Background CMR image acquisition techniques during exercise typically require transient cessation of exercise or complex post-processing analysis, potentially compromising its clinical utility. We evaluated the feasibility and reproducibility of a novel image acquisition Method for the assessment of biventricular physiological response during continuous physical exercise.

Methods 10 healthy volunteers (80% men, age 25±2 years) underwent supine cycle ergometer (Lode) induced exercise CMR (Philips 1.5T Ingenia) on two separate occasions using a free-breathing, multi-shot, navigated, balanced steady-state free precession cine pulse sequence. Individual target heart rates (HR) for both moderate and high-intensity exercise were prescribed based on a prior supine cardiopulmonary exercise test (CPET). The scan protocol included a short axis ventricular volume stack and a 40 phase 4-chamber cine. Images were acquired at baseline, and during steady-state moderate and high-intensity exercise (55% and 75% maximal heart rate, respectively). Data were analysed by two independent observers and left and right ventricular (LV, RV) indices calculated.

Results End-diastolic volume (EDV) of both LV and RV decreased during moderate and high-intensity exercise, although the reduction in indexed RVEDV (RVEDVi) was only observed during maximal exercise (table 1). Similarly, a significant reduction in end-systolic volumes (ESV) was seen in both ventricles, whilst the reduction in indexed LVESV (LVESVi) was only evident during high-intensity exercise. Ejection fraction (EF) increased from rest to moderate and high intensity exercise in the LV (LVEF 58±5% vs 61±8% vs 68±3%, respectively; p<0.001), whereas RVEF was only significantly higher during high-intensity exercise (RVEF 58±7% vs 62±7% vs 66±4%; p<0.01). A biphasic change in global longitudinal strain (GLS) was observed; there was a significant increase in GLS during moderate-intensity exercise which appeared to plateau at maximal exercise. A similar biphasic change was observed for GLS rate (table 1).

Intra-observer reproducibility of LV parameters was excellent at all three stages (Table 2), although measurements of RVESV were more variable. The reproducibility of both RVEF and RV cardiac indexes was however excellent. Similarly, inter-observer reproducibility of LV volumes, EF and cardiac indexes was excellent. Inter-scan LV and RV ejection fraction were highly reproducible at all 3 stages; RVESVi reproducibility was suboptimal.

Conclusion This exercise CMR protocol using a novel free-breathing, multi-shot, navigated imaging Method allows simultaneous assessment of the left and right ventricular response to continuous exercise. Intra and inter-observer reproducibility were excellent. Clinical feasibility and utility now needs to be established.