Abstract
Background
Nearly one-third of heart failure (HF) patients do not respond to cardiac resynchronization therapy (CRT) despite having left bundle branch block (LBBB). The aim of the study was to investigate a novel method of quantifying left ventricular (LV) contractile asymmetry in HF.
Methods
Patients with HF and LBBB undergoing CRT (n = 89, 37.1% females, 68 ± 9 years, ischemic etiology in 61%, LV ejection fraction 27.1 ± 7.1%) were analyzed. LV longitudinal systolic strain rate values were extracted from curved anatomical M-mode plots of standard long-axis 2D-echocardiography images and cubic spline interpolation was used to generate a 3D-phantom. Index of contractile asymmetry (ICA) was calculated based on standard deviation of differences in strain rate of opposing walls. Average ICA was individually assessed pairwise in 12 opposing 30-degree LV sectors. Reduction in LV end-systolic volume (ESV) ≥15% after 6 months was considered as positive response to CRT.
Results
CRT response was found in 66 (74.2%) patients. Responders with both ischemic and non-ischemic cardiomyopathy had a higher and more extensive contractile asymmetry at baseline and achieved a greater ICA reduction after CRT than non-responders. Higher baseline ICA predicted higher degree and wider extent of ICA improvement. Also, both ICA at baseline and reduction of ICA correlated with the degree of ESV reduction after CRT.
Conclusions
Quantification of asymmetrical LV activation in 3D by ICA provides valuable insights into LV contraction in case of LBBB and is a promising tool for improved patient selection for CRT.
Can machine learning improve patient selection for cardiac resynchronization therapy?