Angiographic Techniques in Neuroradiology

In recent years various new cross-sectional imaging techniques have evolved, which makes choosing the right angiographic technique for a given purpose more and more complicated. As various techniques, either single or combined, may provide the clinically relevant information, the final diagnostic algorithms applied will be determined by various factors such as technical infrastructure (scanner, workstation), access throughout the week, clinical pathways presently used, and the hospital’s size and stroke care level (stroke unit, neurosurgery). As a result of the ongoing technical improvements, only specialized investigators, e.g., neuroradiologists or vascular neurologists, are now able to tailor the imaging protocol to the specific clinical questions of the transferring physician as well as to the patient’s needs.

In this chapter, the angiographic methods in clinical use— digital subtraction angiography (DSA), MR angiography (MRA), and CT angiography (CTA)—are presented, considering their historical development, technical aspects, and their main advantages and disadvantages. In addition, image data postprocessing techniques such as multiplanar reformatting (MPR), maximum intensity projection (MIP), and volume rendering (VR) have become indispensable for extracting and visualizing pertinent clinical information that is not available from numerous cross-sectional images. Yet, neuroangiographic data acquisition (CTA, MRA, and DSA) as well as postprocessing techniques (MPR, MIP, and VR) may be misleading in various ways due to inherent pitfalls and limitations. Image artifacts may mimic high-grade vessel stenosis where there is none, whereas inappropriate data postprocessing may hide such stenosis. Whenever treatment decisions are based on cross-sectional imaging, a thorough knowledge of these limitations is of paramount importance.

The current status of neuroimaging in regard to the major issues of stroke, intracranial hemorrhage, vessel wall pathology and stenoses as well as in sinus venous thrombosis is discussed in the context of the different imaging techniques. Further angiographic aspects are focused upon within the selected case histories.

Conventional angiography is a technique that uses radiographs to visualize the lumen of blood-filled structures, such as the cervical and cerebral arteries. The term angiography is derived from the Greek words angeion, meaning “vessel,” and graphien, meaning “to write” or “to record.” The term angiograph or, more commonly nowadays, angiogram, denotes the radiographic image. The use of a radiodense intravascular contrast agent is required to outline the vessel structures; an intravenously administered contrast medium containing nonionic iodine is now commonly used.

Egas Moniz, a Portuguese neurologist who was one of the most important pioneers in this field, developed cerebral angiography in 1927 as a way of using contrasted cerebral X-ray angiograms for the assessment of various central nervous system (CNS) diseases of neoplastic and vascular origin (Petit-Dutaillis 1954). The first cerebral angiographies were performed in 1896 by cadaver testing, as no contrast medium suitable for use in living patients was available at that time. When Moniz performed his first angiograms, the carotid artery had to be opened for access. In 1929 the German physician Werner Forssmann, who worked at the Charité hospital in Berlin, performed the first cardiac catheterization in a self-experiment and received the Nobel Prize in 1956 for this ground-breaking discovery (Forssmann 1954).

The use of a small intravascular tube as well as direct percutaneous vessel puncture, introduced by the Swedish radiologist Sven Ivar Seldinger in 1953, are the hallmarks of modern angiography, as no sharp and potentially harmful introductory devices need to be left inside the vessel lumen (Seldinger 1953).

Briefly, access is achieved by puncture of the femoral artery using the Seldinger technique. After intravascular placement of the introducer sheath, a hydrophilic guide wire is advanced to the level of the aortic arch, followed by a catheter. Assisted by the wire, an endhole-catheter is then moved upwards for diagnostic angiography purposes into the internal carotid and/or vertebral arteries, depending on the particular clinical question. If anatomic variants or significant proximal supra-aortic vessel pathology are expected, an aortic arch angiogram by using the so-called pigtail catheter may be obtained first.

Although the spatial resolution of DSA with a pixel size of ~0.3 mm (Villablanca et al 2007) is closely approximated by multislice CTA (0.35 mm) contrast-to-noise ratio is superior for DSA, thus improving delineation of very small vessels. In particular, primary angiitis of the CNS tends to affect medium-sized and smaller vessels, making the best possible vessel delineation of major importance.

The total amount of injected iodinated nonionic contrast medium varies according to the particular procedure, e.g., whether or not a brachiocephalic angiogram (BCAG) is required. A maximum of ~80 mL contrast medium (300 mg I/mL) is required for a four-vessel angiogram including BCAG.

DSA not only provides pathomorphologic information, but visualizes intracerebral hemodynamics (e.g., collateral blood flow or arteriovenous shunting) by repeatedly acquiring multiple images per second.

There are different ways of dissolving a thrombotic intracranial vessel occlusion, which may be subdivided into intravenous and intra-arterial pharmacological clot lysis as well as mechanical thrombectomy.

Since the results of the National Institute of Neurological Disorders and Stroke t-PA study, followed by its U.S. Food and Drug Administration (FDA) approval in 1996, intravenous thrombolysis has remained the mainstay of stroke treatment for almost two decades. Yet, it has become obvious that certain clot types (defined by clot burden, site of vessel occlusion, clot consistency, clot age, etc.) are less prone to intravenous lysis and consequently associated with worse clinical outcome. Occlusion of major vessels such as the M1-middle cerebral artery (MCA) segment as well as carotid and basilar artery occlusions have been recognized as critical sites, requiring additional approaches such as intra-arterial lysis with or without mechanical clot disruption using a microwire or micro-catheter.

Even then, a significant number of clots could not be dissolved, triggering the development of other mechanical thrombectomy devices, of which the first-generation retrieval system MERCI, as well as the second-generation stent retrievers TREVO and SOLITAIRE, have gained FDA approval. TREVO and SOLITAIRE have proved superior to the MERCI retrieval system in recent studies (SWIFT, TREVO-2) (Nogueira et al 2012, Sheth et al 2015). In February 2015, results of the first randomized clinical trial showing efficacy and safety of mechanical thrombectomy as compared with intravenous lysis solely in the proximal anterior circulation were published, representing a breakthrough for endovascular stroke treatment (Berkhemer et al 2015). Soon after that, various other studies (ESCAPE, SWIFT PRIME, EXTEND-IA) were halted because interim analysis significantly favored endovascular stroke treatment by thrombectomy (Campbell et al 2015, Goyal et al 2015, Saver et al 2015).

Apart from clot manipulation by retrieval systems, suction thrombectomy (also called the direct aspiration first pass technique [ADAPT], with or without the use of an external pump system, has regained attention. Improved catheter technologies make it possible to place large-bore catheters at or right in front of the occlusion site, with average recanalization times of ~30 minutes (Park 2015) or even less. At present, a sequential approach using the ADAPT technique, followed by stent retriever thrombectomy if ADAPT fails, seems to be the most promising and cost-effective procedure (Turk et al 2014).

DSA is considered to be the gold standard whenever the best possible detail resolution and/or assessment of cerebral hemodynamics are required.

Despite being regarded as state-of-the art, the spatial resolution of flat-panel DSA still does not compete with flat-panel volume CT systems. The latter are based on a CT gantry, yielding spatial resolutions as low as 150–200 mm. Flat-panel DSA is restricted to inferior spatial resolution because of geometric inaccuracies, blurring due to the X-ray converter and reconstruction filter, and dose considerations. A CT slip-ring technique is not available for 3D rotational angiography, thus disabling continuous C-arm rotation for time-resolved imaging in flat-panel volume CT.

As described previously, vascular imaging issues are constantly changing and cerebrovascular hemodynamics may now be assessed noninvasively by time-resolved MRA or CTA techniques. Still, none of these alternative four-dimensional (4D) techniques can provide spatial and temporal resolution comparable to DSA, as might be required in the assessment of complex AVMs. If interventional procedures are necessary within a short period of time, such as in patients with intracranial major artery occlusion, invasive catheter angiography remains the first-line diagnostic and therapeutic modality.

Contraindications to DSA are comparable to CTA as they are generally related to the use of iodinated contrast medium as well as ionizing radiation. Renal insufficiency, hyperthyroidism, and iodine allergy are common although not absolute contraindications to DSA (as well as CTA). Pregnancy is also a contraindication for DSA unless the angiography is of vital importance for the mother.

Procedural complications have been reported to amount to a 1–2.3% overall incidence of neurologic deficit and a 0.4–0.5% incidence of persistent deficits following cerebral angiography (Heiserman et al 1994, Leffers and Wagner 2000). However, non-neurologic complications such as local hematomas were observed in 14.7% of procedures in the Leffers and Wagner study. Clinically silent embolisms were encountered in up to 44% of patients undergoing DSA, if suffering from vascular risk factors (Bendszus et al 1999). These figures might seem significant, but in the Leffers and Wagner study, the majority of non-neurologic complications were minor groin hematomas. In addition, Burger et al (2006) showed in a study of DSA-related complications in pediatric neuroangiography, which is technically more demanding than in adults, that in experienced hands this is a low-risk procedure (they reported no intraprocedural complications in 241 consecutive pediatric cerebral angiograms). Also other studies have proven DSA to be a safe procedure, especially if performed in centers that carry out large numbers of procedures (Fifiet al 2009, Thiex et al 2010).

When compared with biplane DSA, rotational 3D DSA resulted in significantly lower skin doses; up to four times lower peak skin doses were reported by Schueler et al (2005). ED measurements in flat-panel rotational DSA and multislice CT showed identical dose figures for 3D angiography as compared with biplane (2D) DSA, and comparable ED figure ranges for multislice and flat-panel cranial CT (Struffert et al 2014). The dose received during CTA has also been described as equivalent to ~15 minutes of fluoroscopy time, e.g., somewhat greater than typically required for routine diagnostic DSA but not outside the safe limits for diagnostic radiologic assessments (Chappell et al 2003).

DSA may also be performed by intravenous injection and this technique was first developed in the late 1970s. This type of DSA compares an X-ray image of a region of the body before and after a radio-opaque iodine-based dye has been injected intravenously into the body. Tissues and blood vessels on the first image are digitally subtracted from the second image, leaving a picture of the artery of interest. Despite its obvious advantages compared with the intra-arterial catheter technique it has not gained much significance mainly because of its inferior resolution compared with conventional DSA and the advantages of MRA, CTA, and duplex sonography as alternative noninvasive procedures. Recently, however, a DSA technique with intravenous contrast medium injection has been proposed, using a biplane flat-detector angiographic system; this technique provided high-resolution assessment of the intracranial vasculature in general (Saake et al 2013) and also showed promising results in evaluating aneurysm remnants after neurosurgical clipping (Gölitz et al 2012).

Nuclear magnetic resonance (NMR) imaging, the original term for MRI, designates a radiation-free imaging technique based on the proton nucleus resonance in response to a radio-frequency pulse, emitted and received by so-called “coils” within a dedicated scanner. The basic physical principle was described by Bloch and Purcell in 1946, but it was Paul Lauterbur and Peter Mansfield (the two shared the Nobel Prize in 2003 for their MRI achievements) who in the early 1980s developed the underlying principle into a technique that allowed the generation of images of the human body (Andrew 1992). For image generation and spatial encoding, magnetic gradients are applied together with the radio-frequency pulse. The resulting data are recorded in a 2D or 3D image matrix and the image itself then is created by applying an algorithm called Fourier transformation. By variation of scanning parameters, tissue contrast can be altered and enhanced in various ways to assess different properties.

In MR angiography (MRA) the introduction of FLASH sequences (Fast Low Angle Shot) in 1985 by Frahm et al (1986) allowed significant shortening of MRI measurement times, by combining a gradient echo sequence (using a low-flip angle pulse) with a rapid sequence repetition. In 1996 time-resolved ce 3D MR angiography (also called 4D-MRA) was introduced (Korosec et al 1996), with ce 4D MRA now being a widely used noninvasive dynamic angiography technique.

In recent years advances in scanner and software technology, as well as health-care reimbursement issues, have shifted boundaries in regard to selecting the most cost-effective procedure.