AI for Longitudinal Medical Data: From Trajectories to Insights

Led by the EPFL AI Center, SDSC, HES-SO Wallis and in collaboration with CHUV, HUG, USZ and PHRT, this mini-symposium aims to foster cross-disciplinary exchanges within the local Swiss ecosystem and to highlight emerging opportunities for AI to inform personalized, time-resolved medicine.

Understanding health as a dynamic process requires models that capture how biological, clinical, and behavioral variables evolve over time. This mini-symposium brings together researchers at the intersection of artificial intelligence, biomedical data science, medical imaging, and clinical research to discuss recent advances in modeling longitudinal medical data, with a strong focus on longitudinal imaging such as PET, CT, and MRI.

The workshop will take place exclusively on-site on November 20, 2025, at the EPFL AI Center.

Program

13:30 – 14:00 – Check-in
14:00 – 14:10 – Welcome
Dr. Dorina Thanou, EPFL
Prof. Adrien Depeursinge, HES-SO, CHUV
14:10 – 14:30 – “AI in Longitudinal Molecular Imaging: Bridging Clinical Needs and Quantitative Insights” – Mario Jreige, CHUV
14:30 – 14:50 – “Learning tumor evolution dynamics and early markers of immunotherapy response from PET/CT imaging” – Jeremy Baffou, EPFL
14:50 – 15:10 – “Segmentation and Tracking for Melanoma Lesions from PET/CT imaging with Promptable Foundation Models” – Xiaoran Chen, SDSC
15:10 – 15:40 – Coffee break
15:40 – 16:00 – “Introduction of the World’s Largest Temporal PET/CT Oligometastatic Disease Dataset and Required Tools”- Maksym Fritsak, USZ
16:00 – 16:20 – “The SwissHeart Study – From Spins to Pictures to Digital Organs” – Sebastian Kozerke, ETH Zurich
16:20 – 16:40 – ”Learning Disease Trajectories: AI for Personalized Management of Brain Metastases” – Vincent Andrearczyk, HES-SO Valais-Wallis
16:40 – 17:00 – “Towards Longitudinal Characterization of Multiple Sclerosis” – Pedro Gordaliza, CIBM SP CHUV-UNIL
17:00 – 19:00 – Networking apéro

Topics covered

Temporal representation learning,
Predictive modeling of disease trajectories,
Segmentation and lesion tracking,
Cross-modality harmonization,
Translation of AI-derived insights into clinical practice.

Registration

This mini-symposium aims to foster cross-disciplinary exchanges within the local Swiss ecosystem and to highlight emerging opportunities for AI to inform personalized, time-resolved medicine.

Due to limited capacity, registration is mandatory.

This mini-symposium is organized in the context of the LETITIA project (supported by PHRT and SDSC).

Organisers

Dr. Dorina Thanou (LTS4 Lab, EPFL AI Center)
Prof. Adrien Depeursinge (HES-SO, CHUV)

Venue

This workshop is hosted by the EPFL AI Center.

EPFL Campus, ELE 117
Lausanne, Switzerland.

Contact

For enquiries, please contact: epflai@epfl.ch.