The health research agency within HHS has tasked UNC and three tech companies with addressing the genetic testing bottleneck through an ambitious, AI-enhanced data infrastructure and new tools.
The Advanced Research Projects Agency for Health has announced four contracts, valued up to $98.5 million total over nearly five years, to advance artificial intelligence technologies and break through infrastructure challenges hindering discovery and diagnosis of rare diseases.
One key goal is to develop representative datasets and build AI-powered diagnostics for earlier identification and treatment, ARPA-H leaders say.
WHY IT MATTERS
Because patients with rare diseases go through years of care and disease progression, and face extreme medical costs, the Rare Disease AI/ML for Precision Integrated Diagnostics program, or RAPID, aims to broaden access to diagnostic expertise, according to the ARPA-H announcement this week.
The effort would connect rare disease test results to siloed clinical data, speeding up patient diagnoses and fueling more effective and appropriate treatment paths.
By unifying and standardizing dispersed data and developing scalable approaches to safe data sharing, consent management and interoperability, the RAPID program will nurture an AI-ready, real-world rare diseases data resource that links longitudinal clinical records with genomic, patient-reported and other emerging sources of health data.
“Rare diseases represent one of the greatest unmet needs in medicine, and one of the most important frontiers for AI,” said Scott Gorman, RAPID program manager, in a statement.
“By building AI-ready datasets and infrastructure, along with technologies designed to learn from sparse and complex rare disease data, RAPID aims to improve outcomes for patients while also advancing precision medicine more broadly.”
The four teams will use the RAPID data infrastructure to develop AI-driven tools that significantly reduce the time to diagnosis and accelerate clinical trials and drug development through more precise cohort identification – while hopefully lowering healthcare costs through earlier intervention.
The University of North Carolina will use its RAPID award to build the largest real-world rare disease dataset as a privacy-preserving foundation for research and innovation. A broad group of partner organizations spanning academic and research institutions, health data and technology companies, and rare disease patient advocacy and research organizations will support UNC in its efforts to integrate clinical, genomic and patient-reported data.
Sage Bionetworks, a data company, will build a secure platform for transparent performance assessment, common evaluation standards and shared infrastructure to accelerate innovation. The platform will harmonize multimodal rare disease data to enable the development, evaluation and benchmarking of new AI approaches for rare disease diagnosis and discovery.
FDNA, an AI imaging analytics vendor, will develop tools used in clinical and direct-to-patient settings to collect longitudinal health data at national scale, including photos, videos, voice and patient-reported concerns.
Probably Genetic, an AI research platform vendor, will generate high-fidelity synthetic datasets to support rare disease research and develop patient-facing AI tools that collect multimodal data – photos, videos and information from wearables, functional assessments, clinical records and genomic sources – and enable earlier rare disease identification. These tools will also connect patients with confirmatory testing, care navigation and clinical trial opportunities, said ARPA-H in the announcement.
The four RAPID performer teams will have the support of rare disease advocacy groups and tech giants.
Global Genes and the National Organization for Rare Disorders will help drive rare disease patient involvement in RAPID program outputs, while the Lawrence Berkeley National Laboratory, the National Institutes of Health’s All of Us Center for Linkage and Acquisition of Data, and the Monarch Initiative will offer their support.
Meanwhile, AI leaders including OpenAI, Anthropic, Amazon Web Services and Google have pledged in-kind resources for RAPID, including computing credits for large language models, subject-matter expertise and engineering development.
Researchers and organizations beyond the core performer teams will also be able to leverage RAPID’s secure data infrastructure and benchmarks through Rare Challenges, an open competition platform supported by the program teams and NASA’s Center of Excellence for Collaborative Innovation.
CoECI will benchmark and accelerate emerging AI approaches across diagnosis, mechanistic understanding and other rare disease priorities, HHS said.
THE LARGER TREND
For more than 15 years, rare disease patients have turned to the internet and groups for support and answers.
As diseases such as rare cancers have challenged pathologists with limited detection and diagnosis capabilities, researchers have looked to AI to solve the conundrum.
But progress in adopting AI that can detect rare diseases has been slow and requires digital support, according to Dr. David Klimstra, cofounder of Paige, a computational pathology platform vendor.
It’s been “limited by the slow adoption of digital pathology platforms that are needed to allow the use of AI tools,” he told Healthcare IT News in 2024. “Digital adoption has been limited by high costs of digitization, logistical challenges and user reluctance after a century of practice using glass slides and microscopes.”
Government funding has helped catalyze some of the world’s most transformative breakthroughs, and RAPID is poised to assemble the largest AI-ready genetic disease dataset in history, according to Lukas Lange, Probably Genetic’s CEO.
“We believe ARPA-H’s RAPID program has the potential to do the same for precision medicine,” he said in a separate announcement about the company’s $10 million RAPID contract. “If we succeed, the years between 2026 and 2030 will stand out as some of the most meaningful for genetic disease innovation – when we helped build technology capable of changing the lives of 400 million people.”
ON THE RECORD
“For millions of patients and families, living with a rare disease still means years without an accurate diagnosis and too few paths to effective treatment,” said Gorman in a statement. “This bold endeavor reflects ARPA-H’ mission to take on challenges that demand ambitious technology development, bringing together capabilities across sectors that no single organization can develop alone.”
