Correlation of bioinformatics-based m6A methylation regulators with prognosis in oral squamous cell carcinoma...

Oral squamous cell carcinoma (OSCC) is characterized by difficulties in early diagnosis and poor prognosis. m6A methylation regulators are closely associated with tumor progression, yet their prognostic value and mechanisms in OSCC remain unclear. This study was to analyze the expression patterns of m6A regulators in OSCC and their association with prognosis, elucidate their mechanisms in OSCC progression, and provide a basis for diagnosis and treatment.

The expression patterns and prognostic value were analyzed using the following bioinformatics methods: the optimal cluster number k = 2 was determined based on the silhouette coefficient, and K-means clustering was employed for sample subtype classification; the cut-off values for high/low gene expression were defined by combining the ROC curve method and the tertile method, with Kaplan-Meier (KM) survival analysis and the Log-rank test used to compare survival differences between groups; univariate Cox regression was applied to screen for potential prognostic factors, followed by multivariate Cox regression with a forward selection method to identify independent prognostic factors; the limma package was used to identify differentially expressed genes (DEGs), and the biological functions of these DEGs were annotated through GO (biological process/cellular component/molecular function) and KEGG enrichment analyses; clinical, gene, and combined models were constructed, and the C-index and Delong test were utilized to evaluate the predictive performance of the models.

OSCC samples were classified into Group A (n = 142) and Group B (n = 83). The 5-year survival rate (SR) of Group A (70.24 %) was significantly higher than that of Group B (60.76 %) (Hazard ratio (HR) = 1.58, P = 0.001). Highly correlated gene pairs such as methyltransferase-like 3 (METTL3) and METTL14 (r = 0.78) and fat mass and obesity-associated protein (FTO) and AlkB homolog 5 (ALKBH5) (r = 0.65) were consistently validated in an independent dataset. METTL3 (adjusted HR = 1.38, P = 0.008), FTO (adjusted HR = 1.42, P = 0.015), and YTHDF1 (adjusted HR = 1.51, P = 0.004) were identified as independent predictors of poor prognosis. DEGs were enriched in immune response, cell cycle, and PI3K-Akt/MAPK signaling pathways. The combined model incorporating these three genes and clinical variables demonstrated the highest C-index (0.72), with HR fluctuations <15 %, indicating robust results.

Multivariate Cox regression confirmed that METTL3, FTO, and YTHDF1 are associated with an unfavorable prognosis in OSCC patients, suggesting their potential as independent prognostic biomarkers. The m6A methylation regulators may contribute to OSCC progression by modulating immune responses, the cell cycle, and the PI3K-Akt/MAPK signaling pathway, indicating preliminary potential targets for precision therapy. However, further experimental validation is required in the future.