Abstract P24: Multi-Scale Modeling for Prediction of Spatial Single-Cell and Bulk Transcriptomic Profiles from Whole-Slide Images of Colorectal Cancer

Key Points

• Machine learning models significantly enhance transcriptomic profile generation from whole-slide images in colorectal cancer.
• The study shows a robust performance of GNN models compared to existing deep learning approaches for multiple spatial resolutions.
• Models were validated against a public bulk RNA-seq dataset (TCGA-COAD/READ), demonstrating their clinical relevance.
• This novel multi-scale framework highlights the potential of integrating spatial omics with traditional histopathological methods.