Unmet Need:
Sudden cardiac death (SCD) is a leading cause of death in western countries, accounting for 15 – 20% of deaths (Kumar, et al. 2021). Currently, the standard for estimating risk for SCD is based on a left ventricular ejection fraction (LVEF) lower than 30-35% (Russo et al, 2013) which captures only 20% of arrhythmic sudden cardiac deaths (SCDA) (Wellens et al, 2014). There is a need for a more accurate, personalized risk assessment tool.
Technology Overview:
This technology incorporates two deep learning networks. One utilizes raw cardiac contrast-enhanced images and one uses as input standard clinical covariates. When combined within a survival analysis, these neural network generate a patient-specific survival curves over a 10-year time period as well as the uncertainty in the predicted times to SCDA. This approach is unlike any currently existing and makes major improvements to the personalized assessment of risk for SCDA.
Stage of Development:
Validated on patients in clinical trials.
Publication:
Popescu, D.M., Shade, J.K., Lai, C. et al. Arrhythmic sudden death survival prediction using deep learning analysis of scarring in the heart. Nat Cardiovasc Res 1, 334–343 (2022). https://doi.org/10.1038/s44161-022-00041-9
