Diagnostic Imaging
December 2000

CT Advanced: Cerebrovasculature

The use of noninvasive, low-cost imaging to evaluate patients with cerebrovascular diseases has increased rapidly.1 Diagnosis, characterization, and treatment planning for patients who present with intracranial aneurysms is one of the most important uses of noninvasive imaging. Intracranial aneurysms are common,2 and when complicated by rupture and aneurysmal subarachnoid hemorrhage, carry a poor prognosis, with a worldwide fatality rate of about 45%.3 Little is known about the natural history of aneurysms, but studies reported in 1998 suggest the risk of spontaneous rupture may be lower than previously thought.4

No matter what treatment option is selected, brain aneurysms require skill and judgment to treat successfully. The traditional strategy has been surgical exposure of the region via a craniotomy and neurosurgical clipping of the aneurysm neck under direct visualization. The morbidity and mortality associated with this surgery is considerable, ranging from 4% to 13%.5

In the last decade, development of embolic materials that can be deposited into the aneurysm sac via a percutaneous, transarterial approach has opened up the possibility of treatment using this less invasive method. The embolic material replaces the blood in the sac lumen, leading to stasis of flow and, ideally, complete and permanent aneurysm occlusion. This approach is also associated with some patient risks and morbidity.

The criteria for selecting patients for surgical or endovascular treatment are complex and constantly evolving and, therefore, remain controversial. Accurate quantification and complete aneurysm characterization are emerging as important prerequisites to determining suitability for either treatment option. For cases treated by endovascular coiling, prior work has shown that the frequency of complete aneurysm occlusion is related to the density of coil packing. This density is in turn strongly dependent on the geometry of the aneurysm sac.6 Aneurysms with neck sizes of 5 mm or greater or dome-to-neck ratios of less than two have complete occlusion rates of only 16%.7 Aneurysms that incorporate important arteries into the sac may not be treatable by endovascular means, because occlusion of the sac would necessarily occlude the ostium of the incorporated artery, leading to infarction of the brain tissue supplied by that vessel.

The absence of extensive calcium or thrombus at the aneurysm neck, an accessible location, and absence of multiple incorporated arterial segments are generally requirements for standard neurosurgical clipping.8 Direct clipping of aneurysms exhibiting mural thrombus or calcium may be hazardous. Calcification of the aneurysm wall at the neck and mural thrombus within the aneurysm lumen act as physical barriers that may prevent a neurosurgical clip from closing completely about the neck of the aneurysm. In addition, calcifications may fracture and lead to aneurysm or parent artery laceration. In these cases, the patient can be treated surgically with a bypass procedure or undergo embolization if considered a good candidate for this approach. Arterial bypass with trapping of the aneurysm is also necessary in cases in which the aneurysm cannot be safely collapsed, when the neck is inaccessible, or when the parent artery cannot be reconstructed. Some patients with complicated lesions are untreatable by surgical or endovascular means.

DSA LIMITATIONS

The traditional tool for triage of patients with suspected intracranial aneurysm has been catheter angiography.9 Digital subtraction angiography (DSA) has serious limitations, including the need for arterial puncture and intra-arterial catheter manipulation, and difficulty in obtaining the ideal angiographic projection showing the desired vascular anatomy is common. This occurs when the contrast-opacified sac obscures the neck, when normal vessels are superimposed over the region of interest, or when the C-arm fluoroscope reaches its rotational limits. DSA is also not reliable for detection of thrombus or calcification of the aneurysm sac. Because the presence of these materials at the neck can alter the treatment method or change the type of surgery performed, this information is essential in appropriate patient triage.

Alternatives to catheter angiography include CT angiography (CTA) and three-dimensional time-of-flight MR angiography (TOF MRA). Since the first published reports, CTA has made rapid improvements in image resolution and scan volume.10 The advent of 3-D CTA has made important contributions,11 and volume rendering has been shown to be superior to surface shaded display and maximum intensity projection techniques.12

Using 2-D multiplanar reformatting (MPR) and volume rendering techniques, we report a sensitivity and specificity of 100% for aneurysm detection, both higher than previously reported in the literature.13 The reason for the improved performance of CTA is multifactorial and primarily due to the evolution of both image acquisition protocols and postprocessing methods. Common limitations of prior published CTA series include thick slice collimation, large field-of-view, excessively high pitch, low contrast-infusion rates, limited scan coverage, failure to use a timing injection, use of only limited viewing angles during 3-D image postprocessing, and failure to use submillimeter reconstruction intervals.

Three-D TOF MRA has been substantially improved and optimized by the use of larger imaging matrices, magnetization transfer saturation, and variable flip angle excitation techniques. These changes have had the effect of allowing smaller voxel size with decreased phase dispersion, enhanced contrast between brain and arteries, and an increase in the signal intensity of distal arteries as they pass through an imaged volume.14,15 Studies report sensitivities compatible with DSA and CTA, although specificity is generally lower, when reported.

Volume rendered spiral CT angiography is a technique that uses volumetric CT image data to create high-resolution 2-D and 3-D images of the intracranial vessels. It uses a timed injection of nonionic contrast material administered through a small angiocatheter placed in a peripheral vein. The contrast material is used to opacify the brain vessels. The technique is very fast, generally requiring 60 seconds or less to perform. Typical coverage is from the high neck to the mid skull with single-detector CT systems, a distance that encompasses 99% of all berry aneurysms, and data volume is sliced into approximately 200 contiguous slices, each 0.5 mm thick. The imaged volume provides a contrast-opacified "cast" of the intracranial vessels.

The digital images are then transferred to a clinical image-processing workstation where they are automatically reformatted into 2-D multiplanar images and volume rendered 3-D objects for visual inspection. The rendering process is accomplished in less than one minute, and data inspection requires five to 25 minutes, depending on lesion complexity and the number of aneurysms present.

The 2-D and 3-D images can be used to rapidly assess the complex aneurysmal geometry and the relationship of the aneurysm to the parent artery. The 2-D images are generally used to perform submillimeter measurements of relevant aneurysm features.

CTA VS. DSA

Using this approach in a population of complex aneurysms, CTA was found to equal DSA in the detection of intracranial aneurysms, including very small lesions (Figure 1). More important, CTA was found to be superior to DSA and MRA in the detailed characterization of intracranial aneurysms.13 This assessment includes a statistically significant superiority of CTA over DSA and MRA in the characterization of arterial branching pattern at the neck, neck geometry, arterial branch incorporation into the sac or neck, mural thrombus, and mural calcification.

CTA was also found to be superior to MRA in the visualization of the superficial temporal arteries (STA). This finding is valuable in cases in which an intracranial to extracranial bypass procedure may be necessary,13 as an STA is often the donor vessel to the middle cerebral artery (MCA).

Figure 1 shows an example of the ability of CTA to visualize incorporated arterial segments. An angiogram was performed with the intention of coiling the lesion. Before any aneurysm can be treated by coiling or clipping, however, incorporation of the artery into the aneurysm sac must be excluded. Only then can the neck of the aneurysm be clipped or the aneurysm sac embolized with coils, without concern for inadvertent occlusion of the vessel. A CT angiogram was performed and clearly showed the origin of the temporal artery branch as the medial wall of the aneurysm, therefore excluding the patient as an endovascular candidate. She is being followed noninvasively by CTA for interval changes.

Figure 2 shows the comparative advantages of volume rendered spiral CTA versus TOF MRA. Because the neurosurgeon was alerted to the presence of the angular branch on the 3-D CTA images, he was able to identify and preserve the vessel and achieved successful clipping by leaving a small neck remnant at the origin of the angular artery.

Mural calcification shown at the neck in Figure 3 renders primary neurosurgical clipping hazardous, and this patient underwent an extracranial/intracranial (EC/IC) bypass procedure with successful preservation of the anterior branch of the middle cerebral artery. In the case shown in Figure 4, intraluminal thrombus extends to the aneurysm neck, indicating that primary clipping of the aneurysm neck would be difficult. The craniotomy was moved more inferiorly, so that ligation of the involved vertebral artery could be performed. The decreased intra-aneurysmal flow achieved by the ligation led to eventual complete intra-aneurysmal thrombosis. Only CTA is capable of accurately showing the location and extent of aneurysmal intraluminal thrombus and mural calcification.

We have shown that the impact of the additional information provided by volume rendered CTA images on patient management is considerable. Management decisions were affected in 85% of patients, and unique information was provided by CTA in 81% of cases.13 In 19% of cases, CTA provided no additional information over that provided by DSA and MRA. Where surgery was performed and the aneurysm visualized directly, the CTA images of the aneurysms were found to be identical or nearly identical in 94% of cases, with no major variations.13 The CTA images did not provide erroneous or misleading information.

MRA LIMITATIONS

The limitations of MRA in aneurysm evaluations are crucial and generally technical. Specifically, TOF MRA depends on uniform flow velocity and is relatively insensitive to slow, turbulent, limited, or complex flow.14 All of these conditions are present in and around aneurysms.16 Since TOF techniques are T1-weighted, fat, blood degradation products, and proteinaceous and other short T1 materials may all appear hyperintense on MRA and thus mimic flow. Under normal conditions, the smaller matrix sizes, larger fields-of-view, and asymmetric voxels required to perform routine MRA provide lower image resolution compared with DSA and CTA. Patient motion can also degrade image quality.

Finally, MR techniques poorly identify calcium from thrombus. Gadolinium-enhanced MRA has improved lesion detection and reduced signal loss due to complex and in-plane flow, but in these cases the technique suffers from frequent interference from normally enhancing venous structures, turbulent flow, and vessel margin obscuration from short T1 substances.

ASSESSMENT OF CTA

The most important of CTA's several limitations is an inability to visualize extremely small arteries, including many anterior choroidal arteries, and nearly all medial and lateral lenticulostriate and thalamoperforators. When these arteries are suspected to arise from an aneurysm, DSA must be performed. Other limitations include inability to visualize collateral flow dynamics and intra-aneurysmal flow patterns and potentially, the simultaneous visualization of arteries and veins in a single region of interest.

Multislice CT technology will further increase craniocaudad scan coverage, allowing the entire neck and head to be scanned in single high-resolution study and further enhancing the diagnostic utility and power of CTA.

Both CTA and DSA offer high sensitivity to aneurysm detection compared with MRA. CTA may have a higher specificity than MRA, particularly for small aneurysms. This difference in specificity is mediated primarily by technical factors relating to MR signal loss in areas of complex flow. The limitations of MRA appear to be only partially mitigated by the infusion of intravenous contrast material.

We have shown spiral CTA with volume rendering to be superior to both DSA and 3-D TOF MRA in the characterization of intracranial aneurysms. The ability of CTA to completely characterize an aneurysm opens the possibility that a single noninvasive study may be the only one necessary for diagnosis, triage, and treatment planning in patients with intracranial aneurysms.

DR. VILLABLANCA is director of clinical image processing service in the department of radiological sciences at the University of California, Los Angeles.

---

References

1. Schwartz RB, Tice HM, Hooten SM, et al. Evaluation of cerebral aneuyrsms with helical CT: correlation with conventional angiography and MR angiography. Radiology 1994;192:712-722.

2. Jellinger K. Pathology of intracerebral hemorrhage. Zentralbl Neurochir 1977;38:29-42.

3. Hop JW, Rinkel GJE, Algra A, et al. Case fatality rates and functional outcome after subarachnoid hemorrhage: a systematic review. Stroke 1997;

28:660-664.

4. The international study of unruptured intracranial aneurysm investigators. Unruptured intracranial aneurysms-risk of rupture and risk of surgical intervention. NEJM 1998;339:1725-1773.

5. Raaymakers TWM, Rinkel GJ, Limburg M, et al. Mortality and morbidity of surgery for unruptured intracranial aneueyrms: a meta-analysis. Stroke 1998;29:1531-1538.

6. Debrun GM, Aletich VA, Kehrli P, et al. Selection of cerebral aneurysm treatment using Guglielmi detachable coils: the preliminary University of Illinois at Chicago experience. Neurosurgery 1998;43:1281-1295.

7. Fernandez-Zubillaga A, Guglielmi G, Vinuela F, et al. Endovascular occlusion of intracranial aneurysms with electrolytically detachable coils: correlation of aneurysm size and treatment results. AJNR 1994;15:815-820.

8. Martin NA. Arterial bypass for the treatment of giant and fusiform intracranial aneurysms. Tech in Neurosurg 1998;4:153-178.

9. Rinkel GJ, van Gijn J, Wikdicks EF. In: Yanagihara T, Piepgrass DG, eds. Handbook of subarachnoid hemorrhage. New York: Decker, 1997:139-158.

10. Velthuis BK, van Leeuwen M, Witkamp T, et al. Computerized tomography angiography in patients with subarachnoid hemorrhage: from aneurysm detection to treatment without conventional angiography. J Neurosurg 1999;91:761-767.

11. Kogori Y, Takahashi M, Katada K, et al. Intracranial aneurysms: detection with three dimensional CT angiography with volume rendering-comparison with conventional angiographic and surgical findings. Radiology 1999;211:497-506.

12. Johnson PT, Halpern EJ, Kuszyk BS, et al. Renal artery stenois: CT angiography-comparison of real time volume rendering and maximal intensity projection algorithms. Radiology 1999;211:337-342.

13. Villablanca JP, Martin NA, Jahan R, et al. Volume rendered helical computerized tomography angiography in the detection and characterization of intracranial aneurysms. J Neurosurg 2000;93:254-264.

14. Chung TS, Joo JY, Lee SK, et al. Evaluation of cerebral aneuyrsms with high resolution MR angiography using section-interpolation technique: correlation with digital subtraction angiography. AJNR 1999;20:229-235.

15. Ida M, Kurisu Y, Yamashita M. MR angiography of ruptured aneurysms in acute subarachnoid hemorrhage. AJNR 1997;18:1025-1032.

16. Huston J III, Nichols DA, Leutner PH, et al. Blinded prospective evaluation of sensitivity of MR angiography to known intracranial aneurysms: importance of aneurysm size. AJNR 1994;15:1607-1614.

http://diagnosticimaging.com

© 2000 CMP Media, LLC.
12/1/00, Issue # 2212, page CT15.