Identification of an optimal method for echocardiographic right ventricular mass estimation in a swine model of pulmonary hypertension

Abstract Background Right ventricular (RV) mass and Fulton Index (FI) - the ratio of right to left ventricular mass - are useful markers of RV remodeling and predictors of mortality in pulmonary hypertension (PH). Cardiac magnetic resonance imaging provides an accurate estimation of RV mass but cost, availability and patient contraindications limit broad application. While echocardiography is universally available and offers feasible longitudinal assessments, current guidelines lack a standardized method for RV mass determination. Purpose We compared the performance of four previously described echocardiographic models of RV mass to estimate actual RV weight and FI measured post-mortem in a swine model of PH. Methods Echocardiography was performed in a Yorkshire pig model of post-capillary PH induced by pulmonary vein banding and healthy animals using a Philips iE-33 ultrasound system (n=70). RV mass was measured using the following models: trapezoid method (1), exact radius method (2), longitudinal plane method (3), and an algorithm (4) proposed by Kochav's group. Figure 1 depicts pertinent echocardiographic views and parameters. A two-dimensional area-length method was applied to estimate left ventricular mass and FI was calculated. On the day of echocardiography, pigs were euthanized, hearts were explanted, and ventricular weights were recorded. Pearson’s correlation, linear regression and Bland-Altman analysis were employed to evaluate different methods compared to recorded weights and Fulton Index. Results PH animals presented with significant RV remodeling and higher RV weights (42.4g, range 18.4-93.6g) compared to control pigs (30.5g, range 11.1-46.0g), allowing for examination of echocardiographic methods across a wide span of RV weights. FI ranged from 0.29 to 3.60 with a mean of 0.70. All methods demonstrated a strong linear relationship with post-mortem RV weight (r(1) = 0.88, r(2) = 0.88, r(3) = 0.84 r(4) = 0.82, p<0.0001). Bland-Altmann analysis revealed good agreement with recorded RV mass for trapezoid, exact radius and longitudinal plane method (bias(1) = 4.5±7.5g, bias(2) = -5.9±6.6g, bias(3) = -2.7±8.4g) while Kochav’s algorithm tended to underestimate RV weights (bias(4) = -18.5±8.7g). Figure 2 illustrates the performance of exact radius and longitudinal plane methods. Measured left ventricular mass correlated strongly with post-mortem left ventricular weight and reached good agreement (r = 0.91, p<0.0001; bias = -4.0±9.8g). FI was best approximated by exact radius (bias = -0.04±0.12) and longitudinal plane (bias = 0.01±0.18) methods. The trapezoid model tended to overestimate FI (bias = 0.10±0.11) and Kochav’s algorithm underestimated (bias = -0.23±0.12). Conclusion RV mass and FI can be successfully approximated using echocardiography in a clinically relevant porcine PH model. Longitudinal plane and exact radius methods performed best and can be suggested for evaluating RV mass in clinical PH patients.Figure 1Figure 2