Assessment of Right Ventricular Volume in Pediatric Patients

The knowledge of right ventricular (RV) volumes is important from a clinical prospective. Traditionally, these volumes have been measured by cardiac magnetic resonance (CMR) because of its accuracy, reproducibility, and freedom from geometric assumptions. More and more, however, 3D echo has been used for assessment of ventricular volumes, first for left ventricle, now increasingly for RV. Most publication are basing RV volumes on images acquired from the apical window. However, some data have called the practice of basing these images on apical window into question. Moreover, despite our best efforts, we have been unable, in our laboratory to reliably visualize RV outflow from the apical view. An alternative view that can be used for RV 3D volume acquisition (first pioneered at BCH) is the subcostal view. Therefore, we plan to conduct a head-to-head comparison of RV volumes acquired from these two windows (apical and subcostal), with comparison to gold-standard CMR.

Hypothesis: RVEDV measured by the subcostal window is more accurate than RV volume measured in the apical window in a cohort of pediatric patients.

Analytic plan: Echocardiograms will be performed on the same day as the CMR, and will be performed by a group of sonographers trained in 3D volume acquisition. Descriptive statistics will include mean and median absolute % error and coefficients of variation. A paired t-test will be used to compare raw differences in %error for apical vs. subcostal window measurements. If the %error is highly skewed, a Wilcoxon signed-rank test will be used or the %error values will be log-transformed before application of the paired t-test (if no zero %error values exist). Linear regression will be used to estimate the window-differences in %error or log(%error) controlling for age and/or sedation status. Assessments of feasibility for both windows will be performed as described by Renella et al; and reproducibility of measurements made with echo vs. CMR will be also be graphically assessed using Bland-Altman plots of the raw measurements for data derived from each window. As a secondary (stratified) analysis, we anticipate dividing the cohort into 2 groups of 25 patients each (25 < 10y; 25 > 10y). Patients under 10y will typically be under anesthesia. Exploratory analyses will investigate subgroup differences defined by demographic/ surgical variables with respect to the magnitude of the difference between the 2 techniques. Of note the analyses performed above will also be performed for LV (in addition to RV) as a secondary analysis (with comparison of LV-specific vs. non-LV specific software, as appropriate).

Sample size/Power: funding is provided for 50 patients and this will be our target sample size. The analysis will be paired, to compare the %error (echo relative to MRI) measurements from apical vs. subcostal windows). To detect a 0.5 SD difference in apical vs. subcostal windows %error with 85% power, 38 subjects with echo & MRI are required. If the comparisons are performed stratified by age (25 subjects per age group), there is 80% power to detect a 0.58 SD difference in apical vs. subcostal windows %error.

Limitations: CMR, while considered a gold-standard, has a certain amount of intrinsic variability in volume measurement. Moreover, a statistically significant difference may not necessary be a clinically important difference. Patients will not have echo and CMR performed at the exact same time.