Researchers at the University of Texas MD Anderson Cancer Center have demonstrated that artificial intelligence (AI)-based tumor biopsy analysis can predict response to immunotherapy in a study of patients with rare cancers. Journal of Cancer Immunotherapy.
The analysis, led by Aung Naing, M.D., Ph.D., professor of Investigative Cancer Therapeutics, builds on recently published research that identified features of the tumor microenvironment that predict immunotherapy response in patients with rare cancers, even in patients without known markers of immunotherapy response.
AI-based pathology has the potential to provide clinicians with useful information about both the tumor and its surrounding microenvironment, helping guide individualized treatment decisions for patients undergoing immunotherapy. ”
Aung Naing, MD, Professor of Clinical Trial Cancer Therapeutics
How does this AI tool work and what benefits does it offer in guiding immunotherapy treatments for rare cancers?
Naing’s previous publication identified two characteristics that may best indicate whether a patient is responding to immunotherapy. These include changes in the number of immune cells present in the tumor before treatment and changes in immune cell infiltration into the tumor during treatment.
Manually counting individual immune and cancer cells on pathology slides requires significant effort, especially when trying to scale the effort to large numbers of slides and patients, but AI tools can quickly accomplish this. In the current study, AI-based analysis rapidly generated these measurements and tracked changes over time across multiple biopsies from the same patient. It is also worth noting that this approach utilizes standard pathology slides that are already routinely collected.
How was this approach implemented and what will this research do next?
Although increased tumor immune infiltration and decreased tumor content were both predictive indicators on their own, these individual signals are more powerful when combined. This pattern reflects both a vigorous immune response and a reduction in tumor burden.
Patients with good signal had a 64% lower risk of disease progression or death and lived almost four times longer on average (median survival 42 months vs. 10 months) compared to patients without these markers.
Although these results are promising, this approach requires validation in a larger patient population before it can guide treatment decisions in the clinic.
“While this AI-powered approach requires validation, this is an exciting step forward as it shows that meaningful insights can be extracted from routine pathology samples across a diverse group of rare cancers,” said Nine.
sauce:
University of Texas MD Anderson Cancer Center
Reference magazines:
Delbara, M.H. Others. (2026). Artificial intelligence-based analysis of tumor microenvironment predicts response to pembrolizumab in rare tumors. Journal of Cancer Immunotherapy. DOI: 10.1136/jitc-2025-014768. https://jitc.bmj.com/content/14/6/e014768
