A pediatric simulated dosimetry platform for clinical use
Presentation of the problem and objective of the experiment
Personalized internal dosimetry is of great interest in children undergoing medical imaging examinations with ionizing irradiation. PediDose aims to exploit advanced Monte Carlo simulations, anthropomorphic computational phantoms, and Artificial Intelligence techniques to develop a realistic, simulated dosimetry database. The goal is to develop a novel software that will offer clinicians the possibility to assess internal pediatric dosimetry and optimize Nuclear Medical Imaging clinical protocols in terms of personalized prediction models.
Short description of the experiment
MC simulations in combination with anthropomorphic digital phantoms can provide accurate dosimetry estimations for clinical acquisitions (gold standard for dosimetry). Such simulations, with low statistical uncertainty, are highly intensive in terms of computational needs. A realistic simulated dosimetry database is going to be produced using a digital pediatric population. ML/DL algorithms will be trained on the simulated data to develop a prediction dosimetry model. The tools will be integrated in a novel software product, which could be exploited in clinical practice. PediDose brings together 2 innovative SMEs (BIOEMTECH & IKnowHow) experts in medical physics & bioinformatics and 1 HPC expert (GRNET) to efficiently implement the project.