Advanced fetal brain MRI processing and analysis: Methods and Applications

Presenter: Kiho Im, PhD
Abstract:
In vivo fetal MRI can improve the diagnostic accuracy for fetal brain abnormalities, leading to better prenatal counseling and clinical management. However, fetal brain structural changes are often too subtle to be detected by qualitative visual MRI inspection. Our research goal is to develop advanced fetal MRI processing and analysis techniques to provide quantitative and biologically relevant imaging biomarkers that help us to better understand prenatal brain development and aid in the detection of disease. We focus on developing a fully automatic pipeline for fetal brain MRI processing including automatic fetal brain segmentation; MRI quality assessment; motion correction; brain tissue segmentation using state-of-the-art medical image processing and deep learning techniques. We have also proposed several advanced methodologies for quantifying and investigating local and regional cortical surface growth; sulcal pits and patterns development; biological fetal brain age. We have applied our technologies to typically developing fetuses and fetuses with cortical malformations, agenesis of corpus callosum, congenital heart disease, Down syndrome, and ventriculomegaly. Our prior studies have successfully shown the potential to identify individuals with abnormal brain structure early in utero and predict their postnatal clinical outcomes, which would allow better fetal care and future fetal interventions.

Biography:
Dr. Kiho Im received his Ph.D. in Biomedical Engineering in 2009 from Hanyang University, Korea, and began his postdoctoral training as a Research Fellow in the FNNDSC, BCH in 2010. He was promoted to Instructor in 2012 and he has been Assistant professor of Pediatrics at Harvard Medical School, BCH since 2016.

He has expertise in quantitative neuroimage analysis using structural and diffusion MRI data. His research goal is to provide unique and biologically relevant imaging biomarkers that not only help us to better understand normal and abnormal brain development, but also aid in the detection and diagnosis of disease. In particular, his team focuses on quantitative analysis of sulcal pits and patterns; gyral based structural brain connectivity/network analysis; genetic and environmental effects on brain development; and advanced fetal brain MRI processing and analysis using deep learning.

Dr. Im has demonstrated an exceptionally high level of productivity for a young scientist in the field of neuroimaging. He has published 58 peer-reviewed research articles in the high-level neuroimaging/neuroscience journals such as Neuroimage, Cerebral Cortex, and Neurology. His scientific articles have been cited on almost 2,700 occasions since 2007 (h-index: 26). Due to the biological relevance of his analyses, he has collaborated with world-renowned neuroscientists and geneticists such as Drs. Christopher Walsh and Elizabeth Engle at Harvard Medical School, providing a more accurate unbiased characterization of brain folding and connectivity. These collaborations have resulted in co-authored papers in Science (2014), Nature (2018), and Neuron (2018), where Dr. Im’s contribution is his unique image analysis. Based on his distinguished expertise, Dr. Im received Faculty Career Development Fellowship funded by Boston Children’s Hospital and two NIH R21 grants. As a PI on NIH R01 and American Heart Association grants, he is currently leading an innovative project that examines genetic and hemodynamic effects on fetal cortical development and develops a predictive model of fetal to neonatal brain development using deep learning in congenital heart disease.