How to measure anisotropy in HR-PCT dicoms?

[Drew Best]

Hi folks,

Thanks for any help with this. Here's what I'm trying to do:

I have image stacks of calcanei and metatarsals, obtained from high-res pct. As I understand it, BoneJ is meant to work with micro-pct images, which are already in the format of square ROI's, yes? I, however, have slices of whole bone, with empty space (well, soft tissue) surrounding the bone. So, I need to crop out lots of stuff, leaving as much trabecular bone as possible. (Ideally I'd be measuring anisotropy in the entire trabecular area of the bone.)

Creating a single, rectangular ROI, and then cropping all of the slices, leaves me with a tiny ROI to sample. I can make it bigger by interpolating that ROI so that it shifts with each slice (moving in the X plane?) thus giving me a bigger, not-perfectly-vertical ROI through the stack. However, this doesn't seem to be working.

Can I create several ROI's and then crop/ run aniostropy on each individually, then average the degree of anisotropy values together to get a bigger sampling area for the bone? Is this a method that folks employ? In short, I want to be following established protocol for this, but I can't seem to find any .

Thanks for any help!
Drew