Can we detect biomarkers of oral squamous cell carcinoma from saliva or mouth swabs?

Abstract

Aim To determine the accuracy of using saliva and oral cytology swabs in the diagnosis of oral squamous cell carcinoma (OSCC) by detecting aberrantly hypermethylated DNA.

Data sources Electronic databases including PubMed/Medline, Embase, Cochrane Library, Scopus and Web of Science with no language or article restrictions. Additionally, LILACS database, OpenGrey and Google Scholar were searched.

Study selection Studies published since the first report proposing DNA hypermethylation for head and neck carcinomas in 2001 until 2020 were included. The diagnoses of oral cancer were limited to OSCC. Authors screened titles and abstracts for relevance, before further screening of full texts and a consensus for inclusion was reached. Qualitative analysis was conducted on 22 studies, and 11 were selected for meta-analysis.

Data analysis Diagnostic test accuracy meta-analysis was performed using a series of investigations including the Haldane-Anscombe correction, forest plots, receiver operator characteristic (ROC) curves and Deeks' funnel plots. Differences in the pooled estimates of the application of both single and combined hypermethylation markers were assessed using Cochran's Q test and Higgins' I2 test. Random-effects meta-regression analysis was used to evaluate the effects of variation in sensitivity and false-positive rates, and to identify sources of heterogeneity. Sensitivity analyses removed outliers.

Results All studies suggest that DNA hypermethylation can discriminate between OSCC and premalignant conditions or normal mucosa. Using summary receiver operator characteristic (SROC) curves, the sensitivity of combined markers was higher than single markers, and specificity of both combined and single markers was comparable. The biomarkers evaluated had fair-to-excellent sensitivity and good-to-excellent specificity for discriminating OSCC from premalignant and normal mucosa. Four studies included in the review suggest that this method of detection is more applicable to patients who smoke due to increased hypermethylation rates.

Conclusion Hypermethylation markers using saliva and oral swabs are more specific than sensitive for OSCC diagnosis. Combining different genes within the biomarker panel can improve diagnostic test accuracy. However, more blinded evaluation study designs with less bias which replicate real-world application are required to endorse the use of saliva sampling and oral swabs in oral oncology.