Description du projet
IRM moléculaire pour l’imagerie des tumeurs cérébrales chez l’enfant
L’imagerie par résonance magnétique (IRM) est la modalité de référence pour le diagnostic et le suivi des tumeurs cérébrales. Cependant, les données IRM standard sont qualitatives, manquent de précision et ne permettent pas le suivi des réponses au traitement. Financé par le Conseil européen de la recherche, le projet BabyMagnet entend résoudre ce problème par le biais d’une technologie d’IRM moléculaire qui permet un suivi rapide du traitement du cancer du cerveau chez l’enfant sans recourir à des agents de contraste. La technologie, assistée par l’IA, s’appuie sur les changements de pH et de concentration de protéines dans le cerveau en tant que biomarqueurs du cancer. Cette recherche pourrait révolutionner l’imagerie du cancer en automatisant l’optimisation des protocoles d’IRM, en accélérant les examens et en proposant une médecine de précision pour le traitement des cancers pédiatriques.
Objectif
Despite vast drug development efforts, brain tumors remain the leading cause of pediatric cancer deaths. Noninvasive monitoring of treatment response is crucial to reveal the mechanisms behind tumor-drug interactions and optimize patient care. However, standard magnetic resonance imaging (MRI) methods involve injecting metals, have severe difficulties in differentiating treatment response from tumor progression, are qualitative, and mandate prolonged anesthesia due to the lengthy acquisition. I propose to develop a transformative molecular MRI technology, based on the chemical exchange saturation transfer (CEST) contrast mechanism that will enable specific, quantitative, rapid, contrast-material free, treatment monitoring of pediatric brain cancer. Recently I revealed that a combination of mathematical CEST models and AI can generate quantitative biomarker maps of pH and protein concentration changes across the brain, two known hallmarks of cancer. Inspired by these results, I now propose to adopt a previously unconsidered perspective and to represent the underlying physics of CEST MRI as a computational graph, enabling an automatic AI-based optimization of molecular imaging. I hypothesize that the combination of biophysical models with a new AI framework, and their synergetic integration throughout the entire imaging pipeline will provide accurate noninvasive treatment monitoring. First, I will establish a method for automated optimization of MRI protocols for early determination of the tumor response to mainstream chemotherapy. Next, I will shorten the 3D scan time by an order of magnitude and quantify the response to next generation immunotherapy. Third, I will translate the method to clinical scanners and validate it in a human pediatric pilot study. This research will yield a fundamental understanding of the molecular mechanisms underlying treatment response and establish an innovative precision medicine methodology that will transform pediatric cancer imaging.
Champ scientifique
- medical and health sciencesbasic medicinepharmacology and pharmacydrug discovery
- medical and health sciencesclinical medicineoncology
- medical and health scienceshealth sciencespersonalized medicine
- engineering and technologymedical engineeringdiagnostic imagingmagnetic resonance imaging
- medical and health sciencesbasic medicineimmunologyimmunotherapy
Mots‑clés
Programme(s)
- HORIZON.1.1 - European Research Council (ERC) Main Programme
Thème(s)
Régime de financement
HORIZON-ERC - HORIZON ERC GrantsInstitution d’accueil
69978 Tel Aviv
Israël