FDA assisting with new imaging 'marketplace' designed to improve AI
The Advanced Research Projects Agency for Health (ARPA-H) has partnered with the U.S. Food and Drug Administration’s (FDA) Center for Devices and Radiological Health to improve the quality and accessibility of medical imaging data for the advancement of artificial intelligence (AI).
According to a statement, the agencies are creating a “medical imaging data marketplace” that will bring in images from all modalities, as well as digital pathology, to a central location. This database will be accessible to AI developers and researchers working to enhance models used in clinical diagnostics and patient care.
The platform aims to improve the quantity and reliability of imaging data, access to which remains a barrier to the advancement of machine-learning algorithms, upon its launch. The database will carry a cost to utilize, but it will be regulated by the FDA and comply with HIPAA privacy standards, ensuring anonymity and security to protect patients.
“How many times have we struggled, as researchers, to locate and curate datasets that were representative of the condition we were looking to diagnose? And how many companies have had to go back to the drawing board because their FDA pre-market submission was returned with requests for broader diversity in the training and/or testing datasets?” ARPA-H Program Manager Ileana Hancu, PhD said in the statement. “The medical imaging data marketplace aims to make this a thing of the past by ensuring fast access to research data and facilitating more streamlined regulatory approvals.”
All stakeholders—developers, researchers and regulatory experts—with an interest in the project are invited to participate in a survey that will shape its development. The deadline to provide feedback is April 16.
ARPA-H said it will aggregate survey results and share a report summarizing the findings in an effort to keep the public informed on the status of the initiative.