Comprehensive bioinformatics analysis combined with experimental validation to screen biomarkers for malignant transformation of oral leukoplakia

Oral leukoplakia (OLK) is the most common potentially malignant disorders in the oral cavity. This study aimed to screen the key genes of OLK malignant transformation using the Gene Expression Omnibus (GEO) database and experiments. In this study, the GEO database was employed to screen OLK malignant transformation-related genes, which were subsequently identified with a series of bioinformatic analyses. External validation showed that the model based on LAPTM4B, NR3C1, and COX6A1 had high accuracy in diagnosing OLK malignant transformation. Furthermore, the DMBA-induced potentially malignant disorders and OSCC models in vivo and real-time PCR experiment in vitro further verified the database analysis results. In conclusion, three key genes (LAPTM4B, NR3C1, and COX6A1) were screened as potential biomarkers for the diagnosis and treatment of OLK malignant transformation.