Deep learning to estimate cardiac magnetic resonance-derived left ventricular mass

Shaan Khurshid; Samuel Freesun Friedman; James P Pirruccello; Paolo Di Achille; Nathaniel Diamant; Christopher D Anderson; Patrick T Ellinor; Puneet Batra; Jennifer E Ho; Anthony A Philippakis; Steven A Lubitz

doi:10.1016/j.cvdhj.2021.03.001

Deep learning to estimate cardiac magnetic resonance-derived left ventricular mass

Cardiovasc Digit Health J. 2021 Mar 17;2(2):109-117. doi: 10.1016/j.cvdhj.2021.03.001. eCollection 2021 Apr.

Authors

Shaan Khurshid^{1

2}, Samuel Freesun Friedman³, James P Pirruccello^{1

2}, Paolo Di Achille³, Nathaniel Diamant³, Christopher D Anderson^{2

4

5}, Patrick T Ellinor^{2

6}, Puneet Batra³, Jennifer E Ho^{1

2}, Anthony A Philippakis³, Steven A Lubitz^{2

6}

Affiliations

¹ Division of Cardiology, Massachusetts General Hospital, Boston, Massachusetts.
² Cardiovascular Disease Initiative, Broad Institute of Harvard and the Massachusetts Institute of Technology, Cambridge, Massachusetts.
³ Data Sciences Platform, Broad Institute of Harvard and the Massachusetts Institute of Technology, Cambridge, Massachusetts.
⁴ Center for Genomic Medicine, Massachusetts General Hospital, Boston, Massachusetts.
⁵ Henry and Allison McCance Center for Brain Health, Massachusetts General Hospital, Boston, Massachusetts.
⁶ Cardiac Arrhythmia Service, Massachusetts General Hospital, Boston, Massachusetts.

Abstract

Background: Cardiac magnetic resonance (CMR) is the gold standard for left ventricular hypertrophy (LVH) diagnosis. CMR-derived LV mass can be estimated using proprietary algorithms (eg, InlineVF), but their accuracy and availability may be limited.

Objective: To develop an open-source deep learning model to estimate CMR-derived LV mass.

Methods: Within participants of the UK Biobank prospective cohort undergoing CMR, we trained 2 convolutional neural networks to estimate LV mass. The first (ML4H_reg) performed regression informed by manually labeled LV mass (available in 5065 individuals), while the second (ML4H_seg) performed LV segmentation informed by InlineVF (version D13A) contours. We compared ML4H_reg, ML4H_seg, and InlineVF against manually labeled LV mass within an independent holdout set using Pearson correlation and mean absolute error (MAE). We assessed associations between CMR-derived LVH and prevalent cardiovascular disease using logistic regression adjusted for age and sex.

Results: We generated CMR-derived LV mass estimates within 38,574 individuals. Among 891 individuals in the holdout set, ML4H_seg reproduced manually labeled LV mass more accurately (r = 0.864, 95% confidence interval [CI] 0.847-0.880; MAE 10.41 g, 95% CI 9.82-10.99) than ML4H_reg (r = 0.843, 95% CI 0.823-0.861; MAE 10.51, 95% CI 9.86-11.15, P = .01) and InlineVF (r = 0.795, 95% CI 0.770-0.818; MAE 14.30, 95% CI 13.46-11.01, P < .01). LVH defined using ML4H_seg demonstrated the strongest associations with hypertension (odds ratio 2.76, 95% CI 2.51-3.04), atrial fibrillation (1.75, 95% CI 1.37-2.20), and heart failure (4.67, 95% CI 3.28-6.49).

Conclusions: ML4H_seg is an open-source deep learning model providing automated quantification of CMR-derived LV mass. Deep learning models characterizing cardiac structure may facilitate broad cardiovascular discovery.

Keywords: Cardiovascular disease; Convolutional neural network; Deep learning; Left ventricular mass; Machine learning.

Abstract

Grants and funding