Back

Heart failure grading using single-lead electrocardiography

Hasumi, E.; Fujiu, K.; Chen, Y.; Shimizu, Y.; Oshima, T.; Matsunaga, H.; Matsuda, J.; Matsubara, T. J.; Fukuma, N.; Yuxiang, L.; Sugita, J.; Nakayama, Y.; Saga, A.; Oguri, G.; Kojima, T.; Maru, Y.; Shoda, M.; Komuro, I.

2020-10-13 cardiovascular medicine
10.1101/2020.10.08.20209700 medRxiv
Show abstract

Convolutional neural networks (CNNs) applied to electrocardiograms (ECGs) have been showing utility for detecting left ventricular (LV) dysfunction1. Although early detection of reduced LV ejection fraction (rEF) could improve handling of heart failure (HF) with rEF (HFrEF), it is not sufficient to detect HF with preserved EF (HFpEF). Here we developed a CNN algorithm to classify the severity of HF based on single-lead ECG data, irrespective of EF. We trained a CNN using ECG data and the HF classification from 7,865 patients with HF. The CNN achieved an area under the receiver-operating characteristic curve (AUC) of 0.996 for distinguishing patients with HF of various severity from healthy controls. It is anticipated that early detection of HF and therapeutic management of HF patients can be improved by employing this CNN with a single-lead ECG device.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.