MRI and MR angiography of vertebral artery dissection

A review of 4,500 angiograms yielded 11 patients with dissection of the vertebral arteries who had MRI and (in 4 patients) MR angiography (MRA) in the acute phase of stroke. One patient with incidental discovery at arteriography of asymptomatic vertebral artery dissection and two patients with acute strokes with MRI and MRA findings consistent with vertebral artery dissection were included. Dissection occurred after neck trauma or chiropractic manipulation in 4 patients and was spontaneous in 10. Dissection involved the extracranial vertebral artery in 9 patients, the extra-intracranial junction in 1, and the intracranial artery in 4. MRI demonstrated infarcts in the brain stem, cerebellum, thalamus or temporo-occipital regions in 7 patients with extra- or extra-intracranial dissections and a solitary lateral medullary infarct in 4 patients (3 with intracranial and 1 with extra-intracranial dissection). In 2 patients no brain abnormality related to vertebral artery dissection was found and in one MRI did not show subarachnoid haemorrhage revealed by CT. Intramural dissecting haematoma appeared as crescentic or rounded high signal on T1-weighted images in 10 patients examined 3-20 days after the onset of symptoms. The abnormal vessel stood out in the low signal cerebrospinal fluid in intracranial dissections, whereas it was more difficult to detect in extracranial dissections because of the intermediate-to-high signal of the normal perivascular structures and slow flow proximal and distal to the dissection. In two patients examined within 36 h of the onset, mural thickening was of intermediate signal intensity on T1-weighted images and high signal on spin-density and T2-weighted images. MRA showed abrupt stenosis in 2 patients and disappearance of flow signal at and distal to the dissection in 5. Follow-up arteriography, MRI or MRA showed findings consistent with occlusion of the dissected vessel in 6 of 8 patients.     

Cerebral artery dissection

The main symptom of arterial dissection is intense acute unilateral headache. The pain is commonly located around the eye, in the temple or the front with a carotid artery dissection [CAD] and in the posterior neck and occiput with a vertebral artery dissection [VAD]. Transient or persistent cerebral ischemic symptoms are similarly frequent but usually occur later in the time course. Horner's syndrome indicating a lesion of perivascular sympathetic fibres represents the third leading symptom and occurs in more than one third of the patients. Compression of local structures such as lower nerve or radicular palsies is rare. This constellation of symptoms in a young patient without vascular risk factors should rise suspicion of a dissection, in particular, if there is a preceding 'trivial' trauma. Characteristic features on Doppler/duplex sonography provide the diagnosis of dissection in almost all CAD and the majority of VAD. MRI demonstrating the mural hematoma allows reliable confirmation of the suspected diagnosis. Angiography is necessary only in selected cases, more often in VAD than in CAD. Brain infarction may be prevented, if premonitory symptoms, which occur in 60 to 80% of the patients, are recognized as such. Therefore, if there is clinical and sonographic suspicion of CAD or VAD, anticoagulation therapy with heparin should should be started before other imaging procedures finally prove the diagnosis. Because immediate anticoagulation may prevent cerebral embolism, this treatment strategy seems appropriate, although its efficacy has not been established by a controlled study. Anticoagulation should be continued until resolution of the dissection.     

Magnetic resonance angiography demonstrates vascular healing of carotid and vertebral artery dissections

BACKGROUND AND PURPOSE: Dissection of the carotid and vertebral arteries is most accurately diagnosed with conventional angiography. MR techniques are sensitive for detecting the abnormalities associated with dissection but may lack specificity. We hypothesized that MR may be useful for serial monitoring of dissection and may therefore guide therapy. METHODS: All patients with angiographically proven carotid and/or vertebral artery dissection from July 1994 to June 1996 were followed for a median duration of 10.5 months. Of these 29 patients (44 vessels), 18 were concurrently evaluated with MR, and a target group of 9 patients (17 vessels) was prospectively followed with MR at 3-month intervals. RESULTS: In the 18 patients with both imaging studies at baseline, angiography revealed 30 dissected vessels while MR detected 27 (90%). In the target group of 9 patients, initial MR identified 15 of the 17 dissections diagnosed with angiography. Serial MR revealed complete healing in 5 vessels, improvement in 6 vessels, no change in 4 vessels, and worsening in 2 vessels. The radiographic features most likely to resolve were stenosis and mural hematoma, while occlusion and luminal irregularity tended to persist. Late ischemic events occurred in 2 patients, both with persistent MR evidence of dissection, one while subtherapeutic on warfarin therapy and the other occurring 1 week after warfarin was discontinued. CONCLUSIONS: MR is a reliable noninvasive method for following the vascular response to treatment and may guide the course of a clinical trial comparing medical therapies for carotid and vertebral artery dissection.     

Headache and neck pain in spontaneous internal carotid and vertebral artery dissections

We studied the characteristics of headaches in 161 consecutive symptomatic patients with spontaneous dissections of the internal carotid artery (n = 135) or the vertebral artery (n = 26). For patients with internal carotid artery dissection (ICAD), the mean age was 47 years and for those with vertebral artery dissection (VAD), 40.7 years. A history of migraine was present in 18% of the ICAD group and in 23% of the VAD group. Headache was reported by 68% of the patients with ICAD and by 69% of those with VAD, and, when present, it was the initial manifestation in 47% of those with ICAD and in 33% of those with VAD. Ten percent of patients with ICAD had eye, facial, or ear pain without headache. The median interval from onset of headache to development of other neurologic manifestations was 4 days for the ICAD group and 14.5 hours for the VAD group. For all dissections, headaches typically were ipsilateral to the side of dissection. In the ICAD group, headaches were limited to the anterior head in 60% of patients and were steady in 73% and pulsating in 25%. In the VAD group, headaches were distributed posteriorly in 83% of patients and were steady in 56% and pulsating in 44%. Neck pain was present in 26% of patients with ICAD (anterolateral) and in 46% of those with VAD (posterior). The median duration of the headache in patients with VAD and ICAD was 72 hours, but headaches became prolonged, persisting for months to years, in four patients with ICAD.     

Magnetic resonance angiography in patients with acute stroke treated by local thrombolysis

Although magnetic resonance angiography (MRA) is accepted for showing chronic intracranial stenotic or occlusive lesions, the method has not been practically examined in patients with acute cerebral ischaemia. We carried out three-dimensional time-of-flight MRA in six patients with acute ischaemia treated by local thrombolysis, and compared the findings with those of digital subtraction angiography (DSA). In all patients, MRA before thrombolysis clearly demonstrated the occluded arteries, which corresponded precisely to those shown by DSA. In four patients with complete recanalisation of the occluded vessels after thrombolysis, the recanalisation could be demonstrated by postoperative MRA. In one patient with reocclusion of the recanalised artery, repeat MRA also demonstrated the reocclusion, confirmed by DSA. These results suggest that MRA may be helpful for noninvasive investigation before and after thrombolysis.     

Intracranial vascular stenosis and occlusion: comparison of 3D time-of-flight and 3D phase-contrast MR angiography

We compared the value of 3D time-of-flight (TOF) and phase-contrast (PC) MR angiography (MRA) for detection and grading of intracranial vascular steno-occlusive disease. Unenhanced 3D-TOF MRA and 3D-PC MRA (30-60 cm/s velocity encoding) were performed at the level of the circle of Willis in 18 patients, mean age 56 +/- 10 years. Postprocessed images using a maximum-intensity projection reconstruction with multiple targetted projections were analysed. A total of 126 vessels was assessed by PC MRA and 143 by TOF MRA, with digital subtraction angiography (DSA) in 15 patients and/or transcranial Doppler sonography (TCD) in 18 as a standard. Two blinded readers reviewed the MRA, DSA and TCD examinations retrospectively. On DSA and/or TCD the two observers found 32 and 28 steno-occlusive lesions. 3D-TOF MRA was more sensitive than 3D-PC MRA (87% and 86% vs. 65% and 60%) and had a higher negative predictive value (96% vs. 89%). Correct grading of stenoses was achieved in 78 % by 3D-TOF and 65% by 3D-PC MRA.     

Rheumatoid arthritis and vertebral artery occlusion: a case report with angiographic and magnetic resonance demonstration

A 40-year-old woman with rheumatoid arthritis (RA) who presented with vertigo and nausea, was found to have left vertebral artery occlusion on angiography. There was minimal atlantoaxial subluxation and no odontoid herniation. Magnetic resonance imaging (MRI) was suggestive of a left vertebral artery dissection. This is the first report of MRI of vertebral artery occlusion in RA. It will be important to accumulate MR images of this complication in future cases.     

Noninvasive angiography for the diagnosis of vertebral artery disease using Doppler ultrasound (vertebral artery Doppler)

A transoral noninvasive procedure in the oropharynx using local anesthesia was applied to detect flow in the vertebral arteries with a bidirectional continuous-wave Doppler ultrasound system. Common carotid artery compression was used to identify the vertebral artery. Flow direction, amplitude of flow signals, diastolic flow, and reaction of flow on common carotid artery compression served as diagnostic parameters. The procedure was applied in 90 patients of whom 42 underwent angiography. The method has proved to be 82% accurate. It was most reliable in the diagnosis of occlusion or aplasia, subclavian steal and normalcy, and was less reliable in the detection of stenosis or hypoplasia of a vertebral artery. Eleven patients with subclavin steal, five patients with a missing vertebral artery three patients with hypoplasia or stenosis, and 15 patients with normal angiorgraphical findings were correctly diagnosed by Doppler; normal Doppler findings were present in three patients with a mission or stenosed vertebral artery. Those patients (five) with Doppler indications of subclavian steal (one patient), missing vertebral artery (two patients), or stenosis (two patients) had normal angiograms. Application of the Doppler procedure, after 11 subclavin endarterectomies, informed the surgeon immediately about the hemodynamic effect of surgical intervention. Rethrombosis was diagnosed in two patients by postoperative Doppler examination.     

Dissection of the carotid artery and vertebral artery--diagnosis and therapy

Carotid and vertebral artery dissections typically occur in young adults after major trauma, although they can arise spontaneously or after trivial injury. Many patients with carotid dissections have minor symptoms such as a subject bruit or Horner's syndrome. Cephalic pain is also frequent and often inaugural in carotid dissection. However, extracranial dissection is a well recognised cause of ischaemic stroke. The diagnosis of dissection was based on angiographic findings. Noninvasive imaging also allows prompt and reliable diagnosis. Our goal was to demonstrate the spectrum of neuroradiologic (CT, MR and angiographic) findings in craniocervical arterial dissection and compare the diagnostic utility of CT, MR, MR angiography. Clinical data imaging studies, and outcome were reviewed and compared with the results in four patients with carotid artery dissection.     

Ophthalmic manifestations of vertebral artery dissection. Patients seen at the Mayo Clinic from 1976 to 1992

PURPOSE: To determine the ophthalmic manifestations of vertebral artery dissections. METHODS: Fifty-one separate episodes of vertebral artery dissections evaluated at the Mayo Clinic from 1976 through 1992 were studied. In all cases, the diagnosis had been documented with angiography. RESULTS: There were 28 men and 19 women (mean and median age, 39 and 40 years, respectively; range, 8-61 years). There were ophthalmic findings in 86% of the episodes. Visual symptoms, in decreasing order of frequency, were diplopia (45% of the episodes), blurred vision (14%), transient visual dimming (8%), oscillopsia (4%), photophobia (4%), upside-down vision (2%), positional transient visual obscuration (2%), and unilateral dry eye (2%). Ophthalmic signs in decreasing order of frequency were nystagmus (37% of the episodes), ocular misalignment (cranial nerve palsy or skew) (33%), Horner syndrome (27%), decreased corneal sensation (22%), ptosis (16%), visual field defect (10%), abnormal pursuits and saccades (6%), ocular bobbing (4%), internuclear ophthalmoplegia (4%), anisocoria (4%), and pinpoint pupils (2%). CONCLUSION: Vertebral artery dissections occur in a relatively young population. Most of the patients in our study had ophthalmic manifestations at the time of diagnosis. Heightened awareness of this entity should enable the consulting ophthalmologist to have a role in the diagnosis and treatment of vertebral artery dissection.     

A case of recanalized basilar artery following basilar artery occlusion due to spontaneous dissection

An 85-year-old man who did not have any previous history of heart disease was admitted to our hospital with the chief complaints of disturbance of consciousness, tetraparesis, and bilateral Babinski signs. On admission, intravenous digital subtraction angiography (DSA) showed bilatral vertebral arteries without filling the basilar artery. On the second hospital day, MRI showed a septum in the basilar artery, and no infarctions were noted in the brain stem. He gradually improved with heparin therapy. On the tenth hospital day, intravenous DSA disclosed filling of the basilar artery. On the sixteenth hospital day, conventional cerebral angiography revealed filling of the basilar artery from the left vertebral artery, and string sign and linear shadow (intimal flap) was noted in the basilar artery. He was diagnosed to have the basilar artery dissection. The patient became ambulatory and was discharged independently. This is very rare case of the basilar artery dissection with a good prognosis.     

Magnetic resonance coronary angiography in heart transplant recipients

BACKGROUND: Magnetic resonance angiography (MRA) using segmented k-space fast low-angle shot imaging has recently been used to demonstrate the proximal coronary arteries in healthy subjects and in patients with coronary artery disease. We assessed the sensitivity and specificity of coronary MRA in heart transplant recipients and investigated the feasibility of coronary MRA in patients with metallic sutures and clips in the chest. MATERIALS AND METHODS: Sixteen cardiac transplant patients aged 57.2 +/- 7.9 years (mean +/- SD) were recruited. Forty-eight arterial segments were evaluated, including the left main artery (LMA), left anterior descending artery (LADA) and right coronary artery (RCA). We excluded the left circumflex artery which could not be imaged accurately. The average time between heart transplant operation and MRA was 6 years, whereas that between MRA and X-ray angiography was 4 months. The coronary MRA was interpreted by two experienced investigators who were blinded to the coronary X-ray angiography results. Similarly, the coronary X-ray angiography results were interpreted by two experienced investigators blinded to the MRA results. The coronary arterial segments were classified by MRA as being normal or as having an amount of disease that was significant (> 50% lesion) or insignificant (< 50% lesion). RESULTS: There were 28 true-negative, five true-positive, four false-negative and six false-positive results. Of the 28 true-negative cases, 13 were in the LMA, six in the LADA and nine in the RCA. There was one false-positive LMA, two false-positive LADA and three false-positive RCA stenoses. There were four false-negative results in the LADA and one in the RCA. Clips precluded evaluation in one LMA, one LADA and one RCA. One LMA and one LADA were not evaluated as a result of poor images. One false-positive RCA stenosis was caused by a metallic clip. Three of the false-negative LADA stenoses had lesions in the distal third of the artery. The sensitivity, specificity, negative and positive predictive values were generally poor for the left coronary artery. The best results were for the RCA (sensitivity 100%, specificity 75%, positive predictive value 50% and negative predictive value 100%). The specificity in the left coronary arteries (LMA and LADA) was 86%, but the other indicators were all poorer. For the RCA, LMA and LADA combined, the overall sensitivity was 56%, specificity 82%, predictive accuracy 45% and negative predictive value 88%. In three patients, < 50% RCA lesions were seen in the MRA data, which were all confirmed by angiography. No < 50% lesions were seen in the LMA or in the LADA by MRA or by X-ray angiography. CONCLUSION: Coronary MRA using the segmented fast low-angle shot technique is feasible in heart transplant recipients but the sensitivity and specificity of this method are limited. Further developments in coil design, rapid imaging techniques and respiratory monitoring methods are necessary to improve the accuracy of coronary MRA.     

The hemodynamics and anatomy of the circle of Willis. The technic and clinical value of selective MR angiography

This study evaluates the usefulness of MR angiography in analysing the individual collateral flow dynamics and anatomy of the circle of Willis in 56 patients with high-grade extracranial carotid stenosis or occlusion. Selective MRA of the carotid or vertebrobasilar area was performed by means of presaturation up to the brain-supplying arteries at the level of the middle neck (angled presaturation slabs). Results obtained with selective and non-selective MRA in 56 consecutive patients were compared with the findings at transcranial Doppler ultrasonography and arterial angiography. Ischaemic cerebral infarctions were classified by computerized tomography and correlated with the results of collateral flow analysis: Sensitivity of selective MRA in detecting intracranial collateral flow via anterior or posterior communicating artery was 96 and 97%, respectively; sensitivity in depicting extracranial transorbital flow was lower (71%). Non-selective MRA was 100% sensitive in detecting a non-filling of the horizontal (A1) segment of the anterior cerebral artery and in identifying an origin of the posterior cerebral artery from the intracranial carotid artery. Slow flow infarctions occurred more frequently in patients with transorbital and posterior-to-anterior collateral flow than in patients with collateral flow via anterior communicating artery.     

Diagnostic reliability of the percutaneous ultrasonic Doppler technique for vertebral arterial occlusive diseases

There is little data on the diagnostic reliability of the ultrasonic Doppler technique for vertebral arterial occlusive lesions. Percutaneous vertebral Doppler examination and the vertebral angiograms were compared to determine the diagnostic reliability of this technique in 64 vertebral arteries of 53 patients with cerebrovascular disease. The percutaneous vertebral Doppler findings were quantitatively analyzed using a sound spectrograph and were classified into three types: no flow signal type, poor flow type and normal flow type. In nine patients with the no flow type the angiograms revealed vertebral occlusion or a missing vertebral artery in six, giving a diagnostic reliability of 67%. In 17 patients with poor flow type the angiograms revealed vertebral occlusion or a missing vertebral artery in five, terminal narrowing of the artery in nine, and hypoplasia in two giving a diagnostic reliability of 94%. For all vertebral arteries examined with this technique, including normal ones, the diagnostic reliability was 92% (59/64). Percutaneous vertebral Doppler examination has clinical usefulness as a screening test for occlusive vertebral arterial diseases.     

Complete MR angiography and Doppler ultrasound as the sole imaging modalities prior to carotid endarterectomy

OBJECTIVE: To assess the combination of duplex Doppler ultrasound (DUS) and complete carotid magnetic resonance angiography (MRA) for the non-invasive imaging of carotid disease and their effect on outcomes. Determine inter-reader agreement of carotid MRA. MATERIALS AND METHODS: One-hundred and ten carotid bifurcations were evaluated using DUS, 2D and 3D time-of-flight MRA from the aortic arch to the Circle of Willis in 55 patients. Percentage stenoses were determined by two blinded readers using standardized criteria. Clinical follow-up was by chart review. RESULTS: Correlation of Doppler and MRA was excellent (r=0.903, P<0.001). Inter-reader agreement (K) for MRA was good: internal carotid artery (ICA) (0.750), external carotid artery (ECA) (0.674) and common carotid artery (CCA) (0.410). Differences in CCA readings were due to minor differences in categorizing lesions as CCA versus ICA or ECA. MRA and Doppler detected nine occluded ICAs. Two DUS occlusions had ICA flow by MRA; one due to a reconstituted precavernous ICA, one a near occluded vessel. Five patients (9%) had surgical management modified by MRA with four not having surgery: three distal ICA/Siphon occlusions and one less severe stenosis by MRA. One tandem lesion not visualized by DUS was surgically significant. Nine aortic arch abnormalities had no surgical impact, possibly due to small sample size. Of 41 endarterectomies, there were no complications from errors of diagnosis. CONCLUSION: Carotid MRA correlates well with DUS with good inter-reader agreement. MRA confirms Doppler findings, expands anatomical information and identifies tandem lesions from the aortic arch to the Circle of Willis which can affect surgical management. This approach to carotid artery imaging appears to have no negative effect on surgical outcome.     

Magnetic resonance angiography: fields of exploration, main indications and limitations

Magnetic resonance angiography (MRA) has become a widely accepted technique with regards to the other available noninvasive techniques in the diagnosis of vascular disease. This paper proposes a review of the different indications of this technique in arterial and venous diseases. Among several MRA pulse sequences, the most frequently used until today consisted of a time-of-flight technique which provided angiograms without any injection of intravascular contrast medium. It required to be performed in a plane perpendicular to the main axis of the vessel to be optimal. New techniques, such as contrast medium bolus-enhanced acquisitions allow examination of vascular segments in a plane parallel to their course (coronal for the aorta and lower limb arteries). An increasing number of clinical applications has raised since the implementation of MRA techniques on MR devices; some of them are widely accepted, whereas some others remain under the scope of extensive validation. With a high level of accuracy in grading carotid artery stenosis, MRA is now routinely used in cerebral arterial occlusive diseases and has in part replaced contrast angiography. MRA of the venous system of the brain plays a major role in the diagnosis and follow up of dural venous thrombosis. Other vascular brain diseases, such as vascular malformations, yet have limited uses. Carotid artery dissections are fairly demonstrated with MRA, which can be used for diagnosis as well as for follow-up. The accuracy of MRA in the diagnosis of venous thrombosis of the cervical/mediastinal veins has been reported as high as 100%. Moreover, MRA allows a precise assessment of collateral vessels in case of complete cervical/mediastinal venous thrombosis.     

Magnetic resonance angiography with TOF-TONE tridimensional sequences in the study of steno-occlusive diseases of the renal arteries. Comparison with digital angiography

The authors examined 32 patients (68 renal arteries in all) with suspected renovascular hypertension with MRA (3D TOF-TONE sequences) and DSA. MRA visualization of the renal vessels was considered good in 96.8%, 91.6%, 76.6% and 16.6% of cases for the ostium, the proximal third, the distal third and the hilum, respectively. MRA correctly assessed the 4 occlusions in our series and 19/20 atherosclerotic stenoses, all in proximal site. In proximal stenosis detection, MRA had 100% sensitivity and 98% specificity in atherosclerotic sclerosis-occlusion grading considering a 50% caliber reduction as the cut-off value. The two techniques were in agreement in 68% of cases; MRA overestimated stenosis grade in 25% of cases. The 3D TOF-TONE sequence is a reliable diagnostic tool in the study of the proximal tract of the renal arteries, but its major limitation remains in distal tract studies. Nevertheless, this technique has elective indications in a selected group of patients, namely the elderly with hypertension and impaired renal function.     

Ability of commercial demineralized freeze-dried bone allograft to induce new bone formation is dependent on donor age but not gender

OBJECTIVE: To describe the clinical features of patients with occlusive disease of the proximal (V1) segment of the vertebral artery. DESIGN AND PATIENTS: Patients with either occlusion or high-grade stenosis involving the V1 segment were chosen for study from the New England Medical Center Posterior Circulation Registry. The registry is a consecutive series of patients with signs and symptoms of posterior circulation ischemia seen at the New England Medical Center, Boston, Mass, during a 10-year period. Clinical features, radiographic findings, and patient outcome were reviewed. RESULTS: Of the 407 patients in the registry, 80 (20%) had V1 segment lesions. Patients could be classified into 5 groups: (1) V1 disease and coexistent severe intracranial occlusive disease of the posterior circulation (n=22); (2) V1 disease with evidence of artery-to-artery embolism (n=19); (3) suspected V1 disease with artery-to-artery embolism, but with other potential causes of stroke or less certain vascular diagnosis (n=20); (4) V1 disease associated with hemodynamic transient ischemic attacks (n=13); and (5) proximal vertebral arterial dissection (n=6). Hypertension, cigarette smoking, and coronary artery disease were common risk factors. Clinical features, location of infarct, and outcome differed between groups and reflected the presumed mechanisms of stroke. CONCLUSIONS: Occlusive disease involving the V1 segment of the vertebral artery is common in patients with posterior circulation ischemia, but is often associated with other potential mechanisms of stroke. However, in a series of patients seen at a tertiary referral center, occlusive disease of the V1 segment was the primary mechanism of ischemia in 9% of patients.     

MRI and MRA for diagnosis and follow-up of cerebral venous thrombosis (CVT)

OBJECTIVE: To assess the sensitivity of magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) in the diagnosis and the follow-up of cerebral venous thrombosis (CVT). To describe MRI and MRA features in CVT. MATERIAL AND METHODS: The clinical and neuroradiological studies of 20 patients with proven CVT were reviewed. All patients underwent computed tomography (CT), MRI (with MRA for 15 patients) and digital substraction angiography (DSA). Eleven patients had follow-up after treatment with MRI (nine with MRA). RESULTS: MRI and MRA together provided the diagnosis of CVT in all cases. The sensitivity of MRI alone was 90%. MRA showed abnormalities in all cases of CVT. Progressive sinus recanalization was demonstrated by follow-up with MRI and MRA at least 15 days after diagnosis and treatment. CONCLUSION: In most of cases, the combination of MRI and MRA is sufficient to allow diagnosis of CVT and obviates the need for invasive angiography. MR studies are also useful for the follow-up of CVT.     

Combined use of duplex imaging and magnetic resonance angiography for evaluation of patients with symptomatic ipsilateral high-grade carotid stenosis

PURPOSE: Advances in cerebral vascular imaging suggest that patients with critical levels of carotid artery stenosis (> 70%) who have symptoms can be identified accurately and necessary information about the intracranial and extracranial circulation obtained before surgery without conventional angiography. We have used carotid duplex imaging in combination with magnetic resonance angiography (MRA) to evaluate 20 patients with symptomatic ipsilateral high-grade carotid stenosis. METHODS: All patients underwent CT and magnetic resonance imaging brain scans, as well as MRA and conventional arteriography of the cerebral circulation. Magnetic resonance angiograms were obtained with two-dimensional phase contrast and time-of-flight techniques. Phase contrast was used for intracranial vascular imaging and for determining qualitative flow velocities and the direction of blood flow in the circle of Willis. Two-dimensional time of flight was used to assess the carotid bifurcations. RESULTS: Twenty patients with symptoms (six with strokes, 11 with transient ischemic attacks, and three with amaurosis fugax) had duplex evidence of high-grade carotid stenoses. Computed tomographic and magnetic resonance brain scans were positive for cerebral infarction in six patients with clinical strokes. Comparison of MRA with conventional angiography was 91% accurate for high-grade stenoses and occlusions (sensitivity 100% and specificity 90% for stenosis; sensitivity/specificity was 100% for complete occlusion). Comparison of duplex imaging with conventional angiography demonstrated 86% accuracy for detection of severe stenosis or occlusion (sensitivity 94% and specificity 89% for stenosis; sensitivity and specificity were 100% for complete occlusion). CONCLUSIONS: This study suggests that combined use of MRA and duplex imaging is accurate for detection and evaluation of high-grade carotid stenoses in patients with symptoms.