Applied Radiology

Dr. Andersen is a third-year Radiology Resident at the Mallinckrodt Institute of Radiology in St. Louis, MO. He received his medical degree in 1998 from Duke University, Durham, NC. He plans to pursue further training as a Neuroradiology Fellow at Mallinckrodt beginning in 2003.

Cerebral aneurysms represent a significant public health risk due to their potential for rupture and the life-threatening sequelae of subarachnoid hemorrhage. Early detection and treatment of unruptured aneurysms is necessary to reduce the incidence of aneurysm rupture. Magnetic resonance angiography and computed tomography angiography are two highly sensitive, noninvasive imaging modalities that are likely to play a role in the future screening and management of unruptured cerebral aneurysms.

Based on reports of autopsy series, cerebral aneurysms occur with an incidence of approximately 5% in the general population. Thus, approximately 10 to 15 million Americans may be harboring an aneurysm. These aneurysms are of concern due to the 30% to 40% mortality rate associated with their rupture, as well as the approximate 30% incidence of permanent neurologic disability among survivors. ¹ While the risk of rupture remains controversial, approximately 27,000 cases of subarachnoid hemorrhage (SAH) due to aneurysm rupture are seen in the United States every year. Given the prevalence of cerebral aneurysms and the life-threatening impact of their rupture, early detection and treatment of unruptured aneurysms have significant potential to improve public health.

The rate of identification of unruptured aneurysms has increased dramatically in recent years as a result of the increased use and improved resolution of computed tomography (CT) and magnetic resonance imaging (MRI). CT and MRI studies are now commonly used for a wide variety of neurologic indications, including headache, dizziness, stroke/transient ischemic attack, trauma, cranial nerve pathologies, and sinus disease, and the detection of unruptured cerebral aneurysms is often an incidental finding on such studies.

Conventional CT and MRI techniques are not well suited for the detection of cerebral aneurysms, however, since standard protocols do not provide the necessary high-resolution imaging and reconstruction of vessels. The gold standard for vascular investigation remains direct catheter cerebral angiography. This technique traces its roots back to Moniz's first description of a method to opacify the carotid arteries and vessels of the brain in 1927. ² Angiographic techniques have been refined to the increasingly safe and effective procedure known as intra-arterial digital subtraction angiography (DSA) (Figure 1).

While DSA remains the most accurate modality for vascular imaging, its disadvantages include a small risk of neurological complication (approximately 0.07% ³ ), high cost, use of iodinated contrast media, and the fact that it is a time-consuming and labor-intensive technique. Thus, the search for other imaging options continues. Computed tomography angiography (CTA) and magnetic resonance angiography (MRA) are the two most notable modalities for noninvasive imaging of the cerebral vasculature. The role of these modalities in the identification and management of unruptured intracranial aneurysms still needs to be defined and should be done in terms of decreasing the incidence of SAH and its associated morbidity and mortality.

CTA

CTA is a relatively new technique, made possible by the improved resolution and rapidly increasing acquisition speeds of newer CT scanners. In particular, multidetector scanners have made vascular applications such as pulmonary angiography, aortic aneurysm modeling, and cerebral angiography viable clinical tools. Current techniques involve multidetector scanners employing narrow collimation (down to 1 mm), high kVp technique, bolus tracking for image timing, and advanced reconstruction algorithms. These scanners also allow for the use of improved postprocessing techniques for rapid online manipulation and interpretation of the three-dimensional (3D) data set (Figure 2).

There are several advantages of CTA over DSA and MRA. While the resolution of CTA is not as good as that of DSA, it continues to improve, primarily due to the narrower collimation available on newer multidetector scanners. Perhaps the greatest advantages of CTA are the relatively short imaging times, wide availability, and logistical ease of imaging critically ill patients. Of note, CTA is subject to fewer artifacts relative to MRA and offers extraluminal information not available with DSA. Like MRA, CTA is a less invasive technique than DSA and has multiplanar reconstruction capabilities. Finally, CTA has excellent sensitivity for intracranial hemorrhage, which is critical in the imaging of cerebral aneurysms.

There are few disadvantages of CTA, but they are important to remember. The most significant disadvantage is the use of iodinated contrast media, which cannot be administered in patients with documented allergies or in many patients with significant renal insufficiency. In cases of small aneurysms, the lower resolution of CTA (compared with DSA) and the lack of real-time imaging information (which is provided by DSA) is particularly problematic. Finally, as with DSA, CTA delivers high radiation doses that should be considered on a case-by-case basis.

MRA

Like CTA, MRA is a rapidly evolving modality. The predominant technique used for MRA is time-of-flight imaging performed on 1.5 T magnets. Phase-contrast imaging is used much less frequently. Recent advances include contrast-enhanced 3D MRA imaging (CE MRA), increased magnet strengths of 3.0 T (Figure 3), and improved gradients on newer scanners. Contrast enhancement and increased field strengths increase the signal-to-noise ratio, which may improve the diagnostic information available. Higher performance gradients allow thinner slices, resulting in improved resolution in reconstructed data sets. In addition, faster gradients allow improved in-plane spatial resolution due to improved frequency resolution. MRA will continue to improve as it is developed in a wide variety of clinical applications including head/neck, body, and peripheral angiography studies.

MRA has two unique advantages over CTA and DSA: the lack of iodinated contrast material and the lack of radiation exposure. In particular, patients with a documented contrast allergy may be best served with MRA imaging. MRA, like CTA, offers a less invasive technique, extraluminal imaging, and extensive multiplanar imaging and reconstruction capabilities not available with DSA.

The unique disadvantages of MRA involve artifacts--such as over estimation of stenosis due to turbulent flow--and the logistical difficulties of imaging critically ill patients in an MRI suite. Like CTA, MRA lacks the resolution and real-time imaging of DSA. Finally, although shorter than those for DSA, MRA imaging times are longer than those for CTA.

MRA versus CTA for detection of aneurysms

Multiple studies have been published assessing the relative abilities of CTA and MRA to detect cerebral aneurysms. One such study is the extensive review by White and Wardlaw ⁴ that employed detailed meta-analysis of 38 studies evaluating the relative ability of CTA and MRA versus DSA to detect aneurysms. Most of the studies reviewed involved patients undergoing both DSA and CTA or MRA. Compared with the number of aneurysms detected on DSA, the overall accuracy rates of CTA and MRA for aneurysm detection were 89% and 90%, respectively. ⁴ The sensitivity for aneurysms <3 mm in size was 61% for CTA and 38% for MRA. Meanwhile, the sensitivity for aneurysms >3 mm was 96% for CTA and 94% for MRA. This high detection rate for aneurysms >3 mm approaches that of the gold standard of DSA, which is known to have a small (<5%) but real rate of missed aneurysms. ⁵ White and Wardlaw ⁴ suggest that the detection rates for CTA and MRA may increase as MRA techniques improve and the use of workstations for clinical interpretation becomes more common.

The enthusiasm for these encouraging numbers, however, must be tempered. White and Wardlaw ⁴ point out that some early studies failed to show significantly improved results with the use of CE-MRA. Also, they note that no studies have demonstrated that real-time manipulation and interpretation of data on workstations improves aneurysm detection. Finally, skewed, high-risk populations have been used throughout the literature. In the majority of studies, this high pretest probability of aneurysm presence may induce observer bias and corresponding falsely elevate the detection rates.

The merits of CTA and MRA for cerebral vascular imaging were confirmed by a prospective study by White et al ⁶ in 2001. For aneurysms >5 mm in size, they reported sensitivities of 94% and 86%, respectively, for CTA and MRA. Again, for smaller aneurysms (<5 mm), the sensitivities were much lower, 57% for CTA and 35% for MRA. Overall, White and Wardlaw ⁴ and White et al ⁶ have found that CTA and MRA are both highly sensitive techniques for identifying the majority of aneurysms, but both remain limited in their ability to detect smaller aneurysms.

While it is fairly well established that CTA and MRA are effective imaging tools for all but small aneurysms (<5 mm), the more important question is how to use such tools for screening to help prevent SAH due to aneurysm rupture. Guidelines must be developed for when, how, and for whom screening should be employed. The data required to establish such screening guidelines include identification of those at high-risk for aneurysm development, the risk of rupture for each aneurysm, and the risks of potential treatment options.

A substantial amount of data suggests how to determine which individuals are most at risk for the development of cerebral aneurysms. Genetic predisposition to the development of aneurysms has been established in a number of connective tissue disorders including Ehlers-Danlos Type IV, Marfan's syndrome, neurofibromatosis Type I, and autosomal dominant polycystic kidney disease. ⁷ Smoking, female gender, and a family history of cerebral aneurysms are all risk factors for aneurysm formation. Finally, cerebral aneurysms are most prevalent in middle-aged patients. This list of known risk factors allows for selection of individuals who would benefit most from screening for the presence of an aneurysm.

There is conflicting evidence in the literature with regard to the risk of rupture of any given cerebral aneurysm. Historically, it has been reported that the risk of aneurysm rupture is approximately 1% per year per aneurysm (not including giant aneurysms, which are known to have a 6% per year rate of rupture). ⁵ However, this notion was brought into question by the International Study of Unruptured Intracranial Aneurysms (ISUIA) Investigators data. ⁸ The ISUIA stated that rates of rupture were as low as 0.05% per year per aneurysm (for aneurysms <10 mm in size in patients with no prior history of SAH). Thus, some have suggested that screening with MRA and/or CTA may be acceptable as both modalities are highly effective at identifying aneurysms >10 mm.

A number of criticisms have been leveled at the ISUIA study, however. It was a retrospective study and subject to selection bias, since only untreated aneurysms were followed. In addition, if the 0.05% rate were accurate, the prevalence of aneurysms in people of 30 years of age would have to be much higher than 5%, given the annual rates of subarachnoid hemorrhage. Finally, the majority of ruptured aneurysms treated today are in fact <10 mm in size. As an example, Forget et al ⁹ recently published a report showing that 85.6% of all ruptured aneurysms seen in their clinical setting were <10 mm in size. Some have argued that this discrepancy is due to larger aneurysms contracting after rupturing, thus appearing smaller at the time of diagnosis. However, this theory has not been proven. ¹⁰ Also, risk factors other than aneurysm size must be investigated further, including the type of aneurysm (eg, saccular or fusiform), morphology, location, presence/absence of calcification or clot, and underlying blood pressure.

These confusing and incomplete data on the risks of aneurysm rupture provide no basis for screening and management guidelines for the majority of unruptured cerebral aneurysms. Therefore, the need to detect aneurysms (other than giant aneurysms) is of questionable utility for patient management. Further large-scale, multicenter prospective studies are needed to produce the necessary data on risks for and causes of aneurysm rupture. Also, some attempt should be made to quantify observer bias by performing screening on a portion of the general population at low-risk. Such studies are challenging given the reliance on DSA for aneurysm identification. CTA and MRA may facilitate future research of cerebral aneurysms by allowing safer, easier imaging for following unruptured aneurysms prospectively. As CTA and MRA become more widely available and logistically easier, performing a large multicenter screening study will become less complicated. Thus, CTA and MRA can be expected to play an increasing role in the future detection and management of cerebral aneurysms.

Conclusion

MRA and CTA both have been proven to be highly effective in the detection of cerebral aneurysms larger than 3 to 5 mm in size. This fact, combined with the increasing availability of MR and CT imaging modalities, easier imaging protocols, and continually improving technique, will make both CTA and MRA important tools in the effort to detect and manage unruptured cerebral aneurysms. More work is necessary to reduce the incidence of SAH and its associated devastating morbidity and mortality.

Acknowledgments

The author wishes to thank Dr. Colin Derdeyn for providing the pictures used in this manuscript and for his excellent expert advice and Dr. Robert McKinstry for his expert knowledge and support.