Skip to main content
Go to the home page of the European Commission (opens in new window)
English en
CORDIS - EU research results
CORDIS

​Multi-modal, Multi-site, Multi-omic, Multi-AGent AI framework for the Clinical management of ILC

Objective

Invasive lobular carcinoma (ILC) is the second most common histological subtype of breast cancer, yet it remains underrepresented in clinical trials due to its distinct clinical course and metastatic behavior. Radiological detection of ILC is particularly challenging, as lesions often present with low contrast and diffuse infiltration. In pathology, overlapping morphological features between classic and non-classic ILC and other breast cancer subtypes lead to diagnostic uncertainty and high inter-pathologist variability. These limitations extend to molecular assays and AI models, which have largely been trained on cohorts with low ILC representation and typically underperform for this subtype. M4GIC-ILC—a Multi-modal, Multi-site, Multi-omic, Multi-AGent AI framework for the clinical management of ILC—addresses these challenges by unifying representation learning, generative AI, and supervised prediction across pathology, radiology, molecular, and clinical data. Foundation-model backbones are fine-tuned for ILC to learn batch-aware, biologically grounded embeddings, while generative AI provides calibrated proxies for missing modalities (virtual IHC, virtual MGS, and imaging views), enabling clinicians to interpret cases in familiar diagnostic settings. On top, task-specific predictors estimate diagnosis (ILC vs other; classic vs non-classic), staging (T/N, focality, laterality), relapse risk, and treatment benefit (endocrine, CDK4/6i, chemotherapy), with calibrated uncertainty and abstention when confidence is low. The framework is orchestrated by a multi-agent system that integrates these tools into interactive, natural-language reports with visual explanations and full provenance. Development and validation occur through federated learning on a harmonized, FAIR multisite repository. M4GIC-ILC will bridge the current diagnostic and therapeutic gap in ILC, delivering trustworthy, explainable AI that improves patient stratification and personalizes breast cancer care.

Fields of science (EuroSciVoc)

This project has not yet been classified with EuroSciVoc.
Be the first one to suggest relevant scientific fields and help us improve our classification service

Keywords

Programme(s)

Topic(s)

Funding Scheme

HORIZON-EIC - HORIZON EIC Grants

See all projects funded under this funding scheme

Call for proposal

(opens in new window) HORIZON-EIC-2025-PATHFINDERCHALLENGES-01

See all projects funded under this call

Coordinator

KATHOLIEKE UNIVERSITEIT LEUVEN
Net EU contribution
€ 1 424 705,50
Address
OUDE MARKT 13
3000 Leuven
Belgium
Region
Vlaams Gewest Prov. Vlaams-Brabant Arr. Leuven
Activity type
Higher or Secondary Education Establishments
Links
Total cost
No data

Participants (5)