Standardized Python Pipeline for Evaluating Perfusion Imaging Challenge

This project will develop a standardized Python pipeline to evaluate submissions to OSIPI perfusion imaging challenges for ASL and DSC DCE MRI methods. Researchers use different computational approaches to generate perfusion parameter maps, but there is a need for a consistent system that can validate submitted datasets run submitted code and calculate evaluation metrics in a reliable and reproducible way. The proposed pipeline will accept challenge submissions that include parameter maps and the code used to generate them. It will first validate that submitted files follow expected structures and medical imaging formats such as NIfTI and BIDS. The system will then run the submitted workflows in a controlled environment and calculate evaluation metrics such as RMSE bias and coefficient of variation. Results will be organized into reports and visualizations that allow researchers to compare the performance of different methods against reference datasets. The pipeline will be implemented in Python and designed to support both ASL and DSC DCE challenge frameworks. The architecture will focus on modular design so new datasets metrics and evaluation procedures can be added in the future. Automated testing validation checks and documentation will be included to ensure the system is reliable and easy for researchers to use and extend. This project will help improve reproducibility and transparency in perfusion imaging research by providing a clear standardized framework for benchmarking computational methods used in quantitative MRI.