Grant: Project Grant

As one of the most common cardiac conditions, heart failure (HF) is a clinical syndrome that is associated with a high degree of disability and death in Canada and worldwide. The prevalence increases with age and mortality is high, with only 50% of the patients remaining alive at five years following diagnosis. Fortunately, many medical, surgical, and device treatments can improve symptoms and prognosis in patients with HF. The timely initiation of appropriate lifestyle, medication, and device treatments can be crucial to achieve favourable clinical outcomes. Comprehensive echocardiography is the clinical standard for the diagnosis of heart failure. However, since it heavily relies on expert analysis of video and blood flow measurements, less experienced users have been reluctant to adopt cardiac ultrasound into practice. As the population ages, so does the demand for comprehensive echocardiography exams, resulting in inappropriately prolonged wait time, which can result in significant delays in the treatment of HF. We propose empowering technology for the evaluation of HF in both the lab and point of care settings for automated detection and classification of left ventricular function. We aim to demonstrate that clinical HF can be detected and categorized with high sensitivity and specificity through direct analysis of ultrasound data using machine learning. We expect the proposed technology will allow front-line physicians to perform point-of-care ultrasounds in patients suspected of HF and triage patients who should be referred for a comprehensive echocardiogram, specialist consultation, and/or interventional procedures.