Skip to content

JSIERO/ClinicalASL

Repository files navigation

ClinicalASL Tool - Release v1

Repository: https://github.com/JSIERO/ClinicalASL
Author: Jeroen Siero, UMC Utrecht
Contact: j.c.w.siero@umcutrecht.nl
Release Date: 2025-06-24
Version: v1.0.3

Description

Official ClinicalASL processing pipeline for quantitative multi-delay ASL MRI data — designed for clinical use with PACS export and visual reporting.

  • DICOM to NIfTI conversion
  • Look-Locker correction
  • Motion correction
  • Outlier removal
  • Quantitative perfusion parameter estimation:
    • Cerebral Blood Flow (CBF)
    • Arterial Arrival Time (AAT)
    • Arterial Transit Artefacts (ATA)
    • Cerebrovascular Reactivity (CVR)
  • Export of results as NIfTI, PNG, and DICOM (PACS-ready) formats.

Supports baseline / stimulus ASL protocols (acetazolamide/diamox, hypercapnia, breath-hold, etc.)

Intended Use

This software is provided for post-processing of ASL MRI data for research and clinical evaluation purposes.
It is not CE marked or FDA cleared

Installation

See INSTALL.md or follow instructions in the GitHub repository.

Configuration

Example config file containing scan and analysis parameters: config/config_default.json
You can supply your own config.json per acquisition protocol.

example config/config_default.json:

{
    "version": "v1.1-MRI_DIAMOX_MDLL_preACZ_postACZ-2025",
    "ASL scan": "multi-delay Look-Locker",
    "tau": 2,
    "N_BS": 4,
    "readout": "2D",
    "labeleff": 0.85,
    "lambda": 0.9,
    "T1t": 1.3,
    "T1b": 1.65,
    "FWHM": 6,
    "outlier_factor": 2.5,
    "range_cvr": [-50, 50],
    "range_AAT": [0, 3],
    "range_ATA": [0, 125],
    "inference_method": "ssvb",
    "device": "cpu",
    "ASL_CONTEXT": ["baseline", "stimulus"],
    "context_study_tags": ["preACZ", "postACZ"],
    "dicomseries_description_patterns": ["*SOURCE*ASL*"]
}

Dependencies

  • Python 3.11+
  • QASL by Quantified Imaging (license key required)
  • dcm2niix (v1.0.20230411 or newer)
  • HD-BET (for brain extraction)
  • ANTsPy v 0.5.4 (for image registration) - version 0.6.1 works bad!
  • Additional Python packages (see requirements.txt)

Installation

Complete installation instructions are available in INSTALL.md.

Quick Start (QI Conda Environment)

# Clone repositories
git clone https://github.com/JSIERO/ClinicalASL.git
git clone https://bitbucket.org/quantified-imaging/qasl_setup.git

# Install QASL (requires license key)
cd qasl_setup
./qasl_setup --yes --key=<your-qasl-key>

# Set up environment
conda activate qi
conda install -n qi -c conda-forge pip dcm2niix

# Install HD-BET
git clone https://github.com/MIC-DKFZ/HD-BET.git
cd HD-BET
pip install -e .

# Install ClinicalASL requirements
cd ../ClinicalASL/python
pip install -r requirements.txt

## Running the pipeline
Example command line:

```bash
python run_pipeline.py /path/to/DICOM_INPUT /path/to/OUTPUT_FOLDER \
    --inference-method [vaby|ssvb] \
    --config /path/to/config.json

##Output Structure The pipeline generates:

  • NIfTI images: CBF, AAT, CVR, ATA maps
  • Quality control PNGs: Visualizations for review
  • DICOM files: PACS-compatible outputs
  • Log file: clinicalasl.log with processing details
  • Configuration copy: config_used.json

License

BSD 3-Clause License. See LICENSE for details.

About

ASL and CVR assessment using (multidelay) ASL data

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors