Prenatal diagnosis of hypoplastic left heart syndrome on ultrasound using artificial intelligence: How does performance compare to a current screening programme?

Background Artificial intelligence (AI) has the potential to improve prenatal detection of congenital heart disease. We analysed the performance of the current national screening programme in detecting hypoplastic left heart syndrome (HLHS) to compare with our own AI model. Methods Current screening programme performance was calculated from local and national sources. AI models were trained using four‐chamber ultrasound views of the fetal heart, using a ResNet classifier. Results Estimated current fetal screening programme sensitivity and specificity for HLHS were 94.3% and 99.985%, respectively. Depending on calibration, AI models to detect HLHS were either highly sensitive (sensitivity 100%, specificity 94.0%) or highly specific (sensitivity 93.3%, specificity 100%). Our analysis suggests that our highly sensitive model would generate 45,134 screen positive results for a gain of 14 additional HLHS cases. Our highly specific model would be associated with two fewer detected HLHS cases, and 118 fewer false positives. Conclusion If used independently, our AI model performance is slightly worse than the performance level of the current screening programme in detecting HLHS, and this performance is likely to deteriorate further when used prospectively. This demonstrates that collaboration between humans and AI will be key for effective future clinical use. Key points What is already known on this topic? Artificial intelligence (AI) can be used to interpret medical images and make diagnoses, including detecting fetal congenital heart disease (CHD) by ultrasound. The sensitivity of the current English screening programme for fetal cardiac malformations is publicly available, but specificity is not reported. What this study adds? The current screening programme in our region is operating at a very high specificity for fetal hypoplastic left heart syndrome (HLHS). Using a curated retrospective dataset, it is possible to train AI models to detect HLHS with a performance approaching that of the current screening programme. Current AI models do not have high enough specificity to be used independently for screening for fetal CHD, meaning that human‐AI interaction when performing or interpreting ultrasound will be important to select cases for specialist referral.