In this issue of the Journal of Neuro-Ophthalmology, Chaudhary et al (1) point out that expert interpretation of modern noninvasive neurovascular imaging should detect nearly all brain aneurysms that measure 4 mm in diameter or more. Multidetector CT angiography may be slightly better than MRA, but either CT or MRA is very sensitive to aneurysms of this size.
This capability has changed the guidelines for evaluating patients with isolated nontraumatic third cranial nerve palsies. Not so long ago, when noninvasive intracranial vascular imaging did not exist or was too crude to detect aneurysms, we were forced to decide between observation and catheter angiography with its attendant risks. So we developed clinical criteria to help us determine the likelihood of an aneurysm, but these criteria are far from perfect (2-5).
We could safely exclude from angiography those patients with isolated mydriasis or ptosis. At first, we even excluded those with ductional deficits and ptosis in whom the iris sphincter is spared. However, further study disclosed that such “pupil sparing” is not an entirely safe haven, especially if the ptosis or extraocular muscle palsies is incomplete or patchy (2). And because most patients with isolated nontraumatic third cranial nerve palsies have incomplete ptosis or extraocular muscle palsies, they would have to be considered candidates for catheter angiography.
With the advent of noninvasive vascular imaging, that dilemma is gone but we still have issues. Is aneurysm a common enough cause to warrant the added expenses of noninvasive vascular imaging in everyone with isolated nontraumatic third cranial nerve palsy? How urgent is the noninvasive study? How does one know if the report of a negative noninvasive study is reliable? If reliable noninvasive imaging fails to disclose an aneurysm, could a small (<4 mm diameter) aneurysm-beyond the reach of noninvasive techniques-cause the palsy? If noninvasive vascular imaging fails to show such an aneurysm, would catheter angiography find it?
Is Aneurysm A Common Enough Cause To Warrant Noninvasive Vascular Imaging?
The data here are not robust. The most common cause of isolated nontraumatic third cranial nerve palsy in adults is believed to be microvascular ischemia, but estimates of its relative prevalence are not reliable (6,7). The relative prevalence of aneurysm as a cause of nontraumatic isolated third cranial nerve palsy is reported to range from 14% to 56% (4,8), but those numbers are undoubtedly inflated by reports of patients examined in tertiary care centers. In my experience from tertiary care academic centers, the prevalence is well under 10%.
Do clinical features help to alter the odds of an aneurysm? Not well enough to make one entirely confident. For example, patient age is not a guide; berry aneurysms become clinically manifest between ages 20 and 60-an enormous age span. Younger and older people do not show manifestations very often, but there are several reports of aneurysms occurring at the extremes of age (9).
The presence of arteriosclerotic risk factors favors a microvascular ischemic cause. However, patients with arteriosclerosis can also have aneurysms, and aneurysms expand-and may become manifest-when arteries are battered by hypertension, diabetes, cholesterol, and other factors associated with arteriosclerosis.
New headache, especially if severe, tends to suggest aneurysm, but its absence does not exclude it. Headache and periocular pain are reported in as few as 30% of patients with aneurysmal third cranial nerve palsy (2). Pain excruciating enough to overshadow the ptosis and diplopia is present in at least 50% of patients with microvascular ischemic third cranial nerve palsy (9).
Although expansion of an aneurysm usually produces acute manifestations, subacute and even chronic clinical evolution is well known.
The relationship between the degree of impairment of somatic (external) and parasympathetic (internal) motor functions has been the focus of debate about how to predict the likelihood of aneurysm in third cranial nerve palsy (2-5). Aneurysms are likely to impair iris sphincter function, but pupil-sparing is especially apt to occur with aneurysms if the ophthalmoplegia is incomplete and particularly if the inferior division of the third cranial nerve is spared (2-5). Pupil involvement occurs in as many as 20% of patients with ischemic third cranial nerve palsies, but anisocoria of more than 1.5 mm is exceedingly rare (10).
How Urgent Is the Noninvasive Study?
Its urgency depends on how recent the manifestations are. When an aneurysm causes an isolated third cranial nerve palsy and the patient is otherwise neurologically intact, the aneurysm has not yet ruptured. Rather, it has recently expanded to compress the adjacent third cranial nerve (11). Repair of an expanding but unruptured aneurysm, whether by surgical or endovascular methods, has a high chance of restoring the patient to normal neurologic health (12,13) and about a 50% chance of restoring adequate third cranial nerve function (14). Most importantly, recent expansion often foretells imminent rupture with subarachnoid hemorrhage-within days, even hours (15)-carrying a 66% mortality rate or a high chance of serious neurologic morbidity arising chiefly from vasospasm and secondary stroke (12,13).
Is Noninvasive Vascular Imaging Safe?
Usually yes. The hazards of CTA are radiation exposure and the contrast load. Under most circumstances, those risks are acceptable, except in children, pregnant women, and those with cardiac or renal dysfunction. The hazards of MRA are essentially nonexistent, because the preferred study sequence (time-of-flight) does not require intravenous contrast material.
Is the Report of a Negative Noninvasive Study Reliable?
This is the hard part. As Chaudhary et al (1) emphasize, there are no published guidelines on the minimum standards for imaging acquisition in the detection of aneurysms. Some neuroradiologists insist that CTA images be acquired at no larger than 1.25-mm slice thickness. Appropriate MRA acquisition involves balancing complex variables in addition to slice thickness, such as field of view, matrix size, and scan times.
Once the images are acquired, they must be manipulated (“post-processed”) for optimal visualization of aneurysms. There are no published guidelines on post-processing standards. Some neuroradiologists recommend that CTA images be post-processed in axial, coronal, and sagittal reformatted planes using <3 mm non-gapped reconstructions. MRA studies should be reviewed using raw axial acquisitions and maximum intensity projections. No one knows how often such post-processing standards are being followed.
Even with excellent post-processing, the detection of aneurysms requires time, skill, and experience. As Chaudhary et al (1) point out, the aneurysm may be evident on only one CTA reformat or one MRA source image. Scrutinizing these images one by one is laborious. In comparing the ability of general radiologists and neuroradiologists in the detection of aneurysm, one study (16) found neuroradiologists to be more sensitive. In a study (17) that compared the interpretations of general radiologists and neuroradiologists on head and neck cancer patient imaging, a corrective change in the interpretation by the neuroradiologists occurred in 42% of cases, leading to substantial alteration in management. There are no published data on what proportion of intracranial imaging studies are being interpreted by general radiologists. In this country, they generally receive an average of 4-6 months' exposure to neuroradiology in residency training, during which they probably see at most 10-20 aneurysms.
If Reliable Noninvasive Imaging Fails to Disclose an Aneurysm, Could a Small Aneurysm Be Causing the Palsy?
Yes. The proximity of the affected vessel to the third cranial nerve varies; in some cases, the aneurysm might not have far to go to compress the nerve. Moreover, aneurysms are often partially clotted, and clots interfere with visualization in vascular imaging. Clotted intra-aneurysmal blood is more likely to be visible on MRI than on CT, giving a rationale for performing MRI/MRA if CTA results are negative and clinical suspicion of aneurysm is high.
However, there are now several well-documented reports establishing that when third cranial nerve palsy is caused by aneurysms situated at the junction of the supraclinoid internal carotid and posterior communicating arteries or at the junction of the basilar and posterior cerebral artery or superior cerebellar artery (9), they will appear to be at least 4 mm in diameter on noninvasive studies (4,8,18). There are anecdotes of apparently smaller aneurysms causing third cranial nerve palsy (4), but this phenomenon is uncommon.
If Noninvasive Vascular Imaging Fails to Show an Aneurysm, Would Catheter Angiography Find It?
There are no data on this point. Chaudhary et al (1) affirm that catheter-based three-dimensional (3D) rotational digital subtraction angiography (DSA) is still considered by most radiologists to be the most sensitive means of detecting intracranial aneurysms. Even so, many institutions are relying on expert review of noninvasive studies to exclude a clinically significant aneurysm; some institutions will even skip catheter angiography on the way to clipping the aneurysm (8). That approach is partly based on the awareness that DSA carries a risk of stroke of up to 2%, a proportion that may be higher in older patients with arteriosclerosis (19).
How Should We Evaluate New-Onset Isolated Nontraumatic Third Cranial Palsy?
1. Every nonpregnant adult patient without renal or cardiac failure should undergo emergency CTA, which is more sensitive to aneurysm, more widely available, and faster than MRA. CT should be included, but its ability to exclude nonaneurysmal causes is low, so MRI should be done if CTA results are negative. If the clinical suspicion of aneurysm is high, MRA should be added to MRI, which is more sensitive to intra-aneurysmal blood products than CT.
2. Children, pregnant women, and those with renal or cardiac failure should undergo emergency MRA. MRI should be included because it allows excellent detection of nonaneurysmal causes.
3. If the noninvasive imaging study is interpreted as negative, it should be reviewed by an experienced neuroradiologist before the diagnosis of aneurysm is dismissed. If the noninvasive study was CTA, MRI/MRA should be performed if the clinical suspicion of aneurysm is high. DSA, preferably in the 3D rotational mode, should be performed only if an experienced neuroradiologist has pronounced the noninvasive studies negative, there is a high clinical suspicion of aneurysm, and the patient is considered healthy enough to tolerate catheter angiography.
1. Chaudhary N, Davagnanam I, Ansari SA, et al. Imaging of intracranial aneurysms causing isolated third cranial nerve palsy. J Neuroophthalmol
2. Trobe JD. Third nerve palsy and the pupil. Footnotes to the rule. Arch Ophthalmol
3. Trobe JD. Managing oculomotor nerve palsy. Arch Ophthalmol
4. 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
5. Lee SH, Lee SS, Park KY, et al. Isolated oculomotor nerve palsy: diagnostic approach using the degree of external and internal dysfunction. Clin Neurol Neurosurg
6. Green WR, Hackett ER, Schlezinger NS. Neuro-ophthalmologic evaluation of oculomotor nerve paralysis. Arch Ophthalmol
7. Nadeau SE, Trobe JD. Pupil sparing in oculomotor palsy: a brief review. Ann Neurol
8. Mathew MR, Teasdale E, McFadzean RM. Multidetector computed tomographic angiography in isolated third nerve palsy. Ophthalmology
9. Smith CH. Nuclear and infranuclear ocular motility disorders. In: Miller NR, Newman NJ, eds. Walsh & Hoyt's Clinical Neuro-Ophthalmology
. 5th ed. Baltimore: Williams & Wilkins; 1998;1:1205.
10. Jacobson DM. Pupil involvement in patients with diabetes-associated oculomotor nerve palsy. Arch Ophthalmol
11. Hyland HH, Barnett HJ. The pathogenesis of cranial nerve palsies associated with intracranial aneurysms. Proc R Soc Med
12. International Study of Unruptured Intracranial Aneurysms Investigators. Unruptured intracranial aneurysms-risk of rupture and risks of surgical intervention. N Engl J Med
13. Molyneux A, Kerr R, Stratton I, et al. International Subarachnoid Aneurysm Trial (ISAT) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms: a randomised trial. Lancet
14. Bhatti MT, Peters KR, Firment C, et al. Delayed exacerbation of third nerve palsy due to aneurysmal regrowth after endovascular coil embolization. J Neuroophthalmol
15. Okawara SH. Warning signs prior to rupture of an intracranial aneurysm. J Neurosurg
16. White PM, Wardlaw JM, Lindsay KW, et al. The non-invasive detection of intracranial aneurysms: are neuroradiologists any better than other observers? Eur Radiol
17. 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
18. Kupersmith MJ, Heller G, Cox TA. Magnetic resonance angiography and clinical evaluation of third nerve palsies and posterior communicating artery aneurysms. J Neurosurg
19. Leffers AM, Wagner A. Neurologic complications of cerebral angiography. A retrospective study of complication rate and patient risk factors. Acta Radiol