Intra-tumor DKK1 and CALML5 heterogeneity as a novel biomarker for prognosis and immunotherapy response in head and neck squamous cell carcinoma
Yuhui Li, Rongxi Chen, Lizao Zhang, Yanping Jiang, Siqi Ren, Shuo Li, Hong-Feng Xu, Yongmei Tan, Suling Chen, Xiao Tan, Lianxi Mai, Guoxin Huang, Fan Wu, Xin Liu, Peisheng Liang, Shijia Kuang, Liansheng Wang, Jingkang Liu, Jintao Li, Yanyan Li, Qiuping Xu, Ziyao Su, Zijing Huang, Qunxing Li, Tianjun Lan, Haotian Cao, Huiqian Wu, Jinsong Li
Abstract
Background Head and neck squamous cell carcinoma (HNSCC) is characterized by substantial molecular heterogeneity, and current biomarkers for predicting immunotherapy response, such as PD-L1 expression and tumor mutational burden (TMB), remain inconsistent. This study aimed to identify reliable prognostic and predictive biomarkers for immunotherapy response in HNSCC. Methods A multi-task deep learning framework was applied to transcriptomic data to classify patients into distinct risk subgroups based on survival outcomes. Differentially expressed genes were used to identify hub prognostic genes. An integrated single-cell RNA sequencing (scRNA-seq) dataset was employed to construct a cellular atlas of HNSCC. Spatial transcriptomics and immunofluorescence were conducted to assess immune cell distribution and tumor microenvironment (TME) features. Results DKK1 and CALML5 were identified as hub prognostic genes, both predominantly expressed in tumor cells. DKK1⁺ tumor cells exhibited immunosuppressive characteristics, whereas CALML5⁺ cells were associated with cytotoxic immune infiltration. The ratio of DKK1 to CALML5 expression (DC score) served as a robust biomarker for predicting immunotherapy response. A higher DC score correlated with incomplete pathological response and poorer overall survival (p < 0.05, AUC = 0.9671). Conclusion The DC score represents a novel, reliable biomarker for predicting prognosis and immunotherapy response in HNSCC. This study provides insight into the spatial and molecular heterogeneity of HNSCC and offers a potential tool for informing personalized immunotherapeutic strategies.