Harmonization of multimodal MRI for radiomics modeling of pediatric ependymomas
30 Juin 2022
Catégorie : Post-doctorant
This position is opened within the context of a multidisciplinary project funded by the French National Institute against cancer. The project associates numerous partners in France, both academic, private and clinical. The present position is for the Laboratory of Medical Information Processing (LaTIM), French Institute of Health and Medical Research (INSERM UMR 1101), Brest, France (https://nouveau.univ-brest.fr/latim/fr).
Context: Treatment of childhood ependymoma, the second most frequent pediatric brain tumor, is based on surgery and radiation therapy. However, 50% relapse, mainly locally. Progress in imaging, molecular biology and radiotherapy ballistics has led us to propose the EPENDYMOMICS project, a multi-omics approach using artificial intelligence to detect the predictive characteristics of relapse, and to define innovative radiotherapy targets using multimodal imaging. We previously reported that the relapse sites are mainly located in the high-dose radiotherapy zone and that there appear to be prognostic factors for relapse based on anatomical and functional MRI abnormalities by diffusion and perfusion. In addition, recent studies in molecular biology have identified significant prognostic factors. The challenge now is to use and integrate all these findings in larger cohorts to tackle the radio-resistance of this disease. Our objective is to collate in a single database called NETSPARE (Network to Structure and Share Pediatric data to Accelerate Research on Ependymoma) the clinical, histological, biological, imaging and radiotherapy data from two consecutive studies that included 370 children and adolescents with ependymoma since 2000 in France. The EPENDYMOMICS project comprises a clinical research team, three imaging research teams, two histopathology teams, and a biostatistics team working on NETSPARE. Our goal is to obtain a radiogenomic signature of our data, which will be validated with the English external cohort of 200 patients that is currently being analyzed. The perspective is to optimize the indications and volumes of irradiation that could in the future be used in a European translational research trial to tackle radioresistance.
The present post-doc position will be assigned to the LaTIM partner, for the specific task of harmonization of the highly multicentric imaging dataset (multimodal MRI are acquired and collected in numerous different clinical centers using different types, models and generations of scanners). This harmonization will be carried out both at the radiomic features level (i.e., statistical harmonization of features distributions) and the image level using different strategies: normalization, histogram matching, as well as image synthesis relying on deep learning techniques such as GANs, cGAN, one-short learning, etc.
Location: Laboratory of Medical Information Processing (LaTIM), French Institute of Health and Medical Research (INSERM UMR 1101), Brest, France (https://nouveau.univ-brest.fr/latim/fr)
Duration: 1 year, renewable for 10 more months (total 22 months), starting on September 2022
Salary: ~2500€ per month
Education: The candidate must hold a PhD in physics, computer science, applied mathematics or equivalent and have a substantial expertise and background in deep learning, machine learning and image analysis or computer vision, especially with image synthesis, medical image processing and analysis and an interest in adressing clinical challenges. Experience in radiomics is a plus but not mandatory.
Skills: Scientific programming, deep learning libraries and implementation
Languages: Good proficiency in English (both written and spoken), French optional
Application: Send before the 29th of July (at the latest) your CV, motivation letter and preferably a reference letter by e-mail to: firstname.lastname@example.org