Bladder cancer (BLCA) remains a pressing global health challenge, characterized by high recurrence, progression, and mortality rates. Traditional prognostic tools, such as the TNM classification, show limited accuracy, particularly in the context of immune checkpoint inhibitors (ICI). In this study, we employed comprehensive bioinformatics analyses on the TCGA dataset to identify genes that are differentially expressed and methylated in BLCA.

We analyzed transcriptomic and methylation data from the TCGA BLCA cohort to identify differentially expressed and differentially methylated genes. By integrating both datasets, and using machine learning techniques were employed to generate a 25-gene survival signature, which was validated using independent cohorts, including the IMvigor 210 trial. Functional enrichment and KEGG pathway analyses were performed to elucidate the biological roles of the signature, and in vitro experiments confirmed the involvement of key lipid metabolism genes.

Our analysis identified a 25-gene signature significantly associated with 5-years disease-specific survival (DSS) in BLCA patients(HR:2.3, p<0.001). Functional enrichment analysis highlighted key genes like FASN, SCD within this signature, emphasizing the critical involvement of lipid metabolism pathways, particularly those related to FASN and SCD. The prognostic value of this gene signature was validated across four independent cohorts, including the IMvigor 210 clinical trial dataset, consistently distinguishing high-risk from low-risk patients and correlating with survival outcomes across BLCA stages. In vitro experiments further demonstrated that inhibiting FASN and SCD impairs BLCA cell proliferation and migration, suggesting their potential as therapeutic targets.

This study introduces a novel biomarker for BLCA prognosis while shedding light on its underlying molecular mechanisms. The findings underscore the importance of this biomarker in the clinical epigenetics of BLCA and offer new avenues for personalized treatment strategies