HIMSSCast: How Karmanos Cancer Institute uses AI to find high-risk breast cancer

Detroit-based Karmanos Cancer Institute, a network of 16 treatment locations across Michigan and Ohio that is part of McLaren Health Care, is using artificial intelligence to elevate risk assessment and is surfacing high-risk breast cancer patients who would benefit from supplemental care and imagery.

With women having a one in eight chance of developing breast cancer over their lifetimes, according to the American Cancer Society, finding breast cancer at earlier stages has moved the needle on survival rates, and can also improve patient treatment journeys and outcomes.

“It’s not cliche, it’s a reality,” said Brown.

Breast cancer often remains obscured in the annual mammograms of women with high levels of breast tissue density, which prompts health systems to send letters to patients that explain the challenges in standard screening. However, aiming additional screening efforts at women who have a direct family history has proved to be not enough to increase early-stage breast cancer diagnosis rates.

As Brown notes in this podcast interview, many women who are high risk due to many other factors aside from a family history of breast cancer often have no idea.

“Screening them once a year might not be enough to identify them early,” he said. “Identification of breast cancer early not only means better long-term survival, but many times, it means less treatment to cure that cancer.”

Karmanos is one of the first cancer hospitals to deploy Volpara’s artificial intelligence-driven risk assessment software at the point of mammogram screenings. Brown said that while it has only added two to three additional minutes to patient appointments, the ability of the AI to combine multiple risk assessment tools provides women and their primary care physicians with five-year and lifetime risk scores. 

At the time of the interview, Brown said the institute rolled out the AI risk assessment tool at two centers to test integration with the network’s electronic health records and “work out the kinks” communicating results with PCPs. 

Using the AI tool, Karmanos screened 7,000 patients and identified 700 as high-risk through the software and will continue to roll it out to all screening locations. Those women who come in for their annual mammograms and have risk scores above 20% are funneled into the women’s wellness program at Karmanos where they have a better chance to be diagnosed and treated early – or have the chance to prevent breast cancer – and reach the apex of survivorship.

Like what you hear? Subscribe to the podcast on Apple Podcasts, Spotify or Google Play!

Talking points:

• Partnerships with the Detroit Tigers and others increase screening
• Combining risk assessments into one AI-driven tool
• How capture of additional health history factors improves risk models
• Why scoring women can result in reaching other high-risk family members, including men
• How data-driven assessment unlocks prior authorization for additional imaging

More about this topic:

AI shows it can improve predictions for invasive breast cancer

New York approves AI-driven breast cancer diagnostic

Breast cancer technologies that improve patient outcomes

UPMC is using AI model for breast cancer sentinel node biopsies

AI helps radiologists improve accuracy in breast cancer detection with lesser recalls