EP 76. Quality control of high resolution T1 images in the global repository of the neuroimaging society in amyotrophic lateral sclerosis (NiSALS)

Introduction Pathology in ALS spreads within the brain along tracts which are connected to the primary motor cortex, eventually causing atrophy in several cortical areas. Grey and white matter signals in high resolution T1 MRI can be analyzed using i.e. cortical thickness analyses, voxel-based morph...

Full description

Saved in:

Bibliographic Details
Published in:	Clinical neurophysiology Vol. 127; no. 9; pp. e270 - e271
Main Authors:	Grosskreutz, J, Dahnke, R, Gaser, C, Prell, T, Agosta, F, Bede, P, Benatar, M, de Carvalho, M, Kalra, S, Kassubek, J, Reischauer, C, Turner, M, van Damme, P, van den Berg, L, Weber, M, Filipi, M, Control Group, N. Quality
Format:	Journal Article
Language:	English
Published:	Elsevier B.V 01-09-2016
Subjects:	Neurology
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Introduction Pathology in ALS spreads within the brain along tracts which are connected to the primary motor cortex, eventually causing atrophy in several cortical areas. Grey and white matter signals in high resolution T1 MRI can be analyzed using i.e. cortical thickness analyses, voxel-based morphometry (VBM), voxel-based intensitometry (VBI), deformation based mapping and other T1 tools. These tools are capable of detecting ALS related pathology in groups of patients. In order to utilize these capabilities as a monitoring and exploration tool in large multicenter cohorts, the quality of T1 datasets must be comparable between different scanners and centers. Causes of low quality scans must be identified and corrected for, by changing acquisition, by using software correction, or both. Objective (i) to quantify the systematic differences in high resolution T1 scans in a large cross sectional data set uploaded to the central MRI repository of the Neuroimaging Society in ALS (NiSALS); (ii) to identify the major components contributing to the quality of the scans; (iii) to correct for scanner specific systematic distortions and (iv) define a set of quality markers to pass to allow pooling of data for large multicenter trials. Methods T1 data sets ( n = 500) were uploaded from 20 centers in Europe, the US and Canada. Data sets were analyzed using preprocessing from the VBM12 package within SPM on the Matlab platform. A custom developed set of quality markers (Dahnke&Gaser) was used to describe geometrical and noise distortions in a quantitative manner and correct for these distortions on a per scanner basis. Mahalanovis distance analyses were used to describe the relative quality of the scans within the NISALS cohort, and in comparison to other freely available cohorts. Results Data from 15 centers passed the minimum requirements for pooling using stringent criteria, data from three more centers were allowed to enter when correction measures were finely adjusted. The quality of the NISALS data set was comparable to the ADNI data set which imposed strict measures during data acquisition. Conclusion Quality control of T1 data sets in a centralized repository allows the pooling of large MRI data sets from many different centers, the quantification of scanner specific distortions, the software correction of some of these distortions, and the monitoring of the quality of T1 data as it would be used during a clinical trial. Acknowledgement This research is supported by BMBF (Bundesministerium für Bildung and Forschung) in the framework of the E-RARE programme (PYRAMID) and JPND(SOPHIA) of the European Union.
ISSN:	1388-2457 1872-8952
DOI:	10.1016/j.clinph.2016.05.126