Science & Enterprise logo
Science for business people. Enterprise for scientists.

A.I. Shown to Boost Cancer Clinical Trial Enrollment

Artificial intelligence graphic
(Seanbatty, Pixabay)

9 March 2018. Early results from a system using a supercomputer and artificial intelligence to match breast cancer patients to clinical trials shows faster screening and higher enrollment rates in these studies compared to manual methods. A team from the Mayo Clinic in Rochester, Minnesota described its experiences with the Watson supercomputer system made by IBM in a session of the Healthcare Information and Management Systems Society Annual Conference and Exhibition now underway in Las Vegas.

Clinical trials that recruit humans to test new drugs and medical devices face a chronic problem of enrolling individuals to participate. Data cited by drug maker Eli Lilly and Company show less than 1 percent of the U.S. population takes part in clinical trials, with nearly half of these studies not meeting their enrollment goals. These shortfalls often translate into longer times needed to complete clinical studies, with the vast majority of trials — 86 percent — taking longer than planned, resulting in longer waits for new drugs and higher costs.

This is particularly a problem with cancer trials. Among cancer patients, only about 5 percent enroll in clinical trials, says Tufia Haddad, a cancer specialist leading the project at Mayo Clinic , and who gave the conference presentation. “Novel solutions are necessary to address this unmet clinical need, advance cancer research and treatments, and, in turn, improve the health outcomes of patients” says Haddad in a joint statement.

To address this problem, Mayo Clinic tested IBM’s Watson supercomputer using the company’s software for matching patients to clinical trials. The solution says IBM takes advantage of Watson’s much greater processing speed and capacity, but also its use of natural language processing, a form of artificial intelligence that applies context and intricate rules of semantic understanding to the often unstructured data in physicians’ notes found in medical records, even in electronic form.

Watson processes these data and returns structured lists of patient characteristics that it matches up to the often complex eligibility criteria listed for clinical trials in databases, such as ClinicalTrials.gov. In addition to recommending suitable clinical trials, Watson also outlines its criteria for selecting and not selecting studies in which to participate.

Mayo Clinic and IBM tested Watson and its software with breast cancer patients beginning in July 2016. For its part, Mayo Clinic says it adapted its workflows and screening processes to Watson, including educating their patients about the availability of clinical trials, and including these studies in the patients’ care plans. In their evaluation, Haddad and colleagues calculated enrollment times and patient trial participants using Watson, compared to manual methods, often with spreadsheets.

The results show after 11 months an 80 percent increase in enrollment for clinical trials of systemic breast cancer treatments with Watson compared to manual methods. In addition, less time was needed with Watson to enroll patients in clinical trials.

Haddad says the Mayo Clinic’s patients benefited from this program. “The speed and accuracy of Watson and the team of screening coordinators,” she notes, “allow our physicians to efficiently develop treatment plans for patients that reflect the full range of options available to support their care.”

Mayo Clinic and IBM plan to expand the use of Watson for clinical trial matching to other types of cancer, with staff now trained to apply the technology to patients with lung and gastrointestinal cancers. The parties say they also plan to continue developing the system to cover other types of medical needs, including surgery, radiation, and supportive care.

More from Science & Enterprise:

Disclosure: The author owns shares in IBM.

*     *     *