Voxel-specific brain arterial input functions from dynamic susceptibility contrast MRI and blind deconvolution in a group of healthy males

Abstract

Arterial input functions may differ between brain regions due to delay and dispersion effects in the vascular supply network. Unless corrected for, these differences may degrade quantitative estimations of cerebral blood flow in dynamic susceptibility contrast magnetic resonance perfusion imaging (DSC-MRI). To investigate in a healthy population (n=44) the properties of voxel-specific arterial input functions that were obtained using a recently published blind estimation approach. The voxel-specific arterial input functions were qualitatively and quantitatively assessed, through visual inspection or by comparing time-to-peak (delays) and peak amplitude (dispersion) values between eight regions of the brain. Furthermore, they were compared to arterial input functions selected manually in the middle cerebral artery (MCA), where normally no delay or dispersion of the contrast agent was expected. The estimated voxel-specific arterial input functions varied between brain regions. Differences in delays and dispersion were larger within one brain region among all participants than between regions in one participant. A good correlation was typically found between the estimated voxel-specific arterial input functions and the manually selected arterial input functions in the MCA region. Given knowledge of neurovascular anatomy, the current blind approach seemingly produced reasonable estimates of voxel-specific arterial input functions. In addition to potentially reducing quantification errors in DSC-MRI, these user-independent voxel-specific arterial input functions could be useful for visualizing abnormal blood supply patterns in patients.