Purpose: The relationship between clinical-to-pathologic downstaging and patient outcome following preoperative radiotherapy was examined, focusing on the mechanism (selection vs. treatment effect) responsible for the benefit seen from such downstaging.
Methods and materials: Three hundred and one patients were treated with preoperative radiotherapy plus cystectomy (PREOP) to a median dose of 50 Gy in 25 fractions between 1960-1983. These patients were compared to 225 patients treated with radical cystectomy, with or without chemotherapy (CYST), between 1984-1990. Multiagent chemotherapy was given to 68% of those in the CYST group and was not given to any in the PREOP group. Lymph node involvement was not formally evaluated in the PREOP group, while 20% had pathologic involvement in the CYST group.
Results: Clinical-to-pathologic downstaging (P < T stage) was found in 73% treated with PREOP and 29% treated with CYST (p < 0.0001, chi-square). The only factors that correlated with P < T staging for the PREOP and CYST groups when each was considered separately were clinical stage, blood urea nitrogen level, and creatinine level (p < 0.05, chi-square). Multivariate logistic regression revealed that treatment (PREOP vs. CYST) correlated independently with P < T staging (p < 0.0001). The relationship of actuarial local control to distant metastasis at 5 years in patients that were downstaged, as stratified by clinical stage and treatment, was then examined. Local control rates for P < T staged T2/T3a patients were independent of treatment (PREOP vs. CYST), while distant metastasis rates were significantly greater for those in the PREOP group. In contrast, P < T staged T3b patients in the PREOP group had significantly better local control and distant metastasis rates.
Conclusions: Significantly higher P < T staging rates were observed with PREOP as compared to CYST, and this was a consequence of the radiotherapy given. The relationship of downstaging from radiotherapy to local control and distant metastasis was contingent on clinical stage. The results of Stage T2/T3a and T3b patients were divergent and supported treatment effect, rather than selection, as the mechanism consistent with the patient outcomes observed.