LAUSR

Uncovering the composition and structure of the tumor microenvironment is critical to a mechanistic understanding of tumorigenesis and therapeutic resistance. Spatial transcriptomics (ST) technology has enabled profiling the molecular features of tumor tissue with position information. However, the spatial probe of various ST strategies with a 10~100 μm diameter might capture a mixture of transcripts from multiple cells or cell lineages. Cell type deconvolution in ST data of tumor tissues remains challenging for existing methods, which are designed to decompose general ST or bulk data. We develop the Spatial Cellular Estimator for Tumors (SpaCET) to infer cell identities from tumor ST data. Without the need of inputting cell references, SpaCET estimates malignant and non-malignant cell abundance by using a gene pattern dictionary of copy number alterations in common malignancies and a hierarchical atlas of immune/stromal cells, respectively. SpaCET provides higher accuracy than existing methods based on eight ST datasets on seven cancer types with matched double-blind histopathology annotations as ground truth. Furthermore, SpaCET can reveal the potential intercellular interactions at the tumor-immune interface by integrating inferred cell fractions with the ligand-receptor interaction network. We expect that SpaCET will be a valuable tool for spatial cancer biology. Citation Format: Beibei Ru, Jinlin Huang, Yu Zhang, Kenneth Aldape, Peng Jiang. Exploring the tumor microenvironment in a spatial context with SpaCET [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB151.