Rapid Point-of-Care Exhaled Breath Analysis for Lung Cancer Diagnosis Using a Micro Gas Chromatography System: A Pilot Study

The study investigates the use of volatile organic compounds (VOCs) in exhaled breath as a non-invasive diagnostic tool for lung cancer (LC). Employing a novel micro gas chromatography-micro photoionisation detector (GC-PID) system, we aimed to identify and validate VOCs that could differentiate between LC patients and those with benign pulmonary diseases. The cross-sectional study included 106 participants, categorized into 85 LC patients and 21 benign controls, based on computed tomography and histological assessments. Participants provided breath samples following a standardized protocol, and the GC-PID system, known for its rapid point-of-care capabilities and low detection limits, was utilized for rapid and sensitive online VOC analysis. Through a meticulous process of data analysis, including principal component analysis, single-factor hypothesis testing, orthogonal partial least squares discriminant analysis and various tests of machine learning algorithms, including random forest, k-nearest neighbor, logistic regression, XGBoost, and support vector machine, we finally identified six potential VOC biomarkers, with diagnostic models incorporating these markers achieving high sensitivity (0.95-1.00) and specificity (0.84-0.88), and areas under the receiver operating characteristic curve ranging from 0.79 to 0.91. Moreover, these models were also extended favourably to the recurrence and metastasis of pulmonary cancer and oesophageal cancer. The study demonstrates the potential of GC-PID as a point-of-care tool for LC differential diagnosis, highlighting the need for further validation in larger, multi-centric cohorts to refine the VOC biomarker panel and establish a robust diagnostic framework for clinical application.