Integrative multi-omics analysis and experimental validation identify molecular subtypes, prognostic signature, and...

Oral squamous cell carcinoma (OSCC) is a challenging malignancy with poor prognosis despite therapeutic advancements. This study seeks to derive a precise molecular subtyping and prognostic model for personalized treatment strategies.

Multi-omics data from TCGA cohort was analyzed using consensus clustering algorithms for subtype classification. Based on the classification, a multi-omics cancer subtyping signature (MSCC) model was constructed using machine learning methods. The model's clinical utility was assessed by evaluating immune features and immunotherapy response. Potential therapeutic agents were identified through drug sensitivity analysis.

Three distinct OSCC subtypes with unique genetic and immunological profiles were identified. The MSCC model, developed using the StepCox [both]+plsRcox algorithm, demonstrated superior prognostic performance compared to existing models. High MSCC scores correlated with poor prognosis, reduced immune cell infiltration, and decreased likelihood of benefiting from immune checkpoint inhibitor therapy. Docetaxel and paclitaxel emerged as potential therapeutic candidates. In vitro experiments validated CA9 as a promising therapeutic target, with its knockdown significantly inhibiting OSCC cell proliferation and migration.

This multi-omics analysis unveiled subtype-specific differences in OSCC and established an MSCC model for predicting prognosis and treatment response. These findings provide a foundation for early diagnosis, molecular subtyping, and personalized treatment strategies in OSCC.