Program Information
Experience with a Digital Reference Object for a Diffusion-Weighted Magnetic Resonance Imaging Quality Control Program
A Antolak*, E Jackson , University of Wisconsin, Madison, WI
Presentations
SU-K-708-2 (Sunday, July 30, 2017) 4:00 PM - 6:00 PM Room: 708
Purpose: Longitudinal assessment of quantitative imaging biomarkers requires a comprehensive quality control program for the acquisition and parametric analysis of imaging data. The purpose of this work is to describe our experience using a diffusion-weighted imaging (DWI) digital reference object (DRO) for use in a quality control (QC) program.
Methods: A newly developed DWI DRO, publicly available through the Quantitative Imaging Data Warehouse (RSNA.org/QIDW), consists of 396 zones representing a range of apparent diffusion coefficient (ADC) and signal-to-noise ratio (SNR) values. As an initial evaluation of the utility of using this DRO in the development of a QC program, DWI parameters were estimated using the commercially-available nordicIce software package (NordicNeuroLab, Inc, Milwaukee, WI). Calculated parameters were compared against ground truth using Matlab code developed in-house. Percent bias measures and coefficients of variation (COV) were calculated for each ADC and SNR value. Bland-Altman analyses were performed, resulting in limits-of-agreement for each SNR value. The concordance correlation coefficient (CCC) was calculated to estimate the agreement between the calculated and ground-truth values.
Results: Measurement bias and variability improved with increasing SNR. Excellent agreement with reference values was seen for all ADC values above an SNR of 15 (CCC greater than 0.97). The average COV measured across all SNR (range: 1-100) and ADC values (range: 0.1-3.5x10⁻³ mm²/s) was 30%, with a maximum of 748%. Considering SNR levels of 15 and above, the average COV was 3.68%, with a maximum of 16%.
Conclusion: The recently-developed DWI DRO allows for the evaluation of analysis software bias and variance. Excellent agreement between calculated and reference values was seen for moderate to high SNR levels. Future work will involve evaluating the relevant range of SNR and ADC levels typically observed in clinical diffusion-weighted images to establish the relevant DRO parameter space to be used in ongoing QC programs.
Contact Email: