Hippocampal segmentation for brains with extensive atrophy using three‐dimensional convolutional neural networks

Abstract

Hippocampal volumetry is a critical biomarker of aging and dementia, and it is widely used as a predictor of cognitive performance; however, automated hippocampal segmentation methods are limited because the algorithms are (a) not publicly available, (b) subject to error with significant brain atrophy, cerebrovascular disease and lesions, and/or (c) computationally expensive or require parameter tuning. In this study, we trained a 3D convolutional neural network using 259 bilateral manually delineated segmentations collected from three studies, acquired at multiple sites on different scanners with variable protocols. Our training dataset consisted of elderly cases difficult to segment due to extensive atrophy, vascular disease, and lesions. Our algorithm, (HippMapp3r), was validated against four other publicly available state-of-the-art techniques (HippoDeep, FreeSurfer, SBHV, volBrain, and FIRST). HippMapp3r outperformed the other techniques on all three metrics, generating an average Dice of 0.89 and a correlation coefficient of 0.95. It was two orders of magnitude faster than some of the tested techniques. Further validation was performed on 200 subjects from two other disease populations (frontotemporal dementia and vascular cognitive impairment), highlighting our method's low outlier rate. We finally tested the methods on real and simulated "clinical adversarial" cases to study their robustness to corrupt, low-quality scans. The pipeline and models are available at: https://hippmapp3r.readthedocs.ioto facilitate the study of the hippocampus in large multisite studies.

Whole Brain and Cranial Size Adjustments in Volumetric Brain Analyses of Sex- and Age-Related Trends

“…Any hardware instabilities may lead to geometric distortions in segmentation of the brain structures (Skorupa et al, 2014;Guadalupe et al, 2017). The accuracy and reproducibility of the current automatic brain segmentation algorithms have been widely tested (Pardoe et al, 2009;Nugent et al, 2013;Velasco-Annis et al, 2018;Goubran et al, 2020). Different imaging protocols, scanner brands and models, subject positioning in the MR scanner, image artifacts, and partial volume averaging were found to reduce the reproducibility of the segmentation methods (Clark et al, 2006;Han et al, 2006;Gronenschild et al, 2012;Velasco-Annis et al, 2018).…”

A Large-scale Comparison of Cortical and Subcortical Structural Segmentation Methods in Alzheimer’s Disease: a Statistical Approach

“…Our results showed strong reliability of HIPS, volBrain and CAT. These methods have been successfully applied to brain images from those with AD [76][77][78][79] .…”

Improved segmentation of the intracranial and ventricular volumes in populations with cerebrovascular lesions and atrophy using 3D CNNs

“…The architecture of the network (Figure 1) was based on the original 2D U-net (Çiçek et al 2016;Ronneberger, Fischer, and Brox 2015), with some modifications in our 3D implementation. Similar to our prior work (Goubran et al 2019), residual blocks were added to each encoding layer. Residual blocks resolve the gradient degradation problem that occurs with deeper networks with an increasing number of layers.…”

“…The preprocessing techniques used in this study are described in our prior work (Goubran et al 2019). Briefly, all images used for training were bias-field corrected using N4 (Tustison et al 2010).…”