Utility of Clinical Risk Stratification in the Selection of Muscle-Invasive Bladder Cancer Patients for Neoadjuvant Chemotherapy: A Retrospective Cohort Study

Introduction: Level I evidence supports the use of cisplatin-based neoadjuvant chemotherapy (NAC) for muscle-invasive bladder cancer prior to radical cystectomy (RC). On average, 30–40% of patients achieve a complete pathologic response (i.e., stage pT0) after receiving NAC. Some centers risk-stratify patients, suggesting that there may be a higher-risk population that would derive the most benefit from NAC. Recently, a risk-stratification model developed at M.D. Anderson Cancer Center (MDACC) specified criteria for clinical staging and patient selection for NAC. We applied this model to our own RC patient cohort and evaluated our own experience with clinical risk stratification and the effect of NAC on post treatment risk categories. Methods: We retrospectively reviewed the charts of consecutive patients who underwent RC at two institutions between 2004 and 2014 and noted whether or not they received NAC. We determined the clinical stage by reviewing the exam under anesthesia, transurethral resection biopsy (TURBT) pathology, and preoperative imaging. Patients with cT2-T4a node-negative disease were included. Those with sarcomatoid features or adenocarcinoma were excluded. Patients were classified as high risk if they had tumor-associated hydronephrosis, clinical stage≥T3b-T4a disease, variant histology (i.e., micropapillary or small cell), or lymphovascular invasion (LVI), as specified by the MDACC model. Variables were examined for associations with cancer-specific survival (CSS), overall survival (OS), and risk-category reclassification. Results: We identified 166 patients with a median follow-up time of 22.2 months. In all, 117 patients (70.5%) did not receive NAC, 68 (58.1%) of whom we classified as high risk. Among patients not receiving NAC, CSS and OS were significantly decreased in high-risk patients (log-rank test p = 0.01 for both comparisons). The estimated age-adjusted hazard ratios of high-risk classification for cancer-specific and overall death were 3.2 (95% CI: 1.2 to 8.6) and 2.2 (95% CI: 1.1 to 4.4), respectively. On post-RC final pathology, 23 (46.9%) low-risk patients were up-classified to high risk and 17 (25.0%) high-risk patients were down-classified. Complete pathologic responses (pT0) were achieved in 7 (6.0%) patients and partial responses (pT1, pTa, pTis) were achieved in 28 (23.9%) patients. Of the 49 patients who did receive NAC, 43 (87.8%) received cisplatin-based and six (12.2%) received carboplatin-based regimens. Applying the MDACC model, we categorized 41 (83.7%) patients as high risk prior to NAC treatment. On final pathology, 3 (37.5%) low-risk patients were up-classified and 17 (41.5%) high-risk patients were down-classified. Complete pathologic responses (pT0) were seen in 13 (26.5%) patients and partial responses were seen in 10 (20.4%) patients. Although the utilization of NAC was not statistically significantly associated with CSS or OS (log-rank test p > 0.05 for both comparisons), it was associated with a 1.2 times increased odds (95% CI: 0.4 to 2.1) of post-RC reclassification from high to low risk on age-adjusted logistic regression. Conclusions: We found similar results using the clinical risk-stratification model in our cohort and showed that the high-risk category was associated with lower CSS and OS. NAC was associated with a higher probability of risk reclassification from high to low risk.