931P - Potential role of a clinical, taxonomical classification and RNA expression integrated signature to predict response to neoadjuvant platinum-based chemotherapy in muscle-invasive bladder cancer (MIBC) patients

Abstract

BackgroundNeoadjuvant chemotherapy (NAC) followed by cystectomy is the standard of care in MIBC patients, although a limited survival benefit has been obtained compared with cystectomy alone. Pathologic response has been associated with survival, but, unfortunately, neither baseline clinical or pathological variables have demonstrated ability to predict which patients will benefit from NAC, pointing to the need of predictive biomarkers of NAC response to guide treatment decisions. The objective of this retrospective study was to identify a NAC response prediction signature integrating baseline clinical features, taxonomic subtypes, and RNA expression profile in MIBC patients. MethodsFormalin-fixed paraffin-embedded pre-treatment tumor samples were collected by transurethral resection from 112 patients with MIBC stage T2-4N0/+M0 treated with NAC followed by cystectomy. Immunohistochemical-based taxonomic subtypes (BASQ-like, luminal-like, mixed) were established. Gene expression analysis was performed on the Nano String nCounter platform. A custom code set of a 41-gene panel involved in the DNA damage repair (DDR) and immune response pathways was used. Lasso and elastic net penalized logistic regression were performed to identify a predictive signature. Calculation of the area under the ROC curve (AUC) was used to assess the predictive ability of the signature. ResultsA nine-gene RNA expression signature (RAD51, CXCL9, PARP, 53BP1, HERC2, ERCC2, CHEK1, Ku80 and RNF 168 genes) was associated with pathologic response. The highest predictive ability was observed with the integrated clinical-taxonomical-RNA signature with an AUC of 0.66, in comparison to the clinical-taxonomical classification (AUC=0.58) or the clinical signature alone (AUC=0.52). Furthermore, the integrated signature was significantly associated with overall survival (p=0.013). ConclusionsWe have identified a nine-gene RNA expression signature that can help to predict response in MIBC patients treated with NAC. Prospective studies are warranted to validate these results. Legal entity responsible for the studyThe authors. FundingAECC. DisclosureAll authors have declared no conflicts of interest.