PhD position in pathology foundation models for invasive lobular breast cancer

The Laboratory for Translational Cell and Tissue Research is embedded within the Department of Imaging and Pathology at KU Leuven and is closely affiliated with University Hospitals Leuven. It brings together research-active pathologists, biomedical scientists and computational researchers working at the interface of tissue biology, digital pathology and artificial intelligence. The research teams study human tissue samples using shared histopathology infrastructure spanning biobanking and tissue processing, advanced microscopy, digital pathology and quantitative image analysis. This interdisciplinary environment provides a strong foundation for translating computational methods into clinically relevant applications.

More information: https://www.kuleuven.be/wieiswie/en/unit/regional/50000686

Project

Invasive lobular carcinoma (ILC) is the second most common histological type of breast cancer. However, patients with ILC remain under-represented in public datasets and in the development of artificial-intelligence models. Its subtle infiltrative growth pattern also creates specific challenges for diagnosis, staging and clinical management.

This PhD project is part of M4GIC-ILC, a 36-month EIC Pathfinder project coordinated by KU Leuven. M4GIC-ILC brings together pathology, radiology, molecular and clinical data from more than 6,600 patients to develop clinically trustworthy AI methods tailored to ILC.

The PhD researcher will fine-tune and benchmark pathology foundation models using multi-site H&E and immunohistochemistry whole-slide images. The aim is to learn representations that capture ILC-specific morphology and biology while remaining robust to differences between hospitals, scanners, staining procedures and protocols. The researcher will investigate self-supervised and parameter-efficient fine-tuning, compare complementary foundation models and evaluate their representations through retrieval, clustering, classification and external validation.

The resulting models will be applied to clinically relevant downstream tasks, including virtual generation of diagnostic immunohistochemical stains from H&E, ILC diagnosis and subtype classification, staging, prognosis, relapse prediction and treatment-response modelling. The researcher will collaborate closely with computational scientists, breast pathologists, radiologists and molecular researchers across the international M4GIC-ILC consortium and will contribute reusable software, model documentation and scientific publications.

Profile

We are looking for a motivated and curious researcher with a strong computational background and an interest in applying artificial intelligence to clinically relevant biomedical questions.

You hold a master's degree in computer science, artificial intelligence, bioinformatics, biomedical engineering, computational biology, statistics or a closely related field. You have strong programming skills, preferably in Python, and experience with machine learning or deep learning.

Experience in computer vision, digital pathology, whole-slide image analysis, self-supervised learning, foundation models, multiple-instance learning or high-performance computing is an advantage. Prior knowledge of pathology or breast cancer is welcome but not required.

You are able to work independently while contributing effectively to an interdisciplinary and international team. You are analytically minded, scientifically curious and committed to producing reproducible, well-documented research. You communicate clearly in English, both orally and in writing.

You meet the admission requirements for the KU Leuven doctoral programme or will meet them by the starting date.

Offer

We offer
a full-time PhD position for three years, with an initial probationary period of one year. The preferred starting date is 1 October 2026, but this can be adjusted where necessary to accommodate visa procedures or other practical constraints.

You will receive doctoral training at KU Leuven and develop expertise in digital pathology, foundation-model adaptation, computer vision and clinically oriented artificial intelligence. You will have access to unique multi-site breast cancer datasets and the computational infrastructure required to analyse large collections of whole-slide images.

You will be supervised by Prof. Asier Antoranz, co-supervised by Prof. Giuseppe Floris, and work closely with experts in computational pathology, breast pathology, radiology and biomedical AI. You will become part of an international and interdisciplinary research environment and collaborate with clinical and technical partners in the M4GIC-ILC consortium.

The position provides opportunities to present your research at international conferences, publish in peer-reviewed scientific journals and develop reusable computational tools with potential clinical impact. Your