Installation

Requirements

• Python 3.9 or later

• ANTs (provides DenoiseImage, ImageMath, antsRegistration, etc.)

• Docker (for tumor segmentation)

• GPU with CUDA support (optional — required for ensemble segmentation mode)

ANTs is a required system dependency for anatomical preprocessing (denoising, skull-stripping, registration). Install it via your system package manager, an HPC module (module load ants), or build from source. Ensure the ANTs bin/ directory is on your $PATH.

Quick install

pip install oncoprep

Optional extras

Radiomics

pip install "oncoprep[radiomics]"

VASARI feature extraction

pip install "oncoprep[vasari]"

Installs vasari-auto — a fork of the original VASARI-auto by Ruffle et al. (2024), maintained for OncoPrep integration.

HD-BET skull stripping

pip install "oncoprep[hd-bet]"

Development

pip install "oncoprep[dev]"

From source

git clone https://github.com/nikitas-k/oncoprep.git
cd oncoprep
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"

Docker

If you prefer a fully self-contained environment with ANTs, FSL, and dcm2niix pre-installed:

docker pull nko11/oncoprep:latest

See Docker Usage for detailed Docker usage.

Singularity / Apptainer (HPC)

module load singularity
singularity pull oncoprep.sif docker://nko11/oncoprep:latest

See HPC / Singularity Deployment for HPC deployment instructions.

Verifying the installation

oncoprep --version