Test-retest reliability of left and right ventricular systolic function by new and conventional echocardiographic and cardiac magnetic resonance parameters.

Abstract

Reproducible evaluation of left (LV) and right ventricular (RV) function is crucial for clinical decision-making and risk stratification. We evaluated whether speckle-tracking echocardiography (STE) and cardiac magnetic resonance feature-tracking (cMR-FT) global longitudinal (GLS) and circumferential strains allow better test-retest reproducibility of LV and RV systolic function than conventional cMR and echocardiographic parameters. Thirty healthy volunteers and 20 chronic heart failure patients underwent cMR and STE twice on separate days to evaluate test-retest coefficient of variation (CV), intraclass correlation coefficient (ICC) and estimated sample sizes for significant changes in LV and RV function. Among LV parameters, cMR-left ventricular ejection fraction (LVEF) had the highest reproducibility (CV = 6.7%, ICC = 0.98), significantly better than cMR-FT-GLS (CV = 15.1%, ICC = 0.84), global circumferential strains (CV = 11.5%, ICC = 0.94) and echocardiographic LVEF (CV = 11.3%, ICC = 0.93). STE-LV-GLS (CV = 8.9%, ICC = 0.94) had significantly better reproducibility than cMR-FT-LV-GLS. Among RV parameters, STE-RV-GLS (CV = 7.3%, ICC = 0.93) had significantly better CV than cMR-right ventricular ejection fraction (RVEF) (CV = 13%, ICC = 0.82). cMR-FT-RV-GLS (CV = 43%, ICC = 0.39) performed poorly with significantly lower reproducibility than all other RV parameters. Owing to their superior interstudy reproducibility, cMR-LVEF (n = 12), cMR-RVEF (n = 41), STE-LV-GLS and STE-RV-GLS (both n = 14) were the parameters allowing the lowest calculated sample sizes to detect 10% change in LV or RV systolic function. STE-LV-GLS and STE-RV-GLS showed higher test-retest reliability than other echocardiographic measurements of LV and RV function. They also allowed smaller calculated sample sizes, supporting the use of STE-LV and RV-GLS for longitudinal follow-up of LV and RV function.