Machine learning for multidimensional response and survival after cardiac resynchronization therapy using features from cardiac magnetic resonance

Derek J Bivona; Srikar Tallavajhala; Mohamad Abdi; Pim J A Oomen; Xu Gao; Rohit Malhotra; Andrew E Darby; Oliver J Monfredi; J Michael Mangrum; Pamela K Mason; Sula Mazimba; Michael Salerno; Christopher M Kramer; Frederick H Epstein; Jeffrey W Holmes; Kenneth C Bilchick

doi:10.1016/j.hroo.2022.06.005

Machine learning for multidimensional response and survival after cardiac resynchronization therapy using features from cardiac magnetic resonance

Heart Rhythm O2. 2022 Jun 17;3(5):542-552. doi: 10.1016/j.hroo.2022.06.005. eCollection 2022 Oct.

Authors

Derek J Bivona^{1

2}, Srikar Tallavajhala¹, Mohamad Abdi², Pim J A Oomen³, Xu Gao⁴, Rohit Malhotra¹, Andrew E Darby¹, Oliver J Monfredi¹, J Michael Mangrum¹, Pamela K Mason¹, Sula Mazimba¹, Michael Salerno⁵, Christopher M Kramer^{1

6}, Frederick H Epstein^{2

6}, Jeffrey W Holmes⁷, Kenneth C Bilchick¹

Affiliations

¹ Department of Medicine, University of Virginia Health System, Charlottesville, Virginia.
² Department of Biomedical Engineering, University of Virginia Health System, Charlottesville, Virginia.
³ Department of Biomedical Engineering, University of California, Irvine, California.
⁴ Department of Medicine, Northwestern University, Chicago, Illinois.
⁵ Departments of Medicine and Radiology, Stanford University, Palo Alto, California.
⁶ Department of Radiology and Medical Imaging, University of Virginia Health System, Charlottesville, Virginia.
⁷ Departments of Medicine, Surgery, and Biomedical Engineering, University of Alabama at Birmingham, Birmingham, Alabama.

Abstract

Background: Cardiac resynchronization therapy (CRT) response is complex, and better approaches are required to predict survival and need for advanced therapies.

Objective: The objective was to use machine learning to characterize multidimensional CRT response and its relationship with long-term survival.

Methods: Associations of 39 baseline features (including cardiac magnetic resonance [CMR] findings and clinical parameters such as glomerular filtration rate [GFR]) with a multidimensional CRT response vector (consisting of post-CRT left ventricular end-systolic volume index [LVESVI] fractional change, post-CRT B-type natriuretic peptide, and change in peak VO₂) were evaluated. Machine learning generated response clusters, and cross-validation assessed associations of clusters with 4-year survival.

Results: Among 200 patients (median age 67.4 years, 27.0% women) with CRT and CMR, associations with more than 1 response parameter were noted for the CMR CURE-SVD dyssynchrony parameter (associated with post-CRT brain natriuretic peptide [BNP] and LVESVI fractional change) and GFR (associated with peak VO₂ and post-CRT BNP). Machine learning defined 3 response clusters: cluster 1 (n = 123, 90.2% survival [best]), cluster 2 (n = 45, 60.0% survival [intermediate]), and cluster 3 (n = 32, 34.4% survival [worst]). Adding the 6-month response cluster to baseline features improved the area under the receiver operating characteristic curve for 4-year survival from 0.78 to 0.86 (P = .02). A web-based application was developed for cluster determination in future patients.

Conclusion: Machine learning characterizes distinct CRT response clusters influenced by CMR features, kidney function, and other factors. These clusters have a strong and additive influence on long-term survival relative to baseline features.

Keywords: Cardiac resynchronization therapy; Heart failure; Implantable cardioverter-defibrillator; Machine learning; Magnetic resonance imaging.

Grants and funding

R01 HL159945/HL/NHLBI NIH HHS/United States