ABSTRACT
BACKGROUND: Initial studies of preoperative checkpoint inhibition before radical cystectomy (RC) have shown promising pathologic complete responses. We aimed to analyze the survival outcomes of patients enrolled in the PURE-01 study (NCT02736266). PATIENTS AND METHODS: We report the results of the secondary end points of PURE-01 in the final population of 143 patients. In particular, we report the event-free survival (EFS) outcomes, defined as the time from the first cycle of pembrolizumab to radiographic disease progression precluding RC, initiation of neoadjuvant chemotherapy (NAC), recurrence after RC, or death from any cause. Other end points were recurrence-free survival (RFS) and overall survival (OS). Subgroup analyses were carried out, including pathological response category, clinical complete responses (CR) assessed via multiparametric magnetic resonance imaging (mpMRI), and molecular subtyping. Cox regression analyses for EFS were also carried out. RESULTS: After a median [interquartile range (IQR)] follow-up of 23 (15-29) months, 12- and 24-month EFS were 84.5% [95% confidence interval (CI): 78.5-90.9] and 71.7% (62.7-82). The prognosis was favorable across all the different pathological response subgroups, with the exception of ypN+ (N = 21), showing a 24-month RFS (95% CI) of 39.3% (19.2% to 80.5%). A statistically significant EFS benefit was observed in patients with a clinical CR (P = 0.002). Programmed cell-death-ligand-1 combined positive score was significantly associated with longer EFS in multivariable analyses. Four patients refused RC after clinical evidence of CR, and none of them have recurred after a median follow-up of 10 months (IQR: 11-15). The claudin-low subtype displayed a numerically longer EFS after pembrolizumab and RC compared with the other subtypes. CONCLUSIONS: The EFS results from PURE-01 revealed that the immunotherapy effect was maintained post-RC in most patients. Pembrolizumab compared favorably with neoadjuvant chemotherapy, irrespective of the biomarker status. Molecular subtyping may be a useful tool to select the patients who are predicted to benefit the most from neoadjuvant pembrolizumab.