Hendricks-Kretzer was diagnosed with Stage 2B triple-positive breast cancer after a tumor had grown to grapefruit size and spread to her lymph nodes. On her surgeon’s advice, she joined a UChicago Medicine clinical trial focused on developing more personalized treatments. The trial uses gene-expression data to identify biomarkers that may indicate which therapies a tumor is likely to respond to.

Hendricks-Kretzer received an antibody-drug conjugate that shrank her tumor enough for her to avoid a particularly potent chemotherapy drug, sometimes nicknamed “the red devil” because of its severe side effects. She then chose a less toxic chemotherapy combination for 12 weeks. After 21 weeks of treatment, she was cancer-free, according to her account.

The artificial-intelligence research is a separate step: AI was not used to select Hendricks-Kretzer’s care. Oncologist Frederick Howard, an assistant professor at the university, said his team is using samples from the trial and digital pathology images to train software to look for tumor patterns associated with responses to different treatments.

The broader aim, Howard said, is to give patients the amount of treatment they need without exposing them to unnecessary side effects. If the technology proves reliable, it could also help make treatment decisions faster and less expensive. AI has long been used to assist with breast cancer screening, including interpreting mammograms; predicting which treatment will work best is a newer research goal.

Howard cautioned that the promise of AI may currently outpace what it can deliver. Researchers still need to establish whether these tools add meaningful value and how they should be used before they are adopted routinely, he said.

Hendricks-Kretzer said participating in research gave her a sense of purpose during a difficult period. She credited past clinical-trial participants with helping make today’s standard treatments possible and said she hoped her participation could contribute to future care.