Radiomic biomarkers for platinum‐refractory head and neck cancer in the era of immunotherapy

Abstract Objective Immune checkpoint inhibitors (ICI) are recommended as the first‐line therapy for platinum‐refractory head and neck squamous cell carcinoma (HNSCC), a disease with a poor prognosis. However, biomarkers in this situation are rare. The objective was to identify radiomic features‐associated biomarkers to guide the prognosis and treatment opinions in the era of ICI. Methods A total of 31 platinum‐refractory HNSCC patients were retrospectively enrolled. Of these, 65.5% (20/31) received ICI‐based therapy and 35.5% (11/31) did not. Radiomic features of the primary site at the onset of recurrent metastatic (R/M) status were extracted. Prognostic and predictive radiomic biomarkers were analysed. Results The median overall survival from R/M status (R/M OS) was 9.6 months. Grey‐level co‐occurrence matrix‐associated texture features were the most important in identifying the patients with or without 9‐month R/M death. A radiomic risk‐stratification model was established and equally separated the patients into high‐, intermittent‐ and lower‐risk groups (1‐year R/M death rate, 100.0% vs. 70.8% vs. 27.1%, p = 0.001). Short‐run high grey‐level emphasis (SRHGE) was more suitable than programmed death ligand 1 (PD‐L1) expression in selecting whether patients received ICI‐based therapy. Conclusions Radiomic features were effective prognostic and predictive biomarkers. Future studies are warranted.