- Email: [email protected]
Carotid artery imaging using duplex scanning and bidirectional arteriography: A comparison
ClinicalRadiology(1984)35~101-106 © 1984RoyalCollegeof Radiologists
0009-9260/84/0263101 $02.00
Carotid Artery Imaging Using Duplex Scanning and Bidirectional Arteriography: A Comparison ERIC COLHOUN and DONAL MacERLEAN
Department of Radiology, St Vincent's Hospital, Elm Park, Dublin
Examination of 101 common carotid bifurcations using ultrasonic methods is compared with the findings of transfemoral bidirectional arteriography. In the common and internal carotid arteries, ultrasonic detection of high-grade stenosis (greater than 50% reduction in lumen diameter) was 100% accurate (25/25) and detection of vessel occlusion was 85 % accurate (11/13). In the two cases of occlusion missed on ultrasonography, a diagnosis of greater than 50% stenosis was made. Of lesions producing low-grade stenosis (less than 50% reduction in lumen diameter), 86% (78/91) were accurately detected using duplex imaging. In 33 % (14/43) of common carotids and 13% (4/30) of internal carotids found to be angiographically normal, duplex scanning revealed the presence of plaque producing up to 25% stenosis, even in the absence of Doppler flow abnormality, confirming the sensitivity of high-resolution real-time imaging in detecting minimal disease.
Symptoms caused by carotid artery occlusive disease result from either reduced cerebral blood flow due to stenosis or occlusion or, alternatively, distal embolisation from an ulcerated plaque or soft thrombus (Bartynski et al., 1981). Carotid arteriography is considered the definitive investigation of the extracranial carotid arteries. However, because of the associated risk and expense, interest has long focused on noninvasive investigations to provide clinically useful information without significant patient morbidity (Ackerman, 1979). The latter requirement is most obviously stated in the assessment of asymptomatic patients at high risk for carotid atherosclerosis or in asymptomatic patients found to have a carotid bruit. The non-invasive approach uses either direct or indirect methods. Indirect methods, which include oculoplethysmography (Gee et al., 1974; Kartchner and McRae, 1977) and periorbital directional Doppler ultrasonography (Muller, 1972), detect only those lesions that reduce cerebral blood flow. Methods of direct non-invasive examination include carotid phonoangiography, isotope angiography, digital subtraction angiography and carotid ultrasonography. Excluding sonographic methods, digital subtraction angiography held most promise as an alternative to arteriography; however, its application has been limited by the relative invasiveness of the technique in practical usage. The following study utilises a system that provides both high-resolution real-time images and a range-gated pulsed Doppler facility to analyse flow patterns in imaged vessels; the so-called 'duplex scanner'. Comparison is made between the findings utilising this method and those of bidirectional arteriography.
PATIENTS AND METHODS
One hundred and one carotid bifurcations in 54 patients were examined sonographically and by transfemoral bidirectional arteriography. The mean time interval between both examinations was 10 days. The presentation of patients at referral varied from asymptomatic carotid bruits to completed strokes and included those with transient ischaemic attacks, non-specific neurological symptoms and patients who had undergone endarterectomy. There were 40 males of mean age 63.1 years (range 40-84 years) and 14 females of mean age 60.1 years (range 44-77 years). Ultrasonic examination was performed using a Diasonics RAI scanner. Within a compact imaging probe a 7.5 MHz transducer moves in a reciprocating fashion to produce a real-time image. Offset below the imaging transducer, with its beam axis in the plane of the B-scan image, is a 3 MHz transducer that provides a pulsed Doppler facility. The Doppler beam can be directed at any angle by the operator. Short pulses of ultrasound are emitted from the transducer and by varying the time at which the same transducer operates as a receiver, analysis of flow may be obtained in a small sample volume at varying locations in the B-mode image, i.e. the system is 'range gated'. The sample volume is cylindrical in shape, having a diameter of 2 mm and a length of 3 mm. When real-time imaging is used alone, the frame rate is 20/s. However, in the combined mode of operation it is reduced to 4 frames/s to permit high-speed data collection by the Doppler system. This reduced frame rate is sufficient to allow the operator to check that the Doppler measurements are being made at the selected site. The position of the Doppler beam is indicated by a white line with a small cross representing the sample volume. The 7.5 MHz imaging system produces useful images to a depth of 8 cm, with an axial resolution of 0.25 mm and a lateral resolution of 0.8 mm. The 3 MHz Doppler system reaches vessels to a depth of 5.5 cm. To reduce the incidence of non-visualisation of poorly echogenic plaques and soft thrombus a high value to the time-based gain setting is used. This increased sensitivity may be associated with artefact echoes in the vessel lumen, a feature of some importance in image interpretation. By alteration of the grey-scale curve, post-acquisition processing may assist in image analysis. The B-scan image is recorded on radiographic film using a multiformat camera. The representation of flow information in a given sample volume is by a zero crossing waveform. This illustrates continuously the average velocity of the red blood cells in the sample volume throughout the cardiac cycle. In the normal common and internal carotid
102
CLINICAL RADIOLOGY
In order to obtain the distribution of red blood cell velocities within the Doppler sample volume at any time in the cardiac cycle, the computer performs a fast Fourier transform analysis, which provides the spectrum of red cell velocities. In the presence of a flow-disturbing plaque, the mean red cell velocity may be increased and evidence of turbulent flow at or distal to the lesion may become apparent. A Doppler sample volume, placed in an area of turbulent flow demonstrates a wider than normal range of red cell velocities including reverse flow components - so-called 'spectral broadening' (Fig. 3).
Fig. 1 - Longitudinal scan of normal common carotid, the upper tracing showing the normal Doppler flow pattern (see text). (All scans illustrated are taken in the longitudinal plane.)
arteries, a broad systolic peak and a relatively high diastolic flow pattern is seen, i.e. a high volume circulation with low peripheral resistance (Fig. 1). Conversely, in the low volume high peripheral resistance of the external carotid circulation, the tracing reveals a sharp systolic peak and low diastolic blood flow (Fig. 2). With disease and altered haemodynamic patterns, such differentiation is severely hindered. Fig. 3 - Extensive atheroma in the internal carotid; the Doppler tracing reveals turbulent flow with 'spectral broadening'.
Fig. 2 - Normal external carotid artery, with normal Doppler flow pattern (see text).
In the study group, atheromatous plaques producing greater than 50% stenosis were always associated with significant Doppler flow abnormalities. However, with plaques producing less than 50% stenosis, Doppler abnormality was frequently seen, particularly if the plaque was irregular in shape or its surface appeared ulcerated. If a minimal plaque was associated with marked Doppler flow abnormality, a careful search for surface ulceration was made. Sonographic diagnosis of occlusion was made in vessels seen to be filled with plaques and/or soft thrombus with absence of flow detected at this site and reversal of flow demonstrated distally. The patient is examined in the supine position with the neck extended and the head turned slightly to the contralateral side. The common carotid is examined first in the longitudinal plane, its course being followed up to the bifurcation. The internal and external carotid arteries are next visualised; rarely are both clearly seen in one longitudinal plane (Wolverson et al., 1983). The region of the bifurcation, of primary interest in this study, is examined in at least two longitudinal planes, preferably at right angles to one another, to detect small areas of plaques. Doppler analysis of flow pattern is undertaken in all three vessels and in areas of plaque
ULTRASOUND AND A R T E R I O G R A P H Y IN CAROTID IMAGING
demonstration. The Doppler signal lies in the audible range, a feature of practical use in image interpretation. Transverse scans help to quantify degrees of stenosis in areas of asymmetrical plaque formation and also help to define the relationship of internal and external carotid arteries to one another. Aortic arch arteriography was performed by catheterisation of the femoral artery using standard Seldinger technique, two views of the cervical carotid arteries being taken at right angles to one another. The percentage of stenosis, if present, was determined and presence or absence of ulceration noted.
RESULTS Duplex scanning demonstrated all 101 common carotid arteries and their internal carotid branches. In all but seven cases the external carotid artery was also identified. Sonographic findings were documented first, independently of angiographic findings, with which they were later compared (see Tables 1 and 2). F o u r categories were used to describe the examination findings: normal, less than 50% stenosis, greater than 50% stenosis or occlusion. Table 1 - Common carotid arteries (101)
Angiography
Dupleximaging
103
accurate (21/21) in detecting stenosis of greater than 50%. Two occlusions were missed, a diagnosis of greater than 50% stenosis being made in both cases (see discussion). Detection of stenosis of less than 50% was 80% accurate (32/40). Of 71 internal carotid lesions demonstrated angiographically, 61 were accurately diagnosed using duplex scanning, a sensitivity of 86%. Of the 30 internal carotids felt to be angiographically normal, duplex scanning demonstrated minimal plaque formation in four (13%). Consideration of both common and internal carotid arteries together reveals that 114 of the 129 angiographically demonstrated lesions were detected accurately using duplex scanning, a sensitivity of 88%. If lesions that reduce cerebral blood flow, i.e. occlusive lesions or lesions producing greater than 50% stenosis are considered as a separate group, all were detected (38/38) using duplex imaging. Within this group 36 were accurately categorised: a sensitivity of 95%; two occlusive lesions were mistakenly diagnosed as areas of high-grade stenosis. Of 73 angiographically normal vessels, using duplex imaging 18 were seen to contain minimal plaques which in no case was seen to produce greater than 25% stenosis. Identification of ulcerated plaques by duplex scanning was unreliable and rarely could ulceration be diagnosed with confidence using this method. DISCUSSION
Agrees Disagrees" Occluded Greater than 50% stenosis Less than 50% stenosis Normal
3 4 51 43
3
4 46 29
5* 14t
* Sonographically normal. t Sonographically abnormal. Table 2 - Internal carotid arteries (101)
Angiography
Dupleximaging Agrees Disagrees
Occluded Greater than 50% stenosis Less than 50% stenosis Normal
10 21 40 30
8 21 32 26
2* 8t 4~
* Both diagnosed as greater than 50% stenosis. tSonographically normal. J; Sonographically abnormal.
Duplex scanning of the common carotid arteries detected with 100% accuracy lesions producing occlusion (3/3) or greater than 50% stenosis (4/4). Lesions producing less than 50% stenosis were detected with an accuracy of 90% (46/51). In all five cases felt to be sonographically normal the angiographic lesion was minimal (less than 25% stenosis). Consequently, of 58 COmmon carotid lesions demonstrated angiographically, duplex scanning accurately detected 53, a sensitivity or true positive rate of 93%. Of the 43 common carotids found to be angiographically normal, duplex scanning demonstrated plaques in 14 (33%). Duplex scanning of the internal carotid arteries was 80% accurate (8/10) in detecting occlusion and 100%
Atherosclerosis is the major aetiological factor in the pathogenesis of stroke, the commonest site of involvement being the common carotid bifurcation (Palmer et al., 1977). Techniques such as oculoplethysmography and periorbital directional Doppler ultrasonography detect lesions in the common or internal carotid arteries which reduce cerebral blood flow. Failure to detect the large numbers of lesions that do not reduce distal flow, inability to differentiate between high-grade stenosis and occlusion, failure to distinguish between disease in the ophthalmic and carotid arteries and difficulty in detection of bilateral flow-reducing lesions limit the use of indirect methods. Doppler imaging of the cervical carotid arteries using both continuous wave and pulsed Doppler principles is well established. Blackshear et al. (1979), using pulsed Doppler to produce a flow map and, in addition, frequency spectral analysis of the Doppler signal from small sample volumes, achieved 88% accuracy in detection of lesions producing either occlusion or stenosis of greater than 50%, i.e. flow-reducing lesions. However, in detection of stenosis of less than 50% the accuracy fell to 53%. In addition, as Doppler devices image flowing blood, the images represent the vessel lumen and provide no information as to the nature of the vessel wall in normal or narrowed segments. High-resolution real-time ultrasound imaging, unlike Doppler imaging methods, visualises the echogenic vessel walt but provides no information concerning blood flow. Occasionally, a thin echogenic line parallel to the thicker vessel wall and lying within 1 mm of it is seen. Wolverson et al. (1983) believe that this normal appearance represents specular reflections from the blood-intima interface. Atheromatous plaques encroach on the vessel lumen, their echodensity varying
104
CLINICAL RADIOLOGY
Fig. 6 - Plaque containing areas of calcification with acoustic shadowing. The residual lumen is indicated by the markers. Fig. 4 - Posterior wall plaque, predominantly fibrous material. with their composition, three categories being described (Wolverson et al., 1983). Brightly echogenic plaques without acoustic shadowing contain a large proportion of fibrous material (Fig. 4). Plaques with a high lipid content are poorly echogenic and have an acoustic impedance very similar to that of blood, making detection difficult, even with high-resolution images. Identification is aided by maintaining a high value to the time-based gain setting in the B-mode image construction and by Doppler interrogation of suspicious areas to detect absent flow (Fig. 5). Brightly echogenic plaques with acoustic shadowing contain areas of calcification (Fig. 6). If such a plaque is
present in the superficial wall of a vessel, visualisation of the distal vessel wall and underlying lumen or Doppler analysis of flow in this region may be difficult. Examination of these regions may be facilitated by using alternative planes of approach, provided the calcified plaque is not circumferential. Despite these limitations, Wolverson et al. (1983) in a comparison of real-time ultrasonography with arteriography, achieved an accuracy of 78% in the detection of stenosis of greater than 50% and an 83% accuracy in detecting stenosis of less than 50%. Significantly, only 36% of occlusive lesions (4/11) were detected. James et al. (1982) also found detection of occlusion difficult with this method, achieving an accuracy of only 25% (3/12) in this regard. This study and those of previous authors which
(a) (b) Fig. 5 - (a) Common carotid artery filled with material of low echodensity. The markers outline a small sonolucent area representing partial recanalisatiom (b) Doppler analysis confirms recanalisation, reverse flow being demonstrated.
ULTRASOUND AND ARTER1OGRAPHY IN CAROTID IMAGING
compare duplex imaging methods with arteriography demonstrate clearly the advantages of additional Doppler flow information in the assessment of carotid artery occlusive disease. Blackshear et aI. (1979), using a duplex scanner and pulsed Doppler spectral analysis, demonstrated accurately 75% (3/4) of occlusive lesions, 95% (21/22) of lesions producing greater than 50% stenosis and 60% (21/32) of lesions producing stenosis of less than 50%. The relative inaccuracy in detection of low-grade stenosis may be related to the low frequency used (5 MHz) in B-mode image production. Zbornikova et al. (1982) achieved 93% (14/15) accuracy in detection of stenosis of greater than 50% and the detection of occlusive lesions was 100% (9/9) accurate. The accuracy in differentiation between low-grade stenosis (less than 50%) and normal vessels was not stated. However, the low frequency used in production of the B-mode image (3 MHz) may have precluded accurate demonstration of minimal plaques in the absence of flow abnormality. In our study, the B-mode image transducer utitised a higher frequency (7.5MHz) than those used in the duplex instruments of Blackshear and Zbornikova (5MHz and 3 MHz, respectively). As a consequence, the greater anatomical detail afforded is reflected in the accuracy of detection of angiographically demonstrated minimal disease (86% sensitivity in our series). The detection of minimal plaques in 25% (18/73) of vessels with no angiographically demonstrated abnormality confirms the sensitivity of the imaging method used. Detection of high-grade stenosis (greater than 50 %) was 100% accurate (25/25) and detection of occlusive lesions was 85% accurate (11/13). Both errors occurred in one patient whose bilaterally occluded internal carotid arteries were mistakenly described as containing stenosis of greater than 50% severity. In this patient, branches
Fig. 7 - Arteriogram showing occlusion of the proximal internal carotid artery (curved arrow). The hypertrophiedascendingpharyngeal branch of the external carotid artery (straight arrow) is also shown.
105
of the bilaterally well developed external carotid arteries were incorrectly identified as internal carotid arteries (Fig. 7). Normally, comparison of Doppler signals in identified vessels would have revealed external carotid flow characteristics in both. However, the presence of extensive plaque formation prevented this distinction and the error in vessel identification was not recognised. All lesions that reduced cerebral blood flow, i.e. lesions producing greater than 50% stenosis or occlusion, were detected using duplex imaging. Routine referral of these patients for angiography will detect occasional misclassification within this group, as occurred in 5% (2/38) of cases in our series. The differentiation of high-grade stenosis and occlusion is important as surgery offers no benefit to the latter group of patients. The inability to diagnose with confidence surface ulceration of plaques using the duplex scanner represents a relative disadvantage of the technique. Differentiation of ulceration from areas within mixed plaques of very low echodensity is difficult, even with the high values to the time-based gain settings that were used. Careful scrutiny of demonstrated plaques by moving the sample volume along the surface may detect local turbulence, a feature suggestive of ulceration. However, the size of the sample volume used in our study (9.4mm 3) precluded confident localisation of small areas of turbulence. In addition, as patients with classical transient ischaemic attacks are likely to require rapid surgical intervention, we feel that immediate referral for arteriography following clinical presentation is indicated in this group. Recognition of intraluminal soft thrombus and plaques of low echodensity whose acoustic impedance properties are very similar to blood represents a major problem in methods using B-mode imaging alone. Using duplex scanning, their identification is made more likely if sample volumes reveal absent flow on Doppler 'interrogation' of suspicious areas. In addition, by alteration of the grey-scale curve, the Diasonics RAI provides a facility which allows for B-mode image reconstruction following data acquisition, a process which can be used to highlight areas of low echodensity (Fig. 8). The sensitivity and accuracy of duplex imaging in detection of all grades of disease in the region of the common carotid bifurcation compares favourably with other non-invasive methods, including digital subtraction angiography. In addition, problems specific to digital subtraction angiography, which include motion artefact and the relative invasiveness of the method of examination, are not encountered with sonographic methods. In the evaluation of asymptomatic patients at high risk for atherosclerosis, or asymptomatic patients with carotid bruits, using duplex imaging techniques all patients with atheromatous plaques producing greater than 50% stenosis will be identified. This group and all other patients with disease of similar extent should proceed to arteriography for detailed analysis of the plaques and for confident distinction between highgrade stenosis or occlusion. The early and accurate detection of minimal lesions using duplex imaging methods should permit folIow-up studies to document non-invasively the natural history of the atherosclerotic process with and without therapeutic intervention. In addition, endarterectomy pa-
106
CLINICALRADIOLOGY
(a)
(b)
Fig. 8 - Doppler analysis of flow from the region of the markers reveals turbulent flow (a). Post-processing highlights the causative plaque (b).
tients may be m o n i t o r e d p o s t - o p e r a t i v e l y to e n s u r e p a t e n c y in the short t e r m a n d to detect r e c u r r e n c e of a t h e r o m a in the long term. F e w surgeons w o u l d c o n t e m p l a t e o p e r a t i o n in the territory of the e x t r a c r a n i a l carotid arteries w i t h o u t arteriographic studies. H o w e v e r , the r a t i o n a l use of n o n - i n v a s i v e m e t h o d s such as d u p l e x i m a g i n g in c o n j u n c t i o n with c o m p r e h e n s i v e clinical e v a l u a t i o n should allow for b e t t e r selection of p a t i e n t s for angiographic studies a n d p o t e n t i a l o p e r a t i v e i n t e r v e n tion.
Acknowledgements. The authors would like to thank Miss Joan Miller, Senior Radiographer, for her assistance in this work.
REFERENCES Ackerman, R. H. (1979). A perspective on non-invasive diagnosis of carotid disease. Neurology, 29, 615-621. Bartynski, W. S., Darbouze, P. & Nemir, P. (198l). Significance of ulcerated plaque in transient cerebral ischaemia. American Journal of Surgery, 141,353-357.
Blackshear, W. M., Hirscb, J. H. & Strandness, D. E. (1979). Detection of carotid occlusive disease by ultrasonic imaging and pulsed Doppler spectrum analysis. Surgery, 86, 698-706. Gee, W., Smith, C. A. & Hinson, C. E. (1974). Ocular pneumoplethysmography in carotid artery disease. Medical Instrumentation, 8, 244-248. James, E. M., Earnest, F., Forbes, G. S., Reese, D. F., Houser, D. W. & Folger, W. N. (1982). High resolution dynamic ultrasound imaging of the carotid bifurcation: a prospective evaluation. Radiology, 144, 853-858. Kartchner, M. & McRae, L. (1977). Non-invasive evaluation and management of the 'asymptomatic' carotid bruit. Surgery, 82, 840-847. Muller, H. R. (1972). The diagnosis of internal carotid artery occlusion by directional Doppler sonography of the ophthalmic artery. Neurology, 22, 816-823. Palmer, F. J., Briscoe, P. J., Sorby, W. A., Barry, B. P. & Williams, R. M. (1977). Transfemoral selective angiography in the investigation of cerebral ischaemic disease. Review of 400 consecutive studies. 11. Pathological studies. Australasian Radiology, 21, 116-121. Wolverson, M. K., Heiberg, E., Sundaram, M., Tantanasirviongse, S. & Shields, J. B. (1983). Carotid atherosclerosis: high resolution real-time sonography correlated with angiography. American Journal of Roentgenology, 140, 355-361 Zbornikova, V., Akesson, J. A. & Lassvik, C. (1982). Diagnosis of carotid artery disease - comparison between directional Doppler, duplex scanner and angiography. Acta Neurologica Scandinavica, 65, 335-346.
0009-9260/84/0263101 $02.00
Carotid Artery Imaging Using Duplex Scanning and Bidirectional Arteriography: A Comparison ERIC COLHOUN and DONAL MacERLEAN
Department of Radiology, St Vincent's Hospital, Elm Park, Dublin
Examination of 101 common carotid bifurcations using ultrasonic methods is compared with the findings of transfemoral bidirectional arteriography. In the common and internal carotid arteries, ultrasonic detection of high-grade stenosis (greater than 50% reduction in lumen diameter) was 100% accurate (25/25) and detection of vessel occlusion was 85 % accurate (11/13). In the two cases of occlusion missed on ultrasonography, a diagnosis of greater than 50% stenosis was made. Of lesions producing low-grade stenosis (less than 50% reduction in lumen diameter), 86% (78/91) were accurately detected using duplex imaging. In 33 % (14/43) of common carotids and 13% (4/30) of internal carotids found to be angiographically normal, duplex scanning revealed the presence of plaque producing up to 25% stenosis, even in the absence of Doppler flow abnormality, confirming the sensitivity of high-resolution real-time imaging in detecting minimal disease.
Symptoms caused by carotid artery occlusive disease result from either reduced cerebral blood flow due to stenosis or occlusion or, alternatively, distal embolisation from an ulcerated plaque or soft thrombus (Bartynski et al., 1981). Carotid arteriography is considered the definitive investigation of the extracranial carotid arteries. However, because of the associated risk and expense, interest has long focused on noninvasive investigations to provide clinically useful information without significant patient morbidity (Ackerman, 1979). The latter requirement is most obviously stated in the assessment of asymptomatic patients at high risk for carotid atherosclerosis or in asymptomatic patients found to have a carotid bruit. The non-invasive approach uses either direct or indirect methods. Indirect methods, which include oculoplethysmography (Gee et al., 1974; Kartchner and McRae, 1977) and periorbital directional Doppler ultrasonography (Muller, 1972), detect only those lesions that reduce cerebral blood flow. Methods of direct non-invasive examination include carotid phonoangiography, isotope angiography, digital subtraction angiography and carotid ultrasonography. Excluding sonographic methods, digital subtraction angiography held most promise as an alternative to arteriography; however, its application has been limited by the relative invasiveness of the technique in practical usage. The following study utilises a system that provides both high-resolution real-time images and a range-gated pulsed Doppler facility to analyse flow patterns in imaged vessels; the so-called 'duplex scanner'. Comparison is made between the findings utilising this method and those of bidirectional arteriography.
PATIENTS AND METHODS
One hundred and one carotid bifurcations in 54 patients were examined sonographically and by transfemoral bidirectional arteriography. The mean time interval between both examinations was 10 days. The presentation of patients at referral varied from asymptomatic carotid bruits to completed strokes and included those with transient ischaemic attacks, non-specific neurological symptoms and patients who had undergone endarterectomy. There were 40 males of mean age 63.1 years (range 40-84 years) and 14 females of mean age 60.1 years (range 44-77 years). Ultrasonic examination was performed using a Diasonics RAI scanner. Within a compact imaging probe a 7.5 MHz transducer moves in a reciprocating fashion to produce a real-time image. Offset below the imaging transducer, with its beam axis in the plane of the B-scan image, is a 3 MHz transducer that provides a pulsed Doppler facility. The Doppler beam can be directed at any angle by the operator. Short pulses of ultrasound are emitted from the transducer and by varying the time at which the same transducer operates as a receiver, analysis of flow may be obtained in a small sample volume at varying locations in the B-mode image, i.e. the system is 'range gated'. The sample volume is cylindrical in shape, having a diameter of 2 mm and a length of 3 mm. When real-time imaging is used alone, the frame rate is 20/s. However, in the combined mode of operation it is reduced to 4 frames/s to permit high-speed data collection by the Doppler system. This reduced frame rate is sufficient to allow the operator to check that the Doppler measurements are being made at the selected site. The position of the Doppler beam is indicated by a white line with a small cross representing the sample volume. The 7.5 MHz imaging system produces useful images to a depth of 8 cm, with an axial resolution of 0.25 mm and a lateral resolution of 0.8 mm. The 3 MHz Doppler system reaches vessels to a depth of 5.5 cm. To reduce the incidence of non-visualisation of poorly echogenic plaques and soft thrombus a high value to the time-based gain setting is used. This increased sensitivity may be associated with artefact echoes in the vessel lumen, a feature of some importance in image interpretation. By alteration of the grey-scale curve, post-acquisition processing may assist in image analysis. The B-scan image is recorded on radiographic film using a multiformat camera. The representation of flow information in a given sample volume is by a zero crossing waveform. This illustrates continuously the average velocity of the red blood cells in the sample volume throughout the cardiac cycle. In the normal common and internal carotid
102
CLINICAL RADIOLOGY
In order to obtain the distribution of red blood cell velocities within the Doppler sample volume at any time in the cardiac cycle, the computer performs a fast Fourier transform analysis, which provides the spectrum of red cell velocities. In the presence of a flow-disturbing plaque, the mean red cell velocity may be increased and evidence of turbulent flow at or distal to the lesion may become apparent. A Doppler sample volume, placed in an area of turbulent flow demonstrates a wider than normal range of red cell velocities including reverse flow components - so-called 'spectral broadening' (Fig. 3).
Fig. 1 - Longitudinal scan of normal common carotid, the upper tracing showing the normal Doppler flow pattern (see text). (All scans illustrated are taken in the longitudinal plane.)
arteries, a broad systolic peak and a relatively high diastolic flow pattern is seen, i.e. a high volume circulation with low peripheral resistance (Fig. 1). Conversely, in the low volume high peripheral resistance of the external carotid circulation, the tracing reveals a sharp systolic peak and low diastolic blood flow (Fig. 2). With disease and altered haemodynamic patterns, such differentiation is severely hindered. Fig. 3 - Extensive atheroma in the internal carotid; the Doppler tracing reveals turbulent flow with 'spectral broadening'.
Fig. 2 - Normal external carotid artery, with normal Doppler flow pattern (see text).
In the study group, atheromatous plaques producing greater than 50% stenosis were always associated with significant Doppler flow abnormalities. However, with plaques producing less than 50% stenosis, Doppler abnormality was frequently seen, particularly if the plaque was irregular in shape or its surface appeared ulcerated. If a minimal plaque was associated with marked Doppler flow abnormality, a careful search for surface ulceration was made. Sonographic diagnosis of occlusion was made in vessels seen to be filled with plaques and/or soft thrombus with absence of flow detected at this site and reversal of flow demonstrated distally. The patient is examined in the supine position with the neck extended and the head turned slightly to the contralateral side. The common carotid is examined first in the longitudinal plane, its course being followed up to the bifurcation. The internal and external carotid arteries are next visualised; rarely are both clearly seen in one longitudinal plane (Wolverson et al., 1983). The region of the bifurcation, of primary interest in this study, is examined in at least two longitudinal planes, preferably at right angles to one another, to detect small areas of plaques. Doppler analysis of flow pattern is undertaken in all three vessels and in areas of plaque
ULTRASOUND AND A R T E R I O G R A P H Y IN CAROTID IMAGING
demonstration. The Doppler signal lies in the audible range, a feature of practical use in image interpretation. Transverse scans help to quantify degrees of stenosis in areas of asymmetrical plaque formation and also help to define the relationship of internal and external carotid arteries to one another. Aortic arch arteriography was performed by catheterisation of the femoral artery using standard Seldinger technique, two views of the cervical carotid arteries being taken at right angles to one another. The percentage of stenosis, if present, was determined and presence or absence of ulceration noted.
RESULTS Duplex scanning demonstrated all 101 common carotid arteries and their internal carotid branches. In all but seven cases the external carotid artery was also identified. Sonographic findings were documented first, independently of angiographic findings, with which they were later compared (see Tables 1 and 2). F o u r categories were used to describe the examination findings: normal, less than 50% stenosis, greater than 50% stenosis or occlusion. Table 1 - Common carotid arteries (101)
Angiography
Dupleximaging
103
accurate (21/21) in detecting stenosis of greater than 50%. Two occlusions were missed, a diagnosis of greater than 50% stenosis being made in both cases (see discussion). Detection of stenosis of less than 50% was 80% accurate (32/40). Of 71 internal carotid lesions demonstrated angiographically, 61 were accurately diagnosed using duplex scanning, a sensitivity of 86%. Of the 30 internal carotids felt to be angiographically normal, duplex scanning demonstrated minimal plaque formation in four (13%). Consideration of both common and internal carotid arteries together reveals that 114 of the 129 angiographically demonstrated lesions were detected accurately using duplex scanning, a sensitivity of 88%. If lesions that reduce cerebral blood flow, i.e. occlusive lesions or lesions producing greater than 50% stenosis are considered as a separate group, all were detected (38/38) using duplex imaging. Within this group 36 were accurately categorised: a sensitivity of 95%; two occlusive lesions were mistakenly diagnosed as areas of high-grade stenosis. Of 73 angiographically normal vessels, using duplex imaging 18 were seen to contain minimal plaques which in no case was seen to produce greater than 25% stenosis. Identification of ulcerated plaques by duplex scanning was unreliable and rarely could ulceration be diagnosed with confidence using this method. DISCUSSION
Agrees Disagrees" Occluded Greater than 50% stenosis Less than 50% stenosis Normal
3 4 51 43
3
4 46 29
5* 14t
* Sonographically normal. t Sonographically abnormal. Table 2 - Internal carotid arteries (101)
Angiography
Dupleximaging Agrees Disagrees
Occluded Greater than 50% stenosis Less than 50% stenosis Normal
10 21 40 30
8 21 32 26
2* 8t 4~
* Both diagnosed as greater than 50% stenosis. tSonographically normal. J; Sonographically abnormal.
Duplex scanning of the common carotid arteries detected with 100% accuracy lesions producing occlusion (3/3) or greater than 50% stenosis (4/4). Lesions producing less than 50% stenosis were detected with an accuracy of 90% (46/51). In all five cases felt to be sonographically normal the angiographic lesion was minimal (less than 25% stenosis). Consequently, of 58 COmmon carotid lesions demonstrated angiographically, duplex scanning accurately detected 53, a sensitivity or true positive rate of 93%. Of the 43 common carotids found to be angiographically normal, duplex scanning demonstrated plaques in 14 (33%). Duplex scanning of the internal carotid arteries was 80% accurate (8/10) in detecting occlusion and 100%
Atherosclerosis is the major aetiological factor in the pathogenesis of stroke, the commonest site of involvement being the common carotid bifurcation (Palmer et al., 1977). Techniques such as oculoplethysmography and periorbital directional Doppler ultrasonography detect lesions in the common or internal carotid arteries which reduce cerebral blood flow. Failure to detect the large numbers of lesions that do not reduce distal flow, inability to differentiate between high-grade stenosis and occlusion, failure to distinguish between disease in the ophthalmic and carotid arteries and difficulty in detection of bilateral flow-reducing lesions limit the use of indirect methods. Doppler imaging of the cervical carotid arteries using both continuous wave and pulsed Doppler principles is well established. Blackshear et al. (1979), using pulsed Doppler to produce a flow map and, in addition, frequency spectral analysis of the Doppler signal from small sample volumes, achieved 88% accuracy in detection of lesions producing either occlusion or stenosis of greater than 50%, i.e. flow-reducing lesions. However, in detection of stenosis of less than 50% the accuracy fell to 53%. In addition, as Doppler devices image flowing blood, the images represent the vessel lumen and provide no information as to the nature of the vessel wall in normal or narrowed segments. High-resolution real-time ultrasound imaging, unlike Doppler imaging methods, visualises the echogenic vessel walt but provides no information concerning blood flow. Occasionally, a thin echogenic line parallel to the thicker vessel wall and lying within 1 mm of it is seen. Wolverson et al. (1983) believe that this normal appearance represents specular reflections from the blood-intima interface. Atheromatous plaques encroach on the vessel lumen, their echodensity varying
104
CLINICAL RADIOLOGY
Fig. 6 - Plaque containing areas of calcification with acoustic shadowing. The residual lumen is indicated by the markers. Fig. 4 - Posterior wall plaque, predominantly fibrous material. with their composition, three categories being described (Wolverson et al., 1983). Brightly echogenic plaques without acoustic shadowing contain a large proportion of fibrous material (Fig. 4). Plaques with a high lipid content are poorly echogenic and have an acoustic impedance very similar to that of blood, making detection difficult, even with high-resolution images. Identification is aided by maintaining a high value to the time-based gain setting in the B-mode image construction and by Doppler interrogation of suspicious areas to detect absent flow (Fig. 5). Brightly echogenic plaques with acoustic shadowing contain areas of calcification (Fig. 6). If such a plaque is
present in the superficial wall of a vessel, visualisation of the distal vessel wall and underlying lumen or Doppler analysis of flow in this region may be difficult. Examination of these regions may be facilitated by using alternative planes of approach, provided the calcified plaque is not circumferential. Despite these limitations, Wolverson et al. (1983) in a comparison of real-time ultrasonography with arteriography, achieved an accuracy of 78% in the detection of stenosis of greater than 50% and an 83% accuracy in detecting stenosis of less than 50%. Significantly, only 36% of occlusive lesions (4/11) were detected. James et al. (1982) also found detection of occlusion difficult with this method, achieving an accuracy of only 25% (3/12) in this regard. This study and those of previous authors which
(a) (b) Fig. 5 - (a) Common carotid artery filled with material of low echodensity. The markers outline a small sonolucent area representing partial recanalisatiom (b) Doppler analysis confirms recanalisation, reverse flow being demonstrated.
ULTRASOUND AND ARTER1OGRAPHY IN CAROTID IMAGING
compare duplex imaging methods with arteriography demonstrate clearly the advantages of additional Doppler flow information in the assessment of carotid artery occlusive disease. Blackshear et aI. (1979), using a duplex scanner and pulsed Doppler spectral analysis, demonstrated accurately 75% (3/4) of occlusive lesions, 95% (21/22) of lesions producing greater than 50% stenosis and 60% (21/32) of lesions producing stenosis of less than 50%. The relative inaccuracy in detection of low-grade stenosis may be related to the low frequency used (5 MHz) in B-mode image production. Zbornikova et al. (1982) achieved 93% (14/15) accuracy in detection of stenosis of greater than 50% and the detection of occlusive lesions was 100% (9/9) accurate. The accuracy in differentiation between low-grade stenosis (less than 50%) and normal vessels was not stated. However, the low frequency used in production of the B-mode image (3 MHz) may have precluded accurate demonstration of minimal plaques in the absence of flow abnormality. In our study, the B-mode image transducer utitised a higher frequency (7.5MHz) than those used in the duplex instruments of Blackshear and Zbornikova (5MHz and 3 MHz, respectively). As a consequence, the greater anatomical detail afforded is reflected in the accuracy of detection of angiographically demonstrated minimal disease (86% sensitivity in our series). The detection of minimal plaques in 25% (18/73) of vessels with no angiographically demonstrated abnormality confirms the sensitivity of the imaging method used. Detection of high-grade stenosis (greater than 50 %) was 100% accurate (25/25) and detection of occlusive lesions was 85% accurate (11/13). Both errors occurred in one patient whose bilaterally occluded internal carotid arteries were mistakenly described as containing stenosis of greater than 50% severity. In this patient, branches
Fig. 7 - Arteriogram showing occlusion of the proximal internal carotid artery (curved arrow). The hypertrophiedascendingpharyngeal branch of the external carotid artery (straight arrow) is also shown.
105
of the bilaterally well developed external carotid arteries were incorrectly identified as internal carotid arteries (Fig. 7). Normally, comparison of Doppler signals in identified vessels would have revealed external carotid flow characteristics in both. However, the presence of extensive plaque formation prevented this distinction and the error in vessel identification was not recognised. All lesions that reduced cerebral blood flow, i.e. lesions producing greater than 50% stenosis or occlusion, were detected using duplex imaging. Routine referral of these patients for angiography will detect occasional misclassification within this group, as occurred in 5% (2/38) of cases in our series. The differentiation of high-grade stenosis and occlusion is important as surgery offers no benefit to the latter group of patients. The inability to diagnose with confidence surface ulceration of plaques using the duplex scanner represents a relative disadvantage of the technique. Differentiation of ulceration from areas within mixed plaques of very low echodensity is difficult, even with the high values to the time-based gain settings that were used. Careful scrutiny of demonstrated plaques by moving the sample volume along the surface may detect local turbulence, a feature suggestive of ulceration. However, the size of the sample volume used in our study (9.4mm 3) precluded confident localisation of small areas of turbulence. In addition, as patients with classical transient ischaemic attacks are likely to require rapid surgical intervention, we feel that immediate referral for arteriography following clinical presentation is indicated in this group. Recognition of intraluminal soft thrombus and plaques of low echodensity whose acoustic impedance properties are very similar to blood represents a major problem in methods using B-mode imaging alone. Using duplex scanning, their identification is made more likely if sample volumes reveal absent flow on Doppler 'interrogation' of suspicious areas. In addition, by alteration of the grey-scale curve, the Diasonics RAI provides a facility which allows for B-mode image reconstruction following data acquisition, a process which can be used to highlight areas of low echodensity (Fig. 8). The sensitivity and accuracy of duplex imaging in detection of all grades of disease in the region of the common carotid bifurcation compares favourably with other non-invasive methods, including digital subtraction angiography. In addition, problems specific to digital subtraction angiography, which include motion artefact and the relative invasiveness of the method of examination, are not encountered with sonographic methods. In the evaluation of asymptomatic patients at high risk for atherosclerosis, or asymptomatic patients with carotid bruits, using duplex imaging techniques all patients with atheromatous plaques producing greater than 50% stenosis will be identified. This group and all other patients with disease of similar extent should proceed to arteriography for detailed analysis of the plaques and for confident distinction between highgrade stenosis or occlusion. The early and accurate detection of minimal lesions using duplex imaging methods should permit folIow-up studies to document non-invasively the natural history of the atherosclerotic process with and without therapeutic intervention. In addition, endarterectomy pa-
106
CLINICALRADIOLOGY
(a)
(b)
Fig. 8 - Doppler analysis of flow from the region of the markers reveals turbulent flow (a). Post-processing highlights the causative plaque (b).
tients may be m o n i t o r e d p o s t - o p e r a t i v e l y to e n s u r e p a t e n c y in the short t e r m a n d to detect r e c u r r e n c e of a t h e r o m a in the long term. F e w surgeons w o u l d c o n t e m p l a t e o p e r a t i o n in the territory of the e x t r a c r a n i a l carotid arteries w i t h o u t arteriographic studies. H o w e v e r , the r a t i o n a l use of n o n - i n v a s i v e m e t h o d s such as d u p l e x i m a g i n g in c o n j u n c t i o n with c o m p r e h e n s i v e clinical e v a l u a t i o n should allow for b e t t e r selection of p a t i e n t s for angiographic studies a n d p o t e n t i a l o p e r a t i v e i n t e r v e n tion.
Acknowledgements. The authors would like to thank Miss Joan Miller, Senior Radiographer, for her assistance in this work.
REFERENCES Ackerman, R. H. (1979). A perspective on non-invasive diagnosis of carotid disease. Neurology, 29, 615-621. Bartynski, W. S., Darbouze, P. & Nemir, P. (198l). Significance of ulcerated plaque in transient cerebral ischaemia. American Journal of Surgery, 141,353-357.
Blackshear, W. M., Hirscb, J. H. & Strandness, D. E. (1979). Detection of carotid occlusive disease by ultrasonic imaging and pulsed Doppler spectrum analysis. Surgery, 86, 698-706. Gee, W., Smith, C. A. & Hinson, C. E. (1974). Ocular pneumoplethysmography in carotid artery disease. Medical Instrumentation, 8, 244-248. James, E. M., Earnest, F., Forbes, G. S., Reese, D. F., Houser, D. W. & Folger, W. N. (1982). High resolution dynamic ultrasound imaging of the carotid bifurcation: a prospective evaluation. Radiology, 144, 853-858. Kartchner, M. & McRae, L. (1977). Non-invasive evaluation and management of the 'asymptomatic' carotid bruit. Surgery, 82, 840-847. Muller, H. R. (1972). The diagnosis of internal carotid artery occlusion by directional Doppler sonography of the ophthalmic artery. Neurology, 22, 816-823. Palmer, F. J., Briscoe, P. J., Sorby, W. A., Barry, B. P. & Williams, R. M. (1977). Transfemoral selective angiography in the investigation of cerebral ischaemic disease. Review of 400 consecutive studies. 11. Pathological studies. Australasian Radiology, 21, 116-121. Wolverson, M. K., Heiberg, E., Sundaram, M., Tantanasirviongse, S. & Shields, J. B. (1983). Carotid atherosclerosis: high resolution real-time sonography correlated with angiography. American Journal of Roentgenology, 140, 355-361 Zbornikova, V., Akesson, J. A. & Lassvik, C. (1982). Diagnosis of carotid artery disease - comparison between directional Doppler, duplex scanner and angiography. Acta Neurologica Scandinavica, 65, 335-346.