Three clinical domains.
One platform.
Applied across three disease domains: longitudinal depth turned into personalized insight, and new knowledge.
Oncology
Gastroenterology
Neurology
Predicting chemotherapy resistance in colorectal cancer
Predicting IBD flares before they occur
Living models of brain tumour progression
Fusing whole-slide image pathomics with CT radiomics into a Deep Microenvironment Score that predicts resistance to adjuvant chemotherapy in Stage II to III colorectal cancer, before treatment begins. The aim: non-genetic biomarkers not visible on standard review.
A longitudinal multimodal model, combining structured clinical data, free-text NLP, and patient-reported outcomes, predicts inflammatory bowel disease flares weeks in advance. It enables pre-emptive treatment adjustment and reduces hospitalisation across the hospital-community care continuum.
Mechanistic-neural hybrid models of glioblastoma growth, parameterised from serial MRI, enabling personalised simulation of radiation response and surgical planning. Counterfactual treatment comparison before the first incision. Every patient deserves their own equation.
Clinical Lead: Prof. Shai Rosenberg · Prof. Aron Popovtzer
Pathology: Dr. Tzahi Neuman · Dr. Nir Pillar
Embedding model: Deep Microenvironment Score (DMS) · Multimodal fusion
Clinical Lead: Gastroenterology and Liver Diseases, Hadassah
Data: hospital and community continuum · Hybrid model: mechanistic + temporal embeddings
Clinical Lead: Prof. Shai Rosenberg
Framework: Reaction-Diffusion equations + Neural ODE · Serial MRI parameterisation