Automatic, Opportunistic Surveillance of Hip Bone Fragility in X-ray Images - PROJECT SUMMARY / ABSTRACT Approximately 1 in 3 women and 1 in 5 men over the age of 50 will suffer from a fragility fracture in their remaining lifetime. Fragility hip fracture is one of the most serious and debilitating outcomes of osteoporosis with a 20–40% mortality rate during the first year after the fracture. Hip fracture incidence rates are known to increase exponentially with age in both women and men; and with the rising life expectancy throughout the globe, the number of men and women who will be above the threshold of fragility fracture is expected to almost double, with a prediction of 319 million cases by 2040. Thus, the number of fractures is predicted to double as well. In this Direct Phase II SBIR, BioSensics, in collaboration with orthopaedic, radiology, endocrinology and biomechanics experts at Harvard Medical School, proposes to develop a cloud-based software solution for automatic, opportunistic screening for hip fracture risk using plain X-ray images, called XFx. X-ray studies are ubiquitous in all corners of the world, are inexpensive and provide high resolution studies that offer insight into bone geometry, microstructure and density, at a low ionizing radiation dose. The proposed software solution will include 1) a desktop application for uploading X-ray images and displaying and visualizing XFx results, and 2) a secure cloud-based backend for receiving the uploaded X-ray images and performing the analysis. The software architecture will support on-premise integration with a hospital cloud services (e.g., PACS systems) to enable automatic, opportunistic screening for hip fracture risk using plain X-ray images. The solution will stand by in the central imaging data server of hospitals or clinics, investigate each non-investigated X-ray image, and if recognized to include a proximal femur, automatically execute the AI/ML-based classification scheme to identify patients with osteoporosis or at high risk of hip fracture. If a patient is identified to have osteoporosis or a high risk of fracture, the software will flag the patient. The clinician providing care for the patient will then be prompted to consider ordering an evaluation of fragility fracture risk and receive a full report. This process is reimbursable under the Current Procedural Terminology (CPT) code 76499 “Unlisted Diagnostic Radiographic Procedure.” This code is used when no other specific procedure code exists. The existence of this CPT will support the initial marketing of the proposed solution. In Phase III, we will prepare an application for a new Category III CPT code and submit if for consideration by the American Medical Association (AMA) CPT Editorial Panel. Given the clinical need of the proposed solution, and recent approval of a CPT code for radiology artificial intelligence (code 0691T) for automated analysis of existing imaging studies for vertebral fracture and bone density assessment, our application for a new CPT code should not face any difficulties. In the longer term, the proposed imaging analysis technology can be used for automatic analysis of thousands of medical images that are taken every day in hospitals and clinics. This will enable detection of diseases and conditions at early stages (e.g., bone metastasis and different tumors), thus facilitating preventive measures and better care for those individuals at risk.
