The distribution and size of ischemic lesions after carotid artery angioplasty and stenting: evidence for microembolization to terminal arteries.

BACKGROUND: Much of the brain is perfused by penetrating arteries that are the "single source" of blood to their surrounding tissues. These tissues should be equally vulnerable to ischemia from embolic occlusion, but there are questions about whether emboli have access to the penetrating arteries serving the deep brain tissues. To examine this issue in humans we recorded the number and distribution of new ischemic lesions on diffusion-weighted magnetic resonance imaging (DWMRI) after carotid artery stenting (CAS), a procedure producing showers of numerous small atheroemboli. METHODS: Twenty-nine men (aged 62-81) underwent 30 CAS procedures with distal protection in place, and DWMRI 48 hours after the procedure documented new lesions had developed. Thirteen patients were asymptomatic, and 16 had experienced recent symptoms ipsilateral to the treated carotid stenosis. A DWMRI study was done in each patient ≤72 hours before the procedure. All MRI studies were read by the same neuroradiologist. RESULTS: One patient sustained a minor stroke, which resolved. DWNRI found 131 new lesions (median, 3; range, 1-17; interquartile range, 2-4). Lesion size was <5 mm in 96.6% and 5 to 10 mm in 3.1%. Lesions were ipsilateral in 83.1% and contralateral in 16.9%. Lesions were in the distribution of the middle cerebral artery (91.6%), posterior cerebral artery (6.1%), and superior cerebellar artery subclavian artery (2.0%). Most lesions were in the cortex but at a depth where they were best described as cortical/subcortical (90.8%). The rest were in the periventricular white matter (6.1%) and deep gray matter (3.0%). CONCLUSIONS: The ischemic areas developing after CAS were predominately in the deeper layers of the cortex in the distribution of the middle cerebral artery, but lesions were seen throughout the brain. The distribution of lesions caused by CAS-induced embolization coincided with estimates of blood flow to the respective areas of the brain. These data add to the evidence implicating microemboli in ischemic pathologies throughout the brain.

Distal filtration versus flow reversal: An ex vivo assessment of the choices for carotid embolic protection.

OBJECTIVES: Choices for embolic protection during carotid stent procedures include distal filtration (DF) and proximal occlusion with flow reversal (POFR). DF devices are widely used but have produced only modest improvements in clinical outcomes. There is less experience with POFR devices but single center reports suggest reduced emboli detected by transcranial Doppler (TCD). To determine if POFR offers a significant improvement in embolic protection, we tested five DF devices and two POFR devices with 8F and 10F sheath design in an ex vivo angioplasty system using human carotid plaques excised en bloc. Physiologic pressures and flows were used and the efficiency of plaque fragment removal by these devices compared. METHODS: Thirty-three human carotid plaques removed en bloc were secured in tailored polytetrafluoroethylene (PTFE) grafts. The distal PTFE was either 6 mm or 5 mm inner diameter (ID). Saline was delivered through the excised carotid plaque as follows: a cleaning 50 mL flush was done prior to the angioplasty procedure and discarded; further flushes of forward flow were done with five pressurized "pulsations" of 10 mL each (50 mL), peak pressure 140 mm Hg. Balloon angioplasty was done with a 4 mm and then a 6 mm balloon. DF flushes were applied after each angioplasty and "postprocedure" after the device was removed. With POFR, 50 mL were collected through the sheath after balloon angioplasty by either back-pressure of 20 mm Hg, 40 mm Hg or 60 mm Hg, or by aspiration. Postangioplasty pressurized forward flush of 50 or 100 mL was done as described. Each flush was collected, centrifuged, and examined for plaque fragments. Fragments greater than 60 microns were sized and counted on a 100 micron grid. RESULTS: When DF devices were used in 6 mm lumen PTFE, the percent of fragments trapped was poor (13.7% to 27.8%). There were no statistically significant differences between the devices. The capture of fragments improved (22% vs 51.4%, P < .001) when devices appropriate for a 6 mm lumen were used in a 5 mm PTFE "ICA", functionally over-sizing the devices. POFR efficiency improved with increasing back-pressures and with repeated aspirations. Postprocedure, successive flushes of pressurized forward flow yielded additional plaque fragments and when the efficiency of POFR was assessed with forward flushing volumes similar to those used for DF, the efficiencies were similar, although larger fragments were more efficiently removed with POFR. CONCLUSION: In our model, both protection strategies were less than ideal. For POFR, high back pressures or multiple aspirations improve the efficiency of cerebral protection but additional fragments were released by pressurized flow even after aspiration of 150 mL of saline. DF devices create a pressure gradient and fragments apparently went around the device with pressurized flow in our PTFE lumen. Over-sizing of DF devices partially corrected this problem and increased over all DF efficiency to be comparable to POFR for smaller fragments but not for larger fragments.

Microemboli composed of cholesterol crystals disrupt the blood-brain barrier and reduce cognition.

BACKGROUND AND PURPOSE: Microemboli occur frequently in patients with asymptomatic carotid atherosclerosis. In other vascular beds, microemboli are known to initiate an inflammatory response, causing organ dysfunction. In the current study, we investigated whether emboli composed of cholesterol crystals, a component of human atherosclerotic plaque, could also cause inflammation and brain dysfunction demonstrated by cognitive impairment. METHODS: Cholesterol crystals of 60 to 100 microm were injected via the rat internal carotid artery. T2-weighted magnetic resonance imaging was conducted after 3 days to estimate infarct volume. Brains were examined for matrix metalloproteinase activation at 24 hours and for albumin leakage and microglia and astrocyte activation at 4 days and 1, 2, and 4 weeks after embolization. To determine changes in cognition, behavioral tests including open field, motor learning, and Barnes Maze tests were conducted on young adult and middle-aged rats 4 weeks after either a single injection or after repeated, bilateral injections given at an interval of 2 weeks. RESULTS: Matrix metalloproteinase activation was detected in 50% of the animals examined. Perivascular albumin staining was found at 4 days but rarely persisted beyond 1 week. Activation of microglia and astrocytes occurred in all animals and persisted for up to 8 weeks. Cognitive impairment was observed in middle-aged rats after repeated, bilateral injections but not after single injections. In these animals, areas of inflammation were small and scattered but often involved the striatum and hippocampus. CONCLUSIONS: Cholesterol embolization caused an inflammatory response in the brain with persistent activation of microglia and astrocytes and led to cognitive impairment after repeated injections in middle-aged animals with only small foci of neural injury. These data indicate that microembolization causes inflammation and that minimal neuronal injury can cause cognitive impairment in older animals.

An experimental model of lacunar infarction: embolization of microthrombi.

OBJECTIVES: Microthrombi are undoubtedly the most common embolic material in the cerebral circulation, originating from even minor irregularities of the arterial wall, fibrillating atria, cardiac valves, and patent foramen ovale. Thrombus fragments are globular and likely to completely obstruct terminal vessels. In contrast, previous work with "atheroemboli" of needle-like cholesterol crystals rarely cause occlusions or infarctions instead creating small foci of inflammation. In this work, we asked if microthrombi would occlude terminal vessels and create lacunar type infarctions in the subcortical tissues of the rat brain where, as in human brain, collateral flow is limited relative to the cortex. METHODS: Three treatment groups of adult male Sprague-Dawley rats were studied. All groups underwent general anesthesia with monitoring of temperature and blood pressure during cannulation of the right internal carotid artery. In the group embolized with thrombus fragments (n = 12), animals had injections of 300 fragments of thrombus size 60 to 100 microns, the cholesterol group (n = 6) had injections of 300 cholesterol crystals of similar size, and the control group (n = 4) had injections of saline. Brains were harvested at 4 days with perfusion fixation and were examined by immunohistochemical staining for breaks in the blood brain barrier (BBB) (albumin), microglial activation (CD11b), astrocyte activation (GFAP), and infarction (loss of NeuN staining). Size and location of the areas of injury and infarction were recorded. RESULTS: Clot fragments caused discreet infarcts in 10/12 animals that were 0.1-1.7 mm in diameter and coincided with activation of microglia and astrocytes. In some areas, necrosis was already underway at this early time point. Consistent with our previous work, the infarcts caused by cholesterol crystals were smaller (P = .014). Foci of BBB disruption and microglial activation were distributed throughout the brain whereas areas of infarction were found almost exclusively in subcortical tissues (P = .029). CONCLUSIONS: Injecting microthrombi reproducibly caused areas of necrosis resembling lacunar type infarctions. These were primarily located in the striatum and thalamus presumably because these areas lack the branching, collateral network seen in the cortex. In addition, these data give further evidence that the extent of brain injury from emboli depends upon composition and shape as well as size.

Atheroembolism during percutaneous renal artery revascularization.

INTRODUCTION: Atheroembolization during renal artery angioplasty and stenting (RA-PTAS) has been postulated as a cause for the inferior renal function results observed when compared with those with surgical revascularization. To further characterize procedure-associated atheroembolism, we analyzed recovered atheroembolic debris and clinical data from patients undergoing RA-PTAS with distal embolic protection (DEP). METHODS: RA-PTAS procedures were performed with DEP using a commercially available temporary balloon occlusion and aspiration catheter system between July 2005 and December 2006. Following RA-PTAS but prior to deflation of the distal occlusion balloon, the static column of blood proximal to the balloon was aspirated and submitted for embolic particle analysis. Angiograms, demographics, and laboratory data were reviewed. Glomerular filtration rate (eGFR) was estimated before RA-PTAS and at 4 to 8 weeks postintervention using the abbreviated Modification of Diet in Renal Disease formula. Associations between clinical factors, captured particle counts, and changes in renal function were examined using univariate techniques and multiple linear regression. RESULTS: Twenty-eight RA-PTAS procedures were performed with DEP. Mean total number of embolic particles counted per procedure was 2033 +/- 1553 for particles 20-60 microm and 265 +/- 132 for particles >60 microm. Significant positive associations with quantity of captured particles 20 to 60 microm were observed for African American race (P = .002), predilation (P = .005), and stent diameter (P < .001); a significant negative association was observed for preoperative aspirin use (P =.016). Quantity of captured particles >60 microm was positively associated with ratio of stent to renal artery diameter (P =.009). Change in eGFR was positively associated with preoperative aspirin use (P = .006) and preoperative eGFR (P < .001), while a negative association was observed for captured particle counts >60 microm (P = .015). CONCLUSION: These results demonstrate the liberation of thousands of atheroembolic particles during RA-PTAS. Clinical, anatomic, and device-related factors may be predictive of procedural embolization, and increasing captured particle counts >60 microm were associated with inferior renal function results. Further investigation is warranted to establish relationships between atheroembolism, end organ functional impairment, and clinical responses.

Subclinical embolization after carotid artery stenting: new lesions on diffusion-weighted magnetic resonance imaging occur postprocedure.

OBJECTIVES: The reported rate of subclinical brain injury after carotid artery stenting (CAS) seen on diffusion-weighted magnetic resonance imaging (DWI) varies from 10% to >40%. Data from transcranial Doppler after CAS indicate that embolization may continue for several days, suggesting that that at least some lesions seen on DWI occur postprocedure. Because DWI lesions appear or=60 microm. RESULTS: There were two periprocedural strokes and one transient ischemic attack (TIA), but no strokes or TIAs occurred during follow-up. In the 23 patients imaged 1 hour postprocedure, new lesions were found in two (9%), and 18 (78%) had new lesions at 48 hours (P < .001). For the entire study group, the incidence of new lesions at 48 hours was 67% (36/54). The median number of DWI lesions was four (range, 1 to 17). Every protection device examined had atherosclerotic debris, with a mean of 135 +/- 73 fragments (range, 18 to 310) sized >60 microm and a mean of eight fragments (range, 2 to 21) sized >500 microm. Findings on postprocedure DWI did not correlate with the degree of stenosis, size of angioplasty balloon, or number of inflations, nor with the number or size of fragments retrieved from the protection device. CONCLUSIONS: CAS can be performed with a very low incidence of clinically evident neurologic events; however, it is associated with embolization during and after the procedure. Protection devices effectively prevent clinical and subclinical events during the procedure. Significant embolization continues for at least 48 hours postprocedure, causing lesions on DWI when there is no mechanism for cerebral protection. These data correlate with transcranial Doppler reports of continued embolization after CAS and indicate that DWI should be done as late as possible to accurately assess the rate of subclinical brain injury with CAS procedures.

Ezetimibe inhibits the incorporation of dietary oxidized cholesterol into lipoproteins.

Oxidized cholesterol is present in significant quantities in the typical Western diet. When ingested, oxidized cholesterol is absorbed by the small intestine and incorporated into both chylomicrons and LDL, resulting in LDL that is more susceptible to further oxidation. Feeding studies in animal models and epidemiological studies in humans have suggested that oxidized cholesterol in the diet increases the development of atherosclerosis. In this study, we determined the effect of ezetimibe, a drug that inhibits small intestinal absorption of cholesterol, on the levels of oxidized cholesterol in the serum after a test meal containing oxidized cholesterol. We demonstrate that ezetimibe, 10 mg per day for 1 month, markedly reduced the levels (50% decrease) of oxidized cholesterol in the serum after feeding a test meal containing either alpha-epoxy cholesterol or 7-keto cholesterol, two of the predominant oxidized cholesterols found in the diet. Moreover, the decrease in oxidized cholesterol in the serum was attributable to a decrease in the incorporation of dietary oxidized cholesterol into both chylomicrons and LDL. Because there was no decrease in postprandial triglyceride levels, we conclude that this decrease in oxidized cholesterol levels in the serum is attributable to decreased absorption and not to enhanced clearance. Whether this decrease in oxidized cholesterol absorption prevents or delays the development of atherosclerosis remains to be determined.

The role of dietary oxidized cholesterol and oxidized fatty acids in the development of atherosclerosis.

The etiology of atherosclerosis is complex and multifactorial but there is extensive evidence indicating that oxidized lipoproteins may play a key role. At present, the site and mechanism by which lipoproteins are oxidized are not resolved, and it is not clear if oxidized lipoproteins form locally in the artery wall and/or are sequestered in atherosclerotic lesions following the uptake of circulating oxidized lipoproteins. We have been focusing our studies on demonstrating that such potentially atherogenic oxidized lipoproteins in the circulation are at least partially derived from oxidized lipids in the diet. Thus, the purpose of our work has been to determine in humans whether oxidized dietary oxidized fats such as oxidized fatty acids and oxidized cholesterol are absorbed and contribute to the pool of oxidized lipids in circulating lipoproteins. When a meal containing oxidized linoleic acid was fed to normal subjects, oxidized fatty acids were found only in the postprandial chylomicron/chylomicron remnants (CM/RM) which were cleared from circulation within 8 h. No oxidized fatty acids were detected in low density lipoprotein (LDL) or high density lipoprotein (HDL) fractions at any time. However, when alpha-epoxy cholesterol was fed to human subjects, alpha-epoxy cholesterol in serum was found in CM/RM and also in endogenous very low density lipoprotein, LDL, and HDL and remained in the circulation for 72 h. In vitro incubation of the CM/RM fraction containing alpha-epoxy cholesterol with human LDL and HDL that did not contain alpha-epoxy cholesterol resulted in a rapid transfer of oxidized cholesterol from CM/RM to both LDL and HDL. We have suggested that cholesteryl ester transfer protein is mediating the transfer. Thus, alpha-epoxy cholesterol in the diet is incorporated into CM/RM fraction and then transferred to LDL and HDL contributing to lipoprotein oxidation. We hypothesize that diet-derived oxidized fatty acids in chylomicron remnants and oxidized cholesterol in remnants and LDL accelerate atherosclerosis by increasing oxidized lipid levels in circulating LDL and chylomicron remnants. This hypothesis is supported by our feeding experiments in animals. When rabbits were fed oxidized fatty acids or oxidized cholesterol, the fatty streak lesions in the aorta were increased by 100%. Moreover, dietary oxidized cholesterol significantly increased aortic lesions in apo-E and LDL receptor-deficient mice. A typical Western diet is rich in oxidized fats and therefore could contribute to the increased arterial atherosclerosis in our population.

Atheroemboli during renal artery angioplasty: an ex vivo study.

OBJECTIVE: We hypothesized that atheroemboli released during renal angioplasty could be responsible for the modest functional result of renal angioplasty even after anatomic reduction of renal artery stenosis. To test this hypothesis, we enumerated and sized fragments released during ex vivo angioplasty and stenting of human renal artery atherosclerotic specimens removed during aortorenal endarterectomy. METHODS: Thirty-three intact aortorenal atheroma specimens (16 pairs with adjacent aortic atheroma and one specimen with a single renal artery orifice) were removed from 17 patients with renal artery occlusive disease who underwent renal artery endarterectomy. specimens. Endarterectomy specimens were removed with a ring of aortic plaque and "fitted" with a polytetrafluoroethylene "adventitia". Ex vivo angioplasty was technically successful in 31 of the 33 specimens and was performed by using a 0.018-inch guidewire and 3.0-mm and 5.0-mm angioplasty balloons inflated for 30 seconds at 15 atmospheres pressure. Stenting was performed with either a 5-mm or 6-mm self-expanding Wallstent. Each artery was flushed with 20 mL of saline after guidewire placement, each angioplasty, and stent placement. The effluent was collected for analysis for counting with either a microscope (size >100 microm) or a Coulter counter (size <100 microm). The number and size of embolic fragments in the effluent collected after each manipulation was recorded. RESULTS: Each manipulation of the specimens, including simply advancing the guidewire through the atherosclerotic lesion, released thousands of fragments. The numbers of fragments in each size category increased with decreasing particle size. Positioning and deploying the Wallstent released an additional bolus of fragments similar to that released after balloon angioplasty. CONCLUSIONS: Ex vivo renal angioplasty releases thousands of atherosclerotic fragments of sufficient size to create vascular occlusions and initiate significant renal parenchymal damage. The results of renal angioplasty procedures could be improved by placing distal protection devices to prevent atheroembolization. CLINICAL RELEVANCE: Athero-emboli produce a local arteritis in the kidney and could cause substantial damage to the renal parenchyma. This report explores the quantity of athero-emboli released during ex vivo angioplasty and stenting of renal atheroma specimens. The number of emboli found in this ex vivo study suggest that the use of protection devices may be advisable to protect the end organ, as done with angioplasty of the carotid artery. Of necessity, this was an ex vivo study and direct application to the clinical setting will need further study. Fortunately, multi-center trials examining the value of protection devices are currently in progress.

Contrast material-enhanced MRA overestimates severity of carotid stenosis, compared with 3D time-of-flight MRA.

OBJECTIVE: Non-contrast-enhanced magnetic resonance angiography (MRA) carotid imaging with the time-of-flight (TOF) technique compares favorably with angiography, ultrasound, and excised plaques. However, gadolinium contrast-enhanced MRA (CE-MRA) has almost universally replaced TOF-MRA, because it reduces imaging time (25 seconds vs 10 minutes) and improves signal-to-noise ratio. In our practice we found alarming discrepancies between CE-MRA and TOF-MRA, which was the impetus for this study.Study design To compare the two techniques, we measured stenosis, demonstrated on three-dimensional images obtained at TOF and CE-MRA, in 107 carotid arteries in 58 male patients. The measurements were made on a Cemax workstation equipped with enlargement and measurement tools. Measurements to 0.1 mm were made at 90 degrees to the flow channel at the area of maximal stenosis and distal to the bulb where the borders of the internal carotid artery lumen were judged to be parallel (North American Symptomatic Carotid Endarterectomy Trial criteria). Experiments with carotid phantoms were done to test the contribution of imaging software to image quality. RESULTS: Twelve arteries were occluded. In the remaining 95 arteries, compared with TOF-MRA, CE-MRA demonstrated a greater degree of stenosis in 42 arteries, a lesser degree of stenosis in 14 arteries, and similar (+/-5%) stenosis in 39 arteries (P =.02, chi(2) analysis). The largest discrepancies were arteries with 0% to 70% stenosis. In those arteries in which CE-MRA identified a greater degree of stenosis than shown with TOF-MRA, mean increase was 21% for 0% to 29% stenosis, 36% for 30% to 49% stenosis, and 38% for of 50% to 69% stenosis. The carotid phantom experiments showed that the imaging parameters of CE-MRA, particularly the plane on which frequency encoding gradients were applied, reduced signal acquisition at the area of stenosis. CONCLUSIONS: Collectively these data demonstrate that CE-MRA parameters must be retooled if the method is to be considered reliable for determination of severity of carotid artery stenosis. CE-MRA is an excellent screening technique, but only TOF-MRA should be used to determine degree of carotid artery stenosis.

Cerebral ischemia and infarction from atheroemboli <100 microm in Size.

BACKGROUND AND PURPOSE: To determine the importance of emboli not trapped by carotid angioplasty filtration devices, we examined fragments <100 microm released with ex vivo angioplasty and asked if fragment composition and size correlated with brain injury. METHODS: Human carotid plaques (21) were excised en bloc, and ex vivo carotid angioplasty was performed. Eight plaques were selected as either highly calcified (4) or highly fibrotic (4) by high-resolution MRI (200 microm3). Fragments were counted by a Coulter counter. Before injection into male Sprague-Dawley rats, fragments from calcified and fibrotic plaques were sized with 60-, 100-, and 200-microm filters. Brain ischemia and infarction were assessed by MRI scans (7-T small-bore magnet) and by immunohistologic staining for HSP70 and NueN. RESULTS: All 5 animals injected with 100- to 200-microm calcified fragments had infarctions. One was lethal. After injection of 60- to 100-microm calcified fragments, 7 of 12 animals had cerebral infarctions, whereas only 1 of 11 had infarctions with fibrous fragments (P<0.02). HSP70 staining showed that ischemia was more common and more extensive than infarction. Ischemia was found in 10 of 12 animals after injection of calcified fragments and in 9 of 11 after injection of fibrous fragments. The mean number of 60- to 100-microm fragments released was 375+/-510; the mean number of 20- to 60-microm fragments was 34 196 (range, 2230 to 186 927). CONCLUSIONS: Hundreds of thousands of microemboli can be shed during carotid angioplasty. Fragments from calcified plaques cause greater levels of infarction than fragments from fibrous plaques, although ischemia is common with both fragment types.

Oxidized cholesterol in the diet is a source of oxidized lipoproteins in human serum.

The aim of this study was to determine in humans whether oxidized cholesterol in the diet is absorbed and contributes to the pool of oxidized lipids in circulating lipoproteins. When a meal containing 400 mg cholestan-5alpha,6alpha-epoxy-3beta-ol (alpha-epoxy cholesterol) was fed to six controls and three subjects with Type III hyperlipoproteinemia, alpha-epoxy cholesterol in serum was found in chylomicron/chylomicron remnants (CM/RM) and endogenous (VLDL, LDL, and HDL) lipoproteins. In controls, alpha-epoxy cholesterol in CM/RM was decreased by 10 h, whereas in endogenous lipoproteins it remained in the circulation for 72 h. In subjects with Type III hyperlipoproteinemia, alpha-epoxy cholesterol was mainly in CM/RM. In vitro incubation of the CM/RM fraction containing alpha-epoxy cholesterol with human LDL and HDL that did not contain alpha-epoxy cholesterol resulted in a rapid transfer of oxidized cholesterol from CM/RM to both LDL and HDL. In contrast, no transfer was observed when human serum was substituted with rat serum, suggesting that cholesteryl ester transfer protein is mediating the transfer. Thus, alpha-epoxy cholesterol in the diet is incorporated into the CM/RM fraction and then transferred to LDL and HDL, contributing to lipoprotein oxidation. Moreover, LDL containing alpha-epoxy cholesterol displayed increased susceptibility to further copper oxidation in vitro. It is possible that oxidized cholesterol in the diet accelerates atherosclerosis by increasing oxidized cholesterol levels in circulating LDL and chylomicron remnants.

Major carotid plaque surface irregularities correlate with neurologic symptoms.

PURPOSE: Many studies have linked carotid plaque surface irregularities with stroke risk, but this relationship has been obscured by the limited ability of available imaging modalities to resolve plaque surface morphology. To address this issue, we performed a prospective study correlating the presenting neurologic symptoms of patients with high-resolution magnetic resonance imaging (MRI; 200 microm) studies of ipsilateral plaque surface invaginations and ledges, lumen shape, and the location of the plaque bulk creating the stenosis. METHODS: One hundred patients, 17 women and 83 men, 45 to 81 years old (mean, 68 years) underwent surgery. Forty-five patients had a transient ischemic attack (TIA) or stroke as the indication for surgery, and 55 patients had no symptoms. Angiograms were obtained in 50 patients. Carotid plaques were removed "en bloc" and placed in gadolinium doped saline for imaging in a Siemens Symphony, 1.5T scanner with a custom-built transmit-receive radiofrequency coil. The resulting slice thickness was 200 microm, with 200 microm by 200 microm in plane resolution. The MRI data and angiograms were reviewed by using National Institutes of Health Image software and read by consensus. A surface irregularity was categorized as a ledge or ulcer and measured by using electronic calipers. Luminal shape was determined at the point of maximal stenosis with a "slice" set at 90 degrees to the lumen axis. The location of the maximal stenosis was recorded. In the internal carotid artery, plaque bulk was designated to be on the flow divider wall or non-flow divider wall. RESULTS: The mean maximal stenosis was 81.5% +/- 12.0%. Surface contour irregularities were found in 80 plaques. Thirty-five plaques were graded as having major surface contour irregularities, and 45 plaques were graded as having minor irregularities. There was a significant correlation between major surface irregularity and TIA or stroke (P <.01). Irregular plaques were identified with angiography, but the irregularity in size was underestimated (P <.01). Only 28% of plaques had circular lumens; 50% had elliptical lumens, and 22% had either crescentic or multi-lobular lumens. The maximal stenosis was located in the internal carotid artery in 82 plaques, the bifurcation in 17 plaques, and the common carotid artery in one plaque. CONCLUSION: Surface irregularities were revealed by means of submillimeter resolution of the carotid plaques with MRI to be common, but only the presence of major irregularities correlated with the patient having TIA or stroke. Lumen shape and plaque location did not appear to predict stroke risk, but may effect imaging accuracy in determining the degree of stenosis. These data further define the relationship of plaque irregularity and cerebrovascular symptoms caused by atheroemboli.