A recurrence risk prediction model based on DNA methylation in laryngeal squamous cell carcinoma

Laryngeal squamous cell carcinoma (LSCC) is a heterogeneous malignant tumor with challenging prognostic assessments. This study aimed to identify DNA methylation markers associated with recurrence risk in LSCC. We analyzed 78 LSCC patients from our hospital, dividing them into recurrence (n = 34) and non-recurrence (n = 44) groups based on disease-free survival (DFS). Reduced representation bisulfite sequencing was performed to identify genome-wide DNA methylation patterns. The study identified four differentially methylated regions associated with recurrence risk, located in the promoter regions of RPH3AL, UBE2I, MAPK1, and one distal intergenic region. Based on LASSO logistic regression analysis, we developed a methylation risk signature (MRS) to assess recurrence risk, which achieved an accuracy of 98.2% (54/55) in the training set and 95.7% (22/23) in the validation set. Survival analysis revealed that MRS could significantly stratify DFS (HR = 79.23, P < 0.001) in LSCC patients. No significant correlation was observed between MRS and stage, age, or histological grade in our data. This genome-wide DNA methylation analysis provides a promising prediction model for LSCC recurrence risk.