AI Predicts Rectal Cancer Outcomes from Routine pathology Images

In most cancers, the area around the tumor-the immune landscape-plays a big role in how the cancer grows and how well patients respond to treatment. However, the complex interactions between immune cells, tumor cells, and treatment are often not well understood.

A new study, published in eBioMedicine, used AI to examine routine pathology images. The goal was to measure the types and amount of key immune cells around rectal cancer tumors to predict how this “tumor microenvironment” affects a patient’s survival and if the cancer comes back.

Pathologists currently examine these images, created from tumor tissue biopsies, under a microscope to make a diagnosis. researchers wanted to see if AI could be trained to identify key immune cell patterns in the images and connect them to patient outcomes much faster.

Dr. Charles-Antoine Collins-Fekete, a senior author of the study from UCL Medical Physics & Biomedical Engineering, said, “Pathology slides are already part of routine care, so they’re a readily available source of data. We thought we could extract useful facts about a patient’s tumor from these slides using AI, which has become very good at analyzing medical images, and link this to patient outcomes.”

“We found that AI can identify crucial immune signals from these slides. Importantly, it can do this in minutes, instead of days, which is how long slower and more expensive methods like whole-genome sequencing or spatial transcriptomics take. This could make more personalized diagnosis and treatment practical and affordable, potentially improving patient outcomes.”

The team studied samples from three groups of patients, including those in the ARISTOTLE clinical trial. They found that patients wiht more lymphocytes-immune cells that fight infection and diseases, including cancer-in and around their tumors tended to live longer and were less likely to experience cancer recurrence.

However, patients with more macrophages-another type of immune cell that usually cleans up harmful invaders like viruses but can sometimes help tumors grow-had worse outcomes.

These immune features aren’t currently used when making clinical decisions for rectal cancer, but they could offer a new way to personalize chemoradiotherapy treatment and identify which patients are most likely to benefit.