To learn how Sutter cancer researchers are applying machine learning to cancer care, we interviewed John Chan, M.D., a gynecologic oncologist at Sutter’s California Pacific Medical Center (CPMC) and Palo Alto Medical Foundation, and the Denise & Prentis Cobb Hale Endowed Chair in Gynecologic Oncology Research.

How might artificial intelligence (AI) be used to improve cancer care?

AI builds smart machines capable of performing tasks that typically require human intelligence like learning, reasoning and problem-solving skills. We already see AI playing a key role in our daily routines and our interactions with media, transportation and communications.

With a surge of new knowledge in this area, researchers are applying AI and machine learning to innovate healthcare with improved diagnoses and treatment. As machine learning algorithms are exposed to more data—in some cases gleaned from the electronic heath record (EHR)—they can detect hidden patterns within data that can be used to perform a task without sophisticated programming.

Given the large number of Americans diagnosed with cancer and the huge volume of data generated during cancer treatment, there is a growing interest in using AI and machine learning to improve cancer care.

How are you applying AI and machine learning to your gynecologic oncology practice at Sutter?

Through a collaboration with the University of North Carolina and The Medical College of Georgia at Augusta University, we’re using machine learning to classify cancer patients into high- or low-risk groups, and to personalize sub-groups of uterine cancer patients who may or may not benefit from chemotherapy. Our research resulted in a clinical calculator that was described in Gynecologic Oncology.

The findings suggest a clinical calculator can help predict benefit and risk of chemotherapy treatment in patients with uterine cancer. And now we’re expanding the machine learning tool to study advanced-stage cancers and gynecologic diseases like ovarian cancer.

We plan to share the clinical calculator with providers across Sutter and with peer institutions that care for gynecologic cancer patients who live in more remote communities, or who are otherwise unable to travel to larger cancer centers. That way, more cancer patients may benefit from the research underlying this new AI-based tool. This could help increase access to safer and more effective cancer treatments, and reduce healthcare costs.

What are the potential benefits of using AI and machine learning to guide oncology treatment decisions?

Machine learning can “free think” creatively because it’s not confined by pre-existing human biases that may be present in routine clinical practice. This and other AI approaches like deep learning can help us determine levels of risk and benefit associated with various cancer treatments.

For example, an older patient with ovarian cancer may be at increased risk of treatment-related complications compared with younger patients. AI can identify risk profiles to help guide treatment decisions based on data from similar sub-groups of cancer patients.

In clinical oncology, AI has increasingly been applied to harness the power of the EHR. Specifically, AI-based natural language processing techniques may help predict the development of diseases across large healthcare systems.

At Sutter, our integrated network enables access to EHR and outcomes data from a highly diverse patient base. Machine learning can find associations and calculate risk scores to better predict treatments that will provide optimal benefit and reduced risk of treatment-related complications or adverse effects. This approach may become a new decision tool that we can add to our clinical toolkit and share with collaborators across Sutter.

Cancer care and research at Sutter: High “CQ” using machine learning:

Other Sutter physicians, researchers, and innovators across the system are applying AI to cancer care and elsewhere. Here’s how:

• Earlier this year, Sutter partnered with Bay Area startup Ferrum Health to develop machine learning algorithms for early detection of lung cancer at radiology clinics in Sacramento. Ferrum’s AI platform read 10,000 CT scans and reports. Within 90 days, it flagged 83 cases in which it detected a mass of tissue on the scan that was not mentioned in the report.

• Albert Chan, M.D., M.S., Sutter Health chief of digital patient experience, is leading digital transformation efforts across Sutter. He oversees a unified digital patient engagement and virtual care strategy, including leadership of Sutter’s patient portal My Health Online, telemedicine and artificial intelligence-powered solutions.

• Through a collaboration with Stanford Medicine, Sutter Health systems researchers will launch the Oncoshare Project using a “big data” approach to improve breast cancer care. Oncoshare enables researchers to generate high-resolution maps of breast cancer treatment, and identify care pathways that yield the best outcomes for patients.

Learn more about Sutter research and clinical trials.