Evaluation of an AIF correction algorithm for dynamic susceptibility contrast-enhanced perfusion MRI

Area of Science:

• Neuroimaging
• Medical Physics
• Radiology

Background:

• Reproducibility of perfusion measurements in longitudinal cerebrovascular disease studies using DSC-MRI is a significant challenge.
• Existing DSC-MRI methods are limited by partial volume and saturation effects in the arterial input function (AIF).

Purpose of the Study:

• To evaluate a novel algorithm designed to overcome AIF limitations in DSC-MRI.
• To assess the algorithm's impact on perfusion parameter reproducibility and accuracy in patients with cerebrovascular diseases.

Main Methods:

• A blood flow stimulation study was conducted in 21 subjects using L-arginine.
• Perfusion parameters were calculated using block-circulant singular value decomposition (cSVD) before and after L-arginine administration.
• Arterial input functions (AIFs) were selected under three conditions: experienced rater, beginner rater (raw AIF), and beginner rater (corrected AIF).
• Reproducibility was assessed by repeated baseline measurements in a subgroup of seven subjects.

Main Results:

• Significant changes in regional cerebral blood flow (rCBF) of 9.0% (P < 0.01) were observed only when AIF correction was applied.
• The novel correction algorithm decreased the standard deviation of the difference between baseline measurements by 42% in a subgroup of subjects.

Conclusions:

• The evaluated algorithm effectively overcomes AIF-related limitations in DSC-MRI.
• This novel approach significantly improves the reproducibility and accuracy of perfusion measurements in cerebrovascular disease patients.
• The findings support the clinical utility of this algorithm for longitudinal monitoring in neuroimaging studies.