AI & digital pathology
AI Companion Diagnostic for Colorectal Cancer
An AI that reads both the immune cells and the tissue scaffold to match colorectal-cancer patients to the right treatment.
Two patients with the same diagnosis can need very different treatments, yet today’s tools rarely tell them apart well enough. This project builds an AI-powered companion diagnostic for colorectal cancer that learns from two signals at once: the tumour-infiltrating lymphocytes (immune cells inside the tumor) and the extracellular matrix (the collagen scaffold around it), read through NanoMslide imaging.
The results are striking: patients with the optimal combination of immune infiltration and tissue architecture show dramatically better survival outcomes than those stratified by standard staging alone — a finding that points to a level of precision currently unavailable in routine clinical practice.
Developed through a La Trobe / ONJCRI / AlleSense partnership, the platform has attracted support from patient advocacy organisations, a competitive selection at the Protoaxiom national healthtech competition (judged by leading medtech investors), and the Anthropic AI for Science program. A critical design principle: the platform is built to run on equipment every pathology lab already owns, at a fraction of the cost of existing gold-standard tests — making precision stratification accessible beyond specialist centres.
