fastMRI Breast: A publicly available radial k-space dataset of breast dynamic contrast-enhanced MRI

Demonstration of DCE MRI reconstruction using temporal TV regularization

• Goal: To make a publicly available radial k-space dataset of breast DCE-MRI which will promote development of fast and quantitative breast image reconstruction and machine learning methods.

• Code descripition:

  'loop_single_data.sh' script executes two python scripts, dce_recon.py and dcm_recon.py. The scripts read k-space data from .h5 file stored within a patient folder (e.g., 'fastMRI_breast_003_2') and generate a new reconstructed image series based on user preferences (such as spokes per frame, slice index, and the number of slices). The resulting image series is stored in a new .h5 file ('_processed.h5') and in a DICOM folder ('_processed'), both saved under the patient folder.

• How to run on your local computer?

  To run the bash script 'loop_single_data.sh' on your local computer, you need to firstly install the required python environments, which I suggest to use conda.

  Here are the steps:

  1. create a new conda environment in the terminal:

  conda create -n dce python=3.10
  conda activate dce
  which python  # to validate the python is under the environment

  conda install -c anaconda pip
  python -m pip install torch torchvision torchaudio
  python -m pip install tqdm
  python -m pip install pydicom
  python -m pip install numba
  python -m pip install scipy
  python -m pip install pywavelets
  python -m pip install h5py
  python -m pip install matplotlib
  
  conda install -c conda-forge cupy cudnn cutensor nccl  # if you have GPU
  conda install -c conda-forge numpy=1.24

  2. clone and install sigpy in the terminal:

  git clone https://github.com/ZhengguoTan/sigpy.git
  cd sigpy
  pip install setuptools==58.2.0  # might be needed, try first without
  python -m pip install -e . 

  3. Now you should be able to run the script with four inputs: data, spokes per frame, slice index, number of slices

  bash loop_single_data.sh fastMRI_breast_002_1 72 100 10 

fastMRI Breast dataset:

The data are available for free through: https://fastmri.med.nyu.edu/. After acceptance of the dataset sharing agreement, researchers receive an email containing links to download the data. In order to use an .h5 file with this code, save it inside a folder named with its patient code name (e.g., 'fastMRI_breast_003_2'), as instructed in 'Code description' above. Note that the provided DICOM files are in 4D (x,y,z,time) with 4 time frames. For easier viewing, we recommend using https://firevoxel.org/

Our dataset also includes case-level labels arranged in an excel file (also available here under 'breast_fastMRI_final') indicating patient age, menopause status, lesion status (negative, benign, and malignant), and lesion type for each case.

Citation:

If you use the fastMRI DCE Breast data or code in your research, please cite our paper: https://pubs.rsna.org/doi/10.1148/ryai.240345

References:

• Zhang S, Block KT, Frahm J. Magnetic resonance imaging in real time: advances using radial FLASH. J Magn Reson Imaging 2010;31:101-109.

• Uecker M, Zhang S, Voit D, Karaus A, Merboldt KD, Frahm J. Real-time MRI at a resolution of 20 ms. NMR Biomed 2010;23:986-994.

• Block KT, Chandarana H, Milla S, Bruno M, Mulholland T, Fatterpekar G, Hagiwara M, Grimm R, Geppert C, Kiefer B, Sodickson DK. Towards routine clinical use of radial stack-of-stars 3D gradient-echo sequences for reducing motion sensitivity. J Korean Magn Reson Med 2014;18:87-106.

• Feng L, Grimm R, Block KT, Chandarana H, Kim S, Xu J, Axel L, Sodickson DK, Otazo R. Golden‐angle radial sparse parallel MRI: combination of compressed sensing, parallel imaging, and golden‐angle radial sampling for fast and flexible dynamic volumetric MRI. Magn Reson Med 2014;72:707-717.