Researchers tested a new artificial intelligence (AI) tool designed to help doctors spot and classify a specific type of hip fracture called a femoral neck fracture. The study used over 10,000 hip images from 806 patients treated at four hospitals in China between 2018 and 2024. The AI model was trained to analyze these images and determine if a fracture was present and how severe it was.
In testing, the AI model correctly identified fractures with high accuracy (over 93%) on average. When the model's suggestions were shown to doctors, it seemed to help resident physicians—doctors in training—improve their own diagnostic accuracy. This suggests the tool could potentially help bridge the experience gap between junior and senior clinicians.
No safety issues were reported in this study, as it only looked at image analysis. The main reason for caution is that this was a retrospective study, meaning it analyzed past patient data. The researchers themselves note that prospective studies—where the tool is tested on new patients as they arrive—are needed to confirm its real-world usefulness. For now, this research shows a promising direction for AI-assisted diagnosis, but it is not yet ready to replace a doctor's judgment.