cancer.

Several factors were evaluated for prognostic significance in 104 patients with a history of recurrent superficial bladder cancer treated with a 6-week course of intravesical bacillus Calmette-Guerin. Purified protein derivative skin test reactivity, tumor stage, tumor grade and number of previous tumor recurrences were evaluated in all patients. In addition, the prognostic value of a granulomatous response in the bladder was evaluated in 62 of the 104 patients. A significant correlation was reconfirmed between purified protein derivative reactivity and status free of tumor (p equals 0.041) after additional followup on 62 patients from a previous report. A significant correlation also was observed in the total patient population (p equals 0.054). Over-all, 60 per cent of the 62 patients and 52 per cent of the 104 patients whose purified protein derivative test converted from negative to positive remained free of tumor, compared to only 28 per cent of the 62 patients and 28 per cent of the 104 whose test failed to convert to positive. Mean followup was 29.3 +/- 5.7 months in the 62 patient subgroup and 23.5 +/- 5.8 months in the total 104 patients. Bladder granuloma data were available only for the 62 patient subgroup in the previous report. With extended followup, the significant correlation previously reported between status free of tumor and granulomatous response on bladder biopsy was lost. Over-all, 29 of 37 patients (51 per cent) with granulomas compared to 8 of 25 (32 per cent) without granulomas remained free of tumor (p equals 0.132). Tumor stage and grade, and number of previous tumor recurrences failed to show a significant correlation to status free of tumor. These results show that with extended followup, granulomatous response in the bladder lost its statistical correlation with status free of tumor, while a significant correlation was maintained for purified protein derivative responsiveness. This level of statistical significance was borderline and the purified protein derivative skin test response should not be considered useful as a prognostic indicator in individual patients.