- The study aims to create and validate an AI system for diagnosing vertebral compression fractures (VCFs) using X-ray images from nearly 1900 patients at multiple hospitals.
- A deep learning model based on the ResNet-18 architecture was tested, showing diagnostic accuracy ranging from 0.757 to 0.850 across various test groups, outperforming radiologists with less expertise.
- The results suggest that this AI diagnostic tool could significantly enhance the detection of VCFs in community hospitals, providing a reliable resource for healthcare professionals.