A review of the diagnostic accuracy of fetal cardiac anomalies

Efficacy of Antenatal Ultrasound Examination in Diagnosis of Congenital Cardiac Anomalies in an Unselected Population: Retrospective Study from a Tertiary Centre

Background

In Low- and Middle-Income Countries like India, where the services and surgical care for Congenital Heart Disease (CHD) are available only in selected centres with geographical variations, it is important to detect Heart defects early and give the parents an opportunity to plan ahead for seeking appropriate care at the earliest. Several developments in recent years such as improvement of quality of ultrasound machines, sonographer's experience, skills and better description of cardiac views have contributed to improve detection rate.

Methods

A retrospective study was done between March 2016 and December 2019, and showed ultrasound evidence of CHD was included.

Results

The total number of morphology scans done during study period was 50,435. The number of congenital anomalies detected was 1482, out of which CHD was detected in 334 (22.5%). Outcome of 50 pregnancies were not available while the rest (284) were available for follow up in post-natal period. There were 51 cases of CHD, missed on routine antenatal morphological screening, which were diagnosed in the post-natal period. There were 18 cases of over-diagnosed CHD on antenatal scan, but were found to have normal echo findings after birth.

Conclusion

A systematic approach is crucial for practitioner to determine the patterns of associated defects. Use of step wise strategy helps in determining the correct diagnosis of isolated cardiac defect, associated with other system or a part of syndrome. Systematic audit of morphological scans could play an important role in improving the diagnostic accuracy, which in turn will lead to early detection.

An ensemble of neural networks provides expert-level prenatal detection of complex congenital heart disease

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.