Kwantitatieve hersenmetingen op pediatrische MRI: Een vroege biomarker voor neuro-ontwikkelingsstoornissen

Medical devices for brain MRI quantification provide objective measurements of brain structures and lesion volumes, supporting the diagnosis and monitoring of neurological disorders. However, their clinical application in pediatric populations, particularly during the transition to adulthood, remains limited. In conditions such as cerebral palsy, lesion assessment often relies on time-intensive visual inspection. This PhD thesis presents the development and validation of deep learning-based models for brain structure quantification across age groups, with a focus on pediatric populations. The models were extensively evaluated for accuracy, reproducibility, and clinical applicability. The research was conducted in collaboration with KU Leuven and icometrix within the European H2020 MSCA-ITN-2020 PARENT project (grant 956394). The thesis includes validation of the FDA-approved and CE-cleared device icobrain on pediatric data, extending its application to children with epilepsy. Additionally, the results support the feasibility of automated brain MRI quantification in pediatric clinical settings, including for patients with cerebral palsy.