Tali Mazor, PhD
Associate Director, Knowledge Systems Group, Department of Data Science, Dana-Farber Cancer Institute
One of the most persistent challenges in clinical trial research is recruiting enough patients to effectively evaluate new cancer treatments. Researchers at the Dana-Farber Cancer Institute are tackling this hurdle with MatchMiner, an open-source platform powered by artificial intelligence (AI) that matches patients to clinical trials based on their unique clinical and genomic profiles. Led by Ethan Cerami, PhD, Director of the Knowledge Systems Group at Dana-Farber, MatchMiner automates and scales the trial-matching process, helping clinicians uncover more treatment testing opportunities for patients, particularly those with advanced or rare cancers. This platform, which is currently in use at Dana-Farber as well as at other institutions, is accelerating the pace of precision oncology research.
Streamlining Patient-Trial Matching
Clinical trials require patients to meet complex criteria, including cancer type, disease stage, biomarker status, prior treatments, and comorbidities. Traditionally, identifying eligible patients has been a manual, time-consuming process that often misses potential matches.
MatchMiner was designed to take the molecular fingerprint of a patient’s tumor and connect it to clinical trials targeting those specific mutations.”
Ethan Cerami, PhD
MatchMiner was created to change that. Launched by Cerami in 2016, the platform began as a rules-based engine that matched patients to trials using structured data—primarily genomic biomarkers and cancer type. This approach leveraged Dana-Farber’s strength in offering next-generation sequencing to nearly all patients, along with its portfolio of more than 600 active clinical trials. Cerami worked with Dana-Farber’s Mike Hassett, MD, MPH, Chief Quality Officer and medical oncologist, and Bruce Johnson, MD, Chief Clinical Research Officer and Leader of the Lung Cancer Program at Dana-Farber, to utilize these assets and create a platform that could automate and streamline trial matching, meeting the pressing need that exists.
“MatchMiner was designed to take the molecular fingerprint of a patient’s tumor and connect it to clinical trials targeting those specific mutations,” Cerami explains. “It’s about making it easier and faster for clinicians to identify promising options—and for patients to access cutting-edge treatments.”
How MatchMiner Works—and How AI Expands Its Reach
While the original version of MatchMiner relied on structured data, many critical eligibility criteria—such as prior lines of therapy or disease burden—exist only in free-text clinical notes, which means they weren’t captured by the platform. But the latest version of MatchMiner now incorporates AI to analyze unstructured data in electronic health records (EHRs) and extract key details that are then mapped to trial eligibility criteria.
“We’re now using AI to summarize the clinical narrative,” explains Kenneth Kehl, MD, MPH, Associate Director of Clinical Research for the Thoracic Oncology Program and researcher in Population Sciences at Dana-Farber. “That includes identifying prior treatments, disease stage, and other trial-relevant features that aren’t always captured in structured fields.”
Broadening Access and Impact
These AI-driven enhancements have significantly expanded MatchMiner’s utility. Previously, the platform could only be used for patients who had undergone genomic sequencing and for trials with genomic eligibility. Now, MatchMiner-AI can support all patients at Dana-Farber and includes all treatment trials open at the institution—maximizing trial options for each patient.
“As part of our focus on serving the full Dana-Farber community, we are excited to be piloting MatchMiner-AI with oncologists at our regional campuses,” says Tali Mazor, PhD, Associate Director of the Knowledge Systems Group in Dana-Farber’s Department of Data Science.
Exploring the results of AI Trial Matching
She points out that MatchMiner is already embedded in Dana-Farber’s clinical workflows and has supported over 400 patient enrollments in clinical trials. A 2022 study in NPJ Precision Oncology found that patients matched through MatchMiner enrolled in trials on average 22% faster than those identified through traditional methods.
Another recent study conducted by Cerami and colleagues appeared on arXiv in 2024 finds that integrating AI into MatchMiner significantly improved trial matching. The authors reported that this AI-driven approach enhanced scalability, interpretability, and adaptability—offering a practical and efficient tool for personalized trial matching.
Open-Source and Collaborative by Design
Unlike commercial platforms, MatchMiner (with AI incorporated) is freely available and customizable. This aims to encourage collaboration and innovation across institutions, helping to advance the field of precision oncology. For instance, it’s already in use at Princess Margaret Cancer Centre in Ontario.
Funding and Support
MatchMiner has received funding from Dana-Farber Cancer Institute, the Harvard Business School Kraft Precision Trials Challenge (where it won first place), and the Fund for Innovation in Cancer Informatics (ICI). These investments have supported both the platform’s technical development and its integration into clinical practice. In addition, a grant from Meta enabled the creation of advanced AI models that power the latest platform enhancements.
Expanding Regionally to Improve Clinical Trial Participation
Dana-Farber is now piloting MatchMiner across its regional practice sites. The goal is to gather feedback from clinicians, refine the AI models, and prepare the platform for broader deployment. Future enhancements may include support for imaging data, lab results, and integration with national trial registries. “There’s a tremendous need for new cancer treatments,” says Cerami. “MatchMiner is about giving patients more opportunities to participate—and giving researchers the tools they need to advance medicine.”
This will help bring the right trials to the right patients, faster and more intelligently than ever before.
Team Members
Bruce Johnson, MD
Chief Clinical Research Officer and Leader of the Lung Cancer Program, Dana-Farber Cancer Institute
Kenneth Kehl, MD, MPH
Associate Director of Clinical Research for the Thoracic Oncology Program
Researcher in Population Sciences, Dana-Farber Cancer Institute
Mike Hassett, MD, MPH
Chief Quality Officer
and Medical Oncologist, Dana-Farber Cancer Institute
Ethan Cerami, PhD
Data Scientist, Dana-Farber Cancer Institute
