Identification of urinary volatile organic compounds as a potential non-invasive biomarker for esophageal cancer

Abstract

Early diagnosis of esophageal cancer (EC) is extremely challenging. The study presented herein aimed to assess whether urinary volatile organic compounds (VOCs) may be emerging diagnostic biomarkers for EC. Urine samples were collected from EC patients and healthy controls (HCs). Gas chromatography-ion mobility spectrometry (GC-IMS) was next utilised for volatile organic compound detection and predictive models were constructed using machine learning algorithms. ROC curve analysis indicated that an 8-VOCs based machine learning model could aid the diagnosis of EC, with the Random Forests having a maximum AUC of 0.874 and sensitivities and specificities of 84.2% and 90.6%, respectively. Urine VOC analysis aids in the diagnosis of EC.