The PhysioNet/Computing in Cardiology Challenge 2021 focuses on the identification of cardiac abnormalities in reduced-lead electrocardiogram (ECG) recordings. While the twelve-lead ECG is the standard diagnostic screening system for many cardiac abnormalities, reduced-lead ECGs are more accessible than twelve-lead ECGs and may have similar diagnostic potential in certain contexts. The 2021 Challenge builds on the 2020 Challenge, which only considered twelve-lead ECGs, to reveal the differential utility of reduced-lead ECGs.
For this Challenge, we asked participants to design working, open-source algorithms for identifying cardiac abnormalities in twelve-lead, six-lead, three-lead, and two-lead ECG recordings.
This Challenge provides several innovations. First, we sourced data from several geographically and demographically distinct institutions to assess the generalizability of the Challenge algorithms. A total of 43,101 recordings from five hospital systems across four continents were posted publicly for training the Challenge algorithms; a similar number of recordings, including recordings from hospital systems that are not represented in the public training data, were sequestered for testing. Second, we required Challenge participants to submit code for their models and code for training their models, improving the reproducibility of the resulting codebases.
So far, 113 teams have submitted a total of 438 algorithms, including 165 algorithms that we could successfully run in a reproducible and containerized cloud computing environment. These algorithms represent a diversity of approaches from both academia and industry for identifying cardiac abnormalities from standard twelve-lead and more accessible reduced-lead ECG recordings.