Skip to content

A 3D lesion segmentation method on whole-body PET images including automated quality control.

License

Notifications You must be signed in to change notification settings

KibromBerihu/MIPsegmentatorV1

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

38 Commits
 
 
 
 
 
 

Repository files navigation

MIPsegmentatorV1

License: MIT Docker build passing

Artificial Intelligence for Efficient Learning-based Image Feature Extraction.

Tool for detecting and segmenting lesions in 3D whole-body PET images with automated quality control. It was trained on lymphoma data, including the AutoPET data. It is available in Docker image, easy use case. First, it automatically detects and segments lesion areas on the coronal and sagittal maximum intensity projections (MIP) PET images. A 3D lesion segmentation reconstruction algorithm is then developed to obtain the 3D lesion from the two 2D MIP segmentations using the SUV values from the 3D PET images.

If you use this pipeline, please cite our JNM paper and the autopet data paper. For additional information regarding the AI-based lesion segmentations from the coronal and sagittal maximum intensity projections (MIP) PET images, please refer to the github code.

This tool provides robust 2D MIP and 3D lesion segmentation on whole-body PET images, including automated quality control in Excel files. Surrogate biomarkers are automatically calculated and saved in an Excel file. The input 3D PET image should be in SUV NIFTI format.

Required input folder structure

Please provide all data in a single directory. The method automatically analyses all given data batch-wise.

To run the program, you only need PET scans (CT is not required) of patients in nifty format, where the PET images are coded in SUV units. If your images have already been segmented, you can also provide the mask (ground truth (gt)) as a binary image in nifty format. Suppose you provided ground truth (gt) data; it will print the dice, sensitivity, and specificity metrics between the reference segmentation by the expert (i.e., gt) and the predicted segmentation by the model. If the ground truth is NOT AVAILABLE, the model will only predict the segmentation.

A typical data directory might look like:

|-- input                                        <-- The main folder or all patient folders (Give it any NAME)

|      |-- parent folder (patient_folder_1)             <-- Individual patient folder name with unique id
|           |-- pet                                     <-- The pet folder for the .nii suv file
                 | -- name.nii or name.nii.gz            <-- The pet image in nifti format (Name can be anything)
|           |-- gt                                      <-- The corresponding ground truth folder for the .nii file  
                 | -- name.nii or name.nii.gz            <-- The ground truth (gt) image in nifti format (Name can be anything)
|      |-- parent folder (patient_folder_2)             <-- Individual patient folder name with unique id
|          |-- gt                                     <-- The pet folder for the .nii suv file
                | -- name.nii or name.nii.gz            <-- The pet image in nifti format (Name can be anything)
|         |-- pet                                      <-- The corresponding ground truth folder for the .nii file  
                | -- name.nii or name.nii.gz            <-- The ground truth (gt) image in nifti format (Name can be anything)
|           .
|           .
|           .
|      |-- parent folder (patient_folder_N)             <-- Individual patient folder name with unique id
|           |-- pet                                     <-- The pet folder for the .nii suv file
                | -- name.nii or name.nii.gz            <-- The pet image in nifti format (Name can be anything)
|           |-- gt                                      <-- The corresponding ground truth folder for the .nii file  
                | -- name.nii or name.nii.gz            <-- The ground truth (gt) image in nifti format (Name can be anything)

| -- output 

Note: the folder name for PET images should be pet and for the ground truth gt. All other folder and sub-folder names could be anything.

Usage

For reproducibility purposes, the method is currently available only in a Docker image.

  1. Assuming you already have docker desktop installed. For more information, kindly refer to THIS.
  2. docker pull kibromberihu/mipsegmentator:latest-0
  3. docker run --gpus all --ipc=host --rm -v "/path/to/input_output/input/":"/home/docker_input" -v "/path/to/input_output/output/":"/home/docker_output" kibromberihu/mipsegmentator:latest-0

Note:

  • /path/to/input_output/input/: This is the path to the folder that includes the 3D PET scans in SUV format and the ground truth (gt) files (if available) in NIFTI format, as detailed in the required input folder structure above.
  • /path/to/input_output/output/: This is the path to the folder where the output files will be saved.

Results

  • Predicted results including predicted segmentation masks and calculated surrogate biomarekrs (sTMTV and sDmax) will be saved into the folder output/.*.

  • Predicted 2D MIP masks are saved under the folder name output/predicted_data/final/*.

  • Predicted 3D masks are saved under the folder name output/predicted_data/predicted_pseudo_3d_reconstructed/*.

  • Surrogate biomarkers (sTMTV and sDmax) will be automatically calculated and saved as an EXCEL file inside the folder output/*.csv. Two EXCEL files will be saved. The first one constitutes computed surrogate biomarkers calculated from the segmentation masks predicted from AI with an indicator predicted in the file name. The second EXCEL file would constitute the surrogate biomarkers computed from the reference segmentation masks (i.e., ground truth) from the expert (if available) with an indicator ground_truth in the file name. In addition to the predicted and ground truth indicator names, the CSV file's name also constitutes an automatically generated month, year, and the processing time.

Citations

Please consider citting the following papers if you use this package for your research:

   1. DOI: https://doi.org/10.2967/jnumed.121.263501 
   2. DOI: https://doi.org/10.1109/TMI.2021.3060497
   3. DOI: https://doi.org/10.1038/s41597-022-01718-3

Acknowledgments

We thank you [the reader].