GLIOGEN-X
General Aspects
GLIOGEN-X – “From Magnetic Resonance Imaging to Molecular Signatures: Generative AI for Virtual Biopsy in Gliomas” is a research project aimed at investigating whether brain magnetic resonance imaging (MRI) scans already contain the information needed to characterize the molecular profile of gliomas.
Coordinated by Siena Imaging s.r.l., the project brings together Fondazione Toscana Life Sciences (TLS), The University of Manchester, Fundació per a la Universitat Oberta de Catalunya, and the Northern Care Alliance NHS Foundation Trust (affiliated entity). It is one of seven projects funded by the European Innovation Council (EIC) under the Pathfinder Challenges call “Generative-AI Based Agents to Revolutionize Medical Diagnosis and Treatment of Cancer” within the Horizon Europe programme, dedicated to supporting high-risk, high-gain frontier research.
GLIOGEN-X aims to determine whether MRI data can provide the information needed to identify the molecular characteristics of gliomas, potentially paving the way for AI-powered virtual biopsy approaches.
The project has received €3,996,043.38 in funding and will run for a period of four years.
Project Activities
GLIOGEN-X explores the possibility of extracting clinically relevant molecular information about gliomas directly from magnetic resonance imaging (MRI) scans through the use of generative artificial intelligence. The project’s long-term goal is to contribute to the development of a diagnostic decision-support tool that could eventually complement conventional diagnostic procedures and help reduce the need for invasive diagnostic and surgical interventions.
The project is built around three main pillars:
- Generative experimentation: AI models are trained to generate realistic MRI scans from real imaging data, with the aim of understanding which molecular characteristics of gliomas leave measurable signatures within the images themselves.
- Validation: the project seeks to verify the existence of a detectable molecular “biological fingerprint” that can be identified through MRI imaging.
- Transparent and trustworthy prediction: the development of interpretable “virtual biopsy” algorithms capable of identifying key molecular signatures using MRI images alone, while also providing clear indications of their level of uncertainty and confidence.
This represents a novel approach in the field of neuro-oncology, leveraging artificial intelligence to demonstrate how medical imaging may reveal information about the molecular characteristics of brain tumors. The ultimate objective is to support a more personalized and less invasive diagnostic pathway for patients.
A distinctive feature of GLIOGEN-X is the adoption of a federated learning infrastructure, which allows AI models to be trained and validated without transferring patient data between institutions. This approach ensures full compliance with GDPR requirements while maintaining the highest standards of data privacy and security and enabling effective collaboration among international research and clinical centers.