Aorta PET Segmentation

Installation

To install this library

python -m venv .aortapetseg_venv # (OPTIONAL)
source .aortapetseg_venv/bin/activate # (OPTIONAL)

pip install git+https://github.com/CAAI/AortaPETSeg.git

Set nnU-Net environment variables (in .bashrc) (nnUnet Documentation)

if [ -z ${nnUNet_raw} ]; then export nnUNet_raw="${nnUNet_raw_data_base}/nnUNet_raw"; fi
if [ -z ${nnUNet_preprocessed} ]; then export nnUNet_preprocessed="${nnUNet_raw_data_base}/nnUNet_preprocessed"; fi
if [ -z ${nnUNet_results} ]; then export nnUNet_results="${nnUNet_raw_data_base}/nnUNet_results"; fi

Usage

Preprocessing

Data must be converted to Standardized Uptake Value (SUV) of the first 40s of data.

Citation

Please cite the main method manuscript when using our method.

Andersen TL, Evenstuen N, Ladefoged C, Lindberg U. Contrast agnostic AI-based extraction of image-derived input function from dynamic PET scans. Annual Congress of the European Association of Nuclear Medicine, October 4–8, 2025 | Barcelona, Spain, EANM Innovation, Volume 1, Supplement, 2025, 100004, ISSN 3051-2913, https://doi.org/10.1016/S3051-2913(25)00004-7.