Effects of systematic partial volume errors on the computation of mean gray matter cerebral blood flow with Arterial Spin Labeling MRI


Effects of systematic partial volume errors on the computation of mean gray matter cerebral blood flow with Arterial Spin Labeling MRI

Petr, J.; Mutsaerts, H. J. M. M.; de Vita, E.; Steketee, R. M. E.; Smits, M.; Nederveen, A. J.; Hofheinz, F.; van den Hoff, J.; Asllani, I.

Objectives: Partial volume (PV) correction is an important step in arterial spin labeling (ASL) MRI used to separate perfusion effects from structural, and to calculate the mean gray-matter (GM) perfusion. There are currently three main methods to perform that: (1) including only voxels with GM volume above a preset threshold (GM-Threshold); (2) using weighted voxel contribution combined with thresholding (GM-Weighted); or (3) applying a spatial linear regression algorithm (PVEc). In all cases, GM volume is obtained from PV maps extracted from T1w images. As such, PV maps contain errors due to the difference in readout-type (a major source of geometric distortions) and spatial resolution between ASL and T1w images. Here, we estimated these errors and evaluated their effect on the performance of each PV-correction method.
Materials and Methods: Twenty-two volunteers were scanned using 2D EPI and 3D spiral ASL. For each PV-correction method, GM CBF was computed using PV maps simulated to contain estimated errors due to geometric distortions and resolution mismatch. Results were analyzed to assess the effect of each error on extraction of GM CBF from ASL data.
Results: Geometric distortion had the largest effect on the 2D EPI data whereas resolution mismatch on the 3D spiral. The PVEc method outperformed the GM-Threshold even in the presence of combined errors. The quantitative advantage of PVEc was 16% without and 10% with the combined errors for both readouts. Consistent with theoretical expectations, for error-free PV maps, PVEc method extracted the true GM CBF. In contrast, GM-Weighted overestimated GM CBF by 5% whereas GM-Threshold underestimated it by 16%. The presence of PV-map errors decreased the calculated GM CBF for all methods.
Conclusion: The quality of PV maps presents no argument for preferring the GM-Threshold method to PVEc in clinical applications of ASL.

Downloads

Permalink: https://www.hzdr.de/publications/Publ-27324
Publ.-Id: 27324