BIỂU HIỆN HOÁ MÔ MIỄN DỊCH CỦA CÁC DẤU ẤN SINH HỌC DỰ ĐOÁN THEO PHÂN NHÓM BIỂU HIỆN GEN TRÊN BỆNH NHÂN UNG THƯ BIỂU MÔ TẾ BÀO GAI NIÊM MẠC MIỆNG

Abstract

Introduction: Oral squamous cell carcinoma (OSCC) is a common head and neck cancer with high rates of metastasis and recurrence, necessitating the identification of novel prognostic biomarkers. Our previous study utilized machine learning models based on mRNA sequencing data and clinical-histopathological features to classify OSCC into two subgroups with distinct molecular and clinical characteristics. It proposed 12 differentially expressed genes between these groups as potential biomarkers. Objective: The current study aimed to evaluate the protein expression of these markers using immunohistochemistry (IHC) in an independent cohort of OSCC patients and determine the association with subgroups predicted by SVM and XGBoost models, combined with preliminary survival analysis. Method: A retrospective study was performed on 50 untreated primary OSCC tissue samples from Vietnamese patients. Results showed that 10 out of 12 markers exhibited protein expression. Results: When subgrouping using the SVM model, four markers (SLAIN2, ADNP, SLK, MARCKS) showed statistically significant differences in protein expression between the two predicted subgroups (p<0.05), consistent with gene expression trends: Subgroup 1 showed increased expression of SLAIN2, ADNP, SLK; Subgroup 2 showed increased expression of MARCKS. With the XGBoost model, only ADNP showed a significant difference. Survival analysis on the limited sample size did not reveal significant differences between the groups. Conclusion: In conclusion, the study confirmed via IHC the differential protein-level expression of SLAIN2, ADNP, SLK, and MARCKS between the SVM-predicted OSCC subgroups, indicating their potential as biomarkers reflecting the molecular subgroups of OSCC. Further studies are needed to validate these markers' prognostic and functional roles.