(Sunday, 7/29/2018) 4:00 PM - 4:55 PM
Room: Room 207
Purpose: To characterize a new metric for determining MR image uniformity for phased-array coils, which are intrinsically non-uniform; and to compare its performance to three common image uniformity metrics.
Methods: A uniformity metric that relies on Fourier power spectra differences between a test and reference image was developed. To benchmark uniformity of this and other uniformity metrics, simulated MRI images were generated in Matlab (Mathworks, Natwick, MA) and their uniformity quantified. A series of simulations were performed with coil elements disabled to various degrees. The uniformity of the images using our metric was calculated and compared to using the American College of Radiology MR Accreditation Program uniformity and two methods from the National Electrical Manufacturers Association (NEMA-GU & NAAD). To compare to actual scan images, an ACR phantom was scanned using a GE 3T system and 8-channel phased-array head coil. Elements were disabled by disconnecting particular channels.
Results: Uniformity test of simulated images yielded mean uniformity decreases of 10.0%, 16.0%, 3.8%, and 17.1% from baseline for ACR, NEMA, NAAD, and our method respectively. Qualitatively, acquired images generally match simulations. In particular, the actual images barely passed the ACR uniformity test (83%) with all elements functioning. Our method shows a uniformity decrease of 23.1% from baseline for broken elements with uniformity corrections turned on, compared to 2.4% for the next largest uniformity decrease (ACR). With uniformity corrections turned on, actual images passed the ACR uniformity test with 1-2 disabled elements.
Conclusion: We have demonstrated that our uniformity test shows sensitivity differences between baseline functionality and a variety of coil conditions that may go undetected using standard uniformity tests.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
