In the setting of new nontraumatic isolated third cranial nerve palsy (ITNP), the detection of an aneurysm compressing the nerve is of utmost concern. In most patients, it will be located at the junction of the internal carotid and posterior communicating arteries, at the apex of the basilar artery, or at its junction with the posterior cerebral or superior cerebellar arteries. Causative aneurysms usually measure ≥4 mm in cross section (1). In their detection, conventional cerebral catheter angiography (CCA) is still considered the gold standard. However, CT angiography (CTA) and MRA have been improving in this regard.
BASIC CONCEPTS OF NEUROVASCULAR IMAGING
Although a complete understanding of the physics behind image production using different imaging modalities is beyond the scope of this article, some familiarity with the concepts of spatial and contrast resolution is necessary.
Spatial resolution is the ability to resolve two linear structures of the same density very close to one another. Contrast resolution is the ability to resolve two complex shapes of similar density close to one another. High spatial resolution inherently reduces contrast resolution and hence there is always a trade off. Of all imaging modalities, traditional nondigital plain film radiography has the best spatial resolution. That limit of 0.08 mm is slightly better than the 0.17 mm limit of the digital/computed radiography (2) that is currently used in radiology departments in North America. On the other hand, in digital radiography, the ability to see different densities is relatively low. It produces images that are black (air), white (bone), or a narrow band of gray (soft tissue). Compared with digital radiography, CT and MRI have lower spatial resolution (0.4 mm for CT and 1 mm for MRI) (2) but much greater contrast resolution. MRI has contrast resolution superior to that of CT. CT and MRI technology, which is based on volumetric data acquisition displayed in two-dimensional (2D) format, provides a wide band of gray shades, discerning between fat, muscle. and gray and white matter brain parenchyma. An increase in the number of CT detectors, magnetic field strength, and shorter scan times has recently allowed an increase in spatial and contrast resolution in these modalities.
To visualize neurovascular anatomy, the aim is to make the blood vessels more conspicuous than their surrounding structures. In CCA and CTA, this objective is met by intravascular injection of contrast material. In MRA, conspicuity of signal from cerebral vessels is achieved by damping down signal from adjacent tissue while simultaneously enhancing signal from flowing blood within vessels.
Once the 2D images are acquired by any of these modalities, technicians trained in the use of software on independent workstations post-process these images to display them in a variety of ways. Post-processed images include three-dimensional (3D) surface-rendered or volume-rendered images (CCA and CTA), curved reformats (CTA), and maximum intensity projections (MIPs) (CTA and MRA). A particular density (CT) or signal intensity (MRI) is set as a threshold, and other densities or signal intensities are smoothed out to project only the structures in the field of view at or above the designated threshold (“thresholding”). The thresholding process generates the MIP images.
CTA scans currently issue data as 2D reformatted images in axial, sagittal, and coronal planes. These data are then modified by technicians at workstations to produce images as MIPs, consisting of 12 projections at 15° intervals (Fig. 1) or curved reformats of individual vessels (Fig. 2). Both of these types of reconstructions are done by bone editing that is performed by thresholding and manual cutting. Another display format is surface rendering or volume rendering of images displayed in a 3D format. On average, post-processing at an independent workstation takes 15-30 minutes per case. Its value is limited by the raw data and the experience and skill of the technician. For example, data may be lost or spuriously added, diminishing detection accuracy. These errors are mainly due to vessel tortuosity: a tight loop of vessel can be digitally summated to look like an aneurysm.
Standard time-of-flight (TOF) MRA produces raw data called “source images” in the axial plane only (Fig. 3A). Post-processing of these source images to display composite images of the intracranial cerebral vasculature is based on the same principles as those involved in post-processing of CTA images and usually takes about the same amount of time (Fig. 3B). CTA MIPs are viewed by scrolling through consecutive slices in axial, sagittal, and coronal planes, whereas MRA MIPs are viewed by rotating the composite images.
An inherent drawback of these post-processed image reconstructions is the problem of volume averaging. Structures close to the threshold value and in close spatial proximity to each other will either be eliminated or added on in the final post-processed images. An anatomic variation such as a vascular fenestration, or tortuosity, or close proximity of bone to a vessel, as occurs in the skull base region, can lead to false-positive aneurysm detection. Small aneurysms can be eliminated from the post-processed images to yield false-negative detection.
MODALITIES USED IN NEUROVASCULAR IMAGING
CCA is a form of digitized radiography performed by introducing a catheter into the common femoral artery in the groin and advancing smaller catheters to selected neck vessels that ultimately form the cerebral circulation. The circulation is visualized by first taking a native image, then injecting contrast material and taking images with superimposed blood vessels containing contrast material, and finally electronically subtracting the native images. CCA is costly and invasive, requires hospital admission, and carries a 5% risk of catheter-related complications and a 0.3%-0.5% permanent neurologic morbidity rate (3-5), which may be higher in older patients and when the operator is not experienced.
CCA formerly produced images in 2D format only. Ambiguous areas of the cerebral vasculature were then interrogated with repeated injection of contrast material and image acquisition in different oblique projections. More recently, 3D rotational angiography (3DRA) has become available. With this instrumentation, data are acquired by rotation of the image intensifier through 180° during intra-arterial injection of contrast material. This innovation has brought about almost perfect sensitivity and specificity in the CCA detection of aneurysms as small as 1-2 mm in cross section (6) (Fig 4).
The detection value added by 3DRA was demonstrated by van Rooij et al (6) in a study comparing 3DRA with 2D CCA in a cohort of 350 patients. In that study, 29% of smaller (range 0.5-4 mm diameter) aneurysms detected on 3DRA were missed on 2D CCA. However, the 3DRA technique requires increased use of contrast material, an increased radiation dose, and a longer image acquisition time, necessitating a completely cooperative patient. Moreover, post-processing is required to produce the 3D images. Volume averaging on these images can sometimes lead to false-positive detection of small aneurysms.
CTA involves an x-ray source that circumnavigates the patient's head and acquires images that are mathematically transformed and collated to display digital images on a screen. Because current multi-detector CT scanners cover a large volume of anatomy in one spin of the detector array, scan time has been reduced to a few minutes and spatial resolution has improved. The newly available 128-slice and 256-slice multidetector scanners and the 320-slice Aquilion (Toshiba) scanner hold the promise of improved spatial resolution with further reduction in scan times. The Aquilion scanner has the potential to come close to CCA in providing dynamic real-time angiographic studies.
MRA is based on the flow effects on MRI and can be classified in two major categories: time-of-flight MRA (TOFMRA) and phase-contrast MRA (PCMRA) (7).
In TOF imaging, the area within a slice thickness specific for the scanner is initially fully saturated to give a low signal (uniform dark shade) from surrounding tissue. Unsaturated images, in which flowing blood that enters the slice gives a very high signal, are then acquired. An additional radiofrequency pulse is also given to damp down signal from flow in the veins in the field of view and thus prevent venous contamination. The images collected in 2D format can be summated and displayed in a 3D format via post-processing. Improved spatial resolution can be obtained by acquiring data as a 3D data set or slab (7). 3D acquisition is routinely applied to visualization of the smaller vessels of the cerebral circulation. However, the drawbacks of standard TOFMRA images are long scan times and loss of signal from spin saturation and phase dispersion induced by slow or turbulent flow (8).
As with CCA, PCMRA is based on acquiring two data sets, one that is flow-insensitive and the other that is flow-sensitive. The former image is then subtracted from the latter to produce images of flowing blood. PCMRA has even longer acquisition times than TOFMRA and hence is much more prone to image degradation by patient motion.
TOFMRA produces a heavily T1 study, and hence subacute thrombus may simulate the bright signal of flowing blood. When high signal intensity blood is present on T1 MRI, PCMRA is superior to TOFMRA because the signal from the thrombus will be decreased with use of the former technique. Huston et al (9) reported that 3D PCMRA may be superior to 3D TOFMRA in the detection of larger aneurysms (>15 mm).
The post-processing involved in the production of 3D images from a TOF acquisition is similar to that involved in CTA images. It takes approximately the same amount of time and is fraught with similar problems of volume averaging.
Spatial resolution in MRA can be improved by increasing the magnetic field strength. One study (10) comparing 3D TOF using a 7-T field strength magnet with TOF using a 1.5-T field strength magnet showed that imaging of the aneurysm dome was superior with the 7-T magnet, using CCA as the gold standard.
With either PCMRA or TOFMRA, turbulent flow at the aneurysmal neck, a common phenomenon, can yield spurious absence of signal and nondetection of aneurysms.
In the past, contrast-enhanced MRA (CEMRA) had spatial resolution inferior to that of TOFMRA because of problems with timing of image acquisition to the first pass of contrast material, speed of image acquisition, and coverage of anatomic volume. These drawbacks have now been overcome with use of higher field strength (3-T) magnets and accelerated parallel imaging (faster acquisition and coverage), which has improved spatial resolution (11-13). One study (14) showed that image quality with CEMRA at 3-T with accelerated parallel imaging was comparable to that of 3D multislab TOFMRA with regard to aneurysm detection and that image acquisition was much faster.
CEMRA is used in the follow-up of aneurysms treated with endovascular methods, but its application in detection of suspected intracranial aneurysms has not yet been fully established. At the time of compilation of this manuscript, there are no studies comparing these technical modifications in the detection of aneurysms in ITNP.
INTERPRETATION OF NEUROIMAGING STUDIES
The images displayed by any of the invasive and noninvasive neuroimaging techniques require close, meticulous, and stringent scrutiny by fully trained neuroradiologists with detailed knowledge of cerebral vascular anatomy and anatomic variations to exclude the presence of an aneurysm. Because aneurysms causing ITNP are located near the skull base and because bone density is similar to intravascular contrast density, distinguishing aneurysm from bone on CTA, the most commonly used modality, is a challenge met only by those with skill and experience (15-18). There are no guidelines for the credentials of the personnel interpreting neurovascular imaging studies. The only study (19) comparing the ability of general radiologists and neuroradiologists in the detection of aneurysms found neuroradiolologists to be more accurate. A study (20) comparing imaging interpretation in patients with head and neck cancer by neuroradiologists and general radiologists found that a change in the interpretation occurred in 42% of cases, leading to substantial alterations in management.
According to records from the American Board of Medical Specialties (ABMS) (21), the number of American Board of Radiology certified trainees who obtained a Certificate of Additional Qualification (CAQ) in neuroradiology increased from 211 in 1998 to 1,263 in 2008. However, the number of radiologists with this additional CAQ is small considering that there were 5,708 hospitals registered with the American Hospital Association in 2008 (22). If most of these hospitals have neurovascular imaging capabilities, there may be only one CAQ neuroradiologist for every 4 hospitals generating such studies.
Detection of intracranial aneurysms requires not only expertise but also time. It is crucial to scrutinize the source data in the standard axes in conjunction with the post-processed reconstructed images (Fig. 5). Manual alteration of the window levels and widths on an interpreting workstation can also sometimes make the difference between detection and nondetection of an aneurysm. Even if all of these steps are followed, small (≤3 mm) aneurysms will occasionally be missed on CTA or MRA by optimally trained personnel because of the inherent limitations of these noninvasive imaging methods.
DETECTING ANEURYSMS IN ITNP
There is ample evidence that aneurysms causing ITNP measure at least 4 mm in cross section (1,23-25). Aneurysms of this size should be reliably detected by properly performed and interpreted CCA. However, CCA is an invasive technique with small but serious risks. In patients with ITNP with a low clinical index of suspicion, the risk of harboring an aneurysm is lower than the risk of performing CCA. Can a noninvasive imaging modality provide a reliable substitute?
The sensitivity of CTA in the detection of aneurysms of 5-mm diameter is at least 95% even with 64-slice multi-detector CT scanners (26-30). For aneurysms smaller than 5 mm, the sensitivity drops to about 94%, and for aneurysms smaller than 3 mm, to less than 90% (31) However, one study (32) concluded that improved reconstructive algorithms in CTA allow the display and detection of unruptured aneurysms as small as 1 mm. In a cohort of 27 patients with ITNP presumed to be due to cerebral aneurysms, all causative aneurysms were detected on CTA (33). The aneurysm size ranged between 2 and 18 mm (mean 7.2 mm). In 2 patients who did not undergo treatment, the ITNP resolved within a few months and patients remained asymptomatic on follow-up. Twenty-five aneurysms detected using CTA underwent endovascular coiling, at which time CCA confirmed that they were indeed aneurysms. No additional aneurysms were detected on CCA in these patients. Although the smallest diameter of a causative aneurysm was 3.2 mm, an incidental aneurysm as small 1 mm was also detected using the same imaging and reconstructive algorithms. The authors of this study concluded that CTA should be the first line of investigation in ITNP.
In patients in whom CTA cannot be performed (children and pregnant women) or is contraindicated due to suboptimal renal function or contrast material allergy, MRA is an alternative. According to some studies, the sensitivity of MRA for aneurysms larger than 5 mm has improved to about 97% (24), similar to that of CTA. However, some studies (32,34) have shown it to be less accurate than CTA for smaller aneurysms. For aneurysms of <5 mm diameter, the sensitivity of MRA falls dramatically to 54% (24). Indeed, two 8-mm posterior communicating artery (PCA) aneurysms (34) and one 7-mm PCA aneurysm (35) were missed by MRA and subsequently detected with CCA and CTA, respectively. Thus, MRA seems to be less reliable than CTA when imaging is performed specifically for an aneurysmal cause of ITNP (34,35).
On the other hand, the relatively high contrast resolution of MRI provides valuable information not available from CT, including displacement of adjacent structures and associated intraluminal thrombus. Huston et al (9) reported that three aneurysms of 4 mm diameter were missed on MRA but identified on standard T2 brain MRI based on the presence of adjacent intraluminal blood products. Thus, if MRA is to be performed in the investigation of ITNP, it should be combined with MRI.
In summary, CTA is more reliable than MRA in detecting aneurysms ≥4 mm, the minimal size apparently needed to cause an ITNP. CTA is also faster and more widely available and should therefore be the first choice except when x-ray and dye exposure is an issue, as for pregnant women and children and those with renal or cardiac dysfunction. The imaging should be interpreted by an adequately trained neuroradiologist. If a study is judged adequate and negative, no further studies need be performed unless clinical suspicion of aneurysm is high, in which case MRA/MRI could be performed. If the results are still negative or equivocal, then 3DRA CCA should be performed.
1. Lee AG, Hayman LA, Brazis PW. The evaluation of isolated third nerve palsy revisited: an update on the evolving role of magnetic resonance, computed tomography, and catheter angiography. Surv Ophthalmol
2. Bushberg JT, Seibert JA, Leidholdt EM Jr, et al. The Essential Physics of Medical Imaging
. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2002.
3. Wardlaw JM, White PM. The detection and management of unruptured intracranial aneurysms. Brain
4. White PM, Wardlaw JM. Unruptured intracranial aneurysms. J Neuroradiol
5. Fifi JT, Meyers PM, Lavine SD, et al. Complications of modern diagnostic cerebral angiography in an academic medical center. J Vasc Interv Radiol
6. van Rooij WJ, Sprengers ME, de Gast AN, et al. 3D rotational angiography: the new gold standard in the detection of additional intracranial aneurysms. AJNR Am J Neuroradiol
7. Laub G, Gaa J, Drobintzky M. Magnetic resonance angiography techniques. Electromedica
8. Isoda H, Takehara Y, Isogai S, et al. MRA of intracranial aneurysm models: a comparison of contrast-enhanced three-dimensional MRA with time-of flight MRA. J Comput Assist Tomogr
9. Huston J 3rd, Nichols DA, Leutmet PH, et al. Blinded prospective evaluation of sensitivity of MR angiography to known intracranial aneurysms: importance of aneurysm size. AJNR Am J Neuroradiol
10. Monninghoff C, Maderwald S, Theysohn JM, et al. Evaluation of intracranial aneurysms with 7 T versus 1.5 T time-of-flight MR angiography-initial experience. Rofo
11. Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: sensitivity encoding for fast MRI. Magn Reson Med
12. Sodickson DK, Manning WJ. Simultaneous acquisition of spatial harmonics (SMASH): fast imaging with radiofrequency coil arrays. Magn Reson Med 1997;38:591-603.
13. Griswold MA, Jakob PM, Heidemann RM, et al. Generalized autocalibrating partially parallel acquisitions (GRAPPA). Magn Reson Med
14. Nael K, Villablanca JP, Saleh R, et al. Contrast-enhanced MR angiography at 3T in the evaluation of intracranial aneurysms: a comparison with time-of-flight MR angiography. AJNR Am J Neuroradiol
15. Karamessini MT, Kagadis GC, Petsas T, et al. CT angiography with three-dimensional techniques for the early diagnosis of intracranial aneurysms: comparison with intra-arterial DSA and the surgical findings. Eur J Radiol
16. Hoh BL, Cheung AC, Rabinov JD, et al. Results of a prospective protocol of computed tomographic angiography in place of catheter angiography as the only diagnostic and pretreatment planning study for cerebral aneurysms by a combined neurovascular team. Neurosurgery
17. Wintermark M, Uske A, Chalaron M, et al. Multislice computerized tomography angiography in the evaluation of intracranial aneurysms: a comparison with intraarterial digital subtraction angiography. J Neurosurg
18. Jayaraman MV, Mayo-Smith WW, Tung GA, et al. Detection of intracranial aneurysms: multi-detector row CT angiography compared with DSA. Radiology
19. White PM, Wardlaw JM, Lindsay KM et al. The non-invasive detection of intracranial aneurysms: are neuroradiologists any better than other observers? Eur Radiol
20. Loevner LA, Sonners AI, Schulman BJ et al. Reinterpretation of cross-sectional images in patients with head and neck cancer in the setting of a multidisciplinary cancer center. AJNR Am J Neuroradiol
21. ABMS Certificate Statistics 2008
. Evanston, IL: American Board of Medical Specialties; 2008:12.
22. Total number of all U.S. registered hospitals. Chicago: American Hospital Association; 2008. Available from http://www.aha.org/aha/resource-center/statistics-and-studies/fast-facts.html
23. Trobe JD. Third nerve palsy and the pupil: footnotes to the rule. Arch Ophthalmol
24. Jacobson DM, Trobe JD. The emerging role of magnetic resonance angiography in the management of patients with third cranial nerve palsy. Am J Ophthalmol
25. Kupersmith MJ, Heller G, Cox TA. Magnetic resonance angiography and clinical evaluation of third nerve palsies and posterior communicating artery aneurysms. J Neurosurg
26. Kouskouras C, Charitanti A, Giavroglou C, et al. Intracranial aneurysms: evaluation using CTA and MRA. Correlation with DSA and intra-operative findings. Neuroradiology
27. Villablanca JP, Rodriguez FJ, Stockman T, et al. MDCT angiography for detection and quantification of small intracranial arteries: comparison with conventional catheter angiography. AJR Am J Roentgenol
28. Goddard AJ, Tan G, Becker J. Computed tomography angiography for the detection and characterization of intracranial aneurysms: current status. Clin Radiol
29. McFadzean RM, Teasdale EM. Computerized tomography angiography in isolated third nerve palsies. J Neurosurg
30. Teasdale E, Statham P, Straiton J, et al. Non-invasive radiological investigation for occulomotor palsy. J Neurol Neurosurg Psychiatry
31. McKinney AM, Palmer CS, Truwit CL, et al. Detection of aneurysms by 64-section multidetector CT angiography in patients acutely suspected of having an intracranial aneurysm and comparison with digital subtraction and 3D rotational angiography. AJNR Am J Neuroradiol
32. White PM, Teasdale EM, Wardlaw JM, et al. Intracranial aneurysms: CT angiography and MR angiography for detection-prospective blinded comparison in a large patient cohort. Radiology
33. Mathew MR, Teasdale E, McFadzean RM. Multidetector computed tomographic angiography in isolated third nerve palsy. Ophthalmology
34. Johnson MR, Good CD, Penny WD, et al. Lesson of the week: Playing the odds in clinical decision making: lessons from berry aneurysms undetected by magnetic resonance angiography. BMJ
35. Vaphiades MS, Horton JA. MRA or CTA, that's the question. Surv Ophthalmol