The power of AI to opportunistically detect fracture risk
As a silent disease, osteoporosis often remains undiagnosed until a fracture occurs, highlighting the reactive nature of current risk assessment strategies. By utilizing existing diagnostic images from conventional X-ray and CT acquisitions, TBS Reveal1 offers a proactive, cost-effective method for widespread bone health assessment.
TBS Reveal is Medimaps’ AI X-ray opportunistic detection software. This product is currently in development. The software’s main functionality is to analyze an X-ray that has already been acquired for another medical indication. TBS Reveal automatically performs an analysis of the X-ray acquisition, providing information on bone microarchitecture and flagging patients at the highest risk of fracture.
Preventing fracture through earlier diagnosis
The analysis performed by TBS Reveal on conventional X-ray acquisitions can prompt an earlier and more effective diagnosis of osteoporosis ensuring that patients at the highest risk are identified quicker for the treatment needed to help prevent painful and disabling fractures.
Objectives of TBS Reveal
To assess bone microarchitecture from conventional X-rays
To detect osteoporosis from conventional X-rays
To prompt clinical management of fracture risk
To enable earlier initiation of preventative or treatment strategies to help maintain bone health and prevent future fractures
Status of development
Medimaps has made significant progress in the clinical validation of TBS Reveal software with international, multi-center cohort studies and presentations at major scientific and clinical conferences2,3.
1. Product under development.
2. Gatineau G, Nguyen G, De Gruttola M, Hind K, Kužma M, Payer J, Guglielmi G, Fahrleitner-Pammer A, Hans D. External validation of AI-driven bone fragility detection in radiographs from multinational cohorts. ISCD 2024.
3. Gatineau G, Nguyen G, De Gruttola M, Hind K, Kužma M, Payer J, Guglielmi G, Fahrleitner-Pammer A, Hans D. Automatic Detection of Bone Fragility in Radiographic Images using Deep Learning with Multi-center Cohort Datasets. Oral presentation, ECR 2024
