Abstract
BACKGROUND: Serial assessment of right ventricular ejection fraction (RVEF) predicts the clinical outcome of patients with pulmonary hypertension (PH). Cardiac magnetic resonance imaging (CMRI) enables RVEF monitoring, but its applicability is limited in clinical practice. This study aimed to examine the correlation between changes in CMRI-derived RVEF with those in echocardiographic indices in patients with precapillary PH. METHODS: CMRI and echocardiographic indices of RV systolic function were evaluated at baseline and follow-up in 54 consecutive patients with precapillary PH (pulmonary arterial hypertension (PAH), n = 23; non-PAH, n = 31). During follow-up, medical treatment was optimized according to the guidelines for PH. Using CMRI-derived RVEF as the gold standard, we examined the accuracy of five echocardiographic indices by correlation analysis and receiver operating characteristic (ROC) analysis and by calculating sensitivity, specificity, and positive and negative predictive values. RESULTS: After an average period of 9.5 months, CMRI-derived RVEF improved from 30.2% ± 10.6% at baseline to 41.4% ± 11.3% at follow-up. These changes significantly correlated with those in the five echocardiographic indices, i.e., %RV fractional shortening (r = 0.27), %RV area change (r = 0.46), tricuspid annular plane systolic excursion (TAPSE) (r = 0.84), RV myocardial performance index (RVMPI) (r = -0.72), and systolic lateral tricuspid annular motion velocity (TVlat) (r = 0.66). Of these indices, %RV area change, TAPSE, and TVlat significantly correlated with those of CMRI-derived RVEF in both PAH and non-PAH subgroups. ROC analysis showed that improvement in echocardiographic indices predicted a pre-specified improvement in CMRI-derived RVEF (>2.9%), with TAPSE and TVlat showing better accuracy over the other three indices. CONCLUSIONS: Echocardiographic indices modestly correlate with the changes in CMRI-derived RVEF in precapillary PH patients. Comparison among the five echocardiographic indices revealed that TAPSE and TVlat provide better accuracy than %RV fractional shortening, %RV area change, and RVMPI.