How machine learning is impacting research in atrial fibrillation: implications for risk prediction and future management

Abstract

There has been an exponential growth of artificial intelligence (AI) and machine learning (ML) publications aimed at advancing our understanding of atrial fibrillation (AF), which has been mainly driven by the confluence of two factors: the advances in deep neural networks (DeepNNs) and the availability of large, open access databases. It is observed that most of the attention has centred on applying ML for dvsetecting AF, particularly using electrocardiograms (ECGs) as the main data modality. Nearly a third of them used DeepNNs to minimize or eliminate the need for transforming the ECGs to extract features prior to ML modelling; however, we did not observe a significant advantage in following this approach. We also found a fraction of studies using other data modalities, and others centred in aims, such as risk prediction, AF management, and others. From the clinical perspective, AI/ML can help expand the utility of AF detection and risk prediction, especially for patients with additional comorbidities. The use of AI/ML for detection and risk prediction into applications and smart mobile health (mHealth) technology would enable 'real time' dynamic assessments. AI/ML could also adapt to treatment changes over time, as well as incident risk factors. Incorporation of a dynamic AI/ML model into mHealth technology would facilitate 'real time' assessment of stroke risk, facilitating mitigation of modifiable risk factors (e.g. blood pressure control). Overall, this would lead to an improvement in clinical care for patients with AF.

Keywords: Artificial intelligence; Machine learning; Risk analysis; Wearables; Atrial fibrillation.

Graphical abstract

Figure 1

Growth in the number of ML in AF publications overall and by categories since 2012.

Figure 2

Trends of the ML algorithm families used in AF research. (A) All ML algorithms, with shallow and deep NNs grouped together as Artificial Neural Networks. (B) Separation of shallow and deep NNs.

Figure 3

Possible ML analysis for AF. The data, as input of the analysis, could be in the form of a single or multiple modalities of electronic health records (EHRs), electrocardiograms (ECGs), and/or other waveforms, and medical images, such as cardiac MRI and echocardiograms.

Figure 4

The role of publicly available databases for AF research compared to other (mainly proprietary) databases, in five AF research areas/topics: detection, risk prediction, wearables, management, and other areas.

Figure 5

Main publicly available databases used for AF research. (A) Uses of these databases by AF research areas/topics. (B) Use of these databases for AF detection for studies that use: (i) methods that rely on transformations of the ECG, (ii) methods that require little or no transformation of the ECG, (iii) methods for the detection of new onset AF, and (iv) other approaches for AF detection using ML.