Bone mineral density

Hello. I need to analize bone mineral density in a sheep jaw bone sample. The image secuence is from a CBCT (cone beam computer tomography). The isotropic voxel size is 200µm.
How can I do it? Is it possible to do it with boneJ plugging?

I installed pQCT. But, when I run it, and I run distribution analysis it says: “.TYP file not found”. What does it mean?

How can I analyze BMD?

1 Like

Hello there,
Have you used the ‘BoneJ’ plugin of ImageJ/Fiji at all yet? If you do not have it in your plugins tab then it can be downloaded from the Image/Fiji site. It is made for bone analysis with a lot of tools you will probably need.
Good Luck

I have instaled it, and I have used some tools, but I didnt find one that allows me to calculate bone mineral density. Haw can I do it?

First you need to make sure that your images are calibrated, so that you can convert pixel values into some measure of BMD. Note that all pixels, including soft tissue pixels, with an attenuation coefficient greater than air will be represented in your calibration function, so you will need to segment those out. Otherwise you might measure negative density, which is a physical impossibility because it implies negative mass or volume.

The usual way to do a calibration is to make a curve by imaging some samples with known BMD (phantoms) using the same settings as your experiments, then plotting BMD vs pixel value, calculating a best-fitting function then applying that to the experimental pixel data.

BoneJ itself doesn’t do a ‘BMD’ measurement, because it’s a technique that is fraught with problems. For example, if your sample has ring or streak artefacts, or too much sharpening filter applied (common in clinical CT), that will show up in your results.

Furthermore, low resolution images used for ‘BMD’ calculation are typically measuring BV/TV rather than matrix mineralisation density (i.e. the amount of bone that is there, not the mineralisation of the bone). If you can resolve microstructure (pixel spacing ≲15 µm), just report BV/TV. It’s by far the most mechanically important variable. If you really need tissue mineralisation then you need to go to higher resolution still, for example synchrotron radiation microtomography, or quantitative backscattered scanning electron microscopy (But both of those still need calibration against phantoms and have their own artefacts to contend with).

Thank you very much.