Nov 29, 2024
Version 1
Accurate Digital Organ Models from DICOM V.1
- Takashi Kimura1
- 1Fukushima Medical University
Protocol Citation: Takashi Kimura 2024. Accurate Digital Organ Models from DICOM. protocols.io https://dx.doi.org/10.17504/protocols.io.bp2l6d19dvqe/v1
License: This is an open access protocol distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Protocol status: Working
We use this protocol and it's working
Created: November 28, 2024
Last Modified: November 29, 2024
Protocol Integer ID: 113024
Keywords: accurate digital organ models from dicom, format surface data of organ, surface data of organ, accurate digital organ model, organ data, dicom data, human anatomy, segmentation from scratch, editing dicom data, segmentation in dicom viewer, imaging data, organ, surface data, format surface data, segmentation, 3d, dicom, such as 3d, comprehensive digital dataset, necessary organ, dicom viewer, slicer, data, imaging
Funders Acknowledgements:
JSPS KAKENHI
Grant ID: JP19K09100
Disclaimer
This system is provided as an open-source tool for research and educational purposes only. It is not a certified medical device and should not be used for diagnostic, therapeutic, or other medical purposes. The accuracy, completeness, and reliability of the data processed by this system are not guaranteed.
Users are solely responsible for ensuring compliance with applicable laws and regulations in their jurisdiction and for protecting the confidentiality and privacy of any data they use with this system. The developers shall not be held liable for any direct, indirect, incidental, or consequential damages resulting from the use or misuse of this system.
This system is provided "as is," and no warranties, express or implied, are provided. Updates and ongoing support are not guaranteed.
By using this system, you agree to these terms and conditions.
Abstract
This protocol describes the procedure for editing DICOM data to create precise STL-format surface data of organs. Using this method, surface data of organs obtained through auto-segmentation in DICOM viewers such as 3D-Slicer, along with Multi-Planar Reconstruction (MPR) images generated from DICOM data, can be spatially aligned within Blender. By referencing the imaging data, it becomes possible to refine and process the data obtained via auto-segmentation and to create organ data that cannot be captured through auto-segmentation from scratch. As a result, this approach enables the construction of a comprehensive digital dataset of human anatomy that includes all necessary organs.
Troubleshooting
JPEG format MPR image creation from DICOM
Run the script on Jupyer Lab.
Paste the path of the first file of the DICOM data you want to convert into `path_ct` (enclosed in quotation marks).
Create a folder to save the generated JPEG files, and paste the folder path into `storage_path`.
Retrieve information about the DICOM data each time a kernel is executed.
Execute the fourth kernel and adjust the values of `k` and `l` using the sliders to optimize the image.
Input the `l-k` and `l+k` values obtained from the fourth kernel into `vmin` and `vmax` in the fifth kernel.
When you execute the fifth kernel, JPEG files will be automatically generated and saved in the specified folder.
Auto-segmentation of organ surface data by 3D-Slicer
Launch 3D-Slicer.
Use the Extension Manager to add TotalSegmentator.
Add the data you want to convert via Add Data.
Load the required data through Add DICOM Data by selecting and clicking Load.
Use TotalSegmentator to auto-segment organ surface data. After pressing the Apply button, you will see two options: "Full resolution (~5 to 50 minutes)" and "Fast (~2 minutes)." Choose the appropriate option based on your needs.
Save and extract the organ surface data as an STL file via Segmentations.
- Prepare a folder in advance to save the data.
- Specify the folder in Destination Folder under Export to Files and press the Export button. The data will automatically be saved in the specified folder.
Auto-segmentation of body surface data by Horos
In 3D-Slicer, body surface data is not extracted, so the body surface data will be extracted using Horos.
Launch Horos.
Import the data to be used via Import.
Open the imported data.
Select the 3D Surface Rendering and set the pixel value to -200, then click "OK."
Skin, lungs, and other structures will be extracted. Select Export as STL from Export 3D-SR, specify the file name and save location, and save the file.
Import MPR images into Blender and display them using slider
Launch Blender and select the Script mode.
Click New to prepare for loading a new script.
Obtain the Python script from https://github.com/tk1971-Jpn/Slider-viewer-in-Blender and paste it into Blender's Script mode.
Choose the appropriate script based on the arrangement order of the DICOM data (head-to-foot or foot-to-head) and whether the data is longer in the anterior-posterior or cranial-caudal direction.
Assign the data obtained during the execution of JPEG format MPR image creation from DICOM to the corresponding variables in the Blender script
pixel(Blender) = pixel(Jupyter)
pixel_pitch(Blender) = pixel pitch(Jupyter)
ax_slide_number(Blender) = number of files(Jupyter)
ax_slide_distance(Blender) = slice distance(Jupyter)
Paste the path of the folder* where the JPEG format MPR image data is stored into path_JPEG
∗ A folder containing Folder A, Folder C, and Folder S
When the script is executed, a tag named Image Slider will appear in the sidebar of the Layout mode. You can use the sliders to display any Axial, Coronal, or Sagittal images.
Since the size of the human body far exceeds Blender's workspace, it was scaled down to 1/50 for import.
Import Organ Surface Data in STL Format into Blender
Select Script Mode in Blender
Click New to prepare for loading a new script.
Obtain the Python script from https://github.com/tk1971-Jpn/Import-organ-STL-data-into-Blender/blob/main/Import%20organ%20STL.ipynb and paste it into the Blender's Script mode.
Past the path of the folder where the organ STL data is stored into folder_path.
When the script is executed, all organ STL data are imported into Blender in bulk, and the object names are appropriately converted during the process.
If there is a positional misalignment of the organs (which often occurs), use Blender's "Move" tool to adjust their positions based on the reference image.
(Option)Delete all the imported STL data, set the positions, and re-import them. The data used for position adjustments can be obtained from Blender's sidebar under Transform > Location. Based on this data, assign values to the script's x_move, y_move, and z_move (default is 0).
#Delete all objects: press A and then press X
Since the size of the human body far exceeds Blender's workspace, it was scaled down to 1/50 for import.