Outcomes of radical cystectomy in potential candidates for bladder preservation therapy.

Objective To analyze the outcomes of patients with urothelial-type bladder cancer (UBC) who optimally met selection criteria for bladder preservation therapy (BPT) but were treated with radical cystectomy (RC) instead. Methods We identified consecutive patients with clinical stage cT2N0M0 UBC who underwent RC with curative intent at our center. Patients without carcinoma in situ, hydronephrosis, multifocality, or mixed histology were classified as BPT eligible. Patients with ≥1 contraindications were considered BPT ineligible. Clinicopathologic characteristics and survival outcomes for BPT-eligible patients were compared with those of the ineligible patients. Results Of the 275 patients who had cT2N0M0 UBC, 157 (57.1%) were BPT ineligible (carcinoma in situ = 54; hydronephrosis = 77; multifocality = 29; mixed histology = 55; ≥2 contraindications = 51). BPT-eligible and -ineligible patients did not statistically differ with regard to age, sex, race, or neoadjuvant chemotherapy. Of the BPT-eligible patients, 24.1% had occult positive lymph nodes and 36.4% had pT3 or pT4 tumors at RC. On multivariate analysis, mixed histology (odds ratio = 3.18; 95% confidence interval [CI], 1.18-8.56) and progression from noninvasive disease to cT2 (odds ratio = 4.81 [95% CI, 1.67-13.85]) were independently associated with upstaging. Two-year overall survival was higher in BPT-eligible patients (76.7% vs 57.1%; P  = .003; hazard ratio = 0.48 [95% CI, 0.3-0.78]). BPT-eligible patients also had better cancer-specific mortality on competing risk analysis (sub-hazard ratio = 0.46 [95% CI = 0.29-0.72]; P  = .001). Conclusion Substantial clinicopathologic stage discrepancies occurred even in patients seemingly ideal for BPT, which may provide insight into BPT failures. Furthermore, substantial survival discrepancies existed for BPT-eligible and BPT-ineligible patients, underscoring the heterogeneity of cT2 disease. In the absence of randomized trials, comparisons between RC and BPT must factor in selection bias.