Non-Invasive Detection of Nasopharyngeal Carcinoma Using Volatile Organic Compounds

Background: Nasopharyngeal carcinoma (NPC) is a leading head and neck cancer in Asia, where late-stage presentation contributes to poor survival. Non-invasive diagnostic strategies such as breath analysis may improve early detection. Objectives: This study aimed to investigate whether volatile organic compound (VOC) features in exhaled breath, detected using proton transfer reaction mass spectrometry (PTR-MS), can distinguish NPC patients from healthy controls. Methods: Breath samples were collected from 50 NPC patients and 40 healthy controls. PTR-TOF-MS was used to measure exhaled VOC features. Group comparisons were performed using univariate analysis, while multivariable regression was adjusted for age, sex, BMI, smoking, and medication use. Multivariate methods, including principal component analysis (PCA) and random forest classification, were used to assess discriminatory potential. Results: Seven distinct VOC features (measured as m/z values) showed significant differences between NPC patients and healthy controls, with m089 and m175 emerging as the strongest markers of distinction. PCA after normalization revealed clearer separation between NPC patients and controls. Random forest models incorporating significant VOCs achieved moderate classification accuracy, and the results remained robust after adjusting for confounders. Conclusions: PTR-MS breath analysis can detect disease-specific VOC features in NPC and shows promise as a non-invasive diagnostic tool. Larger validation studies and definitive compound identification are needed to confirm clinical utility.