Automated Classification of First-Trimester Fetal Heart Views Using Ultrasound-Specific Self-Supervised Learning

TL;DR

This study introduces USF-MAE, a self-supervised learning model trained on unlabelled ultrasound images, which significantly improves the automated classification of fetal heart views in first-trimester echocardiography, aiding early detection of congenital heart disease.

Contribution

We developed and validated a self-supervised ultrasound foundation model, USF-MAE, that outperforms supervised models in classifying fetal heart views from ultrasound images.

Findings

USF-MAE achieved over 90% accuracy in classification.

The model outperformed supervised CNN baselines and natural image pretrained models.

Robust performance without extensive image preprocessing.

Abstract

Congenital heart disease remains the most common congenital anomaly and a leading cause of neonatal morbidity and mortality. Although first-trimester fetal echocardiography offers an opportunity for earlier detection, automated analysis at this stage is challenging due to small cardiac structures, low signal-to-noise ratio, and substantial inter-operator variability. In this work, we evaluate a self-supervised ultrasound foundation model, USF-MAE, for first-trimester fetal heart view classification. USF-MAE is pretrained using masked autoencoding modelling on more than 370,000 unlabelled ultrasound images spanning over 40 anatomical regions and is subsequently fine-tuned for downstream classification. As a proof of concept, the pretrained Vision Transformer encoder was fine-tuned on an open-source dataset of 6,720 first-trimester fetal echocardiography images to classify five categories: aorta, atrioventricular flows, V sign, X sign, and Other. Model performance was benchmarked against supervised convolutional neural network baselines (ResNet-18 and ResNet-50) and a Vision Transformer (ViT-B/16) model pretrained on natural images (ImageNet-1k). All models were trained and evaluated using identical preprocessing, data splits, and optimization protocols. On an independent test set, USF-MAE achieved the highest performance across all evaluation metrics, with 90.57% accuracy, 91.15% precision, 90.57% recall, and 90.71% F1-score. This represents an improvement of +2.03% in accuracy and +1.98% in F1-score compared with the strongest baseline, ResNet-18. The proposed approach demonstrated robust performance without reliance on aggressive image preprocessing or region-of-interest cropping and showed improved discrimination of non-diagnostic frames.