Rapid screening of infertility-associated gynecological conditions via ambient glow discharge mass spectrometry utilizing urine metabolic fingerprints.

Infertility presents a widespread challenge for many families worldwide, often arising from various gynecological diseases (GDs) that hinder successful pregnancies. Current diagnostic methods for GDs have disadvantages such as low efficiency, high cost, misdiagnose, invasive injury and etc. This paper introduces a rapid, non-invasive, efficient, and straightforward analytical method that utilizes desorption, separation, and ionization mass spectrometry (DSI-MS) platform in conjunction with machine learning (ML) to detect urine metabolite fingerprints in patients with different GDs. We analyzed 257 samples from patients diagnosed with polycystic ovary syndrome (PCOS), premature ovarian insufficiency (POI), diminished ovarian reserve (DOR), endometriosis (EMS), recurrent pregnancy loss (RPL), recurrent implantation failure (RIF), and 87 samples from healthy control (HC) individuals. We identified metabolite differences and dysregulated pathways through dimensionality reduction methods, with the result of the discovery of 7 potential biomarkers for GDs diagnosis. The ML method effectively distinguished subtle differences in urine metabolite fingerprints. We anticipate that this innovative approach will offer a patient-friendly, rapid screening, and differentiation method for infertility-related GDs patients.