Osteoporosis is the most common bone disease. Measuring bone mineral density (BMD) by dual energy X-ray absorptiometry (DXA) is considered the gold standard of osteoporosis diagnosis. The scanner produces an image from the scanned site and an algorithm calculates BMD. However, BMD's fracture prediction ability is only moderate and other visual information from the DXA image is neglected. This joint proposal between the University of Eastern Finland, Kuopio University Hospital, Lund University, Sweden, and the University of Bern, Switzerland, seeks to discover how deep learning methods could exploit that information to improve osteoporotic fracture risk prediction. It utilizes large image sets produced in other research projects and in public healthcare. This project has potential to improve fracture risk assessment using existing imaging facilities. Therefore, it may have an enormous impact on osteoporotic fracture prevention without increasing the costs to the healthcare system.

2021

2025