A PhD position is available within the TwinCaRT project, funded by the PEPR Digital Health program, focusing on the development of generative AI models for medical imaging and personalized radiotherapy.
Education
- Master’s degree or engineering degree (MSc or equivalent) in Artificial Intelligence, Computer Vision, Machine Learning, Biomedical Engineering, or Applied Mathematics.
Skills
-
Strong knowledge of artificial intelligence and machine learning
-
Experience in image processing or 3D medical imaging
-
Proficiency in scientific programming, particularly Python
Personal qualities
-
Autonomy and initiative
-
Open-mindedness and motivation
-
Ability to work collaboratively
-
Good command of scientific English
The PhD will be conducted at LaTIM (UMR 1101) within the ACTION research team, under the supervision of:
-
Julien Bert, Senior Research Scientist
-
Nassib Abdallah, Research Scientist
The project includes close collaborations with the LTSI laboratory, particularly with Prof. Oscar Acosta for the tumor evolution model, as well as with the Centre Eugène Marquis for access to clinical data.
This PhD position is part of the TwinCaRT project, funded by the PEPR Digital Health program.
The expected start date is Fall 2026, for a duration of three years.
Scientific Context
External radiotherapy is currently one of the most widely used treatments for localized cancers, particularly prostate, head and neck, and cervical cancers. Despite its effectiveness, a significant proportion of patients still experience tumor recurrence, which may be related to tumor location, patient-specific biological characteristics, or treatment parameters.
Recent advances in MRI-guided radiotherapy (MR-Linac) make it possible to adapt treatment in real time to the patient’s anatomy. However, determining the optimal therapeutic strategy for each patient remains complex.
In this context, digital twins and artificial intelligence open new perspectives for modeling tumor evolution and predicting treatment response.
The TwinCaRT project aims to develop a digital twin of cancer patients capable of simulating tumor evolution and response to radiotherapy in order to optimize treatment protocols. Prostate cancer has been selected as the initial application, although the methods developed may be transferred to other pathologies.
PhD Missions
The PhD project will focus on the development of multimodal AI models capable of generating synthetic medical images within a digital twin framework.
The objective is to design and evaluate models that can transform multiparametric MRI, PET, and CT imaging data into synthetic images consistent with simulated tumor evolution.
These models will enable:
-
the generation of pseudo-PET and pseudo-CT images (pPET, pCT) from MRI data
-
the modeling of tumor evolution during treatment
-
the integration of these images into a patient digital twin to simulate different radiotherapy protocols
Scientific Objectives
The PhD research will focus on:
-
developing conditional generative models (e.g., diffusion models) to synthesize future medical imaging volumes
-
learning multimodal representations combining MRI, PET, and CT data
-
exploring AI approaches integrating biological or physical knowledge of tumor evolution
-
developing image-to-image translation models (e.g., generating CT or PET images from MRI)
The models will be applied to pelvic imaging in prostate cancer and integrated into a digital twin–based radiotherapy simulation platform.
To apply, please send your CV to:
Julien Bert — julien.bert@univ-brest.fr
Nassib Abdallah — nassib.abdallah@univ-brest.fr