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

Abstract

Dynamic susceptibility contrast MRI (DSC-MRI) tends to return elevated estimates of cerebral blood flow (CBF) and cerebral blood volume (CBV). In this study, subject-specific calibration factors (CFs), based on steady-state CBV measurements, were applied to rescale the absolute level of DSC-MRI CBF. Twenty healthy volunteers were scanned in a test-retest approach. Independent CBV measurements for calibration were accomplished using a T1-based contrast agent steady-state method (referred to as Bookend), as well as a blood-nulling vascular space occupancy (VASO) approach. Calibrated DSC-MRI was compared with pseudo-continuous arterial spin labeling (pCASL). For segmented grey matter (GM) regions of interests (ROIs), pCASL-based CBF was 63±11ml/(min 100g) (mean±SD). Nominal CBF from non-calibrated DSC-MRI was 277±61ml/(min 100g), while calibrations resulted in 56±23ml/(min 100g) (Bookend) and 52±16ml/(min 100g) (VASO). Calibration tended to eliminate the overestimation, although the repeatability was generally moderate and the correlation between calibrated DSC-MRI and pCASL was low (r<0.25). However, using GM instead of WM ROIs for extraction of CFs resulted in improved repeatability. Both calibration approaches provided reasonable absolute levels of GM CBF, although the calibration methods suffered from low signal-to-noise ratio, resulting in weak repeatability and difficulties in showing high degrees of correlation with pCASL measurements.