Abstract
INTRODUCTION: Focused cardiac ultrasound (FoCUS) has a high diagnostic yield and a rapid theoretical learning curve. FoCUS can be applied in stroke assessments performed by stroke neurologists when a cardioembolic stroke is suspected. PATIENTS AND METHODS: An international multicenter, prospective validation study was conducted to assess neurologists' ability to perform FoCUS. The FoCUS examination was defined as a simplified 2D transthoracic echocardiography. Neurologists and cardiologists performed the FoCUS independently and blinded. A twenty-question test evaluated neurologists' ability to recognize sources of cardioembolic stroke from recorded FoCUS studies. RESULTS: A total of 432 paired studies involving 216 patients were conducted across 11 centers. No significant differences were found between neurologists and cardiologists in detecting: Left Ventricle (LV) dysfunction (7.4% vs 7.9%, p = 0.834), LV dilation (2.8% vs 2.3%, p = 0.766), VC collapsibility (7.2% vs 9.1%, p = 0.501), Right Ventricle dysfunction (0.9% vs 0.9%, p = 0.999), and pericardial effusion (0.5% vs 1.9%, p = 0.212). Cohen Kappa showed substantial agreement for LV dysfunction (0.640), moderate for LV dilation (0.589), and fair for VC collapsibility (0.226). Neurologists demonstrated 93.82% sensitivity and 92.92% specificity for detecting embolic sources. Success rate for LV akinesia was 88% (16/18), LV dysfunction 83% (15/18), complex aortic plaque 88% (16/18), and mitral stenosis 55% (10/18). DISCUSSION AND CONCLUSION: Properly trained neurologists can reliably perform FoCUS, particularly for assessing LV function and dilation, with better results in patients with favorable echocardiographic windows. While VC assessment requires further training, neurologists demonstrated high accuracy in identifying cardioembolic sources (over 90% of cases correctly identified). This study supports implementing standardized FoCUS training for neurologists through collaboration with cardiology specialists to enhance stroke diagnostics and management.