Development of anatomically accurate digital organ models for surgical simulation and training

Area of Science:

• Medical Simulation
• Computational Anatomy
• Surgical Training Technology

Background:

• Surgical advancements necessitate realistic anatomical training environments.
• Current methods for creating anatomical datasets can be time-consuming and costly.

Purpose of the Study:

• To develop an efficient, open-source system for generating accurate anatomical datasets for surgical simulation.
• To enable medical experts to actively participate in refining anatomical data for training.

Main Methods:

• Utilized AutoSegmentator (nnU-Net) for automatic organ extraction from DICOM data.
• Integrated 3D Slicer and Blender with custom Python scripts for data processing and manual refinement.
• Generated 3D surface data (STL) and aligned it with DICOM-derived images in a shared 3D space.

Main Results:

• Successfully extracted surface data for 104 organs, allowing for manual adjustments and additions.
• Created a customizable anatomical dataset applicable to various surgical simulation types (3D printing, AR).
• The system leverages open-source software, ensuring reproducibility and accessibility.

Conclusions:

• The developed system significantly enhances the accuracy and efficiency of anatomical dataset creation for surgical simulations.
• This approach reduces development time and costs, promoting wider adoption of advanced surgical training tools.
• Empowers medical professionals to contribute directly to the creation of tailored surgical simulation resources.