Background Bacterial vaginosis (BV) and vulvovaginal candidiasis (VVC) are common vaginal infections; however, the existing techniques for diagnosing these infections are often subjective and inefficient. This study examined the ability of headspace-gas chromatography-ion mobility spectrometry (HS-GC-IMS) to diagnose BV and VVC by analyzing volatile organic compounds (VOCs) in vaginal swabs. Methods A total of 616 study participants were recruited, comprising 152 patients with BV, 174 patients with VVC, and 290 healthy controls. Participants were randomly allocated into a training cohort (n = 432) and an independent test cohort (n = 184). Vaginal VOC profiles were examined using HS-GC-IMS. A partial least squares discriminant analysis (PLS-DA) model was developed to differentiate the groups and identify unique biomarkers for BV and VVC. Results Fifty-nine VOC peaks were identified among the samples. The PLS-DA model exhibited strong classification efficacy, with overall predictive accuracies of 84.26% on the training cohort and 80.98% on the test cohort. A receiver operating characteristic analysis produced area under the curve values surpassing 0.90 for all groups, signifying high model reliability. Eight distinct VOCs were identified as potential diagnostic biomarkers. Based on these biomarkers, the PLS-DA model had an overall prediction accuracy of 78.25% across the entire cohort of 616 participants. Conclusion HS-GC-IMS provides a rapid, sensitive, and non-invasive method for characterizing vaginal VOCs. The constructed model accurately distinguished BV and VVC, indicating that this technique has considerable potential as an innovative clinical instrument for objectively identifying vaginal infections.
Liu et al. (Tue,) studied this question.