PREOPERATIVE RADIOTHERAPY FOR MUSCLE-INVASIVE BLADDER-CARCINOMA - LONG-TERM FOLLOW-UP AND PROGNOSTIC FACTORS FOR 338 PATIENTS

Background. This study was performed to determine the importance of va rious potential prognostic factors in a large cohort of patients with transitional cell carcinoma of the bladder who were treated relatively uniformly at a single institution. Methods. Between 1960 and 1983, 33 8 patients with muscle-invasive bladder carcinoma received preoperativ e radiotherapy (50 Gy in 25 fractions) followed 4-6 weeks later with r adical cystectomy. Lymph node sampling was performed only when suspici ous adenopathy was encountered. Ninety-eight percent of the patients c ompleted the treatment as planned. The median follow-up for those livi ng was 90 months. Results. Actuarial 5- year pelvic control, disease f ree, and overall survival rates were 84, 51, and 44%, respectively, fo r all patients, and 88, 58, and 50%, respectively, for those who treat ment completed. The overwhelming majority of failures were from distan t metastases (43% at 5 years). The pathologic complete response rate w as 42%, and downstaging was seen in 65% of the patients. Univariate ac tuarial analyses revealed clinical stage, clinical perivesical extensi on, tumor size, pretreatment hemoglobin level, pretreatment blood urea nitrogen (BUN) concentration, results of intravenous pyelography, sex , age, pathologic response, and pathologic complete response, correlat ed with disease outcome, A Cox proportional hazards model showed patho logic response (P < 0.0001), clinical stage (P = 0.01), hemoglobin lev el (P < 0.02), pathologic complete response (P < 0.05), and BUN concen tration (P < 0.05), were correlated significantly with pelvic control. When only pretreatment factors were analyzed, clinical stage, hemoglo bin level, and BUN concentration remained the only factors predictive of pelvic control. Similar results were obtained when overall survival was used as the endpoint, except that pathologic complete response an d BUN concentration were replaced by sex as significant covariates. A Cox proportional hazards model using disease free status as the endpoi nt revealed pathologic response and tumor size to be independent predi ctors of patient outcome. Restricting this analysis only to pretreatme nt factors showed that pretreatment hemoglobin and tumor size were the only factors correlated with disease free status. Conclusions. The mo st significant prognostic factor was pathologic response, which correl ated highly with all disease endpoints investigated. The most consiste ntly significant pretreatment factors were hemoglobin level and clinic al stage, although tumor size, sex, and BUN concentration also were in dependent predictors of patient outcome. These factors should be consi dered in patients receiving radiotherapy for bladder preservation.