Salivary metabolite signatures of oral cancer and leukoplakia

TLDR

Oral cancer is among the six most common human cancers, with a 5‑year survival rate below 50 % and is frequently diagnosed only after reaching an advanced stage. This study investigates whether salivary metabolomics can serve as a diagnostic and stratification tool for oral cancer and leukoplakia, specifically assessing its ability to detect oral squamous cell carcinoma. Saliva from 37 OSCC patients, 32 leukoplakia patients and 34 healthy controls was profiled by UPLC‑QTOF mass spectrometry and analyzed with multivariate statistics, yielding a five‑metabolite panel (γ‑aminobutyric acid, phenylalanine, valine, n‑eicosanoic acid, lactic acid) selected by OPLS‑DA and evaluated via ROC curves. The five‑metabolite panel achieved high accuracy (0.89–0.97), sensitivity (86.5–94.6 %), specificity (82.4–84.4 %) and positive predictive value (81.6–87.5 %) in distinguishing OSCC from controls or leukoplakia, demonstrating that salivary metabolomics can complement clinical detection and improve diagnosis and prognosis.

Abstract

Oral cancer, one of the six most common human cancers with an overall 5-year survival rate of <50%, is often not diagnosed until it has reached an advanced stage. The aim of the current study is to explore salivary metabolomics as a disease diagnostic and stratification tool for oral cancer and leukoplakia and evaluate the potential of salivary metabolome for detection of oral squamous cell carcinoma (OSCC). Saliva metabolite profiling for a group of 37 OSCC patients, 32 oral leukoplakia (OLK) patients and 34 healthy subjects was performed using ultraperformance liquid chromatography coupled with quadrupole/time-of-flight mass spectrometry in conjunction with multivariate statistical analysis. The OSCC, OLK and healthy control groups demonstrate characteristic salivary metabolic signatures. A panel of five salivary metabolites including γ-aminobutyric acid, phenylalanine, valine, n-eicosanoic acid and lactic acid were selected using OPLS-DA model with S-plot. The predictive power of each of the five salivary metabolites was evaluated by receiver operating characteristic curves for OSCC. Valine, lactic acid and phenylalanine in combination yielded satisfactory accuracy (0.89, 0.97), sensitivity (86.5% and 94.6%), specificity (82.4% and 84.4%) and positive predictive value (81.6% and 87.5%) in distinguishing OSCC from the controls or OLK, respectively. The utility of salivary metabolome diagnostics for oral cancer is successfully demonstrated in this study and these results suggest that metabolomics approach complements the clinical detection of OSCC and stratifies the two types of lesions, leading to an improved disease diagnosis and prognosis.