Contrast-agent-based perfusion MRI code repository and testing framework: ISMRM Open Science Initiative for Perfusion...

Area of Science:

• Medical Imaging
• Software Engineering
• Quantitative Analysis

Background:

• Software significantly influences quantitative perfusion MRI values.
• Lack of standardized code sharing and testing hinders reproducibility in perfusion MRI research and clinical trials.

Purpose of the Study:

• To establish a community-driven, centralized repository for open-source code in contrast-based perfusion imaging.
• To integrate an open-source testing framework to ensure code quality and consistency.

Main Methods:

• A repository was created on GitHub (OSIPI website).
• Python was selected as the programming language.
• An automated unit-testing framework was developed to assess code outputs, including visual results.

Main Results:

• The repository contains 86 perfusion processing implementations from 12 contributors, covering DCE- and DSC-MRI.
• Tests were developed for 52 implementations across five analysis steps.
• While T1 mapping and AIF functions showed consistent outputs, pharmacokinetic models exhibited parameter variations across different implementations.

Conclusions:

• The OSIPI repository offers a new model for collaborative code sharing and testing in perfusion imaging.
• This initiative enhances code reuse and benchmarking, ultimately promoting greater reproducibility in quantitative perfusion MRI.