Medical Imaging Data Marketplace

The Advanced Research Projects Agency for Health (ARPA-H) and the Center for Devices and Radiological Health (CDRH) of the Food and Drug Administration (FDA) are partnering to streamline access to affordable, high-quality, regulatory-ready medical imaging data. The agencies are exploring the development of a medical imaging data marketplace (MIDM) to expedite innovation in Software as a Medical Device (SaMD) and other artificial intelligence (AI)/machine learning (ML)-enabled medical devices.

In support of this initiative, the Investor Catalyst Hub issued a Network Survey to collect feedback about specific needs and challenges that medical imaging software and product developers, users of AI and ML medical imaging products, and private and public organizations face with utilizing, managing, and producing data for product development and evaluation.

The Network Survey feedback will be used to inform the model and strategy used for a MIDM and prioritize the types of data it should support. It will also help to develop a viable economic model to ensure the long-term sustainability of the marketplace.

ARPA-H and FDA CDRH hosted two information sessions to share more about the MIDM concept, focus areas, and the Network Survey. View recordings of the March 14 information session and April 1 information session and find answers to FAQs.

The deadline to provide survey feedback was April 16, 2024. ARPA-H will aggregate the provided information and share a report summarizing the findings from the survey.

Why Are ARPA-H and the FDA Partnering on This Initiative?

With the increase in the development of SaMD incorporating artificial intelligence and machine learning, access to high-quality, regulatory-ready medical imaging data for developing and testing new technology is one of the most significant challenges faced by medical device and software developers. This limited access not only hinders innovation but also delays FDA clearance or pre-market authorization due to data quality deficiencies, ultimately limiting the performance and availability of AI/ML software and medical devices.

This joint effort between ARPA-H and FDA seeks to develop a MIDM to organize and manage medical imaging data at scale by connecting existing databases, marketplaces, and data providers to a trusted, streamlined platform that researchers and customers can use to find and affordably access the data needed to develop and test new algorithms. This initiative will include developing new tools that will give users of the marketplace greater confidence that the data accessed will be consistent with the regulatory requirements for future pre-market authorizations.

Who Was Invited To Respond?

This Network Survey sought to collect feedback from three groups:

Medical imaging data users: Researchers (academic or private-sector) who need/use medical imaging data to train and test AI/ML models. We are specifically interested in data users working on both radiology and digital pathology applications.
Medical imaging data managers: Entities that manage, sell, or aggregate medical imaging data for researchers. These organizations can be databases, platforms, or aggregators. They do not necessarily produce these data themselves, but make them available to data users.
Medical imaging data providers: Entities that create the medical imaging data needed by data users. They do not necessarily provide or sell their data to other parties. We are highly interested in providers of radiology and pathology images.

For more detailed definitions and examples of each group, read the Network Survey description.