Machine Learning Quantification of Pulmonary Regurgitation Fraction from Echocardiography.

Assessment of pulmonary regurgitation (PR) guides treatment for patients with congenital heart disease. Quantitative assessment of PR fraction (PRF) by echocardiography is limited. Cardiac MRI (cMRI) is the reference-standard for PRF quantification. We created an algorithm to predict cMRI-quantified PRF from echocardiography using machine learning (ML). We retrospectively performed echocardiographic measurements paired to cMRI within 3 months in patients with ≥ mild PR from 2009 to 2022. Model inputs were vena contracta ratio, PR index, PR pressure half-time, main and branch pulmonary artery diastolic flow reversal (BPAFR), and transannular patch repair. A gradient boosted trees ML algorithm was trained using k-fold cross-validation to predict cMRI PRF by phase contrast imaging as a continuous number and at > mild (PRF ≥ 20%) and severe (PRF ≥ 40%) thresholds. Regression performance was evaluated with mean absolute error (MAE), and at clinical thresholds with area-under-the-receiver-operating-characteristic curve (AUROC). Prediction accuracy was compared to historical clinician accuracy. We externally validated prior reported studies for comparison. We included 243 subjects (median age 21 years, 58% repaired tetralogy of Fallot). The regression MAE = 7.0%. For prediction of > mild PR, AUROC = 0.96, but BPAFR alone outperformed the ML model (sensitivity 94%, specificity 97%). The ML model detection of severe PR had AUROC = 0.86, but in the subgroup with BPAFR, performance dropped (AUROC = 0.73). Accuracy between clinicians and the ML model was similar (70% vs. 69%). There was decrement in performance of prior reported algorithms on external validation in our dataset. A novel ML model for echocardiographic quantification of PRF outperforms prior studies and has comparable overall accuracy to clinicians. BPAFR is an excellent marker for > mild PRF, and has moderate capacity to detect severe PR, but more work is required to distinguish moderate from severe PR. Poor external validation of prior works highlights reproducibility challenges.

Age-related changes in left ventricular vortex and energy loss patterns: from newborns to adults.

Left ventricular vortex formation optimizes the effective transport of blood volume while minimizing energy loss (EL). Vector flow mapping (VFM)-derived EL patterns have not been described in children, especially in those less than 1 yr of age. A prospective cohort of 66 (0 days-22 yr, 14 patients ≤ 2 mo) cardiovascularly normal children was used to determine left ventricular (LV) vortex number, size (mm2), strength (m2/s), and energy loss (mW/m/m2) in systole and diastole and compared across age groups. One early diastolic (ED) vortex at the anterior mitral leaflet and one late diastolic (LD) vortex at the LV outflow tract (LVOT) were seen in all newborns ≤ 2 mo. At >2 mo, two ED vortices and one LD vortex were seen, with 95% of subjects > 2 yr demonstrating this vortex pattern. Peak and average diastolic EL acutely increased in the same 2 mo-2-yr period and then decreased within the adolescent and young adult age groups. Overall, these findings suggest that the growing heart undergoes a transition to adult vortex flow patterns over the first 2 yr of life with a corresponding acute increase in diastolic EL. These findings offer an initial insight into the dynamic changes of LV flow patterns in pediatric patients and can serve to expand our understanding of cardiac efficiency and physiology in children.NEW & NOTEWORTHY This research article demonstrates, for the first time, echocardiographic evidence of a transition in left ventricular vortex patterns from the newborn to the adult period, with an associated change in cardiac efficiency, marked by increased energy loss, during infancy.

Comparison between proximal thoracic vascular measurements obtained by contrast-enhanced magnetic resonance angiography and by transthoracic echocardiography in infants and children with congenital heart disease.

Accurate assessment of the proximal thoracic vasculature in infants and children with congenital heart disease (CHD) is vital for deciding the appropriate surgical or interventional procedure and predicting outcomes. This information usually is obtained by transthoracic echocardiography (TTE). Contrast-enhanced magnetic resonance angiography (CE-MRA) frequently is used to obtain diagnostic data when the image quality by TTE is limited. Calculation of z-scores for measurements obtained by CE-MRA in this population currently is not possible due to the lack of normative data. A reasonable agreement between vessel dimensions by CE-MRA and TTE will allow the use of TTE-based z-scores on measurements from CE-MRA. This study examines the accuracy and agreement of proximal thoracic vascular measurements obtained by CE-MRA versus TTE. Infants and children younger than 3 years with CHD who had a CE-MRA between August 2006 and May 2011 were retrospectively identified. Main and branch pulmonary arteries, ascending aorta, distal transverse arch, and aortic isthmus were measured from CE-MRA and TTE in analogous imaging planes and locations by two investigators blinded to each other. The study enrolled 35 subjects with CHD. The median age was 129 days (range, 0-1077 days), and the median weight was 5.8 kg (range, 2.16-17 kg). The median interval between the two imaging methods was 9 days (range, 0-60 days). Data analysis was performed with 129 of the 210 possible paired measurements. The remaining 81 paired measurements could not be performed due to inaccurate visualization of vessel borders or an unavailable imaging plane from TTE, CE-MRA, or both. The range of vessel sizes measured from 2.8 to 23.4 mm. There was excellent correlation between CE-MRA and TTE (r = 0.94, p < 0.001). The mean difference between the measurements was -0.1 ± 1.2 mm, and the limits of agreement were -2.5 to 2.3 mm. Proximal thoracic vascular measurements obtained by CE-MRA and TTE in infants and children with CHD have a strong correlation. The agreement between these two imaging methods is adequate. Until normative data for vessel size measurements obtained from CE-MRA are available for this population, TTE-based z-scores can be applied to the measurements obtained by CE-MRA.