Endometrial Cancer: AI Identifies High-Risk Subtypes

Endometrial cancer is the most common gynecologic cancer, and a discovery made by researchers at the University of British Columbia could improve treatment for individuals with this disease.

The researchers identified a unique subset of endometrial cancer that places patients at significantly higher risk of recurrence and death but would otherwise go unidentified by conventional pathology and molecular diagnostics by using artificial intelligence (AI) to identify patterns across thousands of cancer cell images.

Physicians will be able to identify patients with high-risk conditions who may benefit from more extensive therapy with the aid of the findings, which were published in Nature Communications.

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

AI-driven precision healthcare

The finding expands on research conducted in 2013 by Dr. McAlpine and associates at the BC Gynecologic Cancer Initiative, a multi-institutional partnership between UBC, BC Cancer, Vancouver Coastal Health, and BC Women’s Hospital. This collaboration helped demonstrate that endometrial cancer could be divided into four subtypes according to the molecular features of cancerous cells, each of which presented a distinct risk to patients.

Subsequently, Dr. McAlpine and colleagues created ProMiSE, a novel molecular diagnostic instrument that is capable of precisely differentiating between the subtypes. Today, the instrument is utilized to inform treatment decisions in British Columbia, other parts of Canada, and abroad. 

However, difficulties still exist. The most common molecular subtype, accounting for around half of all cases, is essentially a catch-all for endometrial malignancies that don’t have any distinguishable molecular characteristics.

“There are patients in this very large category who have extremely good outcomes, and others whose cancer outcomes are highly unfavorable. But until now, we have lacked the tools to identify those at-risk so that we can offer them appropriate treatment,” said Dr. McAlpine.

In an attempt to further segregate the category using cutting-edge AI techniques, Dr. McAlpine turned to his longtime partner and machine learning specialist, Dr. Ali Bashashati, an assistant professor of biomedical engineering, pathology, and laboratory medicine at UBC.

A deep learning artificial intelligence algorithm created by Dr. Bashashati and his colleagues examines photos of tissue samples taken from patients. After examining more than 2,300 photos of cancer tissue, the AI—which had been trained to distinguish between several subtypes—identified the novel subgroup that had noticeably lower survival rates.

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

Educating patients about the finding

Thanks to funding from the Terry Fox Research Institute, the team is currently investigating how the AI tool could be incorporated into clinical practice in addition to conventional molecular and pathology tests.

“The two work hand in hand, with AI providing an additional layer on top of the testing we’re already doing,” said Dr. McAlpine.

The AI-based approach’s cost-effectiveness and ease of deployment across geographical boundaries are two advantages. Even at smaller hospital locations in rural and distant towns, the AI examines photos that are regularly collected by pathologists and healthcare professionals and shared when getting a second opinion on a diagnosis.

In addition to guaranteeing that patients who require therapy at a larger cancer center can receive it, the combined use of molecular and AI-based analysis may enable many patients to undergo less invasive surgery in the comfort of their own homes.

“What is compelling to us is the opportunity for greater equity and access,” said Dr. Bashashati. “The AI doesn’t care if you’re in a large urban center or rural community, it would just be available, so we hope that this could transform how we diagnose and treat endometrial cancer for patients everywhere.”

For more information: AI-based histopathology image analysis reveals a distinct subset of endometrial cancers, Nature Communications, https://doi.org/10.1038/s41467-024-49017-2