Spatial Transcriptomics Tools May Link Tumor Organization to Treatment Response

Spatial transcriptomics has revolutionized cancer research by allowing scientists to visualize not just which genes are active in a tumor, but where that activity occurs within the tumor microenvironment. However, these maps can be challenging to compare across different tumors due to the heterogeneity of the tumor microenvironment.

Researchers at the University of Chicago have now developed a framework to compare spatial maps across various tumor types. In their study titled "Pan-tumor spatial transcriptomics reveals conserved properties of tumor organization," published in Cell Reports Medicine, the team examined spatial transcriptomics data from 262 solid tumors and discovered that tumor microenvironments could be organized into recurring, hierarchically structured multicellular regions, termed "spatial groups" (SGs).

These SGs represent recognizable biological domains, from the overall tissue context to local cellular neighborhoods. When comparing tumors using these spatial groups, the researchers found that the dominant factor separating tumors was the spatial heterogeneity of immune biology. In a retrospective study of 16 non-small cell lung cancer patients treated with immune checkpoint blockade, this classification successfully distinguished clinical responders from nonresponders and captured features associated with sensitivity to immunotherapy.

The study's authors suggest that these spatial groups may be crucial organizing domains within the tumor microenvironment, linking spatial structure, biological function, and treatment response. Dr. Arjun Raman, the study's senior author and assistant professor of pathology at the University of Chicago, compared this concept to understanding a flock of birds: individual cells behave independently but form cohesive structures that influence the tumor's overall behavior.

By treating these subunits as spatial groups, the researchers created a statistical and machine learning framework to compare tumor "floor plans." This approach places tumors in a comparative latent space, where samples are arranged based on spatial similarity. In theory, this framework could enable researchers or clinicians to determine whether a newly profiled tumor resembles those previously associated with response or resistance to specific therapies.

The ultimate goal is to integrate this comparative spatial analysis into precision oncology, allowing clinicians to quickly assess whether a patient's unique tumor profile indicates suitability for a particular treatment regimen. While the immunotherapy analysis involved only 16 non-small cell lung cancer cases, the study demonstrates the potential for spatial transcriptomics to move beyond detailed tumor maps to provide a standardized language for comparing them across tumor types.