ABSTRACT
The diagnosis of vascular diseases of the brain, besides in some cases well-established clinical signs, has largely relied on the use of conventional angiography. Neuroimaging was profoundly revolutionized by the development of CT, then by the advent of fast MRI [6]. Indeed, although most techniques until then had mostly provided indirect signs of pathological conditions, or very semi-invasive like angiography, these two techniques allowed to visualize the brain in situ and in vivo like never before. Although CT remains the cornerstone for most neuroradiological workups, MRI has proven to be a much superior tool for the assessment of diseases of the nervous system. Indeed, due to the absence of X-rays, the capacity to obtain directly multiplanar images and to perform sequences that provide different contrasts of the tissues visualized, the technique has shown itself superior. However, for a rather long time after its introduction, MRI was a rather slow technique and complete examinations of the neuraxis could take up to an hour or even more. It was with the clinical implementation of at first echoplanar techniques, then of parallel imaging schemes that imaging times could be reduced with even an increase in signal and resolution. It is also becoming more clear that higher magnetic fields such as 3 T are the standard requirements for centers dealing with advanced treatment of cerebrovascular disorders. MR, due to its inherent sensitivity to motion, and especially to vascular motion, was predestined for the development of MRA techniques [5]: this happened early on in the course of the history of MR; however, a lot of improvements were necessary because early images took a long time to acquire, required a lot of postprocessing, and did not provide full coverage of the brain. With the development of fast sequences, the development of stronger gradients and improved coils, it was possible to improve resolution to an important degree. Initially, T1 and T2 images were acquired with spinecho sequences that required long acquisition times; time was an even bigger issue if one was to consider multiplanar imaging. Indeed MRI has the distinct
advantage of being able to obtain directly images in multiple planes in any contrast, but this is of course at the cost of additional imaging time; thus, at times in cases where an extensive imaging protocol had to be performed, it was possible for MR examinations to run well beyond an hour. This, together with the very motion-sensitive nature of the technique, would make the examination only accessible to patients who were able to cooperate and remain still for extensive periods of time. Indeed, it was the advent of echo-planarimaging that was a revolution [6] that allowed to speed up imaging and to be able to perform complex examinations within an acceptable period of time. Not only did conventional imaging become faster, all functional techniques that had remained somewhat dormant would develop quickly and become established as clinical tools. Diffusion imaging that was devised by Le Bihan consists of a simple modification of a spinecho sequence, which is sensitized to movement by the application of two gradient pulses [7]: this makes the image extremely sensitive to changes in water motility in the tissue that is being investigated. The first real application is to ischemic stroke. It has then been applied to a multitude of diseases, but its capacity to detect with certainty acute ischemia is unparalleled: indeed because of this it is very often used also to monitor neurovascular interventions such as thrombectomy [8], coiling, or embolization. In additional to diffusion, perfusion techniques were improved to a new point [9]: these could now be performed using either contrast material (with mostly T2* images) or without contrast (with arterial spin-labeling techniques): one could now obtain reliable maps of relative cerebral blood flow and volume. These techniques, in addition to diffusion-weighted imaging (DWI). allowed clinicians to establish an initial, slightly rudimentary, but working model of an ischemic penumbra in human stroke [10]. The advantages of MRI over CT were multiple from the start; besides being nonirradiating, it provided multiplanar images that would cover the whole brain. One technique to further profit from fast imaging has been functional activation imaging of the brain: here T2* images are very sensitive to changes in oxygenation that occur with activation. During the performance of a task (or paradigm), there is an increase in blood flow, whereas extraction remains stable: this will cause a decrease in deoxygenated blood and thus a signal increase on the images-this is the so-called BOLD effect [11]. Functional imaging with MR can now be adapted for any task involving any region of the brain. However, for most therapeutic purposes, motor or language paradigms will be mostly used. Additionally, noncontrast perfusion techniques
can allow to improve the visualization of brain hemodynamics in function and dysfunction [11,12]. The use of conventional gadolinium chelates or of compounds with higher relaxativity may also impact the way we see both the vascular system and the bloodbrain barrier [13]. A further refinement of diffusion imaging, diffusion tensor imaging (DTI), allows to reconstruct the direction and strength of water motion along the cerebral white fibers [14]: this allows to perform tractography that brings another information in addition to anatomy and function that we have just discussed. Further refinements in T2* images such as susceptibility-weighted imaging (SWI) have become a great help in the diagnosis of diseases of the brain [15] where blood degradation products are involved [16]. Taken together, all these techniques allow us to acquire information about anatomy, pathology, and function of the brain.
