AI Can Help Find Trials for Patients with Rare Cancers

Finding patients with rare cancers to enroll in clinical trials can be like searching for a needle in a haystack. That can make researchers reluctant to go through the time and expense of opening trials for highly specialized drugs, since they may not find potential candidates.

But what if that process could be automated? That’s the premise behind a program that uses AI and patient databases to match patients with trials.

Cleveland Clinic researchers collaborated with investigators from other institutions and evaluated a large-scale AI clinical trial screening and matching program for six months and found that it improved enrollment, enrolling an average of one new patient per day. The study shows the potential of technology to improve access to trials, and to make it easier for researchers to find trials for patients, including those with rare diseases who might otherwise be missed, said Dale Shepard, MD, PhD, a Cleveland Clinic oncologist and first author on the study.

“These are really specialized patients in terms of matching,” says Dr. Shepard. “By using technology, we’re able to match patients and trials much more effectively, and open up more treatment options for patients.”

Cleveland Clinic participates in a trial group with a private company that does genetic sequencing and facilitates the rapid opening of clinical trials. The company uses AI to analyze large amounts of patient data, including genetic sequencing information and clinical data for each patient, and cross-references it with eligibility requirements for different clinical trials.

Shepard noted that the program also tracks changes in patient information, such as updates to their scans or lab values, as well as changes to trial protocols. “Patient eligibility for trials is a changing landscape,” says Dr. Shepard. “The trial may change, the patient’s condition may change, and someone who wasn’t eligible before will become eligible to enroll.”

For the six-month period of the study, the program analyzed a database of 840,523 patients and 74 potential clinical trials. Over this time, it tracked more than 3.7 million patient clinical updates and 271 trial updates.

The analysis initially identified more than 350,000 patients as preliminary trial matches, with 24,500 meeting criteria for a nurse screening. Of these, around 4,400 were found to be a match, and 189 went on to sign consent forms and enroll in a trial.

Shepard noted that the group uses a “just in time” process to open trials. This includes a pre-approval process, negotiating contracts, consents and budgets ahead of time so that a trial can be opened quickly once a patient is matched. During the six-month study period, just in time trials had an average activation time of 14 business days, while trials that had not gone through the process took an average of 39 business days to open.

“By using this technology and having this process in place, we’re able to have access to more trials than we ordinarily would, and we’re able to offer those trials to our patients in a timelier way,” he says.

Many of the trials focus on specific mutations. This approach is ideal for trials focused on specific genetic abnormalities in tumors, because it provides a vehicle for researchers to open trials for patients as needed rather than go through the time and expense of setting up trials for rare mutations that may never be seen in the clinic.

“It would be difficult for us to justify opening a trial for something that happens in maybe one half of one percent of the population, so they are important studies that fill an important need, but we’re not necessarily going to open them unless we find the right patient,” says Dr. Shepard.

An important takeaway for community providers is to look for centers with robust trial portfolios when considering referrals. “Don’t assume that because someone has an uncommon type of cancer, that there’s not a trial for them, because with these mechanisms we oftentimes have something available that may not be obvious,” says Dr. Shepard. “This should encourage people not to shy away from looking into trial options.”

The study, “Impact of AI Clinical Trial Program on Screening, Matching and Enrollment of Patients Over Six Months,” was presented at the 2024 European Society for Medical Oncology (ESMO) Congress.