Recreation of Internal Structures and Individual bones

[Clyde]

Hi!

 

I'm wondering if anyone know is there is a way to create a 3D printable model that includes the internal structures of the bone? I'd like to recreate the bone inside a living animal with internal structures so that we may practice cutting into and exploring it as if it were the real thing.

 

Is there a way to have individual bones created as STL's as opposed to them being joined?

 

Thanks for you assistance!

 

Clyde

[Dr. Mike]

Are you looking to create the microarchitecture of the bone, i.e. tiny pores and crevices inside the medullary bone? To start with you will need a super high resolution CT scan - one that uses more radiation than a typical medical CT scan. Also, any STL file will have tremendous surface area and will be huge.

[Clyde]

Thank you for your response, Dr. Mike. Firstly, as an engineer myself I'm sorry if I ask or say anything that truly shows my lack of knowledge in this area, I will try and research everything you ask such that I can communicate effectively.

 

The microarchitecture is probably a bit beyond what I'll be capable of reproducing with my printer, but It would be useful if I could extract the boundary of the compact bone to the medullary cavity/spongy bone. Whilst not truly mimcry, being able to print compact bone in one density and spongy bone in another should serve my purposes. Extracting just this boundary should also serve to reduce the size of the would be STL. Do you think this is achievable?

[mikefazz]

While I am not a fan of thresholding for segmentation, a decent scan should be able to distinguish between cortical and trabecular bone with different thresholding parameters.  You would need to run thresholding 2 separate times with different thresholding ranges.

 

My preferred method involves using region growing, where you could seed the trabecular and cortical bone with separate seeds and get separate entities.

 

 

[valchanov]

There is a histological staining technique for bone tissue, which can be used with confocal microscope for detailed dataset of the haversian model, the lacunae, the canaliculae, the perivascular spaces - every detail you can possibly imagine. Then the dataset can be segmented into a 3D model with the same workflow, which is used for the CT datasets. The problem is that this technique is very demanding, expensive and potentially letal because of the chemical, which are used in the process. I made several models of pyramidal neurons from confocal datasets (here is one on sketchfab, I have another on embodi3d with the dendrite spikes on it) and it's not that bad - after some deconvolution you can segment the dataset with the automatic threshold segmentator or even with the grayscale model maker and you'll get a good model. Sadly, my department didn't received the funding for this technique and I don't have such dataset...

[valchanov]

If you manage to find a good Z-stack (confocal microscopy dataset), you can use the following workflow to make a 3D model:
1. Use Fiji (an open-source software, extended version of ImageJ for scientific analysis) to convert .tiff slide to binary (Process/Binary/Make Binary). Then use the Dilate function to reduce the size (Process/Binary/Dilate) and crop a region of interest (Image/Crop). Those datasets are ridiculously large - 10 Gb and above (this is why I didn't uploaded such dataset here), so if you don't have a workstation, be prepared for a LONG  wait. Finally, save the stack as a .tiff
2.  Use 3D Slicer, Fiji, Osirix, Mimix or whatever segmentation software you prefer to create volume (Volume Rendering), to convert it to a label map and to export it as .stl. The workflow is the same as in the Dr. Mike's basic tutorial. Just don't forget to use the Curvature Anisotropic Diffusion module of Slicer or a similar denoising tool in the software you prefer. It will make your life A LOT easier - the confocal Z-stacks have a lot of noise and still you want to keep the curves of the objects. It's possible to segment a Z-stack with the Region Grow module of Slicer, but I didn't made it myself yet.
3. Finally, use Blender or Meshmixer to remove the artifacts. It's normal to have a lot of them.
This topic is related to my PhD project, on which I'm working right now - "Bioprinting and morphological analysis of a 3D matrix for biosynthetic implants", so I'll post my results after  I manage to add a syringe pump to my 3D printer and make it work...