By Aaron Krol

April 28, 2016 | The scientific team at MolecularMatch, a Houston-based company founded by immunotherapy researcher Jim Welsh of the MD Anderson Cancer Center, think they’ve found a better way to connect critically ill patients with clinical trials. For many patients who have not improved under standard treatment courses, enrolling in a study of an investigational new drug may be the only way to access promising therapies—yet many of these trials fail to meet their enrollment targets on time, while patients in need of new options never learn that they qualify.

On its website, molecularmatch.com, MolecularMatch now catalogues hundreds of clinical trials recruiting patients around the country and overseas, with a specific focus on cancer therapies. Anyone who visits the site can search for trials targeting narrow types and subtypes of cancer; to date, visitors have made over 330,000 of these searches.

This in itself is not a new idea. The National Institutes of Health opened clinicaltrials.gov in 2000 for this exact purpose, and that site has steadily expanded until, today, it includes all prospective studies of drugs and medical devices in the U.S., and many more internationally. MolecularMatch, however, aims to make the user experience for its site both easier and more fruitful for patients. On the front end, it adds features like suggested search parameters that could help users refine their results. A patient searching for breast cancer, for instance, might be prompted to include information on their disease subtype, results from molecular tests on their tumor, and their location, which can help pull up the most likely trials in their area. MolecularMatch also lets users create profiles where they can store relevant health and demographic information, share results with their physicians, and contact clinical trial sponsors.

On the back end, the company has also developed context-aware search functions that are sensitive to the quirks of biological data. The MolecularMatch engine has an internal vocabulary that can tell how different cancers, drugs, and molecular tests relate to one another, and can resolve the different terminologies used across patient records and the scientific literature. “We had to create synonym matching, and search intents—if I misspell a word, or if I’m saying lung cancer but I may be thinking non-small-cell lung cancer,” says Kevin Coker, the CEO. All this information is kept in a structured database that makes searching faster, and the system’s knowledge of biology lets it surface relevant approved drugs in addition to clinical trials.

It’s this search engine that powers MolecularMatch’s commercial offerings, which are separate from its free and public trial-matching site. Last month at the Molecular Medicine Tri Conference in San Francisco, the company released MM Lab, a platform that connects patients’ test results to drug options and ongoing clinical trials. Designed for pathology labs, MM Lab can be fed the results of tests for genetic variants, antibodies, and other biomarkers, and begin the process of interpreting how they might affect patient care.

As more data on the tangled molecular pathways of cancer is collected, tests like these have become increasingly hard for physicians to interpret on their own. The genetic mutations that drive cancer cases can have many downstream effects on RNA and protein levels, which can be difficult to trace back to the root mutations targeted by approved or investigatory drugs. Oncologists also struggle to stay current with clinical trials, making it useful to connect their hospital labs’ own software with the MolecularMatch registry.

MM Lab, Coker says, is a much more powerful program than the one on MolecularMatch’s free trial-matching site, in part because lab customers have more information at hand. “As we work with labs, their information about a patient’s condition and genetic status is highly structured, so we’re able to filter it much better and match much better,” he says. “And the more we know about a patient, like if we can get a better drug history, whether they became resistant to a drug, co-morbidities—all those things help improve the match.”

With both MM Lab and the company’s previous, hospital-oriented product, MM Data, MolecularMatch’s software is plugged into the customer’s normal workflow, sharing data and reports with a laboratory information management system or a hospital’s electronic medical records.

MolecularMatch is also pursuing partnerships with diagnostic companies to expand the reach of its software. This morning, the company announced an agreement with Liquid Biotech USA, which is developing liquid biopsies, or tests to detect trace markers of cancer in simple blood draws. Liquid Biotech USA will use the MolecularMatch platform on the back end of its tests in pilot programs focused on early detection of cancer.

“MolecularMatch’s ability to quickly match test results to treatment, alongside its considerable knowledge base, are a perfect complement, adding additional speed-to-treatment on top of ours,” said Liquid Biotech USA President and CEO Philip Sass in a statement.

Meanwhile, molecularmatch.com, which has been live since October 2014, has given MolecularMatch a unique bird’s-eye view of the patient demand for clinical trials. “This is a sort of real-time surveillance engine,” says Coker. While MolecularMatch does not track individual users’ activity, it does see trends, such as popular searches, regional hotspots for interest in certain trials, or unexpected search terms. For example, when the team at MolecularMatch noticed a search for “PatientsLikeMe,” a patient advocacy group, it reached out to discuss whether that organization’s non-traditional, patient-driven studies might be included in the MolecularMatch database. (So far, they have not been.)

Sharing this anonymous information with clinical trial sponsors, such as pharma companies, encourages them to take an active role in maintaining MolecularMatch’s registry. “We’ll show them how many patients we’ve matched, and how many patients are searching for key terms,” Coker says. “We pull in information from public sources, but we also get information from clinical trial groups, from pharmaceutical companies, from other private groups, and they can input their own information, so they can come into MolecularMatch and update their trials.” That direct participation by sponsors sometimes gives MolecularMatch a depth of information about trials that exceeds the records in clinicaltrials.gov—making it easier to match molecular test data to relevant studies.

According to Coker, dozens of users a week are now creating accounts and requesting enrollment in trials, although the company does not track how many ultimately enter the studies they find. “It’s really amazing, because the amount of search activity has gone up dramatically,” he says. “The thing that drives us all is this clinical trial matching piece.”