Lucas-Mc / PET-CT_ACR_phantom_analysis

Automated analysis of the PET-CT ACR phantom.

PET-CT_ACR_phantom_analysis

Automated SUV analysis of the PET-CT ACR phantom.

Environment: Windows 10, Python 3.7.9

In the command line, run the following:

foo@bar:~$ python3 -m venv env
foo@bar:~$ .\env\Scripts\activate
foo@bar:~$ pip install -r requirements.txt
foo@bar:~$ python PET-CT_analysis.py

Or open up the Jupyter Notebook to look at it by sections.

There are several notes hidden within the scipt file itself, be sure to check these out if you run into any errors.

Some basic notes:

1. I used CT and PET data which was isotropic and 133 slices each, I think it will also work for non-isotropic and unequal slicing due to its reliance on the DICOM parameter ImagePositionPatient which converts everything to a shared coordinate system.
2. My algorithm for finding the circles works great for my sample images (which seem like good representatives of PET/CT data for ACR phantom analysis), however they may not work as well for other acquisitions and the parameters may need to be adjusted, especially the non-radius parameters (i.e., C_PARAM) which extract the contrast differences.
3. My passing criteria is based on the 2010 update to the ACR guidelines, however if new updates arise, then the PASSING_CRITERIA global variable will have to be adjusted accordingly.
4. Double check the global variables first such as CT_SLICE_LOCATION which defines the slice location of the vials in the CT image ... I wish to make this automatic in the future. REFERENCE_VIAL_SIZES should not change unless the ACR decides to update their phantom.
5. Inside the data/CT and data/PET folders should be the DICOM images for each slice (i.e., 133 DICOM images for my case) without any parent folders so data/CT/im1.dcm, data/CT/im2.dcm, etc.