Abstract 15047: The Probability of miRNA Derived From Extracellular Vesicles as Functional Biomarkers for Patients With Lower Extremity Artery Disease

Abstract

Introduction: Lower extremity artery disease (LEAD) is a progressive and systematic manifestation of atherosclerosis. In LEAD, multiple arterial beds are affected increasing the possibility of death due to coronary artery disease (CAD) and cerebrovascular disease (CVD). We have identified functional biomarkers that can be used to predict clinical outcomes in patients with LEAD by analyzing miRNAs contained in the extracellular vesicles (EVs). Methods: A total of 146 patients with diabetes and without end-stage renal disease were enrolled as research subjects between 2019 and 2021. Among them, 118 had LEAD, including 6 asymptomatic cases, 43 cases of claudication, and 69 cases of chronic limb-threatening ischemia (CLTI). The remaining 28 cases were in the Control group, who were systematically screened for atherosclerotic lesions by radiological modalities. We performed miRNA sequencing using the EV-derived miRNAs from patients’ serum. We developed machine learning based models to diagnose LEAD, CLTI, CAD, and CVD. We used Cox models to identify miRNAs associated with the patency rate at the reconstruction site, amputation-free survival (AFS), and major cardiovascular events (MACE). Results: The area under curve (AUC) of the diagnostic models for LEAD, CLTI, CAD, and CVD for the training data were 0.82, 1.00, 1.00, and 1.00, respectively, and for the test data were 0.68, 0.85, 0.71, and 0.51, respectively. The diagnostic model for CLTI showed the highest generalization performance among these models. We performed pathway analysis for miRNAs with fold change > 2 in CLTI using DIANA-miRPath v3.0, which revealed enrichment of proteoglycans in cancer, Wnt signaling, and TGF-beta signaling pathways.We identified 62 miRNAs as significant risk factors affecting the patency. The group with has-miR-6850-3p above the median (78.8%) had significantly higher one-year patency rate compared to the group with has-miR-6850-3p below the median (59.5%) (p = 0.03). In addition, we identified 6 and 4 miRNAs as risk factors related to AFS and MACE, respectively. Conclusions: EV miRNAs have high potential as biomarkers for predicting prognosis, treatment outcome, and other atherosclerotic lesion associated with LEAD in the diabetic patients.