Jesper Fredriksson, Presenter: Employee, Raforninn EHF

Paul Austin Healy, Abstract Co-Author: Research funded, The Phantom Laboratory Research funded, Image Owl, Inc

Christine Wu Nordahl PhD, Abstract Co-Author: Nothing to Disclose

Hildur Olafsdottir, Abstract Co-Author: Research funded, The Phantom Laboratory Research funded, Image Owl, Inc

Smari Kristinsson, Abstract Co-Author: Research funded, The Phantom Laboratory Research funded, Image Owl, Inc

Joshua Levy, Abstract Co-Author: Research funded, The Phantom Laboratory Research funded, Image Owl, Inc President, The Phantom Laboratory

David John Goodenough PhD, Abstract Co-Author: Director, The Institute for Radiological Image Sciences, Inc Consultant, The Phantom Laboratory Consultant, Live Radiology, LLC

To evaluate MRI scanner geometrical distortion over time and as a function of phantom positioning.

The Magphan® Quantitative MRI phantom (the "ADNI phantom") measures geometrical distortion by segmenting and comparing measured locations of 165 spheres inside the phantom to known manufactured locations. The phantom was scanned 171 times over a period of eight months in a 3T MR scanner. The distortion field was calculated for the 165 spheres for each scan and modeled as a smooth third order polynomial of x/y/z coordinates. The measured distortions were compared over time for each (x,y,z) location by the standard deviation. Effect of phantom placement was carried out by shifting the phantom from one end to the other in the scanner. The phantom was also scanned in different positions, covering as much volume as possible, in three different 1.5T MR scanners, using different coils, during single scanning sessions. The single scanning sessions were designed to both measure distortion throughout the largest possible volume of the scanner and investigate the effects of phantom position on the distortion field.

The stability over time was good for all scans acquired at similar positions in the scanner. A shift in the y-direction proved to yield a larger difference than in the x- and z-directions. Measured distortion at a given position in the scanner depends on the location of the phantom.

Geometrical distortion in MR images can be significant and typically increases with distance from the isocenter. The position of the subject affects distortion and must be considered in cases where distortion is important, for example in quantitative imaging. This indicates MR scanners are typically stable with respect to geometrical distortion, but still need close monitoring, especially in connection with service events or upgrades. Also, the positioning in the scanner can be a factor for the resulting geometrical distortion.

Knowledge of potential distortion of an MR scanner image is crucial for many clinical applications including treatment planning.

Fredriksson, J, Healy, P, Wu Nordahl, C, Olafsdottir, H, Kristinsson, S, Levy, J, Goodenough, D, MRI Scanner Geometrical Distortion Variability Depending on Phantom Placement and over Time.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12034101.html