Feasibility and Reproducibility of Three-Dimensional Echocardiographic Assessment of Right Ventricular Size and Function in Pediatric Patients


Quantitative assessment of right ventricular (RV) size and function is important in congenital heart disease. Although three-dimensional (3D) echocardiography overcomes geometric challenges that limit two-dimensional echocardiography, its feasibility and reproducibility in children have not been systematically evaluated. The goal of this study was to determine the feasibility and reproducibility of 3D echocardiographic RV measurements in children.


Patients aged 0 to 20 years were prospectively enrolled. Full-volume 3D echocardiographic RV data sets were acquired in each from the subxiphoid and apical four-chamber views by a trained sonographer. Another sonographer then acquired a second image data set from the same patient. RV volumes and ejection fraction were then measured offline. Outcomes included successful acquisition, measurable data set, and observer and interstudy variability.


Three-dimensional echocardiographic RV data sets were obtainable in 67 of 70 patients (96%) and were measurable from at least one view in 39 (58%). Factors associated with nonmeasurable data sets by univariate analysis included older age, larger body surface area and body mass index, no sedation, and female gender. Multivariate analysis identified older age (odds ratio, 1.34; 95% confidence interval, 1.17–1.55, P < .0001) and female gender (odds ratio, 6.06; 95% confidence interval, 1.45–25.3, P = .0135) as independently associated with nonmeasurable RV data sets. Intraobserver, interobserver, and interstudy agreement was excellent for analyzable RV volumes and modest for RV ejection fraction.


Transthoracic 3D echocardiographic RV volumes were measurable in 58% of children with and without congenital heart disease. Older age and female gender were independently associated with nonanalyzable data sets. When feasible, measurements of RV volumes are highly reproducible.

The right ventricle is frequently involved in congenital heart disease (CHD), and measurements of its size and function are increasingly important for informing clinical decisions. Conventional two-dimensional echocardiographic assessment of right ventricular (RV) volumes is hindered by the chamber’s complex geometry and retrosternal location within the chest. Consequently, in many patients with CHD, cardiac magnetic resonance (CMR) has become the noninvasive imaging modality of choice. However, compared with transthoracic echocardiography, CMR is more expensive, is not portable, and requires sedation in young patients.

Several studies have demonstrated the ability of transthoracic three-dimensional (3D) echocardiography (3DE) to assess RV volumes and function in children and adults. However, most studies have reported results of only successfully analyzed data sets. Hence, the feasibility of 3D echocardiographic quantification of RV size and function in a population of children with a wide range of ages, body size, and cardiac pathology evaluated in a standard clinical setting has not been determined. Thus, the purpose of this prospective study was to determine the feasibility of acquiring diagnostic-quality full-volume RV data sets using transthoracic 3DE in a “real world” practice of pediatric heart disease and CHD. In addition, we sought to determine the interstudy reproducibility of these measurements and to identify factors associated with a low likelihood of obtaining diagnostic 3D echocardiographic RV data sets.



We prospectively enrolled 70 patients fulfilling the following inclusion criteria: (1) age from day of life 1 to 20 years, (2) request for a clinical echocardiographic examination either in the cardiology outpatient clinic or in the inpatient service at Boston Children’s Hospital, and (3) informed consent. Exclusion criteria were (1) inability to attempt the acquisition of a 3D echocardiographic RV data set because of a lack of patient cooperation and (2) absence of an identifiable right ventricle. To ensure a study population representing a wide range of ages and cardiovascular pathology, subjects were recruited by age strata (0–3, 3–8, 8–13, 13–18, and 18–20 years) and by history of cardiac surgery. Subjects were recruited consecutively, but a given age stratum was closed to enrollment when 20 subjects were recruited (10 with histories of cardiac surgery and 10 without). The study protocol was approved by the Boston Children’s Hospital Committee on Clinical Investigations.

Transthoracic 3DE of the Right Ventricle

Echocardiographic studies were performed using iE33 cardiac ultrasound systems (Philips Medical Systems, Andover, MA) according to our published laboratory protocol. When clinically indicated, young (age < 3 years), uncooperative patients were sedated according to our laboratory’s outpatient sedation protocol (oral chloral hydrate at a dose of 80 mg/kg up to a maximum of 1 g). Transducer frequency (X3-1 or X7-2 matrix probe) and acquisition parameters were optimized according to patient size and acoustic windows. Harmonic imaging was often used to optimize signal-to-noise ratio and to enhance visualization of the blood-endocardial boundaries. All data sets were acquired with the transducer in the subxiphoid (SubX) and apical positions, ensuring that the entire right ventricle could be viewed simultaneously in both orthogonal planes. When possible, the patient was asked to suspend respiration and remain motionless during the four-beat acquisition phase. The 3D volume data set images were then evaluated to ensure that the entire chamber was optimally imaged with minimal spatial and temporal artifacts.

After informed consent was obtained and the clinical echocardiographic examination was completed, a sonographer trained in 3DE acquired two full-volume 3D echocardiographic data sets of the right ventricle, one from the SubX acoustic window and another from an apical four-chamber (A4C) view. Once these data sets were acquired, the sonographer left the examination room, and a second sonographer (blinded to the previously acquired 3D echocardiographic data sets) obtained a second pair of 3D echocardiographic RV data sets from the same views. Thus, a total of four data sets of the right ventricle were attempted in each subject. All 11 sonographers who participated in this study underwent prior training in 3DE. The sequence of 3D RV data set acquisition was not scripted and was performed randomly at the discretion of the sonographer. Sonographers were allowed to repeat an acquisition if they were not satisfied with the technical quality of the image data set.

Data Analysis

Volumetric Measurements

Acquisitions were stored in Digital Imaging and Communications in Medicine format and transferred to a dedicated workstation for offline analysis using commercially available software (4-D EchoView; TomTec, Munich, Germany). RV end-diastolic and end-systolic volumes were measured offline from each data set. Each data set was aligned in two orthogonal planes along the long axis of the right ventricle by one investigator (P.R.). The appropriate end-diastolic and end-systolic frames in sequential cross-sectional planes were then selected. The endocardial border in each slice was then manually traced using the method of disk summation ( Figure 1 ). RV ejection fraction was then calculated using the formula 100 × (RV end-diastolic volume − RV end-systolic volume)/RV end-diastolic volume.

Figure 1

Analysis of a real-time 3D echocardiographic image of the RV. (A) Three-dimensional echocardiographic image showing a full-volume 3D image of the right ventricle. (B) Three-dimensional echocardiograhpic image acquisition showing the right ventricle in orthogonal planes. (C) Manual tracing of the RV endocardial borders in long- and short-axis planes.

Image Quality Score

Each of the four data sets from every subject was analyzed by a single investigator (P.R.), scored for quality, and categorized as analyzable or nonanalyzable. A single investigator determined the image quality score. Data sets were assigned an image quality score according the following predetermined criteria: (1) poor (right ventricle not fully included in the data set and/or indefinable blood pool–endocardium border), (2) borderline (complete data set with markedly blurred endocardial border), (3) good (complete data set with mildly blurred endocardial border), and (4) excellent (complete data set with sharp endocardial border). A data set was considered nonanalyzable if its quality score was 1.

Statistical Analysis

Feasibility of 3D echocardiographic RV data set acquisition was determined for each of the two echocardiographic views (SubX and A4C) and on a per patient basis and defined as the percentage of subjects in whom data sets were deemed analyzable. To determine intraobserver reproducibility, the same investigator (P.R.) repeated all measurements on 10 randomly selected patients ≥3 weeks after completing the first set of measurements. For interobserver reproducibility, a second investigator (G.R.M.) analyzed the RV data sets from 10 randomly selected patients without knowledge of the results of the primary investigator. Interstudy reproducibility was determined by one investigator (P.R.), who analyzed the second set of 3D echocardiographic RV data (acquired by the second sonographer) on each subject ≥3 weeks after the original analysis. For each type of variability (intraobserver, interobserver, and interstudy), there were 20 sets of measurements, 10 from the SubX acoustic window and 10 from the A4C window. However, the final number of paired measurements was influenced by missing data (i.e., not all patients had analyzable data sets when repeated measurements were obtained). Paired t tests were used to determine if there were significant differences in measurements between the different observers and across the two studies in the same patient. The absolute difference divided by the mean of the repeated observations was determined. Bland-Altman plots were constructed to provide a visual assessment of the levels of intraobserver, interobserver, and interstudy variability. Interstudy variability was also assessed using intraclass correlation coefficients estimated with variance components models.

To identify factors associated with nonanalyzable 3D echocardiographic RV data sets, candidate predictors were compared between groups using Fisher’s exact test for categorical variables and Wilcoxon’s rank-sum test for continuous variables. Univariate and multivariate logistic regression analyses were performed; a P values ≤ .05 were required for retention in the final model. Statistical analysis was performed using SAS version 9.2 (SAS Institute Inc, Cary, NC).


Patient Characteristics

Patient demographics are summarized in Table 1 , and their primary cardiac diagnoses are detailed in Table 2 . As expected from the study design, the spectrum of ages from birth to 20 years and a broad range of body sizes were well represented. Nearly 20% of the patients had CHD-associated RV dilatation, and 44% had prior cardiac surgery. Of the 70 patients, 16 (23%) required sedation for their clinical echocardiographic studies.

Table 1

Patient characteristics ( n = 70)

Variable Value
Age (y) 7.8 (0.002–20.6)
Female gender 35 (50%)
Weight (kg) 25.5 (2.7–106.5)
Height (cm) 117 (47–191)
Body surface area (m 2 ) 0.9 (0.2–2.4)
Body mass index (kg/m 2 ) 17.9 (12.4–30.3)
Sedation 16 (23%)
Post cardiac surgery 31 (44%)
Dilated right ventricle 13 (19%)

Data are expressed as median (range) or as number (percentage).

Median age of sedated patients, 0.58 years (range, 8 days to 3 years).

May 31, 2018 | Posted by in CARDIOLOGY | Comments Off on Feasibility and Reproducibility of Three-Dimensional Echocardiographic Assessment of Right Ventricular Size and Function in Pediatric Patients

Full access? Get Clinical Tree

Get Clinical Tree app for offline access