Trabecular Number

[Sara M]

Dear Michael,

I am currently doing a project looking at trabecular bone. I was
hoping that you may be able to help me.

I am attempting to find the number of trabeculae within a sphere which
I have created using the ROI manager tool.

I believe that on your website connectivity measures the connectivity
density which can inturn be interpreted as the trabecular number per
mm3.

However under Analyse skeleton I believe that you can measure the
number of branches and junctions. Therefore my question is does the
number of branches also relate to the trabecular number?

Kind Regards,

Sara

[Michael Doube]

Hi Sara,

> I believe that on your website connectivity measures the connectivity
> density which can inturn be interpreted as the trabecular number per
> mm3.

That's correct. In your case however, you will need to correct the
calculation by dividing the number of trabeculae by the volume of your
sphere and not the volume of the stack.

> However under Analyse skeleton I believe that you can measure the
> number of branches and junctions. Therefore my question is does the
> number of branches also relate to the trabecular number?

It depends how you define 'branch' and 'trabecula'. I prefer the
topological approach used in Connectivity, because trabeculae branch in
complicated ways. The number of branches is equal to connectivity in
certain restricted circumstances. In most cases however, the number of
branches in a skeleton will be greater than topological connectivity.

Consider the simple image resulting from this macro:

newImage("circle_line", "8-bit Black", 256, 256, 3);
setSlice(2);
setColor(255);
drawOval(64, 64, 128, 128);
drawLine(64, 128, 192, 128 );
drawLine(128, 64, 128, 192 );
run("Skeletonize 3D");

Connectivity is 4, while the number of branches is 8. The question is,
which of those lines should be categorised as individual trabeculae, and
does it matter if the structure is always loaded as a whole?

Michael