(Saturday, 4/7/2018)
Room: Foyer
Purpose: In this work we present the preliminary results of DWI QA using ACR phantom on a large fleet of clinical MR scanners. We demonstrate sensitivity and feasibility of this technique.
Methods: A DWI QA procedure was integrated into the regular ACR phantom annual QA. The DWI QA protocol was scanned twice for each phase encoding direction (AP,RL). For each acquisition, three sets of images were reconstructed: source images, diffusion weighted images, ADC maps. The temperature of the ACR phantom was measured to calculate the expected ADC value. A total of 15 scanners were surveyed and analyzed. ADC Accuracy, ADC Spatial Variation, DWI Geometry Accuracy (Image Misregistration and Distortion), DWI Percent Signal Ghosting (PSG) were analyzed to evaluate the quality of DWI.
Results: The percent error of ADC value over 15 scanners was analyzed with respect to field strength and bore size. There was no statistical correlation between ADC accuracy and field strength. However, ADC values for the 60cm bore size demonstrated clear underestimation of ADC values, while the 70cm bore scanners overestimated ADC values with a higher absolute error (-0.77% vs 1.12% respectively). Absolute error of image Spatial Variation in vertical (AP) and horizontal (RL) directions were 0.59% and 0.4% respectively. Image Misregistration and Distortion have a better performance in the horizontal (RL) direction compared to the vertical (AP Mean Shift = 0.89mm, RL Mean Shift = 0.62mm, AP Distortion = 0.8mm, and RL Distortion = 0.61mm). Mean PSG for 1.5T scanners was 1.75% while 3T showed significantly better performance with PSG = 0.76%.
Conclusion: To the best of our knowledge, this is the first time a DWI QA procedure was tested on a large number of clinical MRI scanners of different makes and models. It can be readily incorporated in the regular weekly, monthly or annual QA tests.
ePosters
Contact Email
