Head and neck cancers (HNCs) are common malignancies with generally unfavorable prognoses worldwide, among which squamous cell carcinoma (SCC) represents the predominant histological type. The Combined Positive Score (CPS), a key metric for evaluating PD-L1 expression, has demonstrated predictive value for response to immunotherapy across various solid tumors. However, its distribution patterns and clinical significance among tumors originating from different anatomical subsites within the head and neck region remain unclear.This retrospective study included 591 patients with malignant head and neck tumors treated at Beijing Tongren Hospital between January 2022 and November 2024, comprising 316 hypopharyngeal, 202 laryngeal, 45 oropharyngeal, and 28 nasal cavity and paranasal sinus carcinoma. PD-L1 expression was assessed using the PD-L1 IHC 22C3 pharmDx assay, and CPS was calculated accordingly. Results showed that the CPS was significantly lower in hypopharyngeal carcinoma (18.89 ± 21.65) compared to oropharyngeal carcinoma (24.71 ± 25.97; P < 0.05). Furthermore, CPS was negatively correlated with tumor differentiation in hypopharyngeal cancer (r = -0.1434, P = 0.018), with a similar trend observed in laryngeal cancer (P = 0.054), whereas no significant correlation was found in oropharyngeal or nasal cavity and paranasal sinus carcinoma.Survival analysis in the immunotherapy-treated subgroup revealed no statistically significant differences in overall survival (OS) or progression-free survival (PFS) between patients with CPS ≥ 20 and those with CPS < 20. However, in hypopharyngeal cancer, the CPS < 20 group exhibited a hazard ratio (HR) for death of 4.55 (95% CI: 1.019-20.31), suggesting a potential clinically relevant difference.This study highlights significant heterogeneity in the immune microenvironment across different subsites of head and neck squamous cell carcinoma. The relationship between CPS and tumor differentiation is site-specific, indicating that the value of CPS as a biomarker for immunotherapy response should be interpreted in the context of tumor location.
Feng et al. (2026) studied this question.