A salivary protein panel for detecting malignant transformation of oral leukoplakia

Oral leukoplakia (OLK) is the most common oral potentially malignant disorder and carries a lifetime risk of transformation to oral squamous cell carcinoma (OSCC) of 0.1% to 9.8%. The aim of the current study is to explore salivary proteomics as a non-invasive approach for detecting the malignant transformation from OLK to OLK-associated OSCC (OLK-OSCC). We performed mass spectrometry-based proteomic profiling of saliva from 131 patients (48 OLK and 83 OLK-OSCC patients) and integrated the resulting data with machine learning. Salivary protein diversity was significantly increased in OLK-OSCC patients, and differential expression analysis identified 100 upregulated and 36 downregulated salivary proteins of OLK-OSCC compared with OLK. Using XGBoost and LASSO regression, we developed a nomogram based on a five-protein panel (MMP1, SUMO4, SMS, TMBIM1, and MUC5AC), which achieved high discriminatory performance in the training cohort (AUC = 0.94). The biological relevance of this panel was further supported by single-cell RNA sequencing data, a 4-nitroquinoline 1-oxide-induced murine carcinogenesis model (n = 18), and an independent OSCC tissue cohort (n = 39). Functional correlation analyses suggested that these biomarkers may be associated with an immunosuppressive tumor microenvironment by modulating the infiltration pattern of immune cells. Collectively, these findings characterize the salivary proteomic changes associated with OLK-OSCC and identify a five-protein salivary panel with potential value for OLK surveillance and early detection of malignant transformation.