Adipose tissue and myosteatosis measured by artificial intelligence as a predictor of overall survival in bladder cancer

Abstract

Body composition may influence outcomes in muscle-invasive bladder cancer (MIBC), but its prognostic role remains unclear. In particular, the impact of adipose tissue distribution and muscle quality beyond BMI has been insufficiently studied. This study aimed to evaluate whether adipose tissue distribution and myosteatosis assessed by AI-based CT analysis are associated with OS in patients with MIBC treated with neoadjuvant chemotherapy followed by radical cystectomy. In this retrospective, bicentric study, we analyzed data from 87 patients with MIBC treated with neoadjuvant chemotherapy (NAC) followed by radical cystectomy. Artificial intelligence (AI)–based analysis was applied to CT scans obtained before chemotherapy (BC) and before surgery (BS), focusing on the L3 vertebral level. We assessed the association between body composition and OS, and secondarily progression-free survival (PFS). Evaluated parameters included myosteatosis (fat infiltration of skeletal muscle), subcutaneous fat index (SFI), visceral fat index (VFI), and the visceral-to-subcutaneous fat ratio (VSR). Median follow-up was 24.6 months. A high SFI measured before surgery was significantly associated with improved OS (HR 0.41; 95%CI 0.19–0.86; p = 0.02) whereas SFI measured before chemotherapy was not associated with OS (HR 0.68; 95% CI 0.32–1.45). No significant associations were found between SFI and PFS at either timepoint. Similarly, VFI, VSR, and myosteatosis were not significantly associated with OS or PFS. In this exploratory study, AI-derived SFI assessed before surgery was associated with OS. These findings should be interpreted cautiously given the retrospective design and limited sample size and require validation in larger prospective cohorts.