Noninvasive Diagnosis of Parkinson's Disease via High-Coverage Exhaled Metabolomics: A Proof-of-Concept Study

Early and accurate diagnosis of Parkinson's disease (PD) remains a major clinical challenge. Current diagnostic methods, such as positron emission tomography (PET) and cerebrospinal fluid analysis, are often expensive, invasive, and unsuitable for widespread screening. Exhaled breath contains volatile metabolites that reflect systemic metabolic states, offering a promising noninvasive diagnostic medium. This study aimed to explore the feasibility of using high-coverage breath metabolomics for PD diagnosis. Two independent cohorts comprising 342 participants (103 patients with PD and 239 healthy controls) were included. Real-time breath analysis was performed using secondary electrospray ionization high-resolution mass spectrometry (SESI-HRMS). The results revealed significant metabolic dysregulation in PD patients, primarily involving alterations in lipid-related metabolites (e.g., butyric acid, propionic acid, and myristic acid) and amino acid-related metabolites (e.g., valine and aminoacetone). A random forest classifier based on breath metabolic features demonstrated robust diagnostic performance, achieving an area under the receiver operating characteristic curve (AUC) of 0.805 in an independent validation cohort. Furthermore, an exhaled metabolic score (eMS) was developed using 80 key metabolic features. This score not only effectively discriminated PD patients from healthy controls (P < 0.0001) but also showed a significant correlation with the extent of abnormalities observed in PET imaging (P = 0.049), reflecting the burden of neurodegeneration. Our findings demonstrate that breath metabolomics represents a noninvasive, rapid, and scalable approach with substantial clinical potential for PD screening and disease monitoring. The eMS system provides a promising strategy to support early PD diagnosis and dynamic monitoring of disease progression.