Validating Automated Segmentation Tools in the Assessment of Caudate Atrophy in Huntington's Disease

Freesurfer Software Update Significantly Impacts Striatal Volumes in the Huntington's Disease Young Adult Study and Will Influence HD-ISS Staging

Background

The Huntington's Disease Integrated Staging System (HD-ISS) defined disease onset using volumetric cut-offs for caudate and putamen derived from FreeSurfer 6 (FS6). The impact of the latest software update (FS7) on volumes remains unknown. The Huntington's Disease Young Adult Study (HD-YAS) is appropriately positioned to explore differences in FS bias when detecting early atrophy.

Objective

Explore the relationships and differences between raw caudate and putamen volumes, calculated total intracranial volumes (cTICV), and adjusted caudate and putamen volumes, derived from FS6 and FS7, in HD-YAS.

Methods

Images from 123 participants were segmented and quality controlled. Relationships and differences between volumes were explored using intraclass correlation (ICC) and Bland-Altman analysis.

Results

Across the whole cohort, ICC for raw caudate and putamen was 0.99, cTICV 0.93, adjusted caudate 0.87, and adjusted putamen 0.86 (all p < 0.0005). Compared to FS6, FS7 calculated: i) larger raw caudate (+0.8%, p < 0.00005) and putamen (+1.9%, p < 0.00005), with greater difference for larger volumes; and ii) smaller cTICV (-5.1%, p < 0.00005), with greater difference for smaller volumes. The systematic and proportional difference in cTICV was greater than raw volumes. When raw volumes were adjusted for cTICV, these effects compounded (adjusted caudate +7.0%, p < 0.00005; adjusted putamen +8.2%, p < 0.00005), with greater difference for larger volumes.

Conclusions

As new software is released, it is critical that biases are explored since differences have the potential to significantly alter the findings of HD trials. Until conversion factors are defined, the HD-ISS must be applied using FS6. This should be incorporated into the HD-ISS online calculator.

Striatal and thalamic automatic segmentation, morphology, and clinical correlates in Parkinsonism: Parkinson's disease, multiple system atrophy and progressive supranuclear palsy

Parkinson's disease (PD), multisystem atrophy (MSA), and progressive supranuclear palsy (PSP) present similarly with bradykinesia, tremor, rigidity, and cognitive impairments. Neuroimaging studies have found differential changes in the nigrostriatal pathway in these disorders, however whether the volume and shape of specific regions within this pathway can distinguish between atypical Parkinsonian disorders remains to be determined. This paper investigates striatal and thalamic volume and morphology as distinguishing biomarkers, and their relationship to neuropsychiatric symptoms. Automatic segmentation to calculate volume and shape analysis of the caudate nucleus, putamen, and thalamus were performed in 18 PD patients, 12 MSA, 15 PSP, and 20 healthy controls, then correlated with clinical measures. PSP bilateral thalami and right putamen were significantly smaller than controls, but not MSA or PD. The left caudate and putamen significantly correlated with the Neuropsychiatric Inventory total score. Bilateral thalamus, caudate, and left putamen had significantly different morphology between groups, driven by differences between PSP and healthy controls. This study demonstrated that PSP patient striatal and thalamic volume and shape are significantly different when compared with controls. Parkinsonian disorders could not be differentiated on volumetry or morphology, however there are trends for volumetric and morphological changes associated with PD, MSA, and PSP.

Convolutional Neural Networks Enable Robust Automatic Segmentation of the Rat Hippocampus in MRI After Traumatic Brain Injury

Registration-based methods are commonly used in the automatic segmentation of magnetic resonance (MR) brain images. However, these methods are not robust to the presence of gross pathologies that can alter the brain anatomy and affect the alignment of the atlas image with the target image. In this work, we develop a robust algorithm, MU-Net-R, for automatic segmentation of the normal and injured rat hippocampus based on an ensemble of U-net-like Convolutional Neural Networks (CNNs). MU-Net-R was trained on manually segmented MR images of sham-operated rats and rats with traumatic brain injury (TBI) by lateral fluid percussion. The performance of MU-Net-R was quantitatively compared with methods based on single and multi-atlas registration using MR images from two large preclinical cohorts. Automatic segmentations using MU-Net-R and multi-atlas registration were of excellent quality, achieving cross-validated Dice scores above 0.90 despite the presence of brain lesions, atrophy, and ventricular enlargement. In contrast, the performance of single-atlas segmentation was unsatisfactory (cross-validated Dice scores below 0.85). Interestingly, the registration-based methods were better at segmenting the contralateral than the ipsilateral hippocampus, whereas MU-Net-R segmented the contralateral and ipsilateral hippocampus equally well. We assessed the progression of hippocampal damage after TBI by using our automatic segmentation tool. Our data show that the presence of TBI, time after TBI, and whether the hippocampus was ipsilateral or contralateral to the injury were the parameters that explained hippocampal volume.