AI Tool Predicts Which Cancer Tumors Respond to Treatment

Doctors may soon predict which cancer treatments will work for you before you even start them, thanks to a new AI tool that reads tumors like blueprints.

Researchers at the University of Chicago developed a system that maps out how cells organize inside tumors, creating what they call a "floor plan" for each cancer. Just like homes have different layouts with rooms serving different purposes, tumors have unique structures where cells cluster and interact in specific patterns.

The breakthrough solves a problem that has stumped cancer doctors for years. Every tumor creates its own microenvironment, a complex neighborhood of cells that determines how the cancer behaves and responds to treatment. Until now, there was no good way to compare these environments between patients or predict outcomes based on their structure.

Dr. Arjun Raman, who led the study published in Cell Reports Medicine, approached the challenge by thinking about tumors like flocks of birds. Individual birds follow simple rules, but together they create complex flying patterns. Similarly, cancer cells interact to form organized groups, which cluster into larger units, building up to the complete tumor structure.

His team analyzed 262 solid tumors using spatial transcriptomics, a technology that shows not just which genes are active but exactly where that activity happens inside tissue. They developed mathematical formulas to describe how tumor cells arrange themselves, creating measurable floor plans that AI could compare.

The real test came with 16 lung cancer patients receiving immunotherapy. The AI analyzed their tumor floor plans and predicted who would respond to treatment. It worked better than current medical tests.

Why This Inspires

This research transforms impossibly complex cancer data into actionable information that saves lives. Instead of trial and error with treatments that may not work, oncologists could someday scan a tumor's structure and immediately know the best approach for each patient.

Dr. Raman is already expanding the work to ovarian cancer with collaborators at UChicago Medicine. The method works with any tumor type because it focuses on universal patterns of how cancer cells organize themselves.

The tool represents what AI does best, not replacing human creativity but solving problems beyond human capability. Doctors spend years learning to recognize patterns in tumors, but this system can instantly compare a patient's cancer against hundreds of others to find matches and predict outcomes.

When UChicago Medicine opens its new AbbVie Foundation Cancer Pavilion in 2027, technologies like this could be standard practice. Patients would walk in, get their tumor mapped, and receive personalized treatment plans based on thousands of similar cases analyzed in seconds.

For cancer patients facing uncertain treatment journeys, knowing which path will likely work brings both practical benefits and emotional relief.