Towards AI Based Diagnosis of Rheumatic Heart Disease: Data Annotation and View Classification

Abstract

Rheumatic Heart Disease is a cardiovascular disease highly prevalent in developing countries partially because of inadequate healthcare infrastructure to treat Group A streptococcus pharyngitis and thereafter diagnose and document every case of Acute Rheumatic Fever, the immune-mediated antecedent of rheumatic heart disease. Secondary antibiotic treatment with penicillin injections after a diagnosis of Acute Rheumatic Fever and Rheumatic Heart Disease is used to prevent further attacks of Strep A, preferably prior to any heart valve damage. Echocardiographic screening for early detection of Rheumatic Heart Disease has been proposed as a method to improve outcomes but it is time-consuming, costly and few people are skilled enough to reach a correct diagnosis. Machine Learning is an emerging tool in analysing medical images; our aim is to automate the screening process of diagnosing rheumatic heart disease. In this paper, we present a web application to be used to label echocardiography data. These labelled data can then be used to develop machine learning models that can classify echocardiographic views of the heart and damaged valves from the echocardiograms.