Algorithmic matching of patient profiles and clinical trial criteria.
MatchMiner is the first platform for algorithmically matching clinical trials to patient information (and we are in the process of making it fully open-source right now). MatchMiner was developed by Dana-Farber Cancer Institute in collaboration with The Hyve.
How MatchMiner works
It starts by simply going to the website of MatchMiner using a webbrowser installed on your computer. You can fill in patient’s name and click “Search” to search for the clinical trials your patient is eligible for. You can also create a new filtering criterion (e.g., patients need to have a BRAF E433K mutation and have to be older than 18) and click “Search” to see patients passing this filter. This is the layer of MatchMiner that you, as the end user, communicate with, the frontend.
What happens when you click Search?
Your institution has quite a number of patients, and their information is stored in the patient database – their cancer type, age and gender get there from their Electronic Medical Records (EMR), and their genomic information in the form of mutations and structural variations get there from their sequencing data.
Keep in mind that the patient database needs to be updated regularly, as new patients come in over time and provide their data. Your institution also runs (or participates in) clinical trials, and a structured description on each trial is stored in the clinical trials database. Another layer of MatchMiner – the matching engine – runs routinely to find overlaps between the patient and clinical trail databases and, as a result, stores all matches for each patient’s tumour sample and for each clinical trial. So when you click “Search” to find clinical trials for your patient, you receive the results the matching engine prepared for you in a graphic user-friendly interface on MatchMiner website installed within your institution.
Scenario 1: Oncology
You are an oncologist at a cancer center/cancer institute/hospital. You have a lot of patients, their EMR are available, as well as their genomic profiles — mutations, copy number alterations (CNA). For some of them, treatment is not available, however, they are interested in participating in a clinical trial. So you need to look through all available trials (and you yourself need to define “all available” — all available within your institution? In all institutions from your region? Country? All running clinical trials from clinicaltrials.gov?). And as more genomic information becomes available, clinical trials become more complex and elaborate — multi-arm trials, MAMS (multi-arm, multi-stage platform trials), trials including a specific mutation in a specific gene as a filter/criterion. You need to find a match between the profile of each patient and the full list/database of clinical trials. This can definitely be daunting.
How will your life change when you start using MatchMiner?
- MatchMiner has a modern intuitive user-friendly interface. When you type in a patient’s name or unique identifier, MatchMiner will show you a page with your patient’s data (both genomic and clinical) and a list of clinical trials that this patient is eligible for. You can directly contact the clinical trial manager leading a particular trial if your patient is interested in joining. You can also click on that trial to get more information and see, for example, its detailed description or the enrolment pace.
- If sequencing of new patients’ data is regularly performed in your institution, let’s say every Monday, MatchMiner can be set up to run the matching of the updated patients database against the database of clinical trials automatically, e.g., every Tuesday. After the matching is done, you can subscribe to get an email update from MatchMiner for your patients and any clinical trials they are eligible for.
Scenario 2: Clinical trial expert
You are an expert in clinical trials at a cancer center/cancer institute/hospital. You are responsible for curating some of the currently running trials, and you are working on creating a new one. Nowadays clinical trial design gets more and more elaborate, you have to design multi-arm trials with detailed eligibility criteria, including genomic information (e.g., a copy number alteration in the BRCA gene). MAMS (multi-arm, multi-stage platform trials) and busket trials (including multiple histology types) are also becoming more common, which might feel daunting. Another scenario — when your institution is working on starting a new trial, you would like to be able to see how many patients which provided their genomic and EMR data pass this or that filter, you want to see the enrollment rates over the past years to be able to predict (to a certain extent) how many extra patients passing this filter you might expect in the current months. You would also benefit from making the communication between you and oncologists faster and easier — when you see a new patient eligible for your clinical trial, you would like to be able to contact the oncologist directly and as soon as possible.
How will your life change when you start using MatchMiner?
- MatchMiner has a modern intuitive user-friendly interface. When you type in the name of a clinical trial you curate, you can see all the patients eligible for that trial. You can contact their oncologist directly to inform him/her of this patient’s eligibility, so that the oncologist can directly inform the patient.
- When you create a new clinical trial, you have to follow the Clinical Trial Markup Language guidelines, which basically means that, instead of describing your clinical trial criteria in a plain text format, you do that using so called badges. You add a badge “older than 18” in yellow colour, a badge “wild type BRAF, mutation X in KRAS” in brown colour, tumour type “solid tumour Y” in green colour. Multi-arm clinical trials will be visualized as trees, where each new node is an inclusion criterion, and every new branch says whether it is required or not. When a certain branch of a clinical trial does not show any results, it can be removed from the clinical trial tree (this way we facilitate MAMS).
- If sequencing of new patients’ data is regularly performed in your institution, let’s say every Monday, MatchMiner can be set up to run the matching of the updated patients database against the database of clinical trials automatically, e.g., every Tuesday. After the matching is done, you can subscribe to get an email update from MatchMiner for your clinical trials and any patients eligible for them.
For more information on MatchMiner contact us.