Researchers at The University of Texas MD Anderson Cancer Center have developed an artificial intelligence (AI) approach that analyzes digitized tumor biopsy images to measure changes in the tumor microenvironment over time. Published in the Journal for ImmunoTherapy of Cancer, the study demonstrates how AI can extract quantitative information from routine pathology slides that may help researchers identify biomarkers associated with responses to immunotherapy. While additional validation is needed before the approach could be used more broadly in clinical practice, the work highlights the growing role of AI-assisted image analysis in cancer research.
Quantifying changes in the tumor microenvironment
The study was led by Aung Naing, MD, professor of investigational cancer therapeutics at MD Anderson, and builds on previous research identifying characteristics of the tumor microenvironment that may predict patient responses to immunotherapy. The researchers focused on two measurable features: immune cell infiltration within tumors and changes in overall tumor content over the course of treatment.
To evaluate these changes, the team analyzed routine hematoxylin and eosin (H&E)-stained pathology slides collected from multiple biopsies taken from the same patients at different stages of treatment. Using AI-based image analysis, the researchers quantified tissue characteristics across these longitudinal samples and assessed how they changed over time.
Combining biomarkers improves predictive performance
The AI model evaluated two independent indicators: increasing immune cell infiltration and decreasing tumor content. Each metric provided useful information on its own, but the researchers found that combining both measures produced a stronger model for predicting clinical benefit than either feature alone.
Rather than relying on additional specialized imaging techniques or novel laboratory assays, the approach uses routinely collected pathology slides, demonstrating how digital image analysis can generate additional quantitative information from existing tissue samples.
"While this AI-powered approach needs validation, this is an exciting step forward because it shows that meaningful insights can be extracted from routine pathology samples," Naing said.
What this means for laboratory managers
For laboratory managers overseeing pathology or histology laboratories, the study illustrates how AI-based image analysis may complement existing digital pathology workflows. By automatically extracting quantitative features from digitized tissue images, these tools have the potential to support biomarker discovery and provide more consistent analysis of complex pathology datasets.
Although the researchers did not evaluate laboratory workflow efficiency or operational performance, automated image analysis may help laboratories manage increasingly large collections of digital pathology images while applying standardized analytical criteria across samples. As digital pathology continues to expand in research settings, approaches such as this one could provide additional insights from routinely collected tissue specimens while supporting ongoing biomarker development in oncology.
This article was created with the assistance of Generative AI and has undergone editorial review before publishing.
