The ImagiNg Data EXchange (INDEX) program, launched by ARPA-H, seeks to address the critical need for diverse, high-quality medical imaging data to advance the development of artificial intelligence (AI) and machine learning (ML) tools in radiology and pathology. Accurate AI/ML algorithms can assist healthcare providers in making faster and more reliable diagnostic decisions, but their development has been hindered by limited access to representative and diverse imaging datasets. Without adequate training data, these algorithms risk being biased and unsuitable for clinical use.
INDEX aims to create a centralized platform that connects data providers, users, and service providers, enabling the sharing of imaging data to support the development of robust AI/ML models. The program prioritizes increasing the diversity and quality of images available for training, with a focus on geographic, racial, and ethnic representation. To standardize and enhance the usability of imaging datasets, INDEX will also develop tools to harmonize data and support ML model training in cloud environments. In partnership with the FDA’s Office of Science and Engineering Laboratories (OSEL), INDEX will integrate regulatory science toolkits to evaluate data quality and diversity, ensuring the resulting AI tools meet clinical standards.
The initiative seeks to make AI-driven diagnostic tools more reliable and efficient, reducing physician workload and expediting patient diagnoses and treatment plans. By fostering a collaborative ecosystem, ARPA-H aims to overcome technical barriers and create a comprehensive resource for medical imaging data. This program reflects ARPA-H's ability to unite federal regulators, healthcare providers, ML developers, and other stakeholders to drive innovation in medical AI development.
Proposals for the INDEX program should demonstrate innovative approaches to expanding and standardizing imaging datasets while addressing critical gaps in quality and diversity. Prospective proposers can attend an optional hybrid Proposers’ Day on January 7, 2025, in Seattle, WA, to learn more about the program's goals and requirements. Teams with diverse technical expertise are encouraged to form collaborations, with tools like a dedicated teaming page available to facilitate networking.
Key submission deadlines include January 23, 2025, for solution summaries and March 27, 2025, for full proposals. Proposals will be evaluated based on their potential to enhance medical imaging datasets, the feasibility of technical approaches, and their alignment with INDEX’s mission to support high-quality, trustworthy AI development.
By creating a scalable and accessible imaging data exchange, INDEX aims to transform the development of medical AI tools, ensuring they are inclusive, reliable, and capable of meeting the diverse needs of healthcare providers and patients. This initiative exemplifies ARPA-H’s commitment to solving complex healthcare challenges through collaboration and technological innovation.
