MatchMiner: An open-source AI precision medicine trial matching platform

Share :
Published: 20 Apr 2023
Views: 49
Rating:
Save
Dr Harry Klein - Dana Farber Cancer Institute, Boston, USA

Dr Harry Klein speaks to ecancer about MatchMiner which is an open-source AI precision medicine trial matching platform. 

This is a software platform that was launched in 2017 and it matches patients with appropriate clinical trials of targeted therapies.

Dr Klein explains the study design and says that the platform developers will present new information on 256 clinical trial consents facilitated by MatchMiner at the AACR Annual Meeting 2023.

This is an ongoing study and the results will be shared later. 
 

My study was about MatchMiner which is a precision medicine software platform that’s open source and developed at Dana Farber Cancer Institute. It was designed to be an institutional grade software to accelerate enrolment onto precision medicine trials and to provide more precision medicine options for patients. 

In this particular case we did a retrospective study on patients that had enrolled onto precision medicine trials that were facilitated by MatchMiner. So, thus far we have had 256 trial consents that have been facilitated by MatchMiner. We report in our presentation that there were a diverse set of genomics and cancer types – over 50 different genes and about 20 cancer types were enrolled onto precision medicine trials – so this is a fairly comprehensive platform that covers a wide range of patients.  In general MatchMiner matches patients to precision medicine trials, all alive patients in the platform, and, although it matches all patients to all trials, most of those patients are not necessarily ready to enrol into a trial. 

We developed a natural language processing sorting algorithm that essentially determines, based on tumour radiology scan text, if the patient is ready to enrol into a trial. So the model will then filter for patients that are trial-ready. 

We present the outline of the study in our presentation. This is something that we’re going to be running for the remainder of the year and we will have results early quarter one next year.

Though we don’t have any results yet, so we describe our retrospective study in this presentation and some additional results of the consents that we have had through MatchMiner thus far. Then we just outline what the artificial intelligence, natural language processing, sorting looks like but we will have further results on that early next year.

What are the next steps for this study?

MatchMiner is open source and thus far we have had two other institutions that have adopted MatchMiner, either in a research setting or in a clinical setting. We are hoping to expand that by providing more awareness of MatchMiner and its capabilities at AACR. Adding this AI sorting on is the next step in the evolution of the software so we’re looking for more institutions that are interested in adopting MatchMiner and just other general questions that might be able to improve upon the software’s capabilities in the future.