Container applications for the development and integration of virtual imaging platforms.

Area of Science:

• Medical Imaging
• Computational Phantoms
• Virtual Imaging Trials

Background:

• Virtual imaging trials (VIT) leverage realistic anatomy models and simulations for virtual image interpretation.
• Widespread adoption of VIT requires unified methods for reliable application across imaging studies.

Purpose of the Study:

• Develop a containerized environment to improve collaboration and interoperability across VIT platforms.
• Integrate key components of OpenVCT and VICTRE breast imaging platforms for direct module comparison.

Main Methods:

• Utilize Docker for containerized environment creation, ensuring cross-platform integration.
• Develop wrappers for OpenVCT and VICTRE to simplify setup, execution, and compatibility.
• Augment breast anatomy with OpenVCT's simplex-based method for enhanced texture modeling.
• Compare Monte Carlo (VICTRE) and raytracing (OpenVCT) techniques for lesion and calcification simulations.

Main Results:

• Containerized approach simplified simulation platform setup and execution.
• Visually comparable breast images generated using VICTRE and integrated OpenVCT methods.
• Power spectra analysis showed minor differences (0.2) in high-frequency components, with beta estimates near 3.
• Acquisition geometry variations impacted lesion visualization, showing slight differences between Monte Carlo and raytracing.

Conclusions:

• Containerized environments enable robust, reproducible cross-platform comparisons and hybrid approaches in medical imaging.
• Integration of VICTRE and OpenVCT provides data augmentation and method selection resources.
• This work establishes a foundation for credible, reproducible, and accessible future VIT advancements.