Manual correction of semi-automatic three-dimensional echocardiography is needed for right ventricular assessment in adults; validation with cardiac magnetic resonance

BACKGROUND: Three-dimensional echocardiography (3DE) and semi-automatic right ventricular delineation has been proposed as an appropriate method for right ventricle (RV) evaluation. We aimed to examine how manual correction of semi-automatic delineation influences the accuracy of 3DE for RV volumes and function in a clinical adult setting using cardiac magnetic resonance (CMR) as the reference method. We also examined the feasibility of RV visualization with 3DE. METHODS: 62 non-selected patients were examined with 3DE (Sonos 7500 and iE33) and with CMR (1.5T). Endocardial RV contours of 3DE-images were semi-automatically assessed and manually corrected in all patients. End-diastolic (EDV), end-systolic (ESV) volumes, stroke volume (SV) and ejection fraction (EF) were computed. RESULTS: 53 patients (85%) had 3DE-images feasible for examination. Correlation coefficients and Bland Altman biases between 3DE with manual correction and CMR were r = 0.78, -22 ± 27 mL for EDV, r = 0.83, -7 ± 16 mL for ESV, r = 0.60, -12 ± 18 mL for SV and r = 0.60, -2 ± 8% for EF (p < 0.001 for all r-values). Without manual correction r-values were 0.77, 0.77, 0.70 and 0.49 for EDV, ESV, SV and EF, respectively (p < 0.001 for all r-values) and biases were larger for EDV, SV and EF (-32 ± 26 mL, -21 ± 15 mL and - 6 ± 9%, p ≤ 0.01 for all) compared to manual correction. CONCLUSION: Manual correction of the 3DE semi-automatic RV delineation decreases the bias and is needed for acceptable clinical accuracy. 3DE is highly feasible for visualizing the RV in an adult clinical setting.