[Intl_DxMedPhys] MRI medium phantom geometric accuracy

[Bob Kobistek]

Is anyone else occasionally having problems meeting the ±2 mm tolerance on geometric accuracy when using the medium phantom?

I understand the rationale for reducing the action limits from ±3 mm to ±2 mm when adapting the large phantom limits to the medium phantom. The diameter of the large phantom is 190 mm, and the small phantom is 165 mm. 165 mm divided by 190 mm is 0.868, so the tolerance should be reduced by the same factor. That would make it 2.6 mm, which is closer to 3 mm than it is to 2 mm.  Hmmm.

Also, the pixel size is almost 1 mm x 1 mm. So right way, we have a measurement uncertainty of 0.5 mm on each end of the line we draw, or 1 mm in total. How can we be expected to meet a tolerance that is only 2 x our measurement uncertainty? I can measure the phantom diameter, walk away and come back and measure it again, and easily get a value that is different from my first measurement by 1 mm. I can have a different person make the same measurement I made and get a very convincing result that's 1 mm different than mine.

Yesterday I measured the phantom diameter as 162 mm (failing). I repeated the measurement several times and got the same result. Then the tech drew a line that was exactly 163 mm. He dragged the line into position on the phantom, and it also look very much like the actual diameter.


Robert J. Kobistek, MS, FACR, DABR, MRSE(MRSC(tm))
Medical Physicist
RJK Medical Physics, Inc.
440-463-7879

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osu.edu/pipermail/intl_dxmedphys_wd_osu_list/attachments/20260528/c7da730e/attachment.html>