Joshua Nickerson

Brain perfusion imaging using MR is not a new idea, and indeed, both gadolinium-based techniques and noncontrast methods have been well described in the radiology literature for decades. However, in discussion with colleagues, both in academia and in the private practice world, it is apparent that, currently, dynamic perfusion scanning has become part of the routine clinical examination at only a few of the largest centers in the United States. Reasons range from lack of experience, poor postprocessing software, limited display capabilities of PACS, lack of quantification, and confusion over the utility of the information obtained. Nevertheless, DSC and dynamic contrast-enhanced (DCE) sequences have been available commercially from MR vendors for some time and have found a role in some centers for evaluation of tumor, stroke, seizure, and myriad other conditions.

Arterial spin-labeled (ASL) perfusion imaging has also been around for quite a while, but until recently has been largely confined to the research realm. However, advances in data acquisition techniques, MRI technology, and processing software have resulted in the inclusion of ASL sequences in the armamentarium prepackaged by the major MR vendors. ASL has a number of advantages over gadolinium-based methods, including (obviously) the lack of contrast, which allows for short-term repeatability and, also, at least a semi-quantitative measure of cerebral blood flow depending on the technique used. Admittedly, there are disadvantages, including a relatively limited signal-to-noise ratio and as of yet cumbersome postprocessing needed for quantification. But particularly as the latter concern is progressively overcome by software vendors, we may begin to see this technology find more and more applications in the near future—a contrast-free sequence yielding whole-brain perfusion maps that only takes on the order of 3–4 minutes has undeniable appeal!

In brief, ASL is based on magnetic labeling of the intravascular water in the neck. Following a postlabeling delay, the difference in the signal measured in the brain with and without the label is proportional to the cerebral blood flow. Application of crusher gradients may also be applied prior to signal acquisition to null intravascular signal in the imaging plane leaving only signal that has already perfused the brain. While the concept is relatively simple, limitations in labeling efficiency, radiofrequency energy deposition, and signal-to-noise have resulted in iterations beginning with pulsed ASL (PASL) through continuous ASL (CASL) to the current technique most frequently encountered, termed "pseudocontinuous ASL" (pCASL), which attempts to strike a balance between signal demands and specific absorption rate (SAR) limitations.

Here at the University of Vermont MRI Center for Biomedical Imaging we are currently applying pCASL to a number of innovative questions. Lack of gadolinium allows for study of pregnant patients at risk of, or with a history of, pre-eclampsia or eclampsia. At the most recent meeting of the ASNR we presented our data using pCASL to perform quantitative functional MRI studies using standard language and motor tasks. Researchers here at UVM also are using pCASL to evaluate blood flow in addiction states including nicotine and food addiction in obesity. Hockey is big here in the north country, and head injuries, as a result, are also being considered with pCASL perfusion scans.

In this month’s issue of AJNR News Digest, we highlight a few of the most interesting and informative papers from the Journal published during the last few years on ASL. For those who are unfamiliar with ASL, an excellent 3-part review paper by Maldjian et al begins with a primer on the technical considerations and underlying physics of the method, and goes on in the second and third installments to summarize hyperperfusion and hypoperfusion patterns. An interesting hyperperfusion pattern is elucidated by Pollack et al in association with migraine headaches—a clinical indication encountered on a daily basis by neuroradiologists. Another study by Pollack et al examines the loss of autoregulation seen after anoxic insult as manifest by hyperperfusion on ASL. In the continuing quest to use noninvasive means to grade neoplasms, Furtner et al apply ASL to the spectrum of astrocytic neoplasms with promising results. ASL may be used to detect the high flow of arteriovenous malformations and ateriovenous fistulas as described in a study by Le et al. Taking the technique a step further, Hirai et al segment the labeling plane into the individual vessels in the neck to define vascular territories and tumor feeders to optimize surgical planning.

Image modified from: Deibler AR, Pollock JM, Kraft RA, et al. Arterial Spin-Labeling in Routine Clinical Practice, Part 1: Technique and Artifacts.