CancerScan develops a next‑generation pathology slide scanner that integrates slide‑level multi‑omics profiling from tumour biopsies with standard clinical data to (semi)automatically generate a patient‑specific tumour digital twin. By mapping tumour communication and tumour‑microenvironment interactions, reconstructing biomedical knowledge graphs and using AI‑enabled modelling with hardware acceleration, the platform will simulate drug responses and support more personalised decisions. Pancreatic cancer is used as the first model system, combining organoid experiments with clinical validation in collaboration with hospital partners.