How to improve patient selection for neoadjuvant chemotherapy in bladder cancer patients candidate for radical cystectomy and pelvic lymph node dissection.

PURPOSE: To improve patient selection for neoadjuvant chemotherapy (NAC) before radical cystectomy (RC) in bladder cancer patients (BCa). METHODS: Retrospective evaluation of 1057 patients with cT2-4N0M0 BCa treated with RC and pelvic lymph node dissection between 1990 and 2018 at 3 referral centers. Adverse pathologic features (APF) were defined as pT3-pT4/pN + disease at RC. A regression tree model (CART) was used to assess preoperative risk group classes. A multivariable logistic regression (MVA) was performed to identify predictors of APF at RC. RESULTS: Median age was 70 years and most of the patients were men (83%). Of the 1057 patients included in our study, 688 (65%) had APF. CART analysis was able to stratify patients into 3 risk groups: low (cT2 and single disease, odds ratio [OR] 0.62), intermediate (cT2 and multiple disease, OR 1.08), and high (cT3-cT4, OR 1.28). On MVA APF were associated with variant histology (odds ratio [OR] 3.97, 95% confidence interval [CI] 1.46-10.83, p = 0.007), multifocality at TUR (OR 2.56, CI 1.27-5.17, p = 0.09), completeness of resection (OR 0.47, CI 0.23-0.96, p = 0.04) and clinical extravesical disease (OR 3.42, CI 1.63-7.14, p = 0.001). CONCLUSION: We defined three pre-operative risk classes. Our results indicate that patients with a cT3-T4 disease are those who might benefit more from NAC whereas those with T2 single disease should be those to whom NAC probably shouldn't be proposed. Given the high rate of understaging in BCa patients, NAC can be proposed in selected cases of cT2/multifocal disease.