Tim Trio Gradient Non-Linearity Quality Assurance

Gradient non-linearity on the Tim Trio is monitored every 1-2 weeks using the Magphan® Quantitative Imaging Phantom, commonly know as the ADNI (Alzheimer's Disease Neuroimaging Initiative) phantom. The phantom is produced by The Phantom Laboratory.

The ADNI Phantom contains a total of 165 spheres made from polycarbonate plastic. These spheres are filled with copper sulfate and water solutions, and are mounted on a series of polycarbonate plates and posts to maintain their fixed positions inside the phantom. The phantom includes one 6cm diameter sphere, four 3cm diameter spheres, two 1.5cm diameter spheres and one hundred fifty-eight 1cm spheres, which are spread throughout its 20cm ID housing.

The company, Image Owl Inc., offers a software service that analyzes the phantom images. The software precisely measures, maps and documents the volume stability or distortion by comparing the observed position of each sphere with its known location in the phantom. Any changes over time may indicate increased gradient non-linearity or distortion. Abrupt large changes in any one measurement parameter may indicate a problem with the scanner hardware which are investigated with our Siemens engineer.

The photograph shows a view of the ADNI phantom, and beside that is a representative MR image of the phantom. An example of the Image Owl visualization of the phantom after image analysis, and scatter plot analysis are shown below. The scatter plots show the position of each sphere in each axis (vertical in each plot), with its deviation from each axis (horizontal). Lighter-colored points have higher deviation, however all are below 1.5 mm.



We scan the phantom using a FLASH sequence with 0.9 mm spatial resolution, and a multi-echo MPRAGE sequence with 1.2 mm spatial resolution that is based on the parameters commonly used for anatomical scanning at our facility. Current plots of maximum, x, y, and z distortion, and the mean distortion of the worst 10% of the spheres overall and in the x, y, and z axes are shown below. Upper and lower expectation limits (red lines labeled UCL and LCL on the plots below) are those suggested by Image Owl - the green line is just the half-way point between the two limits. As is evident from these plots (and confirmed by Image Owl scientists who have compared our results with those from other scanners), our distortion levels are considerably lower than most similar scanners. Changes in the distortion levels in the first half 2011 indicated a fault with the phantom that was later repaired, and a resulting change in the method Image Owl used to calculate distortions.