Ever wondered why cartilage thickness changes from sequence to sequence on MRI. There are many reasons for this, and difference in slice thickness and the resultant volume averaging is one of the reasons.
If the slice thickness was infinitely small, cartilage thickness would be as accurate as possible based on lack of volume averaging. Of course, in reality, other factors play into measurement error as well, but here we will concentrate on slice thickness only.
So, by how much would apparent cartilage thickness increase if MR slice thickness is increased? The thicker the slices, the more signal gets averaged from adjacent cartilage voxels making cartilage look thicker than it actually is.
With a little bit of high school trigonometry and geometry, we can come up with the following figure of two right triangles.
Now, with some tincture of atan and cos, we can calculate what the difference between “volume averaged” and infinitely thin slice would be.
Here is an example of hip cartilage thickness error estimation tabulated against increasing slice thickness (ST). Average human femoral head is assumed to measure 48.4 mm in diameter with 1.5 mm of cartilage overlying it.
- ST = 0 mm, Error 0%
- ST = 3 mm, Error 11%
- ST = 4 mm, Error 19%
- ST = 5 mm, Error 30%