Body composition (BC) analysis, which captures the physical and constitutional characteristics of the human body, can provide valuable predictive information for various health conditions. In this work, we developed a fully automatic BC measurement pipeline using routine lung screening chest low-dose computed tomography (LDCT). To overcome the systematic field-of-view (FOV) limitations that causing body tissue truncation, we proposed a two-stage method to extend the image border and generate anatomically consistent body tissues in the truncated regions.
A typical result report obtained for a lung cancer screening CT:
The contents covered by this repository, including code and pretrained models in the docker container, are free for noncommercial usage (CC BY-NC 4.0). Please check the LICENSE.md file for more details of the copyright information.
If you find this study can help your work, please cite the following papers:
[1] Kaiwen Xu, Thomas Li, Mirza S. Khan, Riqiang Gao, Sanja L. Antic, Yuankai Huo, Kim L. Sandler, Fabien Maldonado, Bennett A. Landman. (2022). Body Composition Assessment with Limited Field-of-view Computed Tomography: A Semantic Image Extension Perspective. https://arxiv.org/abs/2207.06551. (Submitted to Medical Image Analysis)
[2] Kaiwen Xu, Riqiang Gao, Yucheng Tang, Steve A. Deppen, Kim L. Sandler, Michael N. Kammer, Sanja L. Antic, Fabien Maldonado, Yuankai Huo, Mirza S. Khan, Bennett A. Landman, "Extending the value of routine lung screening CT with quantitative body composition assessment," Proc. SPIE 12032, Medical Imaging 2022: Image Processing, 120321L (4 April 2022); https://doi.org/10.1117/12.2611784
docker pull masidocker/public:lung_body_composition_v1.0.2
User needs to prepare an input folder which contains the CT images in NIfTI format (.nii.gz) and (optionally) a csv file for metadata to provide the height of the patient.
Step.1 Specify the input directory
export INPUT_DIR=/home/input_dir
Step.2 Put NIfTI images under folder $INPUT_DIR/NIFTI without sub-folder
Step.3 Create a metadata.csv under $INPUT_DIR, which contains the following two fields:
- Filename
- HeightMeters
This metadata file is optional. However, the height data are needed to get the normalized BC indexes.
An example input dataset can be obtained via https://doi.org/10.5281/zenodo.6853516
Note: as we do not have the permission to share the lung screening CT data used in our published papers, the example dataset are compiled using four thoracic CT scans selected from the following public available (CC BY 4.0) dataset: The Caner Imaging Archive (TCIA), Chest Imaging with Clinical and Genomic Correlates Representing a Rural COVID-19 Positive Population (COVID-19-AR). These chest CT imaging studies were acquired for COVID-19 patients. Though these scans were not acquired under lung cancer screening protocols, they can still serve for demonstration purposes.
Specify output location
export OUTPUT_DIR=$INPUT_DIR/output
mkdir -p $OUTPUT_DIR
Run docker container with input/output locations
sudo docker run -it --gpus all --rm -v $INPUT_DIR:/Input -v $OUTPUT_DIR:/Output masidocker/public:lung_body_composition_v1.0.2 /app/src/Scripts/docker.sh
For the folder structure of the output directory, please refer to the README.txt file prepared under the output location. Output folder structure:
+ measurements.csv - per image file measurement results, including TCI value (truncation severity) and FOV extension ratio
+ Report_pdf - report in pdf format
+ Temp - all intermediate results
+ log - run time log file for debug
- Ubuntu 20.04/22.04
- cuda 11.2/11.3
- Docker version 20.10.14
- Nvidia-docker version 2.10.0
- GPU memory requirement: 12 GB
sudo apt-get install apt-transport-https ca-certificates curl software-properties-common
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"
sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu bionic stable"
sudo apt-get update
sudo apt-get install docker-ce
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
sudo apt-get update
sudo apt-get install -y nvidia-docker2
dcm2niix was the tool we used in our study to convert DICOM images to NIfTI format. The software is freely available at https://www.nitrc.org/plugins/mwiki/index.php/dcm2nii:MainPage
Use dcm2niix to convert DICOM images to NIfTI images:
dcm2niix -z y -f %j -o /path_to_output /path_to_input_dicom_folder
This will generate a NIfTI format image (.nii.gz) with seriesInstanceUID as the filename, along with a json file
for key metadata.
The version of dcm2niix used in our study: v1.0.20211006 (JP2:OpenJPEG) (JP-LS:CharLS) GCC7.5.0 x86-64 (64-bit Linux)
Convert3D (c3d) is able to convert images generated from DICOM in other format to NIfTI. The input can be any image format that readable by ITK.
For example, convert an image in Nrrd ("nearly raw raster data") format to NIfTI
c3d ./img.nrrd -o ./img.nii.gz
The code and data of this repository are provided to promote reproducible research. They are not intended for clinical care or commercial use.
The software is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the authors or copyright holders be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise, arising from, out of or in connection with the software or the use or other dealings in the software.