SPACER AI Maps the Local Signals That Attract or Exclude Immune Cells
Published in Artificial Intelligence, Bioinformatics, Immunology, Journals.
Published October 14, 2026.
Researchers at The University of Texas MD Anderson Cancer Center and collaborating institutions have introduced SPACER, an interpretable artificial intelligence model designed to identify local signals that attract immune cells to — or exclude them from — specific regions of tissue. The study was published in Nature Biomedical Engineering and announced on October 13.
What SPACER analyzes
SPACER, short for Spatial Analysis of Cellular Engagement and Recruitment, uses spatial transcriptomics: measurements that preserve both gene activity and the location of cells inside intact tissue. Rather than treating each cell as isolated, the model examines neighboring cellular communities and produces biologically interpretable recruitment signals.
The team applied the approach to 37 datasets spanning several cancer types. In the authors’ benchmarking tasks, SPACER was reported to be four to 21 times more effective than existing approaches at identifying biologically meaningful signals linked to cell recruitment. That comparison is specific to the tested datasets and tasks; it is not a measure of clinical performance.
Local signals, antigen presentation and mucins
The analysis suggested that immune-cell positioning is driven largely by signals in the immediate tissue environment. Tumors expressing genes involved in antigen presentation were more likely to attract T cells. The researchers also found that high mucin levels were associated with T-cell exclusion, poorer responses to immune-checkpoint inhibitors and poorer outcomes in several melanoma and gastric-cancer datasets.
Those mucin findings are associations and do not yet prove that mucins directly cause immune exclusion or treatment failure. They nevertheless provide a testable hypothesis for laboratory and clinical research. When SPACER was applied to preclinical datasets of immunotherapy-related myocarditis, it also highlighted molecular regions linked to T-cell recruitment and suggested a potentially important role for CD4-positive T cells.
Why it matters
Spatial datasets can show where cells are located; SPACER aims to help explain why they gather there. If validated prospectively, the method could help researchers prioritize biomarkers and therapeutic targets in cancer, autoimmune disease, tissue repair and other conditions in which cell positioning matters. The current work is a research tool, not a diagnostic test approved for patient care.
Sources: UT MD Anderson Research News, October 13, 2026; peer-reviewed study linked there in Nature Biomedical Engineering.