Clinical and pathologic factors were analyzed in 40 patients with localized muscle-invasive bladder carcinoma treated in a prospective bladder-preserving program consisting of transurethral tumor resection, neoadjuvant chemotherapy (methotrexate, cisplatin, and vinblastine [MCV]), and 4,000 cGy radiotherapy with concurrent cisplatin. Patients with biopsy-proven complete response after chemotherapy and 4,000 cGy radiation received full-dose radiotherapy (6,480 cGy) with cisplatin. Cystectomy was recommended to patients with residual disease. Distant metastasis rate was associated with tumor stage and size: 0% in T2 patients, 39% in T3-4 patients (P = .035), 6% for tumors less than 5 cm, and 59% for tumors greater than or equal to 5 cm (P = .002). Risk of bladder tumor recurrence was higher in patients with tumor-associated carcinoma in situ (CIS; 40%) than those without CIS (6%; P = .075). Papillary tumors and solid tumors both had similar treatment outcomes. By multivariate analysis, tumor stage T2 (P = .04) and absence of CIS (P = .03) were significant predictors of complete response; CIS was predictive of local bladder recurrence (P = .07); and tumor size (P = .03), response after chemoradiotherapy (P = .02), and vascular invasion (P = .08) were associated with distant metastasis. Six of eight local bladder tumor recurrences were superficial tumors. The low actuarial distant metastasis rate of T2 patients (0% at 3 years), the 3-year actuarial overall survival rates for T2 (89%) and T3-4 (50%) patients, and the similar treatment outcomes for papillary versus solid tumors are encouraging when compared with published historical controls. These results provide preliminary evidence (median follow-up, 30 months) that the current chemoradiotherapy regimen may have beneficial effects in the treatment of muscle-invasive bladder carcinoma. The true efficacy of neoadjuvant chemotherapy remains to be proven by ongoing randomized trials.