A dynamic susceptibility contrast MRI digital reference object for testing software with leakage correction: Effect...

Area of Science:

• Medical Imaging
• Biophysics
• Computational Modeling

Background:

• Dynamic Susceptibility Contrast (DSC)-MRI provides quantitative biomarkers like relative cerebral blood volume (rCBV) for brain tumor assessment.
• Post-processing leakage correction is vital for accurate rCBV measurements, especially in tumors where contrast agent extravasation occurs.

Purpose of the Study:

• To develop Digital Reference Objects (DROs) for validating DSC-MRI post-processing software, specifically leakage correction using the Weisskoff model.
• To investigate the impact of different background signal time curves on the accuracy of leakage correction.

Main Methods:

• Generated three DROs using the Weisskoff model with varying rCBV and K2 (leakage parameter) levels.
• Implemented three background types: brain-like, constant signal sphere, and white matter (WM) signal sphere.
• Analyzed DROs with FDA-cleared software, assessing leakage correction with and without brain segmentation.

Main Results:

• Leakage correction accuracy was verifiable with brain-like and WM-simulating sphere backgrounds.
• Constant signal sphere backgrounds yielded poor performance for leakage correction.
• The study confirmed that leakage correction with brain segmentation resulted in the lowest error for the tested software.

Conclusions:

• DSC-MRI DROs, when biophysically matched to post-processing software, are effective for validation.
• Accurate simulation of background signals is essential for generating the required reference time curves.
• The interplay between DRO design and software-based brain segmentation for reference time curve extraction requires careful consideration.