Comparison of Normalized Cerebral Blood Flow Between Different Post-processing Methods of Dynamic Susceptibility Contrast Perfusion-weighted Imaging and Arterial Spin Labeling in Gliomas with Different Grading

Background: Two post-processing methods of dynamic susceptibility contrast perfusion-weighted imaging (DSC-PWI), arterial input function (AIF) and gamma-variate fitting (GVF), can both derive cerebral blood flow (CBF). Moreover, AIF can provide T2* and T1 leakage indicators. This study aimed to compare the consistency of normalized CBF between different post-processing methods of DSC-PWI and arterial spin labeling (ASL) in gliomas, and take the quantitative metrics percentage of signal recovery (PSR) as a reference to verify the value of T2* and T1 leakage indicators in characterizing leakage effect and evaluating the grading of gliomas.

Methods: From 1 January 2020 to 15 December 2023, 56 consecutive inpatients were retrospectively enrolled, comprising 24 patients with low-grade glioma (LGG) and 32 patients with high-grade glioma (HGG). The normalized CBF was derived from AIF, GVF and ASL. The T2* and T1 leakage indicators of AIF were graded by 4-point scale. The consistency and difference of normalized CBF between DSC-PWI and ASL were tested by linear correlation/regression analysis, Bland-Altman plots, and Student's -test. The correlation between the difference of point for both leakage indicators and PSR was tested by Spearman correlation analysis, then inter-group difference of PSR was compared by -test. The intra-group and inter-group differences of point for T2* and T1 leakage indicators were compared by χ/Fisher's exact test.

Results: The normalized CBF derived from AIF and GVF were correlated with ASL in both groups (all r>0.7, all P<0.001), and linear regressions were not significantly different in each group (all P>0.05). The difference of normalized CBF between ASL and AIF in the HGG group was larger than that in the LGG group (P=0.017); however, the difference of normalized CBF between ASL and GVF was not significant (P=0.085). The strong correlation between the difference of point for both leakage indicators and PSR was verified (r=-0.739, P<0.0001), and the HGG group was shown to have higher PSR compared with the LGG group (t=2.043, P=0.04). The comparison of intra-group and inter-group differences in T2* leakage and T1 indicators showed that the HGG group was more prone to T1 leakage than the LGG group (P<0.05), and weight of T1 leakage was greater than that of T2* leakage (χ=11.28, P=0.004).

Conclusions: The normalized CBF derived from DSC-PWI has good consistency with ASL in gliomas, regardless of post-processing methods. GVF can provide less bias of normalized CBF between HGG and LGG groups. However, T2* and T1 leakage indicators of AIF can be utilized as a surrogate of PSR to characterize both leakage effects and evaluate glioma grading.