Abstract 4136692: Spatial Transcriptomics Identifies Novel Clinical Markers to Classify the Etiology of Rare and Aggressive Cardiac Sarcomas

Background: Cardiac sarcomas, such as leiomyosarcoma (LMS), synovial sarcoma (SS), and undifferentiated pleomorphic sarcoma (UPS), lack sufficient diagnostic markers for early diagnosis and tailored therapeutic interventions. Understanding their molecular signatures with spatial-temporal resolution compared to normal heart tissue is crucial. Hypothesis: Spatial transcriptomics will reveal a comprehensive set of novel clinical markers, including changes in cellular composition, stemness, and metabolic profiles unique for each cardiac sarcoma subtypes. Aims: Identify novel molecular markers for the accurate diagnosis of cardiac sarcoma subtypes. Methods: Using 10X Visium spatial transcriptomics, we analyzed three cardiac sarcoma subtypes (LMS, SS, UPS) and compared them with normal heart tissue. Differential expression analysis identified novel target genes suitable for clinical disease stratification. Subsequently, machine learning models reveal an unbiased transcriptomic profile to distinguish normal from sarcoma tissues. Finally, immunohistochemistry analysis validated the differential marker gene expression. Results: We identify several novel clinical markers, such as IGKC, PCOLCE, NET1, TLE2, MDFI, BGN, TNNC1, and CNN3, that classify normal heart tissue from sarcoma subtypes. Machine learning models and immunostaining analysis confirm the transcriptomic signature in each sarcoma subtype. Pathway enrichment analysis show LMS enriched in antigen processing and B cell-mediated immunity, SS in protein folding and response to incorrect proteins, and UPS in axon development and Wnt signaling regulation. Trajectory Reconstruction Analysis indicate higher stemness scores in all sarcomas, with SS highest, followed by UPS and LMS. Additional cell cycling analysis show LMS cells in G2M phase, SS in G1, and UPS in S phase. SS express a high number of cancer stem markers (EZH2, BMI1, KDM5B, LGR5, HIF1A), while UPS express KDM5B and CTNNB1, and LMS show fewer stem-related marker genes (EZH2, BMI1, KDM5B, CTNNB1). Both SS and LMS express CD44. Glucose and oxidative pathways are also differentially regulated in each sarcoma subtype. Conclusion: Spatial transcriptomics uncovers distinct clinical markers, stemness characteristics, and metabolic activity profiles in cardiac sarcomas, offering insights for fast, accurate and improved diagnostics and therapeutics.