MULTI-FACT study

There is a growing awareness of the importance of the early detection of abnormal neurodevelopment and of its major impact throughout life. Affecting around 10% of children, it can lead to neurological disabilities and their associated burdens. Two-thirds of instances of abnormal neurodevelopment may be detectable before birth, but at present the resulting neurobehavioral and cognitive disorders will in the majority of cases only be detected later in childhood, precluding early intervention, and increasing later-life impact. We therefore aim to describe abnormal cortical development in the early fetal stages, developing non-invasive MRI-derived biomarkers and fetal-specific computational tools, to predict those individuals at higher risk of abnormal post-natal development. Quantitative image analysis of the in vivo fetal brain plays a vital role in clinical decision-making and neuroscience research, and the advantages of in utero MRI over ultrasound in the study of brain development have been demonstrated. Since the use of fetal MRI is limited, multiple centres must work together to gather enough subjects, particularly for pathology, which raises significant technical challenges to harmonisation. We now have the expertise to exploit multi-centric images by developing federated learning strategies to apply AI solutions while preserving privacy. Approaches that adult MRI has long benefitted from will be adapted to fetal brain MRI studies by this project for the first time. Intrauterine growth restriction (IUGR), affecting 5%–10% of pregnancies, and corpus callosum agenesis (CCA), affecting 1 in 4000 pregnancies, will be our target to explore fetal cortical development and identify deviations through a joint analysis of an unprecedented large multi-centric dataset (>950 subjects) with dedicated computational tools. In this way, we will provide the global community with standardised, effective tools to transform the pre-natal diagnosis of abnormal neurodevelopment.