Prediction of atrial fibrillation from at-home single-lead ECG signals without arrhythmias

Abstract

Early identification of atrial fibrillation (AF) can reduce the risk of stroke, heart failure, and other serious cardiovascular outcomes. However, paroxysmal AF may not be detected even after a two-week continuous monitoring period. We developed a model to quantify the risk of near-term AF in a two-week period, based on AF-free ECG intervals of up to 24 h from 459,889 patch-based ambulatory single-lead ECG (modified lead II) recordings of up to 14 days. A deep learning model was used to integrate ECG morphology data with demographic and heart rhythm features toward AF prediction. Observing a 1-day AF-free ECG recording, the model with deep learning features produced the most accurate prediction of near-term AF with an area under the curve AUC = 0.80 (95% confidence interval, CI = 0.79–0.81), significantly improving discrimination compared to demographic metrics alone (AUC 0.67; CI = 0.66–0.68). Our model was able to predict incident AF over a two-week time frame with high discrimination, based on AF-free single-lead ECG recordings of various lengths. Application of the model may enable a digital strategy for improving diagnostic capture of AF by risk stratifying individuals with AF-negative ambulatory monitoring for prolonged or recurrent monitoring, potentially leading to more rapid initiation of treatment.

Implantable cardiac monitors: the digital future of risk prediction?

“…an AI model that is trained to directly predict the risk of future HF, as for example has been developed in the context of atrial fibrillation. 5 Finally, the present study showcases the effectiveness of dynamic risk assessment methodologies for rapidly changing conditions like HF. Such an approach has also been proven effective in the randomized SMART-MI trial 6 in patients after myocardial infarction in which ICM-based detection of asymptomatic yet significant arrhythmic events (including atrial fibrillation, higher-degree atrioventricular-block, and non-sustained ventricular tachycardia) allowed for an effective real-time reclassification of risk.…”

An Arrhythmia classification approach via deep learning using single-lead ECG without QRS wave detection

“…Their model achieved a precision of 0.964, recall of 0.948, and F1 score of 0.956. Gadaleta et al predicted the occurrence of AFIB in single lead ECG recordings from home settings using a deep learning model [ 29 ]. In their experiment, AFIB episodes needed to last for more than 30 s, and the model achieved a sensitivity of 0.80, specificity of 0.65, precision of 0.09, and F1 score of 0.17.…”

Citizen data sovereignty is key to wearables and wellness data reuse for the common good

“…Data-collecting sensors are built into modern smart devices, and users are steered towards the activation of recording by default 8 . These data, linked to later identified diseases, can be key to the development and advancement of predictive analytics, e.g., based on AI 9 – 11 . Such models are applicable to rare 12 and common diseases, either acute or chronic 3 , 6 , 7 .…”