Clinical trials for ARPA-H’s ADAPT Program set to begin using AI models developed by MIT Jameel Clinic and Mass General Brigham researchers to personalize cancer treatment
This week the Advanced Research Project Agency for Health (ARPA-H), an agency within the U.S. Department of Health and Human Services, has launched their first set of clinical trials as part of an ARPA-H program called ADAPT (Advanced Analysis for Precision Cancer Therapy). The ADAPT program aims to integrate powerful computational models, novel biomarkers, and an evolutionary clinical trial to allow for the early detection of tumor changes.
In May, a team of MIT and Mass General Brigham researchers called AURORA (AI-driven Understanding of Response Optimization and Resistance Assessment) was selected as one of ten awardees in the ADAPT program. Led by MIT Jameel Clinic AI faculty lead Regina Barzilay, the goal of the AURORA team (which includes MIT Jameel Clinic Principal Investigator Tommi Jaakkola and MIT Assistant Professor Stephen Bates; Genevieve Boland will be the Co-Principal Investigator for the Massachusetts General Hospital team, which includes Justin Gainor and Russell W. Jenkins) is to develop machine learning models that can utilize a variety of data modalities, allowing clinicians to personalize targeted therapies while adjusting care plans based on real-time data and predictive insights on tumor trajectory and tumor resistance.
These machine learning models developed by Team AURORA will now be used to help select more personalized treatments for patients participating in ADAPT’s clinical trials. This effort represents one of the first attempts at combining multimodal biomarkers with machine learning to develop personalized treatment approaches that can also overcome resistance.
ARPA-H has committed up to $142 million for the ADAPT program to transform personalized cancer care by dynamically tracking tumor changes through multiple lines of treatments. If successful, ADAPT’s new tools and protocols will save American lives and reduce the massive economic burden of cancer nationwide.
Approximately two million Americans are diagnosed with cancer each year. Of those individuals, over 600,000 are diagnosed with metastatic and treatment resistant cancers. Because cancer emerges from our cells, which mutate and change over time, tumors can develop resistance to cancer therapy, making the cancer difficult to treat and control. Through the development of powerful computational models, novel biomarkers, and an evolutionary clinical trial, the ADAPT program seeks to improve early detection of tumor changes across an array of genetic profiles, allowing clinicians to proactively make treatment adjustments that can improve patient outcomes.
