A New AI-Powered Atlas Maps Immune Structures in Tumors, Offering Clues to Prognosis and Treatment
Researchers at The University of Texas MD Anderson Cancer Center have created a comprehensive spatial atlas of tertiary lymphoid structures (TLSs) across multiple cancer types. Published in Science, the atlas reveals that the maturation state, spatial location, and composition of TLSs within tumors may offer clinically relevant information about cancer prognosis and treatment response.
The study, led by Linghua Wang, M.D., Ph.D., professor of Genomic Medicine at UT MD Anderson, introduces scalable AI frameworks to detect, profile, and classify these organized immune structures from spatial omics data and routine pathology slides.
What are Tertiary Lymphoid Structures?
TLSs are organized immune structures found within the tumor microenvironment. They function as local command centers, coordinating the body’s antitumor immune response. Prior research has suggested that mature TLSs are often associated with better patient outcomes and improved responses to immunotherapy.
Key Findings from the Pan-Cancer Atlas
The atlas includes 340 samples from 12 different cancer types. The analysis uncovered several critical insights:
- TLSs vary substantially across different tissues. As they mature, they undergo coordinated changes in their immune, stromal, and vascular components.
- The proximity of TLSs to tumor cells is linked to spatial gradients of tumor signaling, suggesting a dynamic interaction between the immune structure and the cancer.
- The AI framework enabled the evaluation of 25,088 TLSs from over 3,000 whole-slide images across 10 independent cohorts.
The TLS composition score, which accounts for both the number and the maturation state of TLSs, significantly outperformed conventional TLS measures in stratifying patients by prognosis and treatment response.
A Novel Scoring System
The team developed a composite scoring system designed to stratify patients more effectively than traditional methods. This system integrates multiple data points from the TLS analysis, allowing for more precise predictions across different cancer types and treatment contexts.
The Path Forward
While the findings are promising, the researchers emphasize that the TLS composite scoring approach requires validation in prospective clinical trials before it can be widely adopted.
Future studies will also investigate how to promote TLS maturation and enhance their spatial interaction with tumor cells, potentially opening new avenues for therapeutic intervention.