The American College of Radiology Data Science Institute (ACR DSI) introduced a new software platform April 5 aimed at better engaging radiologists in the creation, validation and use of AI models.
The platform—deemed the ACR AI-LAB—will allow radiologists to interact with AI using their own data at their own facilities, behind their own secure firewalls, according to a release from the ACR. It's the result of a collaboration by industry heavyweights like NVIDIA, GE Healthcare and Nuance.
“Acquiring the necessary large amounts of patient data for algorithm training has been a huge problem for developers up to this point,” Keith Dreyer, DO, PhD, chief science officer at ACR DSI, said in the release. “Consequently, AI has been slow to develop and has not spread widely to improve patient care. Enabling radiologists to develop AI on-premises, at their own institutions, will unlock massively larger data stores available for the development of AI. This will rapidly expand the AI solutions available to us all.”
The ACR claims its AI-LAB democratizes AI by allowing for the direct participation of radiologists throughout the AI development life cycle. Through the platform, physicians will be able to learn about AI, contribute AI datasets, share algorithms, evaluate peers’ models, develop their own systems and collaborate to serve their individual communities’ needs.
The platform will also reportedly support any future FDA initiatives that use real-world data for local adaptive learning to improve and monitor AI algorithms.
“By enabling radiologists to actively participate in all aspects of AI development, we can expect to see high impact healthcare solutions grow rapidly,” Bibb Allen Jr., MD, chief medical officer for ACR DSI, said. “This new initiative fills existing gaps and permits direct participation throughout the AI development life cycle.”
The pilot version of ACR AI-LAB is scheduled to be presented at the 2019 ACR Annual Meeting in Washington, D.C., from May 18-22.