Receiv*DeceFunde

RadiologPresen

of NeuroFrom

C.E.L.F.pital and

Addrement ofCambrid

© Amer

AJNR Am J Neuroradiol 19:1747–1752, October 1998

Evaluation of Carotid Endarterectomy withSequential MR Perfusion Imaging:

A Preliminary Report

Jonathan H. Gillard, Charles R. Hardingham, Peter J. Kirkpatrick, Nagui M. Antoun,Charles E. L. Freer*, and Paul D. Griffiths

BACKGROUND AND PURPOSE: Current indications for carotid endarterectomy are deter-mined by balancing the relative risks of surgery with the benefits of reduced risk of subsequentstroke. Our purpose was to use MR perfusion imaging to assess patients being considered forcarotid endarterectomy and to monitor sequential changes in MR perfusion characteristicsafter surgery. In particular, we wished to determine whether this technique could be used todetect changes that might be related to post–carotid endarterectomy hyperemia.

METHODS: We used a single-section gradient-recalled echo sequence to investigate 14patients being examined before possible surgery for carotid artery disease. In the 12 patients inwhom carotid endarterectomy was performed, sequential studies were performed 3 to 5 daysafter surgery and at 3 months. Analysis of bolus-arrival-time (BAT) images was performed.

RESULTS: Significant delays in preoperative BAT images of 0.89 seconds (range, 0.05 to 3.22seconds) were apparent between hemispheres. Excluding the two patients with contralateralinternal carotid artery (ICA) occlusion, early arrival, possibly indicating postoperative hyper-emia, was seen in five patients immediately after carotid endarterectomy but resolved within 3to 5 months after surgery.

CONCLUSION: MR perfusion imaging shows differences in BAT between hemispheres inpatients with ICA stenosis. Changes in perfusion characteristics after carotid endarterectomyare complex, and early BAT on the operative side can occur soon after endarterectomy in overhalf those patients without an occluded contralateral vessel. The significance of these find-ings with regard to patient outcome and risk of postoperative hyperemia requires furtherinvestigation.

Carotid endarterectomy has been shown to be effec-tive in reducing the prevalence of stroke and mortal-ity in persons with severe (.70%) carotid stenosis(1). Its efficacy may be due to the normalization ofcerebral blood flow (CBF), to improvement in cere-brovascular reserve (2–4), and/or to reduction in theformation of emboli from carotid plaque. There is,

ed January 22, 1998; accepted after revision June 15.asedd by a Pump Primer grant from the Royal College ofists.ted in part at the annual meeting of the American Societyradiology, Toronto, Ontario, May 1997.the Departments of Radiology (J.H.G., C.R.H., N.M.A.,, P.D.G.) and Neurosurgery (P.J.K.), Addenbrooke’s Hos-University of Cambridge, United Kingdom.

ss reprint requests to Jonathan H. Gillard, MD, Depart-Radiology, Addenbrooke’s Hospital and University of

ge, Cambridge CB2 2QQ, United Kingdom.

ican Society of Neuroradiology

174

however, little evidence as to whether changes inmicrovascular perfusion lead to clinical improvement.

Newly developed dynamic MR imaging techniquescan assess cerebral perfusion (5) using conventionalMR systems and may delineate ischemic tissue. Theobservation of widespread abnormalities in bolus ar-rival time (BAT) and relative cerebral blood volume(rCBV) in territories with normal T2-weighted MRcharacteristics indicates that this technique may de-fine at-risk territories in patients with carotid arterydisease (6).

Current indications for carotid endarterectomy aredetermined by the need to balance the relative risksof surgery with the benefits of a reduced risk ofsubsequent stroke. In recent large-scale trials, theperioperative stroke rate was a key determinate indefining the degree of carotid stenosis that wouldbenefit from carotid endarterectomy as comparedwith medical management. Hemorrhagic stroke,which may be related to postoperative hyperemia, is asignificant problem and occurs in approximately 0.6%

7

1748 GILLARD AJNR: 19, October 1998

of carotid endarterectomies (7). The overall periop-erative mortality and morbidity of carotid endarter-ectomy is about 5.8% in skilled hands (1). Anymethod that can be used to predict which patients willbe at increased risk of post–carotid endarterectomyhyperperfusion and possible stroke may ultimatelyreduce the postoperative stroke rate.

The aim of this study was to use MR perfusionimaging to assess patients being evaluated for possi-ble carotid endarterectomy and to monitor changes inMR perfusion characteristics soon after surgery andagain at 3 months. In particular, we wished to deter-mine whether this method could detect changes thatmight be related to postoperative hyperemia.

MethodsFourteen patients, referred for the investigation of likely

carotid stenosis, were studied prospectively. All MR imagingand perfusion studies were carried out at 1.5 T. Five series wereobtained as follows. Sagittal T1 localizer images were acquiredwith parameters of 200/9/1 (TR/TE/excitations), a 25-cm fieldof view (FOV), and a 256 3 128 matrix. Single-section fastspin-echo T2-weighted (3000/17,85/1) images were obtainedwith an echo train length of 8, a 22-cm FOV, 8-mm-thicksections, no gap, and a 256 3 256 matrix. MR gradient-recalledecho perfusion images were acquired with parameters of 34/22/0.75, a flip angle of 20°, a 22-cm FOV, a 256 3 128 matrix,and a section thickness of 10 mm. Forty perfusion images wereobtained at 3.25-second intervals. After 10 baseline imageswere obtained, gadopentetate dimeglumine, in a bolus of 0.2mmol/kg of body weight, was administered by one of twoinvestigators as a rapid hand injection through an antecubitalvein followed by a bolus of normal saline. The duration of thebolus injection was 6 to 8 seconds.

Analysis was performed using the WinFun software package(developed by Mike Hayball, Department of Radiology, Uni-versity of Cambridge) on a conventional personal computer.Images were constructed, pixel by pixel, of BAT, and ellipticalregions of interest (ROIs) were chosen in comparable con-tralateral middle cerebral artery (MCA) territories so thatrelative delays between hemispheres could be calculated. TheROIs were drawn as large as possible while maintaining sym-metry, including the gray and white matter, and avoiding anyarea of infarction. The software simultaneously displayed bothan original image from the perfusion sequence and a calculatedimage. The ROIs were drawn on both images, making it easierto exclude areas with abnormalities or artifacts. Dual-displayimages were also obtained of BAT and enhancement. Thetime-signal data were initially converted into a signal-changecurve using the following formula: Signal change 5 2K.log(S/S0), where K is an arbitrary constant, S is the signal, and S0is the average signal before injection of contrast medium. Thisgenerated a curve with a positive peak, to which a gammavariate curve was fitted (8). In our examples, K 5 1000. TheBAT was calculated as the peak of the gamma variate function,fitted to the curve using the Levenberg-Marquardt method (9).This was performed on a pixel-by-pixel basis, and the mean andstandard deviation within each ROI were used to provide 95%confidence limits. Relative interhemispheric delays were calcu-lated using the side of eventual, or contemplated, carotid end-arterectomy as the reference. A positive delay was equivalentto a delay in BAT on the side of carotid endarterectomyrelative to the contralateral hemisphere.

Studies were performed 1 to 13 days before endarterectomy,3 to 5 days afterward, and 3 months later in the same anatomicplane. All patients had confirmatory preoperative conventionalcatheter X-ray angiography. MR perfusion studies were ana-lyzed by investigators who were blinded to the side of carotid

endarterectomy. Statistical significance was defined as P 5 .05using Student’s t-test. Approval was obtained from the LocalResearch Ethics Committee of Addenbrooke’s Hospital.

Results

Of the 14 patients studied, four were women and 10were men; the mean age was 71 years (range, 51 to 81years). Two patients did not undergo carotid endar-terectomy: one (case 13) because of complete internalcarotid artery (ICA) occlusion shown by MR angiog-raphy and confirmed by X-ray angiography, and theother (case 14), who had nonlateralizing clinicalsymptoms, because of delayed BAT in the hemi-sphere contralateral to the diseased carotid, causedby MCA branch occlusions. One patient withdrewfrom the immediate postoperative study because ofclaustrophobia. Thus, 11 patients underwent pre- andpostendarterectomy studies, nine of whom had latestudies (repeated at 3 months or beyond). BAT im-ages were calculated and are presented as interhemi-spheric delays relative to the side of carotid endarter-ectomy. The average number of pixels in the ROIswas 1249 (range, 270 to 2023). Data sets from twopatients (cases 9 and 11) were of technically poorquality owing to movement artifacts.

Preoperative StudiesAll 12 patients with adequate data sets had relative

preoperative delays in BAT between hemispheres of0.89 seconds (range, 0.05 to 3.22 seconds), P 5 .019(see Table). Among the seven patients with unilateralsevere (.70%) carotid stenosis, the delay was on theside of maximal stenosis in six patients. The remain-ing patient (case 4) had trivial early arrival of 0.05 60.08 seconds (mean 6 SD). One patient (case 13) hadunilateral carotid occlusion with a normal contralat-eral vessel. In two patients with bilateral disease, thedelay in BAT was on the side of carotid occlusion. Intwo patients who had unilateral severe disease withcontralateral occlusions there was relatively earlyBAT ipsilateral to the side of surgery. Representativeimages are shown in Figure 1. Figure 2 shows BATimages in another patient.

Postoperative StudiesThree of the six patients with less than severe

(,70%) contralateral ICA disease had early BAT onthe side of carotid endarterectomy 3 to 5 days aftersurgery. At 3 months, only the two patients withcontralateral ICA stenosis had early BAT ipsilateralto the side of surgery. Of the two patients with uni-lateral ICA disease, one (case 2) had early arrivalimmediately after carotid endarterectomy, and theother (case 3) had some improvement in delay timespostoperatively. Among the patients with bilateral,nonocclusive ICA disease without evidence of earlyBAT at 3 months, three had sustained improvement inBAT (cases 1, 3, and 6), and three had worsening of theipsilateral BAT (cases 2, 4, and 5) at 3 months. Two of

AJNR: 19, October 1998 CAROTID ENDARTERECTOMY 1749

Sequential findings in bolus arrival time before and after carotid endarterectomy (CE)

Patient

Conventional X-rayAngiographic Stenosis (%)

Relative Interhemispheric Bolus Arrival Time(Ipsilateral to Side of CE)

Left Right Before CEMean 6 SD 3–5Days after CE

Mean 6 SD 3Months after CE

Unilateral disease1 99 15 1.24 6 0.10 20.29 6 0.23 0.00 6 0.072 80 0 0.06 6 0.20 20.26 6 0.25 1.47 6 1.103 0 90 1.09 6 0.39 0.34 6 0.25 0.42 6 0.20†4 55 95 20.05 6 0.08 0.63 6 0.08 0.26 6 0.195 85 ,70 0.05 6 0.10 20.06 6 0.1 0.74 6 0.176 50 99 3.22 6 0.09 0.97 6 0.14 1.64 6 0.12

Bilateral disease7 90 100 20.71 6 0.21 21.03 6 0.12 20.92 6 0.138 100 99 23.00 6 0.30 1.32 6 0.44 20.86 6 0.09‡9 80* 70* Patient movement artifact

10 80 90 0.06 6 1.18 21.00 6 0.89 . . .

11 99 75 Patient movement artifactNo postsurgical studies

12 ,40%§ 0 0.55 6 1.09 . . . . . .

No surgery13 100 0 20.13 6 0.18 . . . . . .

14 0 70 0.47 6 0.23 . . . . . .

* Defined by sonography and MR angiography.† Study performed 5 months after CE.‡ Study performed 4 months after CE.§ Symptomatic ulcerating plaque required CE.

these patients (cases 2 and 5) had signs of early arrivalimmediately after carotid endarterectomy.

There was no postoperative mortality or significantmorbidity. A review of the postoperative T2-weightedimages revealed small new areas of hyperintensityipsilateral to the side of carotid endarterectomy infour of the 11 patients (cases 1, 2, 5, and 10). None ofthese lesions were clinically apparent. No T2-weighted changes were found in the hemisphere con-tralateral to the carotid endarterectomy.

DiscussionOur study indicates that cerebral MR perfusion

imaging may be a promising and sensitive marker ofcarotid disease. In all patients with symptomatic uni-lateral carotid disease, maximal delays were ipsilat-eral to the side of endarterectomy (the preoperativeresults in patient 4 most likely being within statisticalerror). The mean asymmetry in BAT between hemi-spheres of 0.89 seconds compares well with the resultsobtained by Nighoghossian et al (10), who found anaverage difference in mean transit time (MTT) of1.15 seconds in 12 patients with unilateral carotidstenosis. In our study, the two patients with early BATpreoperatively had contralateral ICA occlusions. Ex-cluding those patients with preoperative ICA occlu-sions, five (71%) of seven patients were shown tohave early BAT ipsilateral to carotid endarterectomysoon after surgery.

MR perfusion imaging has been used to assessrelative CBF and rCBV in patients with carotid arterystenosis (10, 11). Previous investigators have showneither increased rCBV ipsilateral to carotid stenosis

(11) or no significant asymmetry between hemi-spheres (10). MTT has, however, been shown to besignificantly delayed ipsilateral to the side of carotidstenosis (10). The relationship between the degree ofICA stenosis and CBF is complex, with some studiesshowing some relationship (12) and others showingno relationship (2). Because the majority of patientshad bilateral disease in our study, it is not possible toevaluate the data from this preliminary series in termsof whether a correlation exists between the degree ofICA stenosis and BAT abnormalities.

Patients are currently examined macrovascularlywith duplex sonography or MR angiography, and fre-quently proceed to X-ray angiography before surgery.Other techniques for assessing cerebral perfusion,such as positron emission tomography (PET), xenonCT, or dynamic contrast-enhanced CT (13), are eithernot readily available or require the use of ionizingradiation. Although recent developments in the useecho-planar perfusion techniques can achieve muchimproved temporal resolution, our method is able toshow differences between hemispheres, in keepingwith previous work by Nighoghossian et al (10) andReith et al (14).

Early BAT after carotid endarterectomy may be anindication of hyperperfusion, a phenomenon that hasbeen well described (15–17) and is thought to beischemic in origin, owing to its relationship with ca-rotid cross-clamping time (2) and associated intraop-erative electroencephalographic changes (18, 19).Patients with post–carotid endarterectomy hypoper-fusion syndrome are at risk of intracerebral hemor-rhage (20). Naylor and Ruckley (16) described threevariables in the understanding of this syndrome: rec-

1750 GILLARD AJNR: 19, October 1998

FIG 1. Dual display of bolus arrivaltime (BAT) and enhancement in an81-year-old man (case 1) with 99%stenosis of the left ICA and minimalcontralateral disease.

A, Before carotid endarterectomy,clear asymmetry is noted betweenhemispheres with both delayed ar-rival and reduced enhancement inthe left MCA territory.

B, Corresponding display of signalintensity changes with time showsclear BAT delay on the side of even-tual carotid endarterectomy (CE ) ascompared with the contralateral re-gion of interest (Ref ).

C, Postoperative dual displayshows improvement in perfusioncharacteristics.

FIG 2. Case 6: 74-year-old womanwith 99% stenosis in right ICA and50% stenosis in left.

A, Bolus arrival time (BAT) imageshows delays in the right hemispherepreoperatively.

B, Corresponding graph of preop-erative changes in signal intensitywith time from the MCA territories.

C, The 3- to 5-day postoperativestudy shows an improvement in BAT,particularly in the posterior parietalregion.

D, On the 3-month postoperativestudy, some deterioration in thepostoperative changes is noted, al-though improvement is still signifi-cant relative to preoperative status.

AJNR: 19, October 1998 CAROTID ENDARTERECTOMY 1751

ognition of vulnerable patients; definition of the tem-poral relationship among cerebral hemodynamics, ce-rebral autoregulation, postoperative hypertension,and onset of symptoms; and investigation of the mi-crocirculatory changes that result in cerebral edema,infarct extension, and, ultimately, hemorrhage. Al-though none of our patients had evidence of postop-erative hemorrhage, over half may have been at risk,either immediately after surgery or during the subse-quent 3 months.

Data concerning changes in CBF after carotid end-arterectomy are variable, both in the short (19–23)and long (2) term, with studies reporting increased (4,12, 19, 22, 24, 25), decreased (25), and no (19, 22,25–27) changes in CBF after surgery. Hartl et al (27)demonstrated that while carotid endarterectomy hadno significant effect on CO2 reactivity across all pa-tient groups, improvements were seen in those pa-tients with clear asymmetry in CO2 reactivity beforesurgery. Bishop et al (2) showed that while CBF, asmeasured with xenon-133, returned to preoperativelevels by 6 months after carotid endarterectomy, CO2reactivity continued to improve. This improvement issignificant, as CO2 reactivity is regarded as a satisfac-tory method of assessing hemodynamic reserve (28,29) and is a predictive factor for stroke (30). Whileunilateral changes in CO2 reactivity are perhaps morecommon (31, 32), bilateral improvement in CO2reactivity has also been demonstrated 3 months aftercarotid endarterectomy (3) and 6 months aftercarotid angioplasty (33). In addition, there is evidenceof ischemia prior to carotid endarterectomy insome patients when assessed with proton MR spec-troscopy (34).

Araki et al (21) reported Doppler flow velocitychanges after carotid endarterectomy in patients withmore than 75% stenoses, but no changes in patientswith less than 75% stenoses. Substantial reductions inexternal carotid artery blood flow have also beenshown after carotid endarterectomy (35). Using PET,Powers et al (36) failed to establish that abnormalhemodynamics associated with ICA stenosis signifi-cantly increased the risks of subsequent stroke. Morerecent work has clearly shown that subgroups of pa-tients with carotid artery stenosis, in whom autoreg-ulation is significantly impaired, do exist (37).

There is increasingly compelling evidence that themost cost-effective strategy for the investigation ofsymptomatic carotid stenosis is the routine use of MRangiography and duplex sonography, supplementedby X-ray angiography for cases with disparate results(38). Recent guidelines established by the AmericanHeart Association for the investigation of transientischemic attacks and minor stroke also advocate con-sideration of the use of CT angiography (39). Theemerging understanding of the importance of identi-fying subgroups of patients who are at increased riskafter carotid endarterectomy, or patients who mightpotentially enjoy a greater benefit from surgery (3),produces challenges for the development of optimalimaging regimens. MR perfusion imaging may pro-vide just such a role. While the benefits of carotid

endarterectomy are increasingly well defined via as-sessment of degree of extracranial ICA stenosis (1), itis of some concern that up to 18% of carotid endar-terectomies performed in one region were clearlyinappropriate using the NASCET criteria, and a fur-ther 49% were of uncertain necessity (40). Barnett etal (41), however, continue to suggest caution regard-ing the robustness of current studies in their ability todefine specific groups who would benefit from carotidendarterectomy.

ConclusionThis study shows that the susceptibility weighted

MR perfusion sequence is sensitive to differences inBAT between hemispheres in patients with ICA ste-nosis. In addition, it shows that changes in MR per-fusion characteristics after carotid endarterectomyare complex, but that early BAT ipsilateral to the sideof surgery can occur postoperatively. The significanceof these findings with regard to patient outcome andrisk of post–carotid endarterectomy hyperemia re-quires further investigation.

AcknowledgmentsThe help of Clare Sims and Glynn Johnson of the MRIS unit

of Addenbrooke’s Hospital is gratefully acknowledged.

References1. North American Asymptomatic Carotid Endarterectomy Trial Col-

laborators. Beneficial effect of carotid endarterectomy in symptom-atic patients with high-grade carotid stenosis. N Engl J Med 1991;325:445–453

2. Bishop CCR, Butler L, Hunt T, Burnand KG, Browse NL. Effect ofcarotid endarterectomy on cerebral blood flow and its response tohypercapnia. Br J Surg 1987;74:994–996

3. Visser GH, van Huffelen AC, Wieneke GH, Eikelboom BC. Bilat-eral increase in CO2 reactivity after unilateral carotid endarterec-tomy. Stroke 1997;28:899–905

4. Russell D, Dybevold S, Kjartansson O, Nyberg-Hansen R, Root-welt K, Wiberg J. Cerebral vasoreactivity and blood flow before and3 months after carotid endarterectomy. Stroke 1990;21:1029–1032

5. Edelman RR, Mattle HP, Atkinson DJ, et al. Cerebral blood flow:assessment with dynamic contrast-enhanced T2*-weighted MR im-aging at 1.5T. Radiology 1990;176:211–220

6. Soher BJ, Gillard JH, Bryan RN, Oppenheimer SM, Barker PB.Magnetic resonance perfusion imaging in acute middle cerebralartery stroke: comparison of blood volume and bolus peak arrivaltime. J Stroke Cereb Dis 1998;7:17–23

7. Piepgras DG, Morgan MK, Sundt TM, Yanagihara T, MussmanLM. Intracranial hemorrhage after carotid endarterectomy. J Neu-rosurg 1988;68:532–536

8. Rosen BR, Belliveau JW, Chien D. Perfusion imaging by nuclearmagnetic resonance. Magn Reson Q 1989;5:263–281

9. Press WH, Teukolsky SA, Vettering WT, Flannery BP. NumericalRecipes in C. 2nd ed. Cambridge: Cambridge University Press; 1992

10. Nighoghossian N, Berthezene Y, Philippon B, Adeleine P, FromentJC, Trouillas P. Hemodynamic parameter assessment with dynamicsusceptibility contrast magnetic resonance imaging in unilateralsymptomatic internal carotid artery occlusion. Stroke 1996;27:474–479

11. Guckel FJ, Brix G, Schmiedek P, et al. Cerebrovascular reservecapacity in patients with occlusive cerebrovascular disease: assess-ment with dynamic susceptibility contrast-enhanced MR imagingand the acetazolamide stimulation test. Radiology 1996;201:405–412

12. Vanninen R, Koivisto K, Tulla H, Manninen H, Partanen K. He-modynamic effects of carotid endarterectomy by magnetic reso-nance flow quantification. Stroke 1995;26:84–89

1752 GILLARD AJNR: 19, October 1998

13. Muizelaar JP, Fatouros PP, Schroder ML. A new method for quan-titative regional cerebral blood volume measurements using com-puted tomography. Stroke 1997;28:1998–2005

14. Reith W, Heiland S, Erb G, Benner T, Forsting M, Sartor K.Dynamic contrast-enhanced T2*-weighted MRI in patients withcerebrovascular disease. Neuroradiology 1997;39:250–257

15. Harrison PB, Wong MJ, Belzberg A, Holden J. Hyperperfusionsyndrome after carotid endarterectomy: CT changes. Neuroradiol-ogy 1991;33:106–110

16. Naylor AR, Ruckley CV. The post-carotid endarterectomy hyper-perfusion syndrome. Eur J Vasc Endovasc Surg 1995;9:365–367

17. Spencer MP. Transcranial Doppler monitoring and causes ofstroke from carotid endarterectomy. Stroke 1997;28:685–691

18. Sharbrough FW, Messick JM, Sundt TM. Correlation of continuouselectroencephalograms with cerebral blood flow measurementsduring carotid endarterectomy. In: Langfitt TW, McHenry LC,Reivich M, Wollman H, eds. Cerebral Circulation and Metabolism.New York: Springer; 1975:292–294

19. Algotsson L, Messeter K, Rehncrona S, Skeidsvoll H, Ryding E.Cerebral hemodynamic changes and electroencephalography dur-ing carotid endarterectomy. J Clin Anesth 1990;2:143–151

20. Sundt TM, Sharbrough FW, Poepgras DG, Kearns TP, Messick JM,O’Fallon WM. Correlation of cerebral blood flow and electroen-cephalographic changes during carotid endarterectomy with re-sults of surgery and hemodynamics of cerebral ischemia. Mayo ClinProc 1981;56:533–543

21. Araki CT, Babikian VL, Cantelmo NL, Johnson WC. Cerebrovas-cular hemodynamic changes associated with carotid endarterec-tomy. J Vasc Surg 1991;13:854–860

22. Lord RSA, Reid CVA, Ransey SC, Yeates MG. Unilateral carotidstenosis and impaired cerebral hemispheric vascular reserve. AnnVasc Surg 1992;6:438–442

23. Algotsson L, Ryding E, Rehncrona S, Messeter K. Cerebral bloodflow during carotid endarterectomy determined by three dimen-sional SPECT measurement: relation to preoperative risk assess-ment. Eur J Vasc Surg 1993;7:46–53

24. Herrschaft H, Duus P, Gleim F, Ungeheuer E. Preoperative andpostoperative cerebral blood flow in patients with carotid stenosis.In: Langfitt TW, McHenry LC, Reivich M, Wollman H, eds.Cerebral Circulation and Metabolism. New York: Springer; 1975:292–294

25. Sacca A, Pedrini L, Vitacchiano G, et al. Cerebral SPECT with99mTc-HMPAO in extracranial carotid pathology: evaluation ofchanges in the ischemic area after carotid endarterectomy. IntAngiol 1992;11:117–121

26. O’Brien MD, Veall N, Luck RJ, Irvine WT. Cerebral-cortex per-fusion-rates in extracranial cerebrovascular disease and the effectsof operation. Lancet 1967;2:392–392

27. Hartl WH, Janssen I, Furst H. Effect of carotid endarterectomy on

patterns of cerebrovascular reactivity in patients with unilateralcarotid artery stenosis. Stroke 1994;25:1952–1957

28. Herold S, Brown MM, Frackowiak RSJ, Mansfield AO, ThomasDJ, Marshall J. Assessment of cerebral haemodynamic reserve:correlation between PET parameters and CO2 reactivity measuredby the intravenous 133Xenon injection technique. J Neurol Neuro-surg Psychiatry 1988;51:1045–1050

29. Kuroda S, Kamiyama H, Abe H, Houkin K, Isobe M, Mitsumori K.Acetazolamide test in detecting reduced cerebral perfusion reserveand predicting long-term prognosis in patients with internal ca-rotid artery occlusion. Neurosurgery 1993;32:912–919

30. Kleiser B, Widder B. Course of carotid artery occlusions withimpaired cerebrovascular reactivity. Stroke 1992;23:171–174

31. Cikrit DF, Burt RW, Dalsing MC, et al. Acetazolamide enhancedsingle photon emission computed tomography (SPECT) evaluationof cerebral perfusion before and after carotid endarterectomy. JVasc Surg 1992;15:747–754

32. Takagi Y, Hata T, Ishitobi K, Kitagawa Y. Cerebral blood flow andCO2 reactivity before and after carotid endarterectomy. Acta NeurolScand 1979;60:506–507

33. Markus HS, Clifton A, Buckenham T, Taylor R, Brown MM.Improvement in cerebral hemodynamics after carotid angioplasty.Stroke 1996;27:612–616

34. van der Grond J, Balm R, Klijn CJM, Kappelle LJ, Eikelboom BC,Mali WPTM. Cerebral metabolism of patients with stenosis of theinternal carotid artery before and after endarterectomy. J CerebBlood Flow Metab 1996;16:320–326

35. Gordon IL, Stemmer EA, Wilson SE. Redistribution of blood flowafter carotid endarterectomy. J Vasc Surg 1995;22:349–360

36. Powers WJ, Tempel LW, Grubb RL. Influence of cerebral hemo-dynamics on stroke risk: one-year follow-up of 30 medically treatedpatients. Ann Neurol 1989;25:325–330

37. White RP, Markus HS. Impaired dynamic cerebral autoregulationin carotid stenosis. Stroke 1997;28:1340–1344

38. Kent KC, Kuntz KM, Patel MR, et al. Perioperative imagingstrategies for carotid endarterectomy: an analysis of morbidity andcost-effectiveness in symptomatic patients. JAMA 1995;274:888–893

39. Culebras A, Kase CS, Masdeu JC, et al. Practice guidelines for theuse of imaging in transient ischemic attacks and acute stroke: areport of the Stroke Council, American Heart Association. Stroke1997;28:1480–1497

40. Wong JH, Findlay JM, Suarez-Almazor ME. Regional performanceof carotid endarterectomy. Appropriateness, outcomes, and riskfactors for complications. Stroke 1997;28:891–898

41. Barnett HJM, Eliasziw M, Meldrum HE, Taylor DW. Do the factsand figures warrant a 10-fold increase in the performance ofcarotid endarterectomy on asymptomatic patients? Neurology 1996;46:603–608