To develop and validate a deep learning model for real-time identification and localization of 17 common abdominal anomalies on ultrasound images, providing reliable technical support for primary care sonography.

Retrospectively collected 39 530 B-mode ultrasound images from 18 947 patients with abdominal anomalies at the Second Affiliated Hospital of Army Medical University from January 2017 to September 2024). A stratified random sample of 1 891 images formed an independent testing set; the remaining 37 639 images were split 7:1:2 into training (26 258), validation (3 840), and internal test (7 541) sets. After standardized annotation, RetinaNet, Faster R- CNN, RT-DETR, and YOLOv13 were compared on the internal test set using IoU, precision, recall, F₁-score, and mAP@50. The optimal model was further evaluated against 6 sonographers (3 novice, 3 intermediate) and novice sonographers assisted by the model on the independent test set.

YOLOv13 achieved superior performance on the internal testing set: mAP@50 0.653 (95%CI: 0.608 to 0.669) versus RetinaNet 0.546 (95%CI: 0.511 to 0.563), Faster R-CNN 0.578 (95%CI: 0.531 to 0.591), RT-DETR 0.627 (95%CI: 0.585 to 0.643); precision 0.633 (95%CI: 0.624 to 0.643), recall 0.715 (95%CI: 0.704 to 0.725), F₁-score 0.671 (95%CI: 0.663 to 0.681). On the independent testing set (IoU=0.5), YOLOv13’s F₁-score (0.624) exceeded all novice sonographers’mean F₁-score (0.484; P<0.001). Model assistance improved novice sonographers’mean F1-score from 0.484 to 0.644 (33.1% improvement; P<0.001), approaching intermediate-level performance (0.656); mean precision increased 20.6% (0.529 to o 0.638), mean recall increased 42.8% (0.472 to 0.674), and mean interpretation time decreased 24.4% (213.33 min to 161.33 min).

The YOLOv13 model demonstrates optimal performance for detecting 17 abdominal ultrasound anomalies and significantly enhances diagnostic efficiency of novice sonographers, serving as an effective assistive tool for abdominal ultrasound screening in primary care settings.