Multi-parametric analyses to investigate dependencies of normal left atrial strain by cardiovascular magnetic resonance feature tracking.

ABSTRACT

Left-atrial (LA) strain is the result of complex hemodynamics, which may be better characterized using a multiparametric approach. Cardiovascular magnetic resonance (CMR) feature tracking was used to perform a comprehensive LA strain assessment of 183 enrolled healthy volunteers (11-70 years, 97 females, median 32.9 ± 28.3 years). Novel strain dependencies were assessed using multi-parametric regression (MPR) analyses. LA volumetric data, left ventricular strain, transmitral and pulmonary venous blood flow parameters were utilized to create clusters for MPR of all subjects and a heart rate controlled subgroup (pulse 60-75/min, N = 106). The LA reservoir(r) and conduit(c) strains of the total cohort were significantly elevated (p ≤ 0.001) in women (r 49.7 ± 12.9%, c 32.0 ± 11.0%) compared to men (r 42.9 ± 11.4%, c 26.1 IQ 10.5%). In contrast, there were no gender-specific differences (p > 0.05) for subgroup LA reservoir, conduit and booster(b) strains (all, r 47.3 ± 12.7%; c 29.0 IQ 15.5%; b 17.6 ± 5.4%) and strain rates (all, 2.1 IQ 1.0 s-1; - 2.9 IQ 1.5 s-1; - 2.3 IQ 1.0 s-1). MPR found large effect sizes (|R2|≥ 0.26) for correlations between strain and various cardiac functional parameters. Largest effect size was found for the association between LA conduit strain and LA indexed booster volume, LA total ejection fraction, left ventricular global radial strain and E-wave (|R2|= 0.437). In addition to providing normal values for sex-dependent LA strain and strain rate, no gender differences were found with modified heart rate. MPR analyses of LA strain/strain rate and various cardiac functional parameters revealed that heart rate control improved goodness-of-fit for the overall model.