Synthetic PET from CT

Support Feature requests
1

This paper ( PMID: [38471502] Synthetic PET from CT improves diagnosis and prognosis for lung cancer: Proof of concept - PubMed) describes a process to use a CT to predict what the PET will look like.

Is this a useful module to reproduce in Slicer?
Steps would be:

  1. Import CT
  2. Run algorithms
  3. Export PET to same folder with included translation object to be ready to import into viewing software.

(I am a radiation oncologist without coding skills to implement this, hence my request.)
The purpose is twofold, firstly for my own research use, but secondly also for availability in LMICs which cannot afford PET scans but where CTs are reasonably available.

2

Yes, I think this looks like a very cool model that would be good to add.

It looks like the code and the model weights used to generate the results in the paper are available in github:

Maybe someone here will want to take on packaging this into an extension.

Or @oismaster you may want to contact the people at https://mhub.ai/ for advice.

@fedorov do you have any suggestions?

3

@pieper It would be great to include it in the new ModalityConverter extension. I see that there are still some open issues. Let me know what I can do :slight_smile:

4

@oismaster The model has been integrated into the Modality Converter extension, which can be installed directly from the 3D Slicer Extension Manager for versions >5.9. However, I recommend installing at least 3D Slicer 5.10, as some models are only compatible with that version onwards. I hope this helps!