Congenital heart disease (CHD) is the most common birth defect. Fetal screening ultrasound provides five views of the heart that together can detect 90% of complex CHD, but in practice, sensitivity is as low as 30%. Here, using 107,823 images from 1,326 retrospective echocardiograms and screening ultrasounds from 18- to 24-week fetuses, we trained an ensemble of neural networks to identify recommended cardiac views and distinguish between normal hearts and complex CHD. We also used segmentation models to calculate standard fetal cardiothoracic measurements. In an internal test set of 4,108 fetal surveys (0.9% CHD, >4.4 million images), the model achieved an area under the curve (AUC) of 0.99, 95% sensitivity (95% confidence interval (CI), 84-99%), 96% specificity (95% CI, 95-97%) and 100% negative predictive value in distinguishing normal from abnormal hearts. Model sensitivity was comparable to that of clinicians and remained robust on outside-hospital and lower-quality images. The model's decisions were based on clinically relevant features. Cardiac measurements correlated with reported measures for normal and abnormal hearts. Applied to guideline-recommended imaging, ensemble learning models could significantly improve detection of fetal CHD, a critical and global diagnostic challenge.

Background

The clinical utility of echocardiography for the diagnosis of pulmonary vascular disease (PVD) in former preterm infants with bronchopulmonary dysplasia (BPD) is not established. Elevated pulmonary vascular resistance (PVR) rather than pulmonary artery pressure (PAP) is the hallmark of PVD. We evaluated the utility of echocardiography in infants with BPD in diagnosing pulmonary hypertension and PVD (PVR >3 Wood units × m²) assessed by cardiac catheterization.

Methods

A retrospective single center study of 29 infants born ≤29 weeks of gestational age with BPD who underwent cardiac catheterization and echocardiography was performed. PVD was considered present by echocardiography if the tricuspid valve regurgitation jet peak velocity was >2.9 m/sec, post-tricuspid valve shunt systolic flow velocity estimated a right ventricular systolic pressure >35 mm Hg, or systolic septal flattening was present. The utility (accuracy, sensitivity, and positive predictive value [PPV]) of echocardiography in the diagnosis of PVD was tested. Subgroup analysis in patients without post-tricuspid valve shunts was performed. Echocardiographic estimations of right ventricular pressure, dimensions, function, and pulmonary flow measurements were evaluated for correlation with PVR.

Results

The duration between echocardiography and cardiac catheterization was a median of 1 day (interquartile range, 1-4 days). Accuracy, sensitivity, and PPV of echocardiography in diagnosing PVD were 72%, 90.5%, and 76%, respectively. Accuracy, sensitivity, and PPV increased to 93%, 91.7%, and 100%, respectively, when infants with post-tricuspid valve shunts were excluded. Echocardiography had poor accuracy in estimating the degree of PAP elevation by cardiac catheterization. In infants without post-tricuspid valve shunts, there was moderate to good correlation between indexed PVR and right ventricular myocardial performance index (rho = 0.89, P = .005), systolic to diastolic time index (0.84, P < .001), right to left ventricular diameter ratio at end systole (0.66, P = .003), and pulmonary artery acceleration time (0.48, P = .05).

Conclusions

Echocardiography performs well in screening for PVD in infants with BPD and may be diagnostic in the absence of a post-tricuspid valve shunt. However, cardiac catheterization is needed to assess the degree of PAP elevation and PVR. The diagnostic utility of echocardiographic measurements that correlate with PVR should be evaluated prospectively in this patient population.

Background

Neonates with persistent pulmonary hypertension of the newborn (PPHN) can present with hypoxia and right ventricular dysfunction with resultant inadequate oxygen delivery and end-organ damage. This study describes the use of prostaglandin-E1 (PGE) for ductal patency to preserve right ventricular systolic function and limit afterload in newborns with PPHN.

Methods

This is a retrospective cohort study that follows the hemodynamics, markers of end-organ perfusion, length of therapeutics, and echocardiographic variables of 57 newborns who used prostglandin-E1 in the setting of PPHN.

Results

Tachycardia, lactic acidosis, and supplemental oxygen use improved following PGE initiation. Fractional area change (FAC), to assess right ventricular systolic function, and pulmonary arterial acceleration time indexed to right ventricular ejection time (PAAT/RVET), to assess right ventricular afterload, also improved over three time points relative to PGE use (before, during, and after).

Conclusions

Overall, we described the safety and utility of PGE in newborns with severe PPHN for stabilization while allowing natural disease progression.