Abstract
Motivation: Brain morphometry from magnetic resonance imaging (MRI) is a promising neuroimaging biomarker for the non-invasive diagnosis and monitoring of neurodegenerative and neurological disorders. Current tools for brain morphometry often come with a high computational burden, making them hard to use in clinical routine, where time is often an issue. We propose a deep learning-based approach to predict the volumes of anatomically delineated subcortical regions of interest (ROI), and mean thicknesses and curvatures of cortical parcellations directly from T1-weighted MRI. Advantages are the timely availability of results while maintaining a clinically relevant accuracy.Materials and Methods: An anonymized dataset of 574 subjects (443 healthy controls and 131 patients with epilepsy) was used for the supervised training of a convolutional neural network (CNN). A silver-standard ground truth was generated with FreeSurfer 6.0.Results: The CNN predicts a total of 165 morphometric measures directly from raw MR images. Analysis of the results using intraclass correlation coefficients showed, in general, good correlation with FreeSurfer generated ground truth data, with some of the regions nearly reaching human inter-rater performance (ICC > 0.75). Cortical thicknesses predicted by the CNN showed cross-sectional annual age-related gray matter atrophy rates both globally (thickness change of −0.004 mm/year) and regionally in agreement with the literature. A statistical test to dichotomize patients with epilepsy from healthy controls revealed similar effect sizes for structures affecting all subtypes as reported in a large-scale epilepsy study.Conclusions: We demonstrate the general feasibility of using deep learning to estimate human brain morphometry directly from T1-weighted MRI within seconds. A comparison of the results to other publications shows accuracies of comparable magnitudes for the subcortical volumes and cortical thicknesses.
Highlights
Magnetic resonance imaging (MRI) is the method of choice for non-invasive assessments of brain structure
Images were acquired in sagittal direction and MRI protocols were either MDEFT [49], standard 3D MP-RAGE [50], MPRAGE according to the recommendations of the Alzheimer’s Disease Neuroimaging Initiative [51] or MP-RAGE optimized for gray-white contrast [52]
The final model was trained during 7 days over 4,500 epochs, with the best mean R2 score on the validation set reached at epoch 3,920
Summary
Magnetic resonance imaging (MRI) is the method of choice for non-invasive assessments of brain structure. Clinicians use MRI for diagnosis, disease monitoring, and therapy control in a wide range of neurological and neurogenerative disorders like e.g., epilepsy, multiple sclerosis, Alzheimer’s, Parkinson’s, or Huntington’s disease, which are often associated with structural changes of the brain [1]. Structural MRI including high-resolution T1-weighted (T1w) imaging is part of today’s protocol recommendations for many of these disorders [2,3,4]. Beyond visual assessment by trained experts, quantitative brain morphometry is gaining increasingly more attention for medical applications. Precise and automatic reconstruction of structures from MRI is still a topic of active research. Used methods are voxel-based morphometry (VBM) [5] and surfacebased analysis (SBA) [6]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
