[en] Purpose/Objective: To determine the relationship between pretreatment (basal) apoptosis levels and clinical-to-pathologic downstaging resulting from preoperative radiotherapy. Materials and Methods: Between 1960-1983, 338 patients were dispositioned to receive preoperative radiotherapy and radical cystectomy for muscle-invasive transitional cell carcinoma of the bladder. Of these, adequate hematoxylin and eosin stained tissue sections for morphologic analysis of apoptosis were available in 158 patients. These patients were treated to a mean dose of 49.4 ± 3.0 Gy (± S.D.) and a median dose of 50 Gy. The average fractional dose was 2.0 ± 0.2 Gy with a median of 2.0 Gy. No patient had clinical or radiographic evidence of lymph nodal or distant metastasis, and none received neoadjuvant or adjuvant chemotherapy. The median follow-up for those living was 90 mo. The apoptotic index (AI) was calculated from the ratio of the number of apoptotic cells divided by the total counted and multiplied by 100. The apoptotic cells were counted from several random high powered fields. A minimum of 500 cells were counted from each patient. Results: The average AI for the whole group (n=158) was 2.0 ± 1.3 with a median of 1.8. The association of several potential prognostic factors to AI revealed that AI correlated strongly with clinical stage. The average AI for clinical stage T2 (n=56) was 1.8, for stage T3a (n=51) was 1.9, and for stage T3b (n=51) was 2.4 (p=0.036, Kendall Correlation). The relationship of AI to radiotherapy response also was significant with an average AI of 2.2 for those who were downstaged (n=103), 1.9 for those in whom the stage remained unchanged (n=20), and 1.7 for those who were upstaged (n=35, p=0.054, Kendall Correlation). The only other correlations with AI were for the factors gender (p=0.035) and pretreatment hemoglobin level (p=0.077). The AI was then categorized into 3 groups (≤1, >1 and ≤3, and >3) to examine the prognostic significance of this parameter. This subdivision is such that the middle group approximates the mean ± 1 S.D. The distributions of patients by clinical stage (p=0.014, Mantel-Haenszel), radiotherapy response (p=0.051), hemoglobin level (p=0.051), and gender (0.067) were associated with AI using this grouping. When the analysis of the distribution of patients by radiation response and AI was segregated by stage, a significant correlation was observed only for those with stage T3b disease (p=0.006); 93% of T3b patients with an AI >3 were downstaged, while in 7% the stage remained unchanged and none were upstaged. The relationship of AI to 5 yr actuarial patient outcome was investigated using several endpoints including local control, pelvic control, freedom from distant metastasis, disease freedom, and overall survival. Although AI was not significantly predictive of outcome using these endpoints, a trend was seen for improved survival when AI was >3 for stage T3b patients (71% vs 41%,p =0.09). Conclusion: The AI correlated most strongly with radiotherapy response for patients with clinical stage T3b disease, the one subgroup of patients wherein preoperative radiotherapy is likely to be of the most benefit. The application of AI clinically may be limited to late stage patients, although further studies are needed to more precisely define the utility of this measurement