(Stroke. 1996;27:1672-1678.)
© 1996 American Heart Association, Inc.
Articles |
Morphological and Hemodynamic Assessments of Carotid Stenosis Using Quantitative Digital Subtraction Angiography
the Department of Neurology, Austin and Repatriation Medical Centre, Australia (C.F.B.); the Department of Neurology (A.C.F.C.) and Division of Radiological Sciences (D.J.H., N.I., C.R.H.), United Medical and Dental Schools Guy's Hospital, London; and the Department of Surgery (A.M.S.), Royal Free Hospital, London, UK.
Correspondence to Christopher F. Bladin, MD, FRACP, Department of Neurology, Austin and Repatriation Medical Centre, Burgundy St Heidelberg, 3084, Melbourne, Australia. E-mail 100355.1702@compuserve.com.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Background and Purpose Digital angiography is the best established tool for assessing atheromatous disease of extracranial blood vessels. Advances in computer technology have now made it possible and practicable to extract quantitative information (length, width, cross-sectional area, and flow velocity) from good-quality clinical angiograms, allowing calculation of volume flow and pressure gradient. The technique of quantitative angiography (QA) is used for assessing coronary artery disease, but to date there has been no clinical application in patients with cerebrovascular disease.
Summary of Report We have developed a computer program for off-line analysis of routine digital subtraction angiographic images. From biplanar images, the program reconstructs the angiogram in three dimensions and performs quantitative analysis of each vessel. From this data, the pressure drop from the aortic arch to the circle of Willis is then calculated. We assessed the clinical applicability of QA in five patients investigated for transient ischemic attack. The carotid artery ipsilateral to the symptomatic hemisphere was occluded in one patient and had minor plaque in another. In the remaining three patients, ipsilateral internal carotid artery stenosis was measured by QA as producing area reductions of 55%, 72%, and 88% (equivalent to diameter reductions of 33%, 48%, and 65%, respectively). In these patients, the quantitative stenosis pressure gradients were calculated as 1.2, 3.0, and 3.5 mm Hg, respectively. Further calculation showed that each stenosis contributed to 18%, 24%, and 60%, respectively, of the total carotid pressure gradient from the aortic arch to the circle of Willis. These carotid arteries carried 47%, 42%, and 26%, respectively, of the total cerebral flow. The results of quantitative analysis were validated by comparing, within each patient, the differences in pressure gradients between right and left carotid systems or between right and left vertebral arteries (overall mean difference in pressure gradient, 0.6±0.5 mm Hg; P=NS). Finally, comparison was made of pressure gradients across the circle of Willis between the carotid and vertebrobasilar circulations (mean difference in pressure gradient, 4.1±5.3 mm Hg; P=NS).
Conclusions Quantitative angiography allows determination of the hemodynamic parameters of a vessel or stenosis. It has significant potential, both as a research tool and in routine clinical practice, for the investigation of cerebrovascular disease.
Key Words: angiography • carotid stenosis • hemodynamics
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The development of angiography in the 1950s gave clinicians the ability to display atheromatous disease of blood vessels. The considerable improvements in angiographic techniques and advances in computer technology have resulted in angiography becoming safer and more reliable.
The development of digital angiography meant that quantitative information, such as vessel length, width, cross-sectional area, and flow velocity, could potentially be extracted from the images.1 2 3 This could allow accurate measurement of the dimensions and pressure gradient of a stenosis, as well as the calculation of blood flow under various conditions. To date, the development and clinical application of these methods have been confined mainly to coronary angiography1 4 for assessing the hemodynamic severity of coronary stenosis and the effects of balloon angioplasty. However, the complex calibration and processing needed to achieve good accuracy and reproducibility have thus far kept QA largely as a research tool, with limited clinical applicability.4
No studies have used QA for the assessment of craniocervical vascular disease. Surgery for severe carotid stenosis is well established, but there is continuing controversy over exactly how carotid stenosis should be measured.5 6 7 8 9 10 It is also unclear how to determine whether a stenosis is truly hemodynamic (ie, flow/pressure limiting) or indeed its clinical significance.11 12 As a clinical tool, QA could improve the reliability and accuracy of measuring vessel cross-sectional area and percent stenosis. Pressure gradient (pressure drop) along a vessel or stenosis can also be calculated. Determining the pressure drop along each vessel pathway up to the circle of Willis provides uniquely detailed information about the interaction of proximal stenoses and collateral pathways in individual patients. QA also has potential as a research tool for measuring cerebral blood flow and the cerebral perfusion pressure during different therapeutic interventions.
We have developed a computerized system of QA using routine angiographic procedures that has been fully validated in vitro.13 14 15 16 17 18 In this article, we describe the use of QA in the clinical setting and present the results of analysis of a series of patients investigated for extracranial vascular disease.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Five patients with transient ischemic attacks were referred for angiographic assessment of carotid stenosis. All patients were screened with carotid duplex ultrasound and MRI before angiography. For this study, percent stenosis was calculated from vessel cross-sectional areas using the ICA distal to the diseased section as the reference segment, according to NASCET criteria. We also calculated the equivalent percent diameter reduction.
The technical aspects of our QA methodology have been described in detail elsewhere.17 In brief, patients undergo routine digital subtraction angiography (Siemens, Digitron II) through a femoral puncture by use of the Seldinger technique. The catheter is then positioned in the aortic arch. Biplanar and approximately orthogonal views are taken of the aortic arch, neck, and head, using 60 mL of a noniodinated contrast agent (Iopamerol) for each pair of views (a total of 180 mL of contrast).
Images are obtained at frame rates of 2 per second, apart from one neck and one aortic arch view that are recorded at 25 frames per second (digital cineangiography) for the later calculation of flow velocity. For each biplanar view, the positional coordinates of the x-ray gantry and table are recorded. At the end of the study, for geometric calibration, an image is taken of a Perspex cube with steel markers at known locations for each of the x-ray gantry and table positions, as described previously.18 This geometric calibration removes the need for imaging a washer or catheter of known dimensions. All data are then transferred via ethernet to a SUN computer workstation for 3D reconstruction of the vascular network and quantitative analysis. This is done using SARA, a quantitative angiography computer program developed in Radiological Sciences and Neurology at Guy's Hospital.17
The biplanar views are displayed together on the screen, and the 3x4 element matrix, which relates the 3D Cartesian coordinates of a point in space to the 2D coordinates of the x-ray projection, is computed from the Perspex cube images.18 19
Each vessel is then located in turn, and the approximate vessel centerline is manually identified. The edge detection function in SARA then automatically delineates vessel edges on the basis of the change in contrast density across a vessel (the transverse density profile20 ). Once completed, SARA gives data on the length, widths, and cross-sectional areas of the vessel segment being analyzed. A sliding cursor can be moved along the vessel to a particular point of interest (eg, a stenosis), and the dimensions can be noted. Another component of SARA analyzes the movement of contrast in the cineangiography views and generates a parametric image13 18 21 from which flow velocity is calculated.
All QA measurements of length, area, and flow velocity have been extensively validated in phantom models, pathology specimens, and computer-simulated angiograms containing both circular and irregular cross-section stenoses.13 14 15 16 17 22 23 24 Cross-sectional area is estimated directly from the density of the image; simple width measurements are inadequate when the lumen is highly irregular. The range of error established for each measurement is <2% for path length measurement in 3D, <5% for cross-sectional area measurements in diseased vessels with irregular lumens, and ±3% to 9% for flow velocity. Errors are greater for vessels <2 mm in diameter.
The anterior and posterior circulation pathways, from the aorta to the circle of Willis, were systematically analyzed in each patient. The anterior pathway included the CCA and ICA; the posterior pathway, the vertebral and basilar arteries. Volume flow (milliliters per minute) was calculated as the product of flow velocity and cross-sectional area. For each vessel segment, the length and various widths were determined. The pressure gradient (
P) was then calculated using Poiseuille's equation (P=8µLQ/
r4, where µ is viscosity of blood, Q is volume blood flow, L is vessel length, and r is vessel radius); P was converted to mm Hg.1 The derivation of this equation is based on the simplifying assumptions that there is steady-state laminar flow and that blood behaves as Newtonian fluid25 ; its use for in vivo estimation of mean pressure drop (averaged over time) is justified empirically.1 Vessels were analyzed in segments of approximately constant width, as determined by inspection of the graphs displayed on the computer screen during processing. The number of vessel segments measured was dependent on the extent of atheroma. The carotid system was separated into three or four separate segments, in addition to the stenotic segment if present. For the posterior pathways, the vertebral arteries were split into two or three segments each; the basilar artery was treated as a single segment. Pulsatile changes in caliber are largely averaged by the slow frame rate and the frame averaging used for geometric purposes. Moreover, nonpulsatility is less of a problem with the atheromatous, stiffened vessels of the elderly.25 26
In the presence of a stenosis, calculations for pressure gradient need to account for (1) the viscous properties of blood within the stenosis and (2) the effects of convergence into and divergence of blood out of the stenosis, and flow separation/turbulence downstream. This is done using the equation P=FV+SV2 (as stated by Gould27 ), where F is the coefficient of the linear pressure loss as in Poiseuille's equation, V is the flow velocity averaged over time and across the lumen, and S is the coefficient of pressure loss due to flow separation (turbulence). It has been shown experimentally by a number of authors that the pulsatile components of flow and the details of stenosis geometry (eg, asymmetry, exit angle) do not affect the application of this formula.1 28 29
In the present study, we attempted to validate the use of these formulas in vivo by calculating pressure drop independently along separate vascular pathways that anastomosed at the circle of Willis. If extracranial vessel systems (eg, the right and left CCA/ICA or right and left vertebral arteries) are in direct communication distally (ie, via the anterior communicating artery or vertebrobasilar junction), the pressure gradient along them should be the same. Similarly, if the posterior communicating arteries are functioning normally, there should be no significant pressure differences between the anterior (carotid) and posterior (vertebrobasilar) circulations.30 31 32 Pressure gradients along these different vessel pathways were compared to see whether these relationships held true. The time-averaged differences in pressure along the aorta at the origins of the main arteries to the head are very small and can be ignored.
Statistical analysis was undertaken using the 
2 test for the comparison of counts (as formulated by Armitage33 ) and the t test for the comparison of means.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The results of the vessel analysis for each patient are presented in Table 1
. Since vessel diameters do not remain constant (because of anatomic tapering and/or atherosclerosis), multiple segmental measurements were performed, giving a range of diameters for each vessel. Pressure gradients of each vessel segment were summed to give total pressure drop for each carotid or vertebrobasilar pathway.
|
Of the five patients, two had ICA stenosis, one ICA occlusion, one bilateral carotid disease (left ICA occlusion, right ICA stenosis), and one a nonstenosing plaque. In three patients, there was absence or occlusion of a vertebral artery. Table 2 summarizes the measurements that allow the interpretation of the hemodynamic significance of the focal stenoses and of the supply pathways in each patient. From QA, the three ICA stenoses measured 55%, 72%, and 88% area reduction (equivalent to 33%, 48%, and 65% diameter reduction, respectively) with calculated P of 1.2, 3.0, and 3.5 mm Hg, respectively. The equivalent percent stenosis readings from carotid duplex ultrasound were approximately 40%, 50%, and 70% diameter reduction. The stenoses in these patients contributed 18%, 24%, and 59%, respectively, of the total pressure drop along the CCA/ICA to the circle of Willis. The range of values for total cerebral blood flow calculated as the sum of the vessel flows was from 522 mL/min (patient 3) to 918 mL/min (patient 2). The mean cerebral blood flow was 1098±310 mL/min.
|
Comparison of pressure gradients between left and right vessel pathways was possible only in patients 1 through 4, since both the right ICA and right vertebral were occluded in patient 5 (Table 3). The mean difference in total pressure drop was 0.6±0.5 mm Hg or 6% (P=NS).
|
Comparison between the mean total pressure drop of the anterior (carotid) circulation and the posterior (vertebrobasilar) circulation (Table 4) revealed only minor differences in patients 3, 4, and 5, indicating that the circle of Willis was functioning normally. However, in patients 1 and 2 there were large differences between the anterior and posterior values. Review of the vascular anatomy on MRI and digital subtraction angiographic images showed the reasons for these differences. In patients 3, 4, and 5, the posterior communicating arteries were patent. In patient 1, there was an incomplete circle of Willis (absence of both posterior communicating arteries) with no communication, and therefore no pressure equilibration, between the anterior and posterior circulations. In patient 2, the right posterior communicating artery was large, and there was substantial anterior-posterior flow that accounted for a significant pressure drop along this vessel. The point of equilibration of pressure drop was therefore displaced posteriorly.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The results of this study indicate that QA techniques for measurement of vessel length, width, area, and volume flow, which have been extensively validated experimentally, can be used in good-quality digitally recorded clinical angiograms of the carotid and vertebral pathways. QA also allows calculation of stenosis dimensions and pressure gradients. In the present study, we have shown good internal consistency of pressure drop estimates. However, in these clinical cases the majority of the quantitative measurements were not cross-validated using different techniques. This work represents the first time a systematic analysis has been undertaken of these vessel parameters in a clinical context, providing a complete "hemodynamic map" for the blood supply to the circle of Willis.
We argue that the "hemodynamic significance" of a stenosis can only be interpreted when static and dynamic measurements relating to the stenosis itself can be compared with measurements made in important collateral pathways. In the absence of good collateral supply, there will be a high pressure drop across a significant stenosis, reflecting the low arterial pressure downstream. On the other hand, when anastomoses via the circle of Willis allow a good collateral supply, a geometrically significant stenosis may only cause a modest pressure drop, reflecting the fact that pressure downstream of the stenosis may be well maintained by the collaterals. In the latter circumstances, evolution of the hemodynamic change with increasing stenosis is different: volume flow will be increasingly carried by the collaterals, and pressure drop across the stenosis will rise less quickly in the affected vessel than would be the case in the absence of collaterals.
Our QA technique is based on current angiographic acquisition methods. The key to quantitative analysis is good vessel definition of the extracranial circulation. For the purposes of this clinical study, nonselective aortic arch injections were performed to determine flow and pressure gradients in all vessels from the aorta to the circle of Willis. Better contrast resolution and freedom from overlapping vessels would be provided by selective injection, but it is harder to avoid injection artifacts with flow measurements near the catheter tip. Selective injection could also increase the risks of angiography. If necessary, selective catheterization could be used to provide further views for improved analysis of the dimensions of a particular vessel or stenosis. The time needed for data processing and quantitative measurement of the whole angiogram is about 60 minutes per pathway. The time is largely due to the experimental nature of the software. We have found it very encouraging that with each successive development of the computer program the processing time has shortened.
The limitations of QA are primarily those of angiographic technique. Overlapping or poorly contrast-enhanced vessels can result in difficulty in detecting vessel edges and therefore in calculating vessel width and cross-sectional area. As in any angiographic study, good positioning, with clear views of each vessel, is imperative. The calculation of flow velocity and flow volume is dependent on the ability to visualize approximately 3 cm of vessel (with no branches) to generate a parametric image. The use of cineangiography frame rates faster than 25 per second could theoretically reduce the length of vessel required.
A number of important findings are evident from the results. Vessel widths were quite variable for the ICA and vertebral artery, which are both anatomically tapering vessels subject to extensive atheroma; the width of the CCA, which is least affected by atheroma, varied little. According to Poiseuille's equation, pressure gradient is inversely proportional to the fourth power of the radius, so small changes in vessel width can result in large changes in pressure gradient. In the course of this study, we found that a failure to calculate pressure drop for each segmental change in vessel radius could result in substantial errors in the calculation of total pressure drop. We were able to test the validity of our approach by comparing, within each patient, independent calculations of total pressure drop along two pathways from the aorta to the circle of Willis. It is reassuring that despite the intrinsic error of each measurement, and the multiple segments that contributed to the calculation of total pressure drop in each vessel, there were no significant differences between results when good anastomosis existed at the level of the circle of Willis (Table
s 2 and 3). The excellent correspondence also confirmed that the quantitative techniques, which have been extensively tested experimentally, are applicable to good-quality clinical angiograms.
Our results are also consistent with recent morphometric data and mathematical modeling showing hemodynamic equilibration of flow in the circle of Willis.30 31 32 The results from the two patients who did have significant anterior/posterior discrepancies in pressure drop were explained by the absence of posterior communicating arteries in one case and a carotid-dependent posterior circulation via a large posterior communicating artery in the other.
The results of the NASCET and ECST carotid surgery trials34 35 have emphasized the need for careful quantification of ICA stenosis, particularly as the benefit from surgery is greater with increasing stenosis severity. Depending on the results of ongoing trials,36 such precision may become important in asymptomatic patients as well. These measurements can be made with QA, in addition to being able to calculate volume flow and stenosis pressure drop. As the equation for pressure gradient over a stenosis suggests, there is a complex interplay between flow, width, and length of a stenosis.27 37 It is clearly not sufficient to predict stenosis hemodynamics on the basis of percent luminal reduction alone. It is only when the hemodynamic data are available for the diseased vessel and the other main vessels that the hemodynamic significance can be understood.
Ideally, similar hemodynamic information could be obtained using noninvasive techniques such as duplex sonography or MR angiography. Duplex ultrasound measures blood flow velocity, and current machines give good information on vessel caliber. We found a close relationship between the duplex and QA measurement of percent stenosis. However, despite the improvement in quality of duplex ultrasound measurements, there are still some reservations about the accuracy of measurement of flow volume and pressure gradient.6 38 39 40 This is due to problems with uncertainty of the ultrasound beam flow angle,41 the small lateral and elevation width of the insonating beam (kept narrow in most commercial systems to maintain image quality), problems with shift in the main Doppler frequency (due to the high pass characteristics of the Doppler receiver and the frequency dependence of tissue attenuation), and noise in the Doppler signal itself.38 40 The use of color duplex (as opposed to pulsed Doppler) compounds these problems because it compromises the ability to detect flow velocity to improve the enhancement of spatial resolution.40
MR angiography measurement of flow is a promising technique, but it also suffers from problems of reliability. The accuracy of caliber and flow measurements is limited, particularly in small vessels, and flow void (signal dropout) in tight stenoses is common.10 42 43 The rapid advances in technology will eventually overcome these problems, but at present digital subtraction angiography remains the gold standard for preoperative assessment of carotid stenosis and evaluation of the cerebral circulation.
In conclusion, quantitative estimation of volume blood flow and pressure drop in all four major (carotid and vertebral) supply pathways to the brain assists in the understanding of the hemodynamic significance of focal stenoses. X-ray angiography can provide this information by means of careful acquisition and computer processing techniques, which have been extensively validated in the laboratory but have not previously been used clinically for assessment of extracranial vascular disease. Further effort is needed to streamline these x-ray angiographic techniques and also to develop noninvasive methodology that would make such quantitative data more readily available in the clinical context.
|
Selected Abbreviations and Acronyms |
|---|
|
Received December 21, 1995; revision received May 16, 1996; accepted May 16, 1996.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
1. Brown GB, Bolson E, Frimer M, Dodge HT. Quantitative coronary arteriography: estimation of dimensions, hemodynamic resistance, and atheroma mass of coronary artery lesions using the arteriogram and digital computation. Circulation. 1977;55:329-337.
2. Colchester ACF, Brunt JNH. Measurement of vessel calibre and volume blood flow by dynamic quantitative digital angiography: an initial application showing variation of cerebral artery diameter with PaCO2. J Cereb Blood Flow Metab. 1983;3(suppl):S640-S641.
3. Reiber JHC, Kooijman CJ, Slager CJ. Computer analysis of the severity of obstructions from coronary cineangiograms: a methodological review. Automedica. 1984;5:219-238.
4. Mancini GBJ. Quantitative coronary arteriographic methods in the interventional catheterization laboratory: an update and perspective. J Am Coll Cardiol. 1991;17:23B-33B.
5.
Fox AJ. How to measure carotid stenosis. Radiology. 1993;186:316-318.
6.
Barnett HJM, Warlow CP. Carotid endarterectomy and the measurement of carotid stenosis. Stroke. 1993;24:1281-1284.
7.
Alexandrov AV, Bladin CF, Norris JW. Measuring carotid stenosis: time for a reappraisal. Stroke. 1993;24:1292-1296.
8. Barnett HJM, Warlow CJ. Stroke. 1994;25:710-711. Letter.
9.
Bladin CF, Alexandrov AV, Murphy J, Magissano R, Norris JW. Carotid stenosis index: a new method of measuring internal carotid artery stenosis. Stroke. 1995;26:230-234.
10. Ackerman RH, Candida MR. Identifying clinically relevant carotid disease. Stroke. 1994;25:1-3.[Medline] [Order article via Infotrieve]
11.
DeWeese JA, May AG, Lipchik EO, Rob CG. Anatomic and hemodynamic correlations in carotid artery stenosis. Stroke. 1970;1:149-157.
12. Powers WJ, Press GW, Grubb RL, Gado M, Raichle ME. The effect of hemodynamically significant carotid artery disease on the hemodynamic status of the cerebral circulation. Ann Intern Med. 1987;106:27-35.
13. Seifalian AM, Hawkes DJ, Hardingham CR, Colchester ACF, Reidy JF. Validation of a quantitative radiographic technique to estimate pulsatile blood flow waveforms using digital subtraction angiographic data. J Biomed Eng. 1991;13:225-233.[Medline] [Order article via Infotrieve]
14. Seifalian AM, Hawkes DJ, Colchester ACF, Hobbs KEF. A new algorithm for deriving pulsatile blood flow waveforms tested using simulated dynamic angiographic data. Neuroradiology. 1989;31:263-269.[Medline] [Order article via Infotrieve]
15. Hoffman KR, Doi K, Fencil LE. Determination of instantaneous and average blood flow rates from digital angiograms of vessel phantoms using distance-density curves. Invest Radiol. 1991;26:207-212.[Medline] [Order article via Infotrieve]
16. Colchester ACF, Seifalian A, Hawkes DJ, Brunt J, Wicks D, du Boulay G. Invest Radiol. 1992;27:274. Letter.[Medline] [Order article via Infotrieve]
17. Hawkes DJ, Seifalian AM, Colchester ACF, Iqbal N, Hardingham CR, Bladin CF, Hobbs KEF. Validation of volume blood flow measurements using three-dimensional distance-concentration functions derived from digital x-ray angiograms. Invest Radiol. 1994;29:434-442.[Medline] [Order article via Infotrieve]
18. Colchester ACF, Hawkes DJ, Brunt JNH, du Boulay GH, Wallis A. Pulsatile blood flow measurements with the aid of 3-D reconstruction from dynamic angiographic recordings. In: Bacharach SL, ed. Information Processing in Medical Imaging. Boston, Mass: Martinus Nijhoff Publishers; 1986:247-265.
19. Hawkes DJ, Mol CR, Colchester ACF. The accurate 3D reconstruction of the geometric configuration of vascular trees from x-ray recordings. In: Guzzardi R, ed. Physics and Engineering of Medical Imaging. The Hague, Netherlands: Martinus Nijhoff; 1987:250-256.
20. Hawkes DJ, Colchester ACF, de Belder MA, Norbury RN, Camm AJ, Davies MJ. The measurement of absolute lumen cross sectional area and lumen geometry in quantitative angiography. In: Todd-Pokropek AE, Viergever MA, eds. Medical Images: Formation, Handling and Evaluation. Heidelberg, Germany: Springer; 1992:609-626. NATO ASI Series.
21. Brunt JN, Wicks DA, Hawkes DJ, Seifalian AM, du Boulay GH, Colchester ACF, Wallis A. The measurement of blood flow waveforms from x-ray angiography, I: principles of the method and preliminary validation. Proc Inst Mech Eng Med [H]. 1992;206:73-85.
22. Hawkes DJ, Colchester ACF, Brunt JNH, Wicks DAG, du Boulay GH, Wallis A. Development of a model to predict the potential accuracy of vessel blood flow measurements from dynamic angiographic recordings. In: Viergever MA, Todd-Pokropek AE, eds. Mathematics and Computer Science in Medical Imaging. Heidelberg, Germany: Springer; 1988:469-478.
23. Hawkes DJ, Hardingham CR, Colchester ACF. Matching of geometric models in the analysis of lumen shape in digital angiography. In: Ortendahl D, Llacer J, eds. Information Processing in Medical Imaging. New York, NY: John Wiley & Sons; 1991;11:341-353.
24. du Boulay GH, Brunt JNH, Colchester ACF, Hawkes DJ, Wallis A, Wicks D. Volume flow measurement of pulsatile flow by digitised cine-angiography. Acta Radiol. 1987;369(suppl XIII):59-62.
25. Milnor WR. Hemodynamics. Baltimore, Md: Williams & Wilkins; 1989.
26. Nichols WW, O'Rourke MF. McDonald's Blood Flow in Arteries. 3rd ed. Philadelphia, Pa: Lea & Febiger; 1990.
27.
Gould LK. Pressure-flow characteristics of coronary stenoses in unsedated dogs at rest and during coronary vasodilation. Circ Res. 1978;43:242-253.
28.
Gould KL, Kelley KO, Bolson EL. Experimental validation of quantitative coronary arteriography for determining pressure-flow characteristics of coronary stenosis. Circulation. 1982;66:930-937.
29.
Young DF, Cholvin NR, Kirkeeider RL, Roth AC. Hemodynamics of arterial stenoses at elevated flow rates. Circ Res. 1977;41:99-107.
30. Hillen B, Hoogstraten HW, Post L. A mathematical model of flow in the circle of Willis. J Biomech. 1986;19:187-194.[Medline] [Order article via Infotrieve]
31. Hillen B. The variability of the circle of Willis: univariate and bivariate analysis. Acta Morphol Neerl Scand. 1986;24:87-101.[Medline] [Order article via Infotrieve]
32. Hillen B. The variability of the circulus arteriosus (Willisii): order or anarchy? Acta Anat. 1987;129:74-80.[Medline] [Order article via Infotrieve]
33. Armitage P. Statistical Methods in Medical Research. London, UK: Blackwell Scientific Publishing; 1980:138-140.
34. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade stenosis. N Engl J Med. 1991;325:445-453.[Abstract]
35. European Carotid Surgery Trialists' Collaborative Group. MRC European carotid surgery trial: interim results for symptomatic patients with severe (70-99%) stenosis or with mild (0-29%) stenosis. Lancet. 1991;337:1235-1244.[Medline] [Order article via Infotrieve]
36. Asymptomatic Carotid Atherosclerosis Study Group. Study design for randomized prospective trial of carotid endarterectomy for asymptomatic atherosclerosis. Stroke. 1989;20:844-899.
37. Gould KL. Dynamic coronary stenosis. Am J Cardiol. 1980;45:287-292.
38.
Eliasziw M, Rankin RN, Fox AJ, Haynes RB, Barnett HJM. Accuracy and prognostic consequences of ultrasonography in identifying severe carotid stenosis. Stroke. 1995;26:1747-1752.
39. Gill RW. Measurement of blood flow by ultrasound: accuracy and sources of error. Ultrasound Med Biol. 1985;11:625-641.[Medline] [Order article via Infotrieve]
40. Burns PN. Measuring volume flow with Doppler ultrasound: an old nut. Ultrasound Obstet Gynecol. 1992;2:248-251.[Medline] [Order article via Infotrieve]
41. Ku DN, Giddens DP, Phillips DJ, Strandness DEJ. Hemodynamics of the normal human carotid bifurcation: in vitro and in vivo studies. Ultrasound Med Biol. 1985;11:13-26.[Medline] [Order article via Infotrieve]
42. Mistretta CA. Relative characteristics of MR angiography and competing vascular imaging modalities. J Magn Reson Imaging. 1993;3:685-698.[Medline] [Order article via Infotrieve]
43. Hangiandreou NJ, Rossman PJ, Riederer SJ. Analysis of MR phase-contrast measurements of pulsatile waveforms. J Magn Reson Imaging. 1993;3:387-394.[Medline] [Order article via Infotrieve]
This article has been cited by other articles:
 |
|
 |
|
  M. J. Hartkamp, J. van der Grond, K. J. van Everdingen, B. Hillen, and W. P. T. M. Mali Circle of Willis Collateral Flow Investigated by Magnetic Resonance Angiography Stroke, December 1, 1999; 30(12): 2671 - 2678. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||