Abstract Number: Esoc2026a120 Focus in Stroke: Validation of Neurologist-Performed Cardiac Ultrasound for Embolic Source Detection

Abstract Background and aims Focused Cardiac Ultrasound (FoCUS) is a point-of-care bedside examination performed by non-cardiologists to rapidly identify cardiac abnormalities. We validated neurologist-performed FoCUS for detecting cardiac embolic sources and assessing cardiac parameters in acute stroke patients. Methods Single-center retrospective study at Florence Stroke Unit. Following intensive structured training, a neurologist performed bedside FoCUS on 114 consecutive acute stroke/TIA patients. Images were blindly reviewed by a cardiologist (gold standard). Primary outcome: detection of any cardiac source of embolism (SOE). Secondary outcomes: left ventricular parameters and valve assessment. Statistical analysis included Cohen's kappa and bootstrap confidence intervals. Results Among 100 valid paired assessments, neurologist-performed FoCUS achieved excellent primary outcome results: sensitivity 97.0% (95%CI: 92.6-100%), specificity 85.3% (95%CI: 72.5-96.7%) and Cohen's kappa 0.841. Individual SOE analysis demonstrated high concordance across all embolic sources. For secondary outcomes, substantial-to-almost-perfect agreement was demonstrated for left ventricular size (κ=0.799), systolic function (κ=0.778), segmental wall motion (κ=0.806), and ejection fraction (κ=0.697). Valvular assessment showed almost perfect agreement for aortic regurgitation (κ=0.808), mitral regurgitation (κ=0.801), substantial agreement for aortic stenosis (κ=0.669), and moderate agreement for mitral stenosis (κ=0.551) and wall thickness (κ=0.596). Conclusions Appropriately trained stroke neurologists can perform FoCUS with diagnostic accuracy approaching cardiologist assessment for both embolic source detection and cardiac parameter evaluation, enabling immediate bedside assessment to guide acute management and secondary prevention strategies. Conflict of interest Fiore: Nothing to disclose, Masi: Nothing to disclose, Poggesi: Nothing to disclose, Nencini: Nothing to disclose, D'Alfonso: Nothing to disclose, Pescini: Nothing to disclose Figure 1 - belongs to Methods Figure 2 - belongs to Results Figure 3 - belongs to Conclusions