Assessment of Positive Cardiac Remodeling in Hypertrophic Obstructive Cardiomyopathy Using an Artificial Intelligence–Based Electrocardiographic Platform in Patients Treated With Mavacamten

Comprehensive Summary

This research uses a convolutional neural network (CNN) to analyze the efficacy of the cardiac myosin inhibitor mavacamten in treating patients with obstructive hypertrophic cardiomyopathy (oHCM). A standard 12-lead ECG was taken of sixteen oHCM patients pre- and post- a 12 week treatment of mavacamten. The CNN was developed and internally validated with thousands of both HCM and control ECGs to estimate probabilities of atrial fibrillation (AF), low ejection fraction, and HCM severity. The median risk of AF dropped significantly from a baseline probability of 14.14% to 9.66% post-treatment. Additionally, HCM severity, characterized by left ventricular outflow tract gradient and NTproBNP levels, dropped from 71.65% to 6.76%. In addition to the AI-predictions, mavacamten treatment was associated with non-significant reductions in transverse left atrial diameter (46.81± 4.79 to 44.08± 5.22) and left atrial volume index (42.12± 12.40 to 37.93±12.78).

Outcomes and Implications

Mavacamten is a fairly new first-in-class inhibitor that may eventually provide an alternate treatment option for oHCM rather than the traditional surgical route. However, whether myosin inhibitors can actually remodel cardiac structure and reduce HCM severity is still being explored. This research sheds a hopeful light on myosin inhibitor efficacy, with significant reductions in both risk of AF and HCM severity. However, no statistically significant reductions in left atrial structure were found. More research must be conducted, with larger cohorts and over longer periods of time, to fully explore the effects of mavacamten and its implications for patients with oHCM.