A Collecting Device for Exhaled Breath Detection by Nanostructure-initiator Mass Spectrometry

The metabolites in exhaled breath (EB) dynamically reflect the metabolic status of the human body and can serve as biomarkers for health management and disease diagnosis. However, there is currently a lack of clinical diagnostic methods for rapid metabolite detection in EB samples and metabolomic data processing. In this project, we develop a method that uses a collecting device to efficiently collect exhaled breath samples on perfluoroethylene propylene (FEP) modified silicon nanowires (SiNWs). The sample chips are directly detected by nanostructure-initiator mass spectrometry (NIMS), and the mass spectrometry (MS) results are collected as a dataset and analyzed by machine learning techniques. FEP@SiNWs chips improved metabolite trapping efficiency and laser desorption/ionization efficiency during sample collection and mass spectrometry detection, enabling rapid metabolomic analysis with good stability and detection repeatability. Through machine learning technology, the MS information of metabolites in EB samples of patients with pulmonary nodules and healthy people was processed, the characteristic peaks were selected, and the prediction accuracy (>95.4 %) and sensitivity (92.6 – 98.5 %) of the machine learning model were tested. Based on this work, the collecting device and FEP@SiNWs chips have the potential to be used as home health management devices, and a metabolomic database of respiratory diseases via exhaled breath samples can be established in future research, which can be applied to respiratory disease diagnosis and intelligent healthcare.