TOMOGRAPHY, December 2016, Volume 2, Issue 4: 396-405
QIN DAWG Validation of Gradient Nonlinearity Bias Correction Workflow for Quantitative Diffusion-Weighted Imaging in Multicenter Trials
1Department of Radiology, University of Michigan, Ann Arbor, Michigan; 2Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, California; 3Vanderbilt University (VU) Institute of Imaging Science, VU Medical Center, Nashville, Tennessee; 4Russel H. Morgan Department of Radiology and Radiological Science, John Hopkins University School of Medicine, Baltimore, Maryland; 5Advanced Imaging Research Center, Oregon Health and Science University, Portland, Oregon; 6Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, Massachusetts; 7Translational and Molecular Imaging Institute, Icahn School of Medicine at Mt Sinai, New York, New York; and 8Department of Biomedical Engineering, University of Texas at Austin, Austin, Texas
Previous research has shown that system-dependent gradient nonlinearity (GNL) introduces a significant spatial bias (nonuniformity) in apparent diffusion coefficient (ADC) maps. Here, the feasibility of centralized retrospective system-specific correction of GNL bias for quantitative diffusion-weighted imaging (DWI) in multisite clinical trials is demonstrated across diverse scanners independent of the scanned object. Using corrector maps generated from system characterization by ice-water phantom measurement completed in the previous project phase, GNL bias correction was performed for test ADC measurements from an independent DWI phantom (room temperature agar) at two offset locations in the bore. The precomputed three-dimensional GNL correctors were retrospectively applied to test DWI scans by the central analysis site. The correction was blinded to reference DWI of the agar phantom at magnet isocenter where the GNL bias is negligible. The performance was evaluated from changes in ADC region of interest histogram statistics before and after correction with respect to the unbiased reference ADC values provided by sites. Both absolute error and nonuniformity of the ADC map induced by GNL (median, 12%; range, -35% to +10%) were substantially reduced by correction (7-fold in median and 3-fold in range). The residual ADC nonuniformity errors were attributed to measurement noise and other non-GNL sources. Correction of systematic GNL bias resulted in a 2-fold decrease in technical variability across scanners (down to site temperature range). The described validation of GNL bias correction marks progress toward implementation of this technology in multicenter trials that utilize quantitative DWI.
Supplemental Media 1: Supplemental Material
View this media larger in a new window