Biventricular volumetric assessment of cardiac function by a novel real-time cardiac magnetic resonance technique during strenuous exercise

Background: Changes in cardiac function during exercise are often inferred from indirect measures and there are few modalities which are able to accurately characterise the relative contribution of the left and right ventricles (LV and RV). There is increasing interest in RV function and its role in exercise performance in both health and disease. We sought to directly measure biventricular volumes during strenuous exercise using a novel cardiac magnetic resonance (CMR) imaging sequence. Methods: Twenty healthy and physically active subjects (17 male, 3 female, age 33 ± 9 years) underwent CMR at rest and during supine exercise on a cycle ergometer. Biventricular volumes were obtained at rest and whilst cycling at moderate (112±18 bpm; 172% resting heart rate (HR)) and strenuous (151±15 bpm; 232% resting HR) workload intensities. Images were acquired during exercise (free-breathing) using ungated steady-state free precession sequence cine images in short-axis and horizontal long-axis planes (12-18 contiguous 8mm slices). ECG and respiratory movements were measured and retrospectively synchronysed to the images such that bi-plane endocardial tracings could be performed with compensation for cardiac movement and translation (locally developed software). We sought to assess (1) feasibility, (2) reproducibility by inter-observer variability and (3) accuracy using LV/RV output equality as a surrogate. Results: Feasibility; images were acquired at rest through to strenuous exercise in all subjects and cardiac output (CO) increased by 177% (7.1±1.4 vs. 19.7±4.7 l/min, p<0.0001). A maximal CO of 33.2 l/min was measured in a professional cyclist. Reproducibility; there was excellent interobserver agreement for end-diastolic volume (mean difference 2.6±10.6 ml, inter-class correlation r=0.97, p<0.0001), ejection fraction (0.4±3.5%, r=0.86, p<0.0001) and CO (0.2±1.1 l/min, r=0.98, p