Quantitative cerebral perfusion using dynamic susceptibility contrast MRI: Evaluation of reproducibility and age‐ and gender‐dependence with fully automatic image postprocessing algorithm

Abstract

A novel approach for quantifying cerebral blood flow (CBF) is proposed that combines the bookend technique of calculating cerebral perfusion with an automatic postprocessing algorithm. The reproducibility of the quantitative CBF (qCBF) measurement in healthy controls (N ‫؍‬ 8) showed a higher intraclass correlation coefficient (ICC) and lower coefficient of variation (COV) when calculated with automatic analysis (ICC/COV ‫؍‬ 0.90/0.09) than when compared to conventional manual analysis (ICC/COV ‫؍‬ 0.58/0.19). Also, the reproducibility in patients (N ‫؍‬ 25) was successfully evaluated with the automatic analysis (ICC/COV ‫؍‬ 0.81/0.14). In 175 consecutive clinical scans, we found 3.0% and 7.4% of qCBF decrease per decade in white matter (WM) (21.5 ؎ 6.66 ml/100 g-min) and gray matter (GM) (49.6 ؎ 16.2 ml/100 g-min), respectively. Cerebral blood volume (CBV) showed a significant 3.7% decrease per decade in GM (3.00 ؎ 0.94 ml/100 g) but not in WM (1.69 ؎ 0.40 ml/100 g). Bolus tracking cerebral perfusion has been shown to reflect the underlying pathophysiology in a variety of diseases including cerebrovascular occlusive disease, stroke, central nervous system tumors, and potentially Alzheimer's disease (1-5). Parametric images of cerebral perfusion are calculated by analyzing the tracer kinetics of a known tracer, whether it is: radio-labeled water in positron emission tomography (PET), iodinated contrast agent in computed tomography (CT), spin-labeled water in arterial spin labeling, or a paramagnetic contrast agent in dynamic susceptibility contrast (DSC) MR imaging. While radiolabeled PET is the standard of reference for quantification of cerebral perfusion, the necessity of a cyclotron limits its availability. CT has the potential to quantify perfusion; however, iodinated contrast and large doses of radiation are required. This is particularly problematic for frequent follow-up or in certain populations, such as pediatric patients in which the radiation dose should be minimized MR-based parametric images of relative cerebral blood flow (CBF) and cerebral blood volume (CBV) images using DSC analysis are commonly used to assess changes in cerebral perfusion. The development of quantitative MRbased perfusion (i.e., in ml/100 g-min) has been identified by the American Heart Association as a priority for treatment of acute stroke (6). However, quantitative perfusion MR with DSC analysis has been elusive and challenging (7-10).One approach to quantification of perfusion using MRI has been to assume a population-averaged value for normal white matter (WM) and determine quantitative values relative to "normal-appearing white matter." This method assumes that the average CBF value in WM is not significantly dependent on age, gender, and physiologic conditions, and calculates the population-averaged correction factor by setting WM CBF values to 22 ml/100 g/min (11-13). However, Mukherjee et al. (14) have shown that the individual subject-to-subject correction factor is favored over the population-averaged correction fac...

Absolute quantification of perfusion by dynamic susceptibility contrast MRI using Bookend and VASO steady-state CBV calibration: a comparison with pseudo-continuous ASL

“…Bookend Bland-Altman plots showed larger SDs than in a previous test-retest study by Shin et al [36], who observed an average difference±SD of 1.4±6.9 ml/(min 100g). On the other hand, Crane et al [37] For the Bookend method, the use of GM ROIs resulted in a repeatability that was fairly similar to the previous observations by Shin et al [36].…”

Decreased Frontal Lobe Gray Matter Perfusion in Cognitively Impaired Patients with Secondary-Progressive Multiple Sclerosis Detected by the Bookend Technique

“…Previously published studies examining the variability and repeatability of results generated using the bookend technique have reported a testretest intraclass coefficient of 0.90 and coefficient of variation of 0.09. 15,16,20 Potential limitations include differences in WM properties between patients with MS and healthy controls that could theoretically invalidate the bookend assumptions. However, no significant differences were identified between the expected values from the water correction factor model and the observed data points for healthy controls, impaired patients with MS, and nonimpaired patients with MS.…”

“…[14][15][16] These "bookend" scans allow DSC calibration by quantifying parenchymal and blood pool T1 changes while correcting for intrato-extravascular water exchange. 15 The bookend technique does not require normalization for comparative measurements and is reproducible, reliable, 15,16 and accurate compared with PET. 17 The goal of this study was to quantify differences in bookend-derived cerebral perfusion between cognitively impaired and nonimpaired patients with SPMS.…”

Cerebrovascular Occlusive Disease: Quantitative Cerebral Blood Flow Using Dynamic Susceptibility Contrast MR Imaging Correlates with Quantitative H2[15O] PET

“…The bookend technique has been previously described (12,14,15) but will Dynamic susceptibility contrast (DSC) MR imaging has shown promise in calculating CBF by measuring the susceptibility-induced signal loss with the influx of a contrast agent bolus (7,8). However, due to the inherent complexities of susceptibility imaging, the derived values have only been confirmed to accurately show relative, rather than quantitative, values (7,(9)(10)(11).…”