Abstract 89: TissueTrek: An interactive multimodal web-based platform for exploring spatial morphology-molecular relationships in cancer.

Key Points

• The aim is to develop an accessible platform for exploring spatial morphology-molecular relationships in cancer.
• Built on a cohort of 57 triple-negative breast cancer tissue cores.
• Annotated tumor regions of interest by board-certified pathologists.
• Extracted multiple modalities from H&E images, spatially profiled genes, and proteins.
• Integrated machine learning outputs to enhance interpretability.
• Utilized web technologies for real-time exploration of spatial data.
• TissueTrek offers the first ROI-resolved interface for simultaneous visualization of multiple data types.
• Enables interactive exploration of spatial gene and protein expression alongside pathomic features.
• Facilitates understanding of complex spatial morphology-molecular relationships in tumors.
• Provides interpretable maps that enhance insight into tumor biology.

Abstract