Arterial spin-labeling in routine clinical practice, part 1: technique and artifacts

Abstract

The routine use of arterial spin-labeling (ASL) in a clinical population has led to the depiction of diverse brain pathologic features. Unique challenges in the acquisition, postprocessing, and analysis of cerebral blood flow (CBF) maps are encountered in such a population, and high-quality ASL CBF maps can be generated consistently with attention to quality control and with the use of a dedicated postprocessing pipeline. Familiarity with commonly encountered artifacts can help avoid pitfalls in the interpretation of CBF maps. The purpose of this review was to describe our experience with a heterogeneous collection of ASL perfusion cases with an emphasis on methodology and common artifacts encountered with the technique. In a period of 1 year, more than 3000 pulsed ASL cases were performed as a component of routine clinical brain MR evaluation at both 1.5 and 3T. These ASL studies were analyzed with respect to overall image quality and patterns of perfusion on final gray-scale DICOM images and color Joint Photographic Experts Group (JPEG) CBF maps, and common artifacts and their impact on final image quality were categorized.

Fig 1.

Normal ASL CBF map. Multisection JPEG map with color ramp representing units of mL/100 g tissue/min. No flow asymmetry or artifact is present.

Fig 2.

Robust CBF in a pediatric patient, an 8-year-old boy with elevated CBF values in gray and white matter, a normal finding in this age group. CBF has been reported to peak around 8 years of age before gradually decreasing to normal adult levels.

Fig 3.

Physiologic regional distribution of spin tag. Hyperfrontal (white arrows) and visual cortex (yellow arrow) patterns of signal intensity, normal variants on ASL CBF maps.

Fig 4.

Effect of spin tag decay during image acquisition. ASL CBF map shows decreased global signal intensity in the more rostral images compared with more inferior levels. Images are acquired from inferior to superior.

Fig 5.

Tissue masking artifact. Postgadolinium T1-weighted image reveals a tangle of enhancing vessels in the right posterior thalamic region consistent with arteriovenous malformation (yellow arrow). The lesion is markedly hyperperfused on the ASL CBF map, but a central signal intensity void is seen because of tissue-masking error (white arrow). Corresponding unmasked image confirms the source of the artifact.

Fig 6.

Transit time effects. Axial T2-weighted image through the cavernous sinuses revealed absence of flow void in the right internal carotid artery (not shown), indicating slow flow or occlusion. Restricted diffusion is present, which is consistent with watershed infarct (arrow). ASL CBF map reveals decreased flow in the right posterior watershed zone as well as linear high signal intensity representing slow flow in cortical vessels (arrowheads).

Fig 7.

Susceptibility artifact in a 5-year-old boy with previous resection of a glioma. Focally decreased signal intensity is present in the right parietotemporal region on the ASL CBF maps (arrow). The baseline magnetization map confirms the presence of metallic hardware causing magnetic field distortion (arrow).

Fig 8.

Globally increased ASL signal intensity due to artifact. Perfusion pattern appears normal except for high signal intensity in the lateral ventricles (arrows), due to shinethrough of T2-weighted signal intensity that was not adequately suppressed during only 1 volume (vol 10). Vol 11 represents a normal control-label image subtraction obtained during signal intensity averaging.

Fig 9.

Artifact secondary to motion. A peripheral ring of high signal intensity is a common finding in ASL cases degraded by motion (arrow).

Fig 10.

Artifact secondary to asymmetric coil sensitivity. There is a regional zone of high perfusion signal intensity in the right cerebral hemisphere that corresponds to the same zone of high signal intensity on the diffusion image and the M₀.