Prediction of platinum-based neoadjuvant chemotherapy response in muscle-invasive bladder cancer using a multifaceted approach

Abstract

Radical cystectomy (RC), preceded by platinum-based neoadjuvant chemotherapy (NAC), is recommended for eligible patients with muscle-invasive bladder cancer (MIBC). Although NAC improves survival, many patients do not respond; in non-responders, it may delay surgery and cause toxicity. Predictive biomarkers to guide NAC remain lacking. Here, high-throughput transcriptomics, immunohistochemistry, and cell transfection experiments were applied to identify biomarkers of cisplatin-based NAC response. Two transcriptome analyses of transurethral resection of bladder tumor (TUR-BT) samples from NAC-naïve patients were conducted, followed by feature selection using Stable Iterative Variable Selection (SIVS). Protein validation was performed by immunohistochemistry on tissue microarrays, and functional validation of a candidate long non-coding RNA was assessed by in vitro overexpression in bladder cancer cells. SIVS identified a gene panel distinguishing chemoresistant from chemosensitive patients (AUROC >0.8). The three leading protein-coding genes ( PYGM , FOXQ1 , LGALS7B ) were validated immunohistochemically but lacked sufficient predictive performance at the protein level. The long non-coding RNA DIO3OS showed only a slight, non-significant reduction in cisplatin sensitivity upon overexpression in 5637 cells. These findings underscore the complexity of predicting NAC response in MIBC and suggest that multi-gene RNA/DNA-based panels may offer more robust patient stratification than single biomarkers.