The FDA has permitted marketing of an artificial intelligence-based algorithm for detection of fractures in the wrist.
In a continued focus on improving digital health technology, the United States Food and Drug Administration has permitted marketing of an artificial intelligence (AI)-based algorithm for detection of wrist fractures. The software, known as OsteoDetect, effectively identifies distal radius fractures in two-dimensional X-ray images.
The device is intended as an adjunct and not a replacement for clinician review of radiographs, the FDA noted. In retrospective studies submitted to the FDA for the approval, use of the device increased sensitivity and specificity as well as both positive and negative predictive values when compared with standard methods.
“Artificial intelligence algorithms have tremendous potential to help health care providers diagnose and treat medical conditions,” said Robert Ochs, PhD, the acting deputy director for radiological health in the Office of In Vitro Diagnostics and Radiological Health at the FDA’s Center for Devices and Radiological Health, in a statement. “This software can help providers detect wrist fractures more quickly and aid in the diagnosis of fractures.”
The approval, which was handed down under the FDA's De Novo premarket review program, was based on a retrospective review of 1000 radiograph images that were analyzed by the AI-based algorithm. In this first study, the software, which pinpoints the exact location of the fracture, improved fracture detection performance for 3 board certified orthopedic hand surgeons.
In addition to the retrospective review of 1000 images, a second study showing the accuracy of OsteoDetect was also submitted for the approval. This study included examples from 24 providers across 200 patients’ cases. Again, in this scenario, the AI-based software helped improve health care provider performance.
The OsteoDetect device joins an ever-growing list of FDA-approved AI-enhanced diagnostic tools. In February 2018, the FDA established a new approval pathway for computer-aided triage software devices with the approval of an AI-based software for stroke triage. This was followed in April 2018 by an FDA approval for an AI-based device for detecting diabetic retinopathy.
“Given the enormous promise of artificial intelligence technologies for the future of medicine, our digital health team is actively developing a new regulatory framework to promote and support innovation in this rapidly advancing field,” FDA Commissioner Scott Gottlieb, MD, Tweeted after the OsteoDetect approval. “While our approach to artificial intelligence must be flexible to keep pace with its unique traits, it must establish appropriate guardrails for patients to meet our standards for safety and effectiveness and so we instill confidence in their benefits.”
The approval for the OsteoDetect device was granted to the startup Imagen Technologies, a company that was launched in January 2016 after being founded in 2015. In addition to an initial venture capital investment of $5 million, the company closed its series A financing in 2017 at close to $15 million. The focus of the company is on medical applications for AI.