Research News

A multidisciplinary breast cancer treatment team from Cleveland Clinic's Center for Personalized Genetic Healthcare (CPGH) and Digestive Disease Institute (DDI) is using artificial intelligence to explore alternate methods of providing the education piece of genetic counseling to patients with breast cancer. Their chatbot is designed to provide high-quality care to patients to expand access to genetic counseling, education and testing. 

Charis Eng, MD, PhD, chair of the CPGH and the senior author of the Annals of Surgical Oncology study, says it is important for patients with cancer to get genetic counseling and genetic testing if they believe their cancer could be caused by a gene mutation. Many so-called 'breast cancer genes' carry increased risks of other types of cancer. The most famous genetic cause of breast cancer, BRCA gene mutations, also increases the risk of ovarian cancer.

"If we can confirm a patient's breast cancer is caused by one of these mutations, we can develop personalized treatment and surveillance plans to catch cancers early on, or even prevent a second cancer before it can occur," Dr. Eng says. "We will also know to keep a close eye on the patient's family members if they test positive for the family-specific mutation." 

Genetic counselors educate their patients about these complicated topics so the patient can make informed decisions about whether to pursue any genetic testing. If the patient decides to pursue genetic testing, the genetic counselor will order the tests, interpret them and assist in treatment planning. 

The challenge is making sure genetic counseling remains accessible to everyone in a reasonable timeframe, says cancer genetic counselor and study author Ryan Noss, MS, CGC. 

"It takes years of highly specific training to become a genetic counselor," he says. "Only a couple hundred graduate each year. Only a small subset of these graduates become cancer-specific genetic counsellors." 

Meanwhile, almost 300,000 women receive a breast cancer diagnosis each year. Noss says more of these women are now qualifying for genetic testing – and therefore require appointments with genetic counsellor.

"Guidelines are continually changing to test more and more patients with cancer for gene mutations," he says. "This is not a problem we can hire our way out of – there simply aren't enough genetic counselors in the country to keep up with the increasing demand. We need to find alternatives that don't sacrifice the quality of care a patient receives, before the wait time for an appointment becomes unreasonable." 

Chatbots support providers and patients 

Many parts of a genetic counseling appointment are uniquely tailored to the patient: ordering specific tests, interpreting the results and formulating a plan of action. However, the first appointment where counselors educate patients about genetics and obtain informed consent is not. While a patient may have unique questions, the information taught to them is standardized. 

First author and breast surgery specialist Zahraa Al-Hilli, MD says genetic counseling and education - from humans or from chatbots - is a "win-win" for providers and patients, adding that these resources can revolutionize current genetic counseling and testing practices by broadening access for patients newly diagnosed with breast cancer. 

The breast cancer treatment team, which included Dr. Eng; Dr. Al-Hilli; and Noss, decided to ask if artificial intelligence chatbots could educate patients as well as a genetic counselor could. If patients had the option to learn from a chatbot instead of booking a pre-test counseling appointment, genetic counselors' schedules would have more room for ordering tests and working with patients who have confirmed genetic mutations. This option also has the potential for more patients to receive pre-test education than feasible by a genetic counselor alone. 

For the next 39 patients diagnosed with breast cancer, the team assigned half to receive in-person education from a genetic counselor. The other half were taught the same information by a chatbot. Follow-up questions revealed that the two groups of patients had received comparable levels of education and understood the same aspects of genetic and genetic testing. Patients also reported similar levels of satisfaction regardless of whether they spoke with a human or a chatbot - a proof-of-concept for incorporating the chatbots into care, Noss says. 

The research team hopes other breast cancer providers in Cleveland Clinic will use the AI tool to give genetic counselors more time to work with women who have confirmed breast cancer mutations and allow for more patients to receive testing. After that, they will try to expand the chatbot's use into other types of cancer. 

"It's difficult to challenge the status quo and change how we provide patient care," says Noss, "but if we can get ahead of access issues before they impact our patient's experiences, the change will be worth it."