Hello BJ community
I am using QCT images (ScancoXtreme) to analyse the entheses of the heel (Achilles, plantar fascia, and a reference mid calcaneal site).
Having decided on the volumes of interest, I do the following:
- import scanco ISQ file, specify slice range, and clean image with median filter (I also crop the image to reduce the data load)
- trace out ROI at intervals and interpolate through the stack- up to 110 slices per ROI (and then double check and spline/ fix as required).
- I then use the BoneJ plugin to assess, amongst other things, Bone Volume Fraction.
This is where it gets tricky. The thought was to set the threshold in the Threshold window at a standard percentage (eg 33%), to facilitate comparison across participants. However it appears that setting the threshold by this method is slice specific- and can yield different results depending on which slice in the stack I use. Also, I am not sure if 33% is the same from participant to participant- especially if it is determined from a histogram of that individuals pixel values?
I am aware that BoneJ has an ‘Optimise Threshold’ option, which utilizes information from the whole stack. However, I am not sure how this improves comparability across subjects if I don’t get to standardize the thresholding output?
Any advice on how best to standardize for comparison across subjects would be greatly appreciated (accepting that for a specific individual it may not provide the optimal resolution).