Aims:0„2Cetuximab resistance significantly limits the efficacy of targeted therapy in head and neck squamous cell carcinoma (HNSCC). This study aims to delineate molecular subtypes and identify predictive biomarkers associated with cetuximab response to guide personalized treatment. Methodology:0„2Differentially expressed genes (DEGs) from cetuximab-resistant and -sensitive cohorts (GEO datasets and our independent clinical cohort) were integrated. Consensus clustering was utilized to define molecular subtypes. Comprehensive analyses including tumor microenvironment (TME) profiling, drug sensitivity prediction, and a LASSO-Cox prognostic model were performed. The mechanistic role of0„2THBS10„2in resistance was validated0„2in vitro0„2through overexpression and knockdown models, with drug response, cellular phenotypes, and TGF-¦Â/EMT pathway activation assessed via functional assays, western blotting, and immunofluorescence. Results:0„2Three distinct HNSCC subtypes (C1-C3) were identified. Subtype C3 exhibited the worst overall survival, an immuno-stromal enriched TME, and profound resistance to EGFR inhibitors. Our prognostic model highlighted0„2THBS10„2as a core risk gene, which was significantly upregulated in C3. Functional assays confirmed that0„2THBS10„2overexpression abrogates cetuximab sensitivity, whereas its knockdown restores therapeutic response by inducing apoptosis and attenuating tumor proliferation and motility. Mechanistically,0„2THBS10„2drives cetuximab resistance by activating the TGF-¦Â signaling cascade and inducing epithelial-mesenchymal transition (EMT), thereby bypassing tumor dependence on the EGFR pathway. Conclusion:0„2This study delineates clinically relevant HNSCC subtypes and establishes0„2THBS10„2as a critical driver and predictive biomarker of cetuximab resistance, providing a rationale for stratifying patients and overcoming drug resistance.