The Paradoxical Flow Hypothesis of the Carotid Artery: Supporting Evidence from Phase-contrast Magnetic Resonance Imaging

The Paradoxical Flow Hypothesis of the Carotid Artery: Supporting Evidence from Phase-contrast Magnetic Resonance Imaging Peter J. Yim, PhD, Amish Tilara, BA, and John L. Nosher, MD

Objective: Narrowing of the poststenotic internal carotid artery (ICA) has been found to be associated with reduced risk of ipsilateral stroke. A paradoxical mechanism has been hypothesized to explain this finding: narrowing of the distal-normal (reference) ICA is associated with low blood flow rates (Q) in the stenotic ICA, and lower Q causes lower risk of ipsilateral stroke, perhaps by an associated reduction in mechanical stress on the atherosclerotic plaque. The purpose of this study was to confirm that the reference ICA diameter (RICAD) is indeed predictive of Q, a finding that would indirectly support the hypothesis of a relationship between lower Q and lower risk of ipsilateral stroke. Methods: Magnetic resonance imaging from 38 patients was included in the study. The study included 17 stenotic carotid arteries and 59 normal carotid arteries. All patients underwent contrast-enhanced magnetic resonance angiography from which measurements were obtained of the RICAD and the internal-common carotid diameter ratio. Patients underwent cardiac-gated, velocity-encoded phase-contrast magnetic resonance imaging for measurement of Q. Results: Mean flow rates differed between stenotic (4.3 6 1.7 mL/s) and normal (5.4 6 1.7 mL/s) arteries (P 5 .02). RICAD was found to be a predictor of Q for stenotic arteries (P 5 .009) and for all arteries (P 5.025) but not for the group of normal arteries (P 5 .162). Right-left differences in RICAD were highly predictive of rightleft differences in Q in the subgroup of individuals with normal arteries (P , .001) and in the group of all participants (P , .001). Internal-common carotid diameter ratio was not found to be a statistically significant predictor of Q in the subgroup of stenotic arteries (P 5.156). Conclusions: This study demonstrated that, as hypothesized, RICAD is correlated with Q. Key Words: Carotid stenosis—magnetic resonance imaging—blood flow. Ó 2008 by National Stroke Association

The relationship between blood flow rate (Q) in the internal carotid artery (ICA) and the risk of stroke is poorly understood. Recently, however, Rothwell and Warlow1

From the Department of Radiology, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, New Brunswick, New Jersey. Received August 9, 2007; revision received October 2, 2007; accepted October 12, 2007. Address correspondence to Peter J. Yim, PhD, Department of Radiology, UMDNJ-Robert Wood Johnson Medical School, Medical Education Bldg 404, New Brunswick, NJ 08903. E-mail: yimpj@ umdnj.edu. 1052-3057/$—see front matter Ó 2008 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2007.10.004

found that narrowing of a normal segment of the ICA distal to a severe stenosis significantly reduces the 5-year risk of stroke. They hypothesized that narrowing of the distal normal ICA is linked to reduced risk of stroke via reduced blood flow in the ICA; reduced blood flow causes both the reduction in the distal normal ICA diameter and the reduction in the risk of stroke. The reason why a reduction in Q would reduce the risk of ipsilateral stroke is not clear at this point, but might be the result of an associated reduction in mechanical stress on the atherosclerotic plaque. The hypothesized relationship between low Q and low risk of ipsilateral stroke is consistent with the finding from a retrospective study carried out by Henderson et al2 in which intracerebral collateral-flow pathways were examined in patients with carotid artery stenosis.

Journal of Stroke and Cerebrovascular Diseases, Vol. 17, No. 2 (March-April), 2008: pp 101-108

101

YIM, TILARA, AND NOSHER

102

The study found that patients with more prominent collateral-flow pathways supplying the territory of the stenotic carotid artery were significantly less likely to have had an ipsilateral stroke at 2-year follow-up. Presumably, the collateral-flow pathways supply blood to the territory of the stenotic ICA and, thus, reduce the demand for blood flow through the stenotic ICA. The association between compromised intracerebral collateral flow and recent history of cerebrovascular events has also been established. Hendrikse et al3 found that such events referable to the stenotic carotid artery were strongly associated with the absence of collateral flow through the anterior communicating artery, although the association was not established for collateral flow through the posterior communicating artery. This hypothesized relationship between Q and risk of stroke depends strongly, of course, on the assumption that reference ICA diameter (RICAD) is related to the Q. That assumption is strongly supported by many studies on the phenomena of flow-mediated dilatation.4 However, direct observation of the relationship between flow and ICA diameter in the setting of ICA stenosis has not been obtained previously. Thus, the purpose of the current study was to demonstrate the presence of a flow-diameter relation in the ICA in a population of patients who underwent magnetic resonance (MR) imaging of the ICA because of suggestion of carotid artery stenosis.

Methods Patients Patients were recruited to participate in this study with the approval of our institutional review board and with informed consent from the patients or their authorized representative. Patients were all imaged as inpatients and were all identified based on the MR angiography (MRA) examination. The primary inclusion criterion was that the patient was scheduled to undergo a neck contrastenhanced MRA examination. Patients were excluded if they were in intensive care or required sedation to undergo the MRA examination. The median age of patients included in the study was 66 years with a range of 32 to 94 years. The indication for undergoing the neck MRA examination was stroke, transient ischemic attack, near syncope or syncope, amaurosis fugax, carotid bruit, severe headache, facial pain, visual changes, neck injury, coronary artery disease, and end-stage renal disease. Imaging studies of 76 arteries from 38 patients were included in the final analysis. Three studies were excluded from the final analysis. The study included 17 stenotic carotid arteries and 59 normal carotid arteries. There were 27 patients without carotid artery stenoses and 11 patients with one stenosis or more who were included in the study. One was excluded from the analysis because of the improper field of view of the phase-contrast (PC) MR imaging, a second because

of a severe flickering artifact in the PC MR imaging, and a third because of carotid artery occlusion.

Image Acquisition Contrast-enhanced MRA and velocity-encoded PC MR imaging were obtained with a 1.5-T system (GE Signa Excite, GE Healthcare, Waukesha, WI). The images were acquired with the neurovascular 8-channel array radiofrequency coil. Both the MRA and the PC MR were acquired in the 8NVANGIO MED configuration. MRA was acquired as part of the standard-of-care imaging for the detection of supra-aortic vascular disease. The beginning of the acquisition of the MRA was set to coincide with the peak of the timing bolus using the fluorotrigger method. In this method, the acquisition is initiated manually based on real-time detection of the arrival of contrast media in the common carotid arteries. For the MRA, 0.1 mmol/kg was injected at a rate of between 2 to 3 mL/s depending on the gauge needle of the existing intravenous line. Images were acquired in the coronal orientation. Image acquisition was performed with elliptical centric k-space ordering with echo time (TE)/repetition time (TR) of 1.5 ms/6.6 ms in an inplane image matrix of 288 frequency encoding samples in the right-left direction by 256 phase-encoding samples in the superior-inferior direction. The flip angle was 30 degrees. The inplane field of view was 26 cm with a slice thickness of 1.2 mm. Images were zero-fill interpolated to an inplane image resolution of 512 3 512. PC MR was acquired immediately after the MRA. PC MR was acquired with cardiac gating based on pulse oxymetry using a probe placed on the index finger. The acquisition location was prescribed based on the maximum intensity projection of the MRA. Placement of the PC MR axial cross section was considered at any point along the ICA distal to the carotid siphon where the ICA tends to follow a relatively straight inferior-superior course and where it has diverged from the jugular vein. Acquisition was based on the cine PC protocol in which cardiac gating is applied in a retrospective manner for image reconstruction. PC MR was acquired for 32 cardiac phases. Images were acquired with an inplane field of view of 20 cm and a slice thickness of 10 mm. Velocity encoding of 100 cm/s was applied in the superior-inferior direction. Postprocessing of the PC MR was based on the phasedifference method. TE/TR was 4.2 ms/15 ms and the flip angle was 20 degrees.

Image Analysis Analysis of MRA included the following procedures: (1) classification of the ICA as normal or stenotic; (2) measurement of the common carotid artery diameter; (3) measurement of the distal normal ICA diameter (RICAD); and (4) measurement of the image contrast. All analysis was performed on a workstation (Vitrea, Vital Images

PARADOXICAL FLOW HYPOTHESIS OF THE CAROTID ARTERY

Inc, Minnetonka, Minn). Procedures 1, 2, and 3 were performed before analysis of PC MR and were, thus, blinded to the flow measurements. Procedure 4 was performed after analysis of PC MR. Detection of ICA stenosis was based on real-time volume rendering with the maximum intensity projection. An ICA was considered to be stenotic if there was observed to be any unambiguous narrowing. This analysis was performed independent of the clinical interpretation of the MRA because those interpretations involved numerous readers who had potentially inconsistent criteria and terminology for reporting stenosis. Measurements of arterial diameters were performed so that the measurement of the same artery was performed in sequence for all patients during a given session or in consecutive sessions. Measurements of diameters of different arteries were made with at least 1 day of separation to essentially eliminate any memory effect, particularly for left-to-right comparisons. Diameter measurements of the common carotid diameter were made proximal to widening at the carotid bifurcation and at the same location for both the right and left common carotid arteries. Diameter measurements in the ICA were made at a location in which the ICA was least tortuous and most oriented with the superior-inferior axis. Unless a single superior-inferior position did not apply well to both the left and right ICA, both measurements were made at the same superior-inferior position. All measurements of normal arterial diameter were performed by the same observer (P. J. Y.). Measurements of the diameter at the point of maximal stenosis for calculation of the degree of stenosis were made using the vessel-probe option on the workstation. That measurement was made by a second observer (A. T.). In 3 arteries, there is a flow gap at the point of maximal stenosis and measurements of the degree of stenosis are meaningless. A numeric value of 90% stenosis was assigned to those particular cases for statistical purposes. This numeric value was based on previously reported findings using similar imaging techniques in which the flow gap is typically representative of the most severe level of stenosis of the ICA.5 Each arterial diameter measurement was made in the following manner. Measurements were all made from the axial view with maximal magnification. The window-level settings were based on the maximal intensity within the lumen at the given cross section. Window was set equal to the maximal luminal intensity. Level was set to half of the maximal luminal intensity. The ruler tool was used for measurement of the diameter and the diameter was drawn across the lumen at an orientation where the diameter is minimal to minimize any effect of obliquity of the artery to the axial cross section. Diameter measurements were made on the reformatted axial cross section at 3 different levels. The mean of those 3 measurements was used for all subsequent analyses. For measurement of the internal carotid

103

diameters, sculpting of the MRA was first performed to exclude all arteries except the artery to be measured to avoid possible confusion identifying the correct artery in the axial view. Measurements of flow (Q) were made from PC MR using software (maximum intensity projection AV, National Institutes of Health, Bethesda, Md) on a consumer personal computer. PC MR images were provided in masked form by the MR system. To obtain a velocity-encoded image in units of cm/s, the phase component of the PC MR was divided by the magnitude component and then multiplied by a scale factor of 20. Segmentation of the luminal cross section was performed in the following manner. First, the luminal cross section was contoured in a magnitude-component image in a systolic cardiac phase. The contour was drawn to overestimate the size of the cross section and was projected onto the magnitude-component image in all of the cardiac phases. Then, the magnitude-component image was thresholded at an intensity of 50 to exclude regions with air where velocity encoding was randomized noise. The final segmentation of the luminal cross section was then formed by taking the intersection of the contoured and the thresholded regions. Because the magnitude-component image was, by definition, coregistered with the phase-component image, segmentation of the magnitude-component image applies directly to the velocity-encoded image. The velocities within the segmented region were then summed, multiplied by the pixel area in square centimeters, and divided by the number of cardiac phases (32) to obtain the flow rate (cm3/s). Image contrast of the MRA was measured for every ICA. Image contrast was defined to be the maximum image intensity in the MRA at the axial location in which the ICA diameter was measured. The measurement was made by adjusting the window-level settings until the location of the maximum image intensity was apparent and then by manually reading out the image intensity at that location using the cursor.

Statistical Analysis The relationship between ICA diameter and flow was analyzed using linear regression in software (SPSS, Chicago, Ill). The ICA diameter was analyzed in its absolute measurement, RICAD, and in terms of its diameter relative to that of the common carotid artery, internal-common carotid diameter ratio (ICCDR). The correlation coefficients were tested to determine if the slope of the relations between RICAD and Q and between ICCDR and Q were significantly different from zero. Analysis was performed for the complete set of arteries and for the subgroups of stenotic and normal arteries. To factor out interindividual differences, linear regression analysis was also performed for right-left differences in RICAD and Q and between ICCDR and Q.

104

Multiple regression analysis was also performed to evaluate possible bias as a result of variability in image quality on the relationships among RICAD, ICCDR, and Q. In the multiple regression analysis, image quality, defined as image contrast of the MRA, was included as a possible predictor variable, in addition to measures of ICA diameter. Variables were added to the multiple regression models in a hierarchical manner with image contrast included first. In all statistical analyses, relationships were considered statistically significant if P values were less than or equal to .05.

Results An example of magnetic resonance imaging of the Q and of the ICA anatomy is shown in Figure 1. Vessel diameters were found to be measurable in a reproducible

YIM, TILARA, AND NOSHER

manner from the MRA. The mean SD among 3 measurements made by one observer was 0.14 mm for the diameter of the ICA and 0.21 mm for the diameter of the common carotid artery (Figure 1). The flow rate, Q, was found to be correlated with the degree of stenosis (r2 5 0.0908, P 5 .004) (Figure 2). Statistically significant differences were found in the flow rates of stenotic versus normal arteries. Mean flow rates differed between stenotic (4.3 6 1.7 mL/s) and normal (5.4 6 1.7 mL/s) arteries (P 5 .02). RICAD was found to be a predictor of Q for stenotic arteries (P 5 .009) and for all arteries (P 5 .025) but not for the group of normal arteries (P 5 .162). The relationship between RICAD and Q is shown in Figure 3. RICAD remained a predictor of Q for stenotic arteries (P 5 .017) in the multiple regression model when possible bias as a result of image quality was removed. However, the

Figure 1. Magnetic resonance angiography (MRA) in the maximum intensity projection view (A) and in the axial cross-sectional view used for measurement of the diameter of the right internal carotid artery (B). The location of the axial cross section is indicated by the solid line in (A) and the right internal carotid artery is indicated by the arrowhead in (B). The phase contrast magnetic resonance (PC MR) imaging used for measurement of the flow in the internal carotid artery consists of a magnitude component (C) and a phase component (D). The segmentation of the phase component of the PC MR used for quantification of flow is shown in (E). Dashed line in (A) indicates location at which PC MR was acquired.

PARADOXICAL FLOW HYPOTHESIS OF THE CAROTID ARTERY 12

A 12.0

10 R2 = 0.0908

Normal arteries

10.0

8

Flow (cc/sec)

Flow (cc/sec)

105

6 4 2

8.0 6.0 4.0 2.0

0 0

20

40

60

80

100

0.0

Percent Stenosis

Discussion This study confirms, in a direct manner, the hypothesis that the distal normal diameter of the ICA, RICAD, is correlated with the mean flow rate in the ICA, Q. The study also confirmed the hemodynamic effect of carotid artery stenosis (ie, a reduction in Q). On the other hand, the ratio index (ICCDR) that most directly correlates with the measurements studied by Rothwell and Warlow1 was not found to be a significant predictor of the flow rate although there was a trend toward a relationship. This can be potentially explained by two factors. First, the ICCDR involves two measurements and, thus, the uncertainty in the measurement is proportionally increased. Second, the relationship between ICCDR and Q may be weaker than that between RICAD and Q and was only evident in the study of Rothwell and Warlow1 because of the large number of patients included in their study.

4.0

5.0

6.0

7.0

8.0

7.0

8.0

Diameter (mm) Stenotic Arteries

B 12.0 10.0 8.0

Flow

relationship between RICAD and Q in all arteries did not quite reach the level of statistical significance (P 5 .051) in the multiple regression model. Right-left differences in RICAD were highly predictive of right-left differences in Q in the subgroup of patients with normal arteries (P , .001) and in the group of all patients (P ,.001). These differences remain statistically significant after adjustment for right-left differences in image contrast for group of all patients (P , .001) and the group with normal arteries (P , .001). However, the right-left differences in RICAD were not predictive of right-left differences in Q in the group with one or more stenotic arteries (P 5 .395). A summary of the regression relationships is given in Table 1. ICCDR was not found to be a statistically significant predictor of Q for all arteries or either subgroup of arteries, although there was a trend toward a relationship between ICCAR and Q for the subgroup of stenotic arteries (P 5 .156). The results are summarized in Table 1.

3.0

6.0 4.0 2.0

R2 = 0.3761

0.0 3.0

4.0

5.0

6.0

Diameter (mm)

C

All arteries 12.0 10.0

Flow (cc/sec)

Figure 2. Stenosis-flow relation in ICA. There is statistically significant negative correlation between flow rate and percent stenosis in ICA (P 5 .004).

R2= 0.0339

8.0 6.0 4.0 R2 = 0.0661

2.0 0.0 3.0

4.0

5.0

6.0

7.0

8.0

Diameter (mm) Figure 3. Diameter-flow relation. Correlation of flow, Q, and reference diameter, RICAD, in normal (A), stenotic (B), and all (C) ICA. Correlation is statistically significant for considering group of stenotic arteries alone or group of all arteries.

Adaptation of the vasculature to obstruction of the carotid artery is a phenomena that may be closely associated

YIM, TILARA, AND NOSHER

106

Table 1. Correlations between diameter and flow

Normal arteries Flow v ICA diameter Flow v ICA/CCA diameters Stenotic arteries Flow v ICA diameter Flow v ICA/CCA diameters All arteries Flow v ICA diameter Flow v ICA/CCA diameters Patients with no stenoses Right-left flow v right-left diameter Patients with $1 stenosis Right-left flow v right-left diameter All patients Right-left flow v right-left diameter

n

r2

P value

59 59

0.0340 0.0044

.162 .618

17 17

0.3767 0.1295

.009* .156

76 76

0.0661 0.0163

.025* .272

27

0.4135

,.001*

11

0.0813

.395

38

0.3013

,.001*

Abbreviations: CCA 5 common carotid artery; ICA 5 internal carotid artery. *Statistically significant correlation.

with variation in flow rates in the stenotic ICA. A low rate of flow occurring in the ICA would be consistent with the presence of adequate compensatory mechanisms, such as collateral blood flow, in this case, through the circle of Willis. Studies have begun to address the question of how the adaptability of the cerebral circulation may be affected by carotid artery stenosis and, conversely, how a reduced adaptability of that circulation may complicate carotid artery stenosis. One measure of the adaptability of the cerebral circulation is cerebrovascular reactivity. Cerebrovascular reactivity is a measure of the ability of the vasculature to respond to vasodilatory stimulea such as carbon-dioxide inhalation6 or intravenous administration of acetazolamide.7 The loss of cerebrovascular reactivity indicates that the flow is limited by vascular disease that may consist of diffuse disease of the microvasculature or focal stenoses of the large arteries rather than by the mechanism in the healthy vasculature of vasoconstriction. Thus, an exhausted cerebrovascular reserve or reactivity in the territory of a carotid artery suggests that the stenosis of the carotid artery is flow limiting or, in other words, that for the given severity of stenosis, the flow is maximal. Thus, findings from studies of cerebrovascular reserve can potentially also be interpreted in terms of carotid artery hemodynamics. In the study of Orosz et al,7 asymmetry in cerebrovascular reserve was found to be significant for patients with symptomatic carotid artery stenosis but not for patients with asymptomatic carotid stenosis. This study was designed to test the hypothesis of Rothwell and Warlow1 that poststenotic narrowing of the ICA

is associated with reduced flow. The main findings of the study were, therefore, positive; the flow rate in stenotic ICA was correlated with the distal normal ICA diameter and a similar relation existed for the group that included both stenotic and normal arteries. The directionality of the relation was that predicted by Rothwell and Warlow1; ICA flow and diameter were positively correlated. However, the measure of ICA diameter used in the study of Rothwell and Warlow,1 the ratio of the distal normal ICA diameter to the common carotid diameter, was not found to be correlated with the flow rate. This is considered less critical; the ratio of the ICA to common carotid artery, per se, probably has limited physiologic relevance. The use of the ICA to common carotid artery ratio in the study of Rothwell and Warlow1 was presumably intended to be an indirect measure of absolute ICA diameter given that absolute measurements of ICA diameter are not available from angiography. In the group of healthy individuals in the group of all participants there was a highly significant correlation between right-left differences in ICA diameter and Q that also supports the hypothesis of the diameter-flow relation in the ICA. The imaging technique for measurement of the ICA diameter, contrast-enhanced MRA, is now widely used in the diagnosis of extracranial cerebrovascular disease.8 A major strength of this imaging technique is the insensitivity to the low-flow artifact that is commonly observed in time-of-flight MRA.9 There are known limitations to MRA for accurate delineation of the lumen in the immediate vicinity of severe stenoses as a result of very high blood velocities, shear rates, and turbulent conditions.10,11 Such conditions do not exist in the segments of the ICA in which diameter measurements were made for this study and, thus, the potential adverse impact of this artifact is unlikely. Another possible factor in the accuracy of the measurements of the ICA diameter is the dynamics of the contrast media. The image contrast is roughly proportional to the concentration of the contrast media. In addition, given the dynamic conditions during the image acquisition, variability in the contrast concentration may also affect the image resolution.12 The quality of the contrast-enhanced MRA is dependent on the concentration of the contrast media in the artery at the time of the image acquisition and the concentration of the contrast media may vary from one individual to another and even in different arteries within a given individual. The concentration of the contrast media is a large determinant of the image contrast of the lumen in the MRA. Thus, to correct for possible bias related to the concentration of the contrast, the regression analysis was performed after correcting for variability in image contrast. The principle outcome, the relationship between flow and diameter, remained valid after applying this correction for the group of stenotic arteries and only narrowly did not reach statistical significance for the group of all arteries.

PARADOXICAL FLOW HYPOTHESIS OF THE CAROTID ARTERY

Measurement of arterial diameters was done manually after objectively resetting the window and level settings of the display. Although computational methods for measurement of arterial diameter from contrast-enhanced MRA have been developed,13-16 the standard of truth remains manual measurement. The hypothesized relationship between a low flow rate in the ICA and a reduced risk of stroke is seemingly counterintuitive as stroke is, by definition, the result of inadequate flow of blood to a region of the brain. This apparent contradiction is semantic only. Patients with reduced flow through a stenotic ICA may have entirely adequate blood flow to the brain through collateral flow pathways. On the other hand, in other patients, the presence of relatively high flow rates through a stenotic ICA may be a consequence of compromised collateral flow pathways as seems to be indicated by the study of Hendrikse et al.17 In fact, hemodynamic principles dictate that given an equivalent stenosis geometry, the transstenotic pressure loss is greater for higher flow rates.18 Thus, the cerebral perfusion pressure ipsilateral to a carotid stenosis is inversely related to the flow in the ICA, assuming that the pressure in the proximal carotid artery remains constant. Thus, higher flow rates in the more severely stenotic ICA may even be associated with hypoperfusion in the ipsilateral hemisphere. The hypothesis of a relationship between low flow rate and reduced risk of stroke, as discussed above, could involve various mechanisms that cannot be clarified based on the results of this study. As mentioned above, the flow rates in the ICA could conceivably be an indication of hemispheric hypoperfusion. However, flow rates in the ICA could also be a factor in the destabilization of the atherosclerotic plaque and subsequent production of thromboemboli, which is believed to be the primary mechanism of stroke resulting from carotid artery disease. Studies have shown, indirectly, that such a relationship may exist. Several findings with respect to the analysis of endarterectomy specimens, for example, strongly support the hypothesis of flow-mediated changes in atherosclerotic plaque. In the study of Tricot et al,19 apoptosis was found to be significantly more prevalent in the aspect of the plaque distal to the location of maximal stenosis as compared with the proximal aspect. In the study of Dirksen et al,20 inflammatory cells were found to be more concentrated in the proximal aspect of carotid plaque whereas smooth-muscle cells were found to have a greater concentration in the distal aspect. In the above studies, the differences between the proximal and distal aspects of the plaque are very likely to be flow related. Support for the deleterious effect of high flow rates on atherosclerotic plaque can also be seen in the study of Beach et al.21 In that study, involving patients undergoing carotid endarterectomy, the highest levels of peak systolic and diastolic blood velocities in the ICA, as measured by Doppler ultrasound, were found to be strongly associated with the presence of intraplaque hemorrhage. This potential

107

relationship between flow and the destabilization of atherosclerotic plaque can be studied in more detail with emerging methodologies for computational modeling of the fluid mechanics at arterial stenoses.22

Conclusions The results of this study confirm that the RICAD is correlated with the flow in the ICA (Q) in the presence of carotid artery stenosis. In all probability, Q is also related to the ICCDR, which has been observed to be a predictor of low stroke risk, although the Q-ICCDR relationship was not conclusively demonstrated in this study, possibly because of greater physiologic variability and measurement error associated with ICCDR. This study certainly justifies further examination of the role of Q as a mechanistic factor in stroke from carotid artery disease. Acknowledgment: We would like to acknowledge the technical support of Mark Ambrose and the magnetic resonance imaging staff at the Robert Wood Johnson University Hospital.

References 1. Rothwell PM, Warlow CP. Low risk of ischemic stroke in patients with reduced internal carotid artery lumen diameter distal to severe symptomatic carotid stenosis: Cerebral protection due to low poststenotic flow? On behalf of the European carotid surgery trialists’ collaborative group. Stroke 2000;31(3):622-630. 2. Henderson RD, Eliasziw M, Fox AJ, et al. Angiographically defined collateral circulation and risk of stroke in patients with severe carotid artery stenosis: North American symptomatic carotid endarterectomy trial (NASCET) group. Stroke 2000;31:128-132. 3. Hendrikse J, Eikelboom BC, van der Grond J. Magnetic resonance angiography of collateral compensation in asymptomatic and symptomatic internal carotid artery stenosis. J Vasc Surg 2002;36(4):799-805. 4. Kelm M. Flow-mediated dilatation in human circulation: Diagnostic and therapeutic aspects. Am J Physiol Heart Circ Physiol 2002;282(1):H1-H5. 5. DeMarco JK, Huston J III, Bernstein MA. Evaluation of classic 2D time-of-flight MR angiography in the depiction of severe carotid stenosis. AJR Am J Roentgenol 2004; 183(3):787-793. 6. Baumgartner RW, Regard M. Role of impaired CO2 reactivity in the diagnosis of cerebral low flow infarcts. J Neurol Neurosurg Psychiatry 1994;57:814-817. 7. Orosz L, Fulesdi B, Hoksbergen A, et al. Assessment of cerebrovascular reserve capacity in asymptomatic and symptomatic hemodynamically significant carotid stenoses and occlusions. Surg Neurol 2002;57(5):333-339. 8. U-King-Im JM, Trivedi RA, Graves MJ, et al. Contrastenhanced MR angiography for carotid disease: Diagnostic and potential clinical impact. Neurology 2004; 62(8):1282-1290. 9. Ahn KJ, You WJ, Lee JH, et al. Re-circulation artefact at the carotid bulb can be differentiated from true stenosis. Br J Radiol 2004;77(919):551-556. 10. Lin W, Haacke EM, Smith AS. Lumen definition in MR angiography. J Magn Reson Imaging 1991;1(3):327-336.

YIM, TILARA, AND NOSHER

108 11. Mitsuzaki K, Yamashita Y, Onomichi M, et al. Delineation of simulated vascular stenosis with Gd-DTPA-enhanced 3D gradient echo MR angiography: An experimental study. J Comput Assist Tomogr 2000;24(1):77-82. 12. Fain SB, Riederer SJ, Bernstein MA, et al. Theoretical limits of spatial resolution in elliptical-centric contrastenhanced 3D-MRA. Magn Reson Med 1999; 42(6):1106-1116. 13. Frangi AF, Niessen WJ, Hoogeveen RM, et al. Modelbased quantitation of 3-D magnetic resonance angiographic images. IEEE Trans Med Imaging 1999; 18:946-956. 14. Manniesing R, Velthuis BK, van Leeuwen MS, et al. Level set based cerebral vasculature segmentation and diameter quantification in CT angiography. Med Image Anal 2006;10(2):200-214. 15. Yim PJ, Cebral JJ, Mullick R, et al. Vessel surface reconstruction with a tubular deformable model. IEEE Trans Med Imaging 2001;20(12):1411-1421. 16. Yim PJ, Vasbinder GB, Ho VB, et al. Isosurfaces as deformable models for magnetic resonance angiography. IEEE Trans Med Imaging 2003;22(7):875-881.

17. Hendrikse J, Rutgers DR, Klijn CJ, et al. Effect of carotid endarterectomy on primary collateral blood flow in patients with severe carotid artery lesions. Stroke 2003; 34(7):1650-1654. 18. Young DF, Tsai FY. Flow characteristics in models of arterial stenoses, I: Steady flow. J Biomech 1973;6(4):395-410. 19. Tricot O, Mallat Z, Heymes C, et al. Relation between endothelial cell apoptosis and blood flow direction in human atherosclerotic plaques. Circulation 2000; 101(21):2450-2453. 20. Dirksen MT, van der Wal AC, van den Berg FM, et al. Distribution of inflammatory cells in atherosclerotic plaques relates to the direction of flow. Circulation 1998; 98:2000-2003. 21. Beach KW, Hatsukami T, Detmer PR, et al. Carotid artery intraplaque hemorrhage and stenotic velocity. Stroke 1993;24:314-319. 22. Yim P, Demarco K, Castro MA, et al. Characterization of shear stress on the wall of the carotid artery using magnetic resonance imaging and computational fluid dynamics. Stud Health Technol Inform 2005;113: 412-442.

Correction Meschia J, Safirstein BE, Biller J. Stroke and the american presidency. J Stroke Cerebrovasc Dis 1997;6:141-143. In the above article, President Franklin Delano Roosevelt’s date of death is incorrectly stated as March 29, 1945. President Franklin Delano Roosevelt actually died on April 12, 1945.