Feed

Head and neck squamous cell carcinoma (HNSCC) is a highly heterogeneous malignancy with poor prognosis and limited predictive biomarkers for therapy response. Characterizing malignant cell heterogeneity may improve prognostic and therapeutic stratification. We integrated single-cell RNA sequencing (scRNA-seq) data from 58 HNSCC patients (181,003 cells) to define malignant cell subpopulations, their differentiation states, developmental trajectories, cell-cell interactions, and spatial localization. Coexpression gene modules and meta-programs were identified using hdWGCNA and NMF. These programs were projected onto bulk RNA-seq datasets to classify HNSCC subtypes and examine associations with clinical outcomes, tumor microenvironment (TME), genomic instability, and predicted response to immune checkpoint inhibitors (ICIs). Twelve malignant clusters were identified with distinct clinical and molecular features. MC-5 exhibited a stem-like phenotype associated with poor prognosis, while MC-7 and MC-11 showed high TME communication and immune engagement. Coexpression analysis revealed 16 modules and eight meta-programs encompassing proliferation, differentiation, stress response, and immune activity. Translation to bulk RNA-seq defined three HNSCC subtypes (MS-1, MS-2, MS-3) with divergent survival, immune infiltration, stromal composition, and genomic features. MS-2, an immune-enriched subtype, demonstrated superior survival, high HPV positivity, and predicted ICI responsiveness. A 25-gene malignant cell score (MCScore) robustly predicted both prognosis and immunotherapy response. This study provides a comprehensive map of malignant cell heterogeneity in HNSCC, identifies key functional expression programs, and defines molecular subtypes with clinical and therapeutic relevance. Malignant cell-specific signatures, such as MCScore, offer promising tools for patient stratification and precision immunotherapy.