MASiVar: Multisite, Multiscanner, and Multisubject Acquisitions for Studying Variability in Diffusion Weighted Magnetic Resonance Imaging

Abstract

Diffusion weighted imaging (DWI) allows investigators to identify microstructural differences between subjects, but variability due to session and scanner biases is still a challenge. To investigate DWI variability, we present MASiVar, a multisite dataset consisting of 319 diffusion scans acquired at 3T from b = 1000 to 3000 s/mm2 across 97 different healthy subjects and four different scanners as a publicly available, preprocessed, and de-identified dataset. With these data we characterize variability on the intrasession intrascanner (N = 158), intersession intrascanner (N = 328), intersession interscanner (N = 53), and intersubject intrascanner (N = 80) levels. Our baseline analysis focuses on four common DWI processing approaches: (1) a tensor signal representation, (2) a multi-compartment neurite orientation dispersion and density model, (3) white matter bundle segmentation, and (4) structural connectomics. Respectively, we evaluate region-wise fractional anisotropy (FA), mean diffusivity, and principal eigenvector; region-wise cerebral spinal fluid volume fraction, intracellular volume fraction, and orientation dispersion index; bundle-wise shape, volume, length and FA; and connectome correlation and maximized modularity, global efficiency, and characteristic path length. We plot the scan/re-scan discrepancies in these measures at each level and find that variability generally increases with intrasession to intersession to interscanner to intersubject effects and that sometimes interscanner variability can approach intersubject variability. This baseline study suggests harmonization between scanners for multisite analyses is critical prior to inference of group differences on subjects and demonstrates the potential of MASiVar to investigate DWI variability across multiple levels and processing approaches simultaneously.

Fiber tractography bundle segmentation depends on scanner effects, vendor effects, acquisition resolution, diffusion sampling scheme, diffusion sensitization, and bundle segmentation workflow

“…Here we utilize two open-sourced multi-subject, multi-scanner, and multi-protocol benchmark databases: the MASiVar [59] and MUSHAC datasets [40, 41]. We note that other multi-site databases exist (see Discussion), although they are often limited to investigating differences across subjects and scanners, whereas the two chosen datasets together allow investigation of repeats, scanners, vendors, and acquisition protocols (resolution, direction, b-values).…”

“…The MASiVar database will allow investigation of scan-rescan and cross-scanner effects. Here we used a subset of Cohort II of this database described in [59], which consisted of 5 healthy subjects with 6 unique “datasets”. Each subject was scanned on four scanners: a) 3T Philips Achieva (80 mT/m) and b) a different 3T Philips Achieva (60mT/m) at the same site, c) a 3T General Electric Discovery MR750 Scanner at a different site, and d) a 3T Siemens Skyra scanner at a different site.…”

PreQual: An automated pipeline for integrated preprocessing and quality assurance of diffusion weighted MRI images

“…The first dataset, Dataset A, was acquired on one subject scanned repeatedly on the same scanner over the course of three sessions, one session each consecutive day. 59 Each session consisted of 6 sets of scans. Each set of scans consisted of one 3-direction b = 1000 s/mm 2 image phase encoded in the anterior to posterior direction (APA), one 96-direction b = 1000 s/mm 2 image phase encoded in the posterior to anterior direction (APP), one 96-direction b = 1500 s/mm 2 APP image, one 96-direction b = 2000 s/mm 2 APP image, one 96-direction b = 2500 s/mm 2 APP image, and one 96direction b = 3000 s/mm 2 APP image.…”

BAI-Net: Individualized Anatomical Cerebral Cartography using Graph Neural Network

“…An additional dataset from the Multisite, Multiscanner, and Multisubject Acquisitions for Studying Variability (MASiVar) dataset [47], consisting of 74 scans (8 scans were removed due to incomplete brain tissue data in the diffusion images) from 14 adult subjects (8 males and 6 females, age , in order to evaluate the impacts of multiple sessions, sites, and scanners. The dMRIs of these subjects were acquired from 3 cohorts using four scanners (two 3T Philips Achieva scanners used at two sites, one 3T General Electric Discovery MR750 scanner, and one 3T Siemens Skyra scanner) with at least one T1weighted image for each subject at each session.…”