Artificial intelligence may soon help doctors treat patients with ovarian cancer

UCLA’s Dr. Joshua G. Cohen is the principal investigator on a multi-campus study launching in the spring

UCLA Health will soon launch a study using artificial intelligence to nimbly fine-tune treatment for patients with ovarian cancer, in order to minimize medication side effects and improve quality of life.

Participants will regularly answer questions on an interactive mobile platform about concerns such as pain, fatigue and nausea, as well as about general well-being. The software will then use artificial intelligence to make evidence-based treatment recommendations to their oncologists, said principal investigator Joshua G. Cohen, MD, a member of UCLA’s Jonsson Comprehensive Cancer Center.

"This data is meant to improve the lives of all women with ovarian cancer in the University of California system," says Dr. Joshua Cohen, principal investigator on a study using artificial intelligence to improve treatment. (Photo by Joshua Sudock/UCLA Health)

“That information goes in real time to their primary oncology doctor and the nursing team,” said Dr. Cohen, associate professor of gynecologic oncology at the David Geffen School of Medicine at UCLA. “It would allow us in a very efficient manner to make dose adjustments to the medication or potentially use a different medication.”

In another example, if a patient experiences neuropathy while undergoing chemotherapy, the app could suggest that her doctor recommend nutritional supplements, lower the dosage or refer her to acupuncture for nerve pain.

Artificial intelligence can be used in medicine to detect patterns in large sets of data with use of complex algorithms.

The study will include patients from the five University of California cancer centers starting this spring.

“This data is meant to improve the lives of all women with ovarian cancer in the University of California system,” Dr. Cohen said. “The artificial intelligence incorporates data from all the sites that it gets in real time. It’s learning with us what the right nausea medications are, the right doses, the right timing of supportive medication. We’re hoping to get a much better understanding of what works and what doesn’t for patients in real time.”

Long-term treatment

Ovarian cancer is most often diagnosed at a late stage, so many patients will need long-term treatment, Dr. Cohen said. But because the medications can lead to side effects, identifying them and addressing them quickly can improve outcomes.

“While we’re not able to cure most patients with ovarian cancer, we’re hoping to make things chronic,” he said. “If they’re so fatigued and nauseous that they can’t live their life, they’re going to come off that drug. It offers a significant advantage to survival if we can manage the patient’s side effect profile.”

That, in turn, could allow patients to stay on maintenance medications for longer, while resulting in fewer complications that require medical care and drive up health care costs.

“We can more astutely adjust the medication and address the toxicity of the treatment to avoid ER admissions and hospitalizations,” Dr. Cohen said.

Patient-reported data

The study, which is funded by a grant from GlaxoSmithKline, will recruit 200 patients from UCLA, UC Irvine, UC San Diego, UC San Francisco and UC Davis.

Called the VIRGO Project, which stands for Value-Based Integrated Recommendation software Guiding Ovarian treatment, the study uses the DEARHealth platform. DEARHealth, a health tech company, said all patient data will be securely encrypted to protect privacy.

Participants, who have been diagnosed with new or recurring ovarian, fallopian tube or primary peritoneal cancer, will be followed for one year. Half will use the app for their surveys, while the other half receives standard care with surveys done less frequently on paper.

Patients in both groups will answer questions about topics intended to gauge side effects and measure quality of life, including sleep quality, amount of energy and level of contentment. Quality of life scores will be compared between the groups.

"The collected information will allow us to better identify health equity gaps for ovarian cancer patients enrolled in the trial, and make it possible for health care providers to start to close those gaps," Dr. Cohen said.

Dr. Cohen said patients using the app will devote 20 to 30 minutes a week but can spend more time exploring other educational resources if they choose.

“We would assess, are they able to do the things they want to do and live their life while still taking this medication,” Dr. Cohen said

Researchers will also look at medication adherence, adverse events, hospitalizations and survival rates.

Dr. Cohen said the VIRGO software will be integrated into patients’ electronic medical records so the data they enter will be securely delivered to their doctors through the EMR.

“That information goes in real time to their primary oncology doctor and the nursing team,” he said. “It would allow us in a very efficient manner to make dose adjustments to the medication or potentially use a different medication.”

AI recommendations

Recommendations for treatment will also be provided in the electronic medical records.

“We’ve always felt there needs to be a more effective way to enter data; it needs to be a living entity that provides real-time feedback,” Dr. Cohen said. “The EMR has been a way to store data and make billing more efficient. It’s never been a way to improve patient care in real time. I think providers are excited about this opportunity.”

In addition to medication-related recommendations, if a patient reports feeling sad and in need of help, the app could suggest a referral to a cancer support group.

Doctors must decide whether to go along with the artificial intelligence suggestions.

“It’s of course up to the provider and patient if those recommendations are followed,” Dr. Cohen said.

A Jonsson Comprehensive Cancer Center pilot study of the software showed favorable results. Patients answered a 15-question survey each week, for 12 weeks, on physical symptoms, mental health and daily functioning.

“Patients found the application was accessible,” Dr. Cohen said. “They found that even knowing providers had access to the information improved the outlook of their treatment and increased confidence in the care they were receiving.”

The patient app will provide nutritional information, information on visits and lab results, education and caregiver support. For instance, patients will have access to tailored information to help make decisions about their care such as whether to have surgery or chemotherapy first.

“When someone meets a physician in the office and they hear the word ‘cancer,’ we know they remember about 30% to 40% of what’s discussed, no matter how much time we spend,” Dr. Cohen said. “It allows for education to be given within the application and allows patients to make more informed decisions.”

Collaborators on the study are: UCLA Health co-principal investigator Dr. Cortney Eakin, as well as Dr. Rebecca Brooks, UC Davis; Dr. Jill Tseng, UC Irvine; Dr. Pratibha Binder, UC San Diego; and Dr. Lee-May Chen, UC San Francisco.

Learn more about cancer research at the Jonsson Comprehensive Cancer Center.

Courtney Perkes is the author of this article.