Blog

A discovery by researchers at the University of British Columbia promises to improve the treatment of patients with uterine cancer, the most common gynecological malignancy.

Using artificial intelligence (AI), researchers were able to identify patterns in thousands of images of cancer cells and identify a specific subset of uterine cancer that poses a much higher risk of recurrence and death in patients but would otherwise go undetected by traditional pathology and molecular diagnostics.

The results published today Nature communicationwill help physicians identify patients with high-risk disease who may benefit from more comprehensive treatment.

“Endometrial cancer is a complex disease, and some patients are much more likely to have cancer recur than others,” said Dr. Jessica McAlpine, professor and Dr. Chew Wei Chair in Gynecologic Oncology at UBC and a surgeon and scientist at BC Cancer and Vancouver General Hospital. “It is so important that patients with high-risk disease are identified so that we can intervene and hopefully prevent recurrence. This AI-based approach will help ensure that no patient misses an opportunity for potentially life-saving interventions.”

AI-supported precision medicine

The discovery builds on work by Dr. McAlpine and colleagues at BC’s Gynecologic Cancer Initiative – an inter-institutional collaboration between UBC, BC Cancer, Vancouver Coastal Health and BC Women’s Hospital – who in 2013 helped show that uterine cancer can be classified into four subtypes based on the molecular characteristics of the cancer cells, each of which poses a different risk to patients.

Dr. McAlpine and his team then developed an innovative molecular diagnostic tool called ProMiSE that can accurately distinguish the subtypes. The tool is now used throughout British Columbia, parts of Canada and internationally to support treatment decisions.

However, challenges remain. The most common molecular subtype, comprising approximately 50 percent of all cases, is largely a catch-all category for endometrial cancers without identifiable molecular features.

In this very large group, there are patients who achieve very good results and others whose cancer progression is very unfavorable. However, until now we have lacked the resources to identify those at risk and offer them appropriate treatment.”

Dr. Jessica McAlpine, Professor, University of British Columbia

Dr. McAlpine turned to his long-time collaborator and machine learning expert, Dr. Ali Bashashati, an assistant professor of biomedical engineering, pathology and laboratory medicine at UBC, to try to further segment the category using advanced AI methods.

Dr. Bashashati and his team developed a deep learning AI model that analyzes images of tissue samples taken from patients. The AI ​​was trained to distinguish between different subtypes and, after analyzing over 2,300 cancer tissue images, identified the new subgroup that had significantly worse survival rates.

“The power of AI is that it can objectively look at large volumes of images and identify patterns that human pathologists miss,” said Dr. Bashashati. “It’s like looking for a needle in a haystack. It tells us that this group of cancers with these characteristics are the worst offenders and pose a higher risk to patients.”

Making the discovery accessible to patients

Thanks to funding from the Terry Fox Research Institute, the team is now investigating how the AI ​​tool could be integrated into clinical practice alongside traditional molecular and pathological diagnostics.

“The two work hand in hand, with AI adding an additional layer to the testing we already do,” said Dr. McAlpine.

One advantage of the AI-based approach is that it is cost-effective and easily deployable across geographies. The AI ​​analyzes images that pathologists and healthcare providers routinely collect, even in smaller hospitals in rural and remote communities, and shares them when a second opinion is sought on a diagnosis.

The combined use of molecular and AI-based analytics could allow many patients to stay in their home community for less-invasive surgeries, while ensuring that those who need treatment at a larger cancer center can do so.

“What really excites us is the opportunity for greater equity and access,” said Dr. Bashashati. “AI doesn’t care if you live in a large urban center or a rural community, it would just be available. So we’re hopeful that this could really change the way we diagnose and treat uterine cancer in patients everywhere.”