This study aims to examine the feasibility of ML-assisted salivary-liquid-biopsy platforms using genome-wide methylation analysis at the base-pair and regional resolution for delineating oral squamous cell carcinoma (OSCC) and oral potentially malignant disorders (OPMDs). A nested cohort of patients with OSCC and OPMDs was randomly selected from among patients with oral mucosal diseases. Saliva samples were collected, and DNA extracted from cell pellets was processed for reduced-representation bisulfite sequencing. Reads with a minimum of 10× coverage were used to identify differentially methylated CpG sites (DMCs) and 100 bp regions (DMRs). The performance of eight ML models and three feature-selection methods (ANOVA, MRMR, and LASSO) were then compared to determine the optimal biomarker models based on DMCs and DMRs. A total of 1745 DMCs and 105 DMRs were identified for detecting OSCC. The proportion of hypomethylated and hypermethylated DMCs was similar (51% vs. 49%), while most DMRs were hypermethylated (62.9%). Furthermore, more DMRs than DMCs were annotated to promoter regions (36% vs. 16%) and more DMCs than DMRs were annotated to intergenic regions (50% vs. 36%). Of all the ML models compared, the linear SVM model based on 11 optimal DMRs selected by LASSO had a perfect AUC, recall, specificity, and calibration (1.00) for OSCC detection. Overall, genome-wide DNA methylation techniques can be applied directly to saliva samples for biomarker discovery and ML-based platforms may be useful in stratifying OSCC during disease screening and monitoring.

No abstract

This review sought to determine the range and nature of prospective-sampling and blinding methods for validating nonviral biofluid markers diagnostic of head and neck carcinomas. Electronic database searching was conducted to identify studies published in English from January 1, 2009 to August 1, 2020. Sixteen studies from 17 articles published between 2011 and 2020 were included in this review. We found that about 3 out of 100 studies utilized at least one of the mock testing approaches for biomarker validation. Protein, mRNA, and metabolomic markers also represented the only groups whose validation has been attempted using these methods. Furthermore, studies that utilized both methods were found to have lower bias concerns on the quality assessment of diagnostic accuracy studies (QUADAS-2) tool. Overall, there is a need to include these protocols in research endeavours verifying diagnostic biomarkers for head and neck carcinomas following the preliminary establishment of their classification accuracy.