Abstract 2091: Integrating single-cell spatial whole transcriptome and histopathology to uncover drivers of tumor heterogeneity in lung adenocarcinoma and squamous cell carcinoma

Lung cancer, the leading cause of cancer-related deaths worldwide, primarily consists of non-small cell lung cancer (NSCLC), including subtypes like lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). These subtypes exhibit distinct histopathological features but share overlapping molecular characteristics. Understanding the spatial and molecular heterogeneity of these tumors is critical for identifying novel biomarkers and therapeutic targets. This study utilizes the CosMx® Spatial Molecular Imager (SMI) platform to profile LUAD and LUSC, aiming to uncover shared and unique gene signatures, tumor heterogeneity, and rare tumor cell subtypes, while integrating histopathological insights from H&E staining. Two formalin-fixed, paraffin-embedded (FFPE) lung tumor samples—one LUAD and one LUSC—were profiled using the CosMx Whole Transcriptome (WTX) panel, which profiles approximately 19,000 genes with single-cell subcellular resolution. Spatial transcriptomic data were analyzed using Uniform Manifold Approximation and Projection (UMAP) for dimensionality reduction and Leiden clustering to identify cell clusters. Marker genes were used to annotate cell types and generate a high-resolution cellular atlas of the tumor microenvironment (TME). Hematoxylin and eosin (H&E) staining was performed on the same slides, providing histopathological context and validating CosMx-derived cell types, including immune and tumor hotspots. The alignment of H&E and CosMx SMI data highlights their potential for AI model training that integrates spatial transcriptomics and histopathology. Our analysis revealed shared and unique gene signatures between LUAD and LUSC. Shared signatures included common tumorigenic processes such as cell proliferation, immune modulation, and angiogenesis. Unique signatures distinguished each subtype, with LUAD enriched in epithelial-to-mesenchymal transition pathways, while LUSC showed upregulation of squamous differentiation markers. UMAP and Leiden clustering identified distinct tumor and immune cell subpopulations, demonstrating spatial and molecular heterogeneity with clear immune and tumor hotspots. CosMx SMI data and H&E staining corresponded well, suggesting their potential for AI model integration. Integrating spatial transcriptomics and histopathology using CosMx SMI provided valuable insights into the molecular and spatial complexity of LUAD and LUSC. This study identified rare tumor cell subtypes and mechanisms of tumor progression, underscoring the -Bruker Confidential-potential for AI-driven biomarker discovery and therapeutic strategies in NSCLC and beyond. -Bruker Confidential- Citation Format: Michael Patrick, Liang Zhang, Patrick Danaher, Shanshan He, Sierra Mckinzie, Haiyan Zhai, Michael Rhodes, Joseph Beechem. Integrating single-cell spatial whole transcriptome and histopathology to uncover drivers of tumor heterogeneity in lung adenocarcinoma and squamous cell carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2091.