Machine Learning and the Future of Cardiovascular Care: JACC State-of-the-Art Review

Abstract

The role of physicians has always been to synthesize the data available to them to identify diagnostic patterns that guide treatment and follow response. Today, increasingly sophisticated machine learning algorithms may grow to support clinical experts in some of these tasks. Machine learning has the potential to benefit patients and cardiologists, but only if clinicians take an active role in bringing these new algorithms into practice. The aim of this review is to introduce clinicians who are not data science experts to key concepts in machine learning that will allow them to better understand the field and evaluate new literature and developments. The current published data in machine learning for cardiovascular disease is then summarized, using both a bibliometric survey, with code publicly available to enable similar analysis for any research topic of interest, and select case studies. Finally, several ways that clinicians can and must be involved in this emerging field are presented.

Keywords: artificial intelligence; bibliometric analysis; cardiology; deep learning; literature search; machine learning.

Figure 1.. Marked growth in cardiac AI/ML studies.

The plots in the first row show numbers of publications per year; red denotes original research, while pink denotes non-original articles such as reviews and commentary. The second row shows the total number of papers (any topic) added to each of the three collections, while the third row presents the percentage of papers on cardiac AI/ML. The insets for Pubmed in the first two rows show the cumulative number of papers across all previous years, while in the third row it shows the cumulative percentages considering all papers till that year. *2020 numbers are partial.

Figure 2.. Content in cardiac AI/ML publications on PubMed.

(a) Distribution of publications by disease category. (b) Distribution of publications by data modality. (c) Number of papers studying various cardiac disease categories by data type (waveform data includes catheterization, arterial pulse waveforms, plethysmography, and other waveform data but does not include ECG; atherosclerosis includes dyslipidemia, peripheral vascular disease and cerebrovascular disease). (d) Distribution of publications by goal. (e) Distribution of publications by machine learning method.

Figure 3.. An international and collaborative effort.

(a) Total publications on cardiac AI/ML in PubMed, worldwide and in the United States (note log scale). (b) Publications per capita worldwide and in the United States (note log scale). (c) Collaborations measured by author locations on each publication, presented worldwide and statewide (for clarity, only top ten countries and/or states are labeled).

Central Illustration.. Marked growth in cardiac AI/ML studies span several disease focus areas and several data types.

Publications on machine learning in cardiology continue to grow (inset shows cumulative number of cardiac AI/ML papers). Several disease categories and data modalities have been studied in these publications, but several areas remain open for further exploration. Asterisk indicates partial information for the year 2020.