Fixation Duration Effects on Microstructural Changes in Disease: Insight from Ex-vivo Diffusion MRI

291 

Abstract Submission 

Zaki Alasmar^1,2, Roqaie Moqadam^2,3, Liana Sanches², Yashar Zeighami^2,4, Mahsa Dadar^2,4

¹Integrated Program in Neuroscience, McGill University, Montreal, Quebec, Canada, ²Cerebral Imaging Centre, Douglas Mental Health University Institute, Verdun, Montreal, Quebec, Canada, ³Faculty of Medicine - Department of Medicine, University of Montreal, Montreal, Quebec, Canada, ⁴Department of Psychiatry, McGill University, Montreal, Quebec, Canada

Zaki Alasmar  
Integrated Program in Neuroscience, McGill University|Cerebral Imaging Centre, Douglas Mental Health University Institute
Montreal, Quebec, Canada|Verdun, Montreal, Quebec, Canada

Roqaie Moqadam  
Cerebral Imaging Centre, Douglas Mental Health University Institute|Faculty of Medicine - Department of Medicine, University of Montreal
Verdun, Montreal, Quebec, Canada|Montreal, Quebec, Canada

Liana Sanches, PhD  
Cerebral Imaging Centre, Douglas Mental Health University Institute
Verdun, Montreal, Quebec, Canada

Yashar Zeighami  
Cerebral Imaging Centre, Douglas Mental Health University Institute|Department of Psychiatry, McGill University
Verdun, Montreal, Quebec, Canada|Montreal, Quebec, Canada

Mahsa Dadar  
Cerebral Imaging Centre, Douglas Mental Health University Institute|Department of Psychiatry, McGill University
Verdun, Montreal, Quebec, Canada|Montreal, Quebec, Canada

Diffusion MRI (dMRI) is a common tool to assess the brain's microstructure (1). Through dMRI, we can extract metrics assessing fibre integrity in various neurological conditions (e.g. Alzheimer's and Parkinson's Disease [AD & PD]). The Douglas-Bell Canada Brain Bank (DBCBB) is the largest Canadian repository of donated human brains from various sources including multiple neurodegenerative conditions. However, the fixation of these brains changes the tissue composition and their magnetic properties, complicating post-mortem inference of pathology and preventing ante-/post-mortem comparative mapping of the same microstructural characteristics (2). We aimed to validate in-vivo diffusion measures when used ex-vivo by assessing the impact of sample fixation on these measures, in AD and PD, and then qualitatively compare their distribution patterns between ante- and post-mortem.

We acquired 3T structural MRI (T1w & T2w) and dMRI (32 directions, 2 phase-encoding directions, b-value = 1000 s/mm²) of 45 donated specimens (mean age at death = 81 years, mean fixation duration =13.8 years, 20 females). We used BISON (3) to segment cortical grey and white matter and subcortical structures on the structural scans. We adapted Mrtrix3 preprocessing tools to denoise, correct eddy-currents, and bias correct (i.e., inhomogeneity correction) the dMRI (4). From the preprocessed dMRI, we computed the tensor model and extracted its metrics (fractional anisotropy [FA], mean [MD], axial [AxD] diffusivity). We first modelled the change in dMRI microstructural metrics using linear regression models, including age at death and sex as covariates. We then qualitatively compared the microstructural patterns in a subset of participants that also had in-vivo imaging. We performed the statistical analysis in the whole sample, and in AD, PD and other neuropsychiatric conditions separately (including Amyotrophic lateral sclerosis (ALS), mood and vascular disorders, and other dementias).

We found differential patterns of associations between DTI metrics and fixation time, that differed between brain regions and across diagnoses (all diagnoses: f(FA-GM)=3.73, R2=0.38, p-val=0.03, f(FA-WM)=2.94, R2=0.33, p-val=0.03; AD: f(FA-GM)=5.56, R2=0.217, p-val=0.028, PD: f(MD-subcortex)=7.7, R2=0.56, p-val<0.001, f(AxD-Total)=11.4, R2=0.65, p-val<0.001; fig.1). This effect was consistent in AD and PD specimens. Taken together, these results reflect a strongly altered microstructural environment in fixed brains, that differed between AD, PD, and other neuropsychiatric conditions, and between directional and other diffusion metrics. Based on this, we performed in-vivo and ex-vivo diffusion tensor computation in one AD subject with ante- and post-mortem scans to qualitatively test the distribution patterns of the microstructural metrics. We were able to identify several regions of the prefrontal lobe that show altered microstructure (fig.2). Overall, the associations between in-vivo and ex-vivo metrics were r(FA)=0.49, r(MD)=0.42, r(AxD)=0.43. The difference observed could be due to the localization of neuropathology that accelerated the effect of fixation on the diffusion metrics. This alteration was observed to be lower for directional FA rather than MD and RD.

We show a strong association between brain fixation duration and the neural microstructure across several neurological diseases using directional and other diffusion measures. We demonstrated that ex-vivo diffusion scans can serve as extensions to in-vivo protocols, which could help unveil disease signature maps and neuropathology. This could enrich previous ex-vivo studies by supplementing invasive histological analyses with non-invasive neuroimaging that could then be used to compare pathological states across disease stages. Future studies should investigate these effects in other neuropsychiatric conditions such as ALS and mood disorders, and examine the effect of vasculature change in these conditions.

Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) ¹

Diffusion MRI Modeling and Analysis ²

Univariate Modeling

Diffusion MRI

Multi-Modal Imaging

Acquisition

Degenerative Disease

Modeling

MRI

Psychiatric Disorders

STRUCTURAL MRI

WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC