- Email: [email protected]
Noninvasive assessment of normal carotid bifurcation hemodynamics with color-flow ultrasound imaging
Ultrasound in Med & Biol Vol, 13, No. 8, pp. 471-476, 1987 Pnnted in the U.S.A.
0301-5629/87 $3.00 + .00 © 1987 Pergamon Journals Ltd.
OOriginal Contribution NONINVASIVE A S S E S S M E N T OF N O R M A L CAROTID BIFURCATION H E M O D Y N A M I C S WITH COLOR-FLOW U L T R A S O U N D IMAGING R . E U G E N E ZIERLER, D A V I D J. PHILLIPS, K I R K W . BEACH, JEAN F. P R I M O Z I C H and D. E. STRANDNESS, JR. Division of Vascular Surgery, Department of Surgery RF-25, University of Washington School of Medicine, Seattle, WA 98195, U.S.A. (Received in first and final form 24 March 1987) Abstract--The combination of a B-mode imaging system and a single range-gate pulsed Doppler flow velocity detector (duplex scanner) has become the standard noninvasive method for assessing the extracranial carotid artery. However, a significant limitation of this approach is the sniall area of vessel lumen that can be evaluated at any one time. This report describes a new duplex instrument that displays blood flow as colors superimposed on a real-time B-mode image. Returning echoes from a linear array of transducers are continuously processed for amplitude and phase. Changes in phase are produced by tissue motion and are used to calculate Doppler shift frequency. This results in a color assignment: red and blue indicate direction of flow with respect to the ultrasound beam, and lighter shades represent higher velocities. The carotid bifurcations of 10 normal subjects were studied. Changes in flow velocities across the arterial lumen were clearly visualized as varying shades of red or blue during the cardiac cycle. A region of flow separation was observed in all proximal internal carotids as a blue area located along the outer wall of the bulb. Thus, it is possible to detect the localized flow patterns that characterize normal carotid arteries. Other advantages of color-flow imaging include the ability to rapidly identify the carotid bifurcation branches and any associated anatomic variations.
Key Words: Duplex scanning, Carotid artery, Flow separation, Pulsed Doppler, Spectrum analysis, Linear array.
1985). When duplex scanning was initially used to evaluate carotid arteries the specificity of the test, or the ability to identify normal vessels, was very low (Fell et al., 1981). This was because the normal complex flow disturbances in the carotid bulb were being interpreted as abnormal. It was not until the work of Phillips et al. (1983) that the nature of normal carotid bifurcation flow was fully appreciated. Recognition of these complex normal flow patterns in the carotid bulb and increased reliance on the c o m m o n carotid flow pattern improved the specificity to 84% (Roederer et al., 1982); however, many who use duplex scanning have difficulty identifying these complicated three-dimensional flow patterns. Most duplex instruments employ a pulsed Doppler with a single range-gate that can be positioned at any site in the B-mode image. While this facilitates selective sampling of flow patterns, the region of the vessel lumen that can be assessed at any one time is relatively small. Thus, flow disturbances that are limited to a short arterial segment or a discrete section of the flow stream may be overlooked. This report describes a new duplex approach that presents a real-time color
INTRODUCTION
In the past decade, ultrasonic duplex scanning, which c o m b i n e s real-time B - m o d e imaging and pulsed Doppler flow velocity detection with spectral waveform analysis, has become the standard method for noninvasive evaluation of carotid bifurcation disease (Strandness, 1986). This combination of methods has been valuable for both clinical diagnosis and research. A d v a n c e s in u l t r a s o u n d h a r d w a r e and Doppler signal processing techniques have continuously improved the diagnostic accuracy of duplex scanning when compared to contrast arteriography (Langlois et al., 1983). Experience with the duplex approach has also provided a better understanding of the complex nature of flow in the carotid bulb. For example, it is now recognized that a region of boundary layer separation is present in normal carotid bifurcations (Ku et al., Address correspondenceto: R. EugeneZierler, M.D., Department of Surgery (112), Seattle VA Medical Center, 1660 South Columbian Way, Seattle, Washington 98108. 471
472
Ultrasound in Medicine and Biology
image of blood flow within a two-dimensional Bmode image of tissue. Since this method displays a component of the speed and direction of blood flow in color, it is possible to demonstrate many of the complex flow patterns in the carotid bifurcation with a degree of spatial and temporal detail not available by other noninvasive or invasive methods. In order to establish a standard examination technique, 10 subjects with normal carotid bifurcations were studied. METHODS
Instrumentation Examinations were performed with a real-time B-mode and color Doppler duplex scanner, t The scan head for this instrument contains a linear array of ultrasound transducers approximately 3 cm long that operate at 7.5 MHz for both B-mode imaging and Doppler functions. This 3-cm configuration permits easy access to the cervical carotid arteries and provides a rectangular scanning field 3 cm wide and either 4 cm or 6.7 cm deep. Electronic phasing of the array elements produces dynamic continuous focus over the scanning field to provide both improved lateral resolution and smaller Doppler sample volumes at all depths in the image. The system permits rapid visualization of arterial and venous flow patterns by superimposing a real-time color image showing regions of blood flow on a two-dimensional B-mode tissue image. This combination is made possible by high speed, electronic array processing of ultrasound signals. Returning echoes from the scan field are continuously processed for amplitude and phase. Changes in phase are used to calculate a Doppler shift frequency that provides information on the speed and direction of blood flow at a particular site in tissue. Echoes from stationary or very slow moving tissue interfaces are used to create the B-mode image. The real-time m onoc hr om e B-mode image is generated in the conventional way by amplitude detection of the received echoes. For each transmitted ultrasound pulse, the echo amplitude is sampled once every 0.26 us to give an axial depth resolution of approximately 0.6 mm at 40 dB. A gray-scale value related to this amplitude is then used to construct the tissue image. The color-flow image is produced by analyzing the phase changes between echoes from each ultrasound scan line. To generate one scan line in the color image, a series of several pulse-echo cycles are t AngioDynograph 1 Vascular Imaging System, Quantum Medical Systems, Inc., Issaquah, Washington.
August 1987, Volume 13, Number 8
required. For each depth along the scan line, the phase change of the echo from pulse to pulse is measured. The changes in phase between successive echoes from the same depth in tissue are then used to calculate a Doppler frequency shift. The axial depth resolution of the Doppler system is the same as that for the B-mode image. When a Doppler shift is detected, a color is assigned to the corresponding depth along the scan line. The hue and intensity of the color are determined by the direction and magnitude of the Doppler frequency shift. Shades of red and blue are used to show the direction of flow approaching or receding from the transducer; however, the relationship between the red or blue color and a particular flow direction can be specified by the examiner. The video display of the instrument has been set up so the subject's head is on the left and the primary direction of arterial flow is from fight to left. For the examinations described in this report, red generally represents flow away from the heart and blue represents flow toward the heart. Increases in Doppler shift frequency produce discrete changes in color desaturation for both directions of flow: lighter shades indicate higher flow velocities with the maximum velocities shown as white. The B-mode and Doppler information within the scan field is updated 18 times per second for the shortest field of view. This results in a real-time display of the flow patterns in the twodimensional tissue image. In addition to the composite B-mode and colorflow image, a single range-gate pulsed Doppler sample volume can be used to evaluate the flow pattern at selected arterial sites. This Doppler signal is processed by a real-time spectral waveform analyzer and displayed with time on the horizontal axis, frequency or velocity on the vertical axis, and amplitude represented by a gray-scale. The angle between the ultrasound beam and the direction of blood flow can be determined for each selected sample site by placement of a cursor and used to scale the velocity according to the Doppler equation. In these calculations it is generally assumed that blood flow is parallel to the vessel walls. For the examinations reported here, this Doppler angle correction was only available for signals obtained with the single sample volume. The instrument has now been modified to allow Doppler angle adjustment for the color-flow image.
Examination technique The carotid bifurcations of 10 presumed normal subjects were examined. Subjects ranged in age from 23 to 41 and were without signs or symptoms of cerebrovascular disease. An 18 degree wedge-shaped standoff was attached to the transducer array to as-
Normal carotid bifurcation • R. E. ZIERLERel al.
473
Center Stream
Systole
Adjacent to Wall
Diastole
Fig. 1. Common carotid artery color-flow images in systole and diastole with pulsed Doppler spectral waveforms obtained in the center stream and adjacent to the arterial wall. Center stream flow is forward throughout the cardiac cycle; a brief period of flow reversal is seen in diastole near the arterial wall. Arrows indicate location of Doppler sample volume (circle) and corresponding spectral waveform features.
Fig. 2. Color-flow image of a carotid bifurcation at peak systole. A separation zone (blue) is present along the outer wall of the bulb. Forward flow (red and white) is seen along the flow divider. The Doppler sample volume (circle) is positioned near the flow divider, and the corresponding spectral waveform shows forward flow throughout the cardiac cycle.
Fig. 3. Color-flow image similar to that shown in Fig. 2, but with Doppler sample volume (circle) positioned in the separation zone. Spectral waveform shows forward and reverse flow during systole with minimal flow during diastole.
474
Ultrasound in Medicineand Biology
August 1987, Volume 13, Number 8
Fig. 4. Color-flow image of a carotid bifurcation in diastole. Separation zone now appears predominantly black because the very low flow velocities do not register in the color image.
Fig. 5. Color-flow image of a carotid bifurcation showing flow separation (blue or black) at the origins of the internal and external carotid arteries.
sure a favorable Doppler angle for vessels parallel to the skin surface. The scan head was coupled to the skin with acoustic gel. Scans were performed with the subjects supine and the head rotated away from the side being examined. The carotid system was imaged in the longitudinal plane from the proximal c o m m o n carotid artery just above the clavicle to the distal internal carotid above the angle of the mandible. Particular care was taken to visualize the flow pattern in the carotid bulb or the proximal segment of the internal carotid artery. The external carotid and its flow pattern were also identified. Each examination was recorded on videotape and 35-mm slide film. A complete examination of both carotid bifurcations required approximately 30 minutes.
tern was characterized by a more uniform distribution of colors indicating a blunt or plug-like flow profile. A feature of flow in the c o m m o n carotid artery that is not widely recognized is the presence o f a small, transient region of reverse flow near the vessel wall occurring in late systole or early diastole. This was observed bilaterally in five of the 10 subjects as a narrow blue area along the vessel wall, usually in the proximal segment of the c o m m o n carotid. Figure 1 illustrates these normal c o m m o n carotid color-flow patterns. Typical examples of systolic and diastolic flow patterns in the proximal internal carotid artery are shown in Figs. 2, 3 and 4. The normal internal carotid origin is characterized by a bulb of variable size (Bharadvaj et al., 1982). A zone of boundary layer separation was present in the proximal internal carotid arteries of all l0 subjects. This zone was located along the outer wall of the bulb on the side opposite the flow divider, as predicted by model studies (Ku and Giddens, 1983; LoGerfo et al., 1985) and clinical pulsed Doppler studies (Phillips et al., 1983; Ku et al., 1985). The length and width of this separation zone fluctuated during the cardiac cycle but were largest around peak systole. Figures 2 and 3 show the separa-
RESULTS The common, internal, and external carotid arteries were readily identified in all subjects. During systole the flow velocity in the c o m m o n carotid was higher in the center of the stream than near the vessel wall, as indicated by the predominance of white in the center of the vessel lumen. The diastolic flow pat-
Normal carotid bifurcation • R. E. ZIERLERet al. tion zone during systole as an oval-shaped blue area. Forward flow along the inner wall of the bulb adjacent to the flow divider was clearly visualized by shades of red. During diastole a dark space often appeared in the bulb between the forward m o v i n g stream and the outer wall, indicating extremely slow or stationary flow in the separation zone (Fig. 4). These zones of flow separation were limited to the proximal internal carotid or bulb segment and were not observed in the distal internal carotid artery. A smaller zone of flow separation was noted near the origins of six external carotid arteries in four subjects (Fig. 5). The single range-gate pulsed Doppler was used to evaluate spectral waveforms at selected sites within the color-flow image. Center stream spectral waveforms were normal according to the criteria based on a 5 M H z transmitting frequency and a 60 degree Doppler angle described by Fell et al. (1981), with peak velocities less than 120 cm/sec and minimal spectral broadening. Typical c o m m o n carotid artery spectral waveforms are shown in Fig. 1. In the center stream o f the c o m m o n carotid, forward flow is present t h r o u g h o u t the cardiac cycle, and the Doppler shift frequencies or velocities remain above the baseline. However, a brief period of reverse flow in diastole is indicated by the presence of frequencies below the baseline in waveforms taken near the vessel wall. Spectral waveforms from the separation zone in the bulb were characterized by a prominent period of flow reversal during systole (Fig. 3), while spectral waveforms obtained near the flow divider showed forward flow throughout systole and diastole (Fig. 2).
DISCUSSION The distinctive feature of color-flow ultrasound imaging is the ability to display in real-time one vector c o m p o n e n t o f the flow velocities within a two-dimensional tissue image. This composite image facilitates visualization of the complex flow patterns at the carotid bifurcation. Although it is possible to characterize these flow patterns with a conventional duplex system using a single range-gate pulsed Doppler, such an approach requires more time and experience on the part of the examiner. With the color-flow instrument, the area o f boundary layer separation near the outer wall of the normal carotid bulb described by Phillips et al. (1983) is easily recognized. Both the spatial and temporal characteristics of the velocity distribution in this area can now be visualized in real-time.
475
In addition to the area of flow separation in the bulb, color-flow imaging revealed several other important features of normal carotid flow. Half of the subjects studied showed flow reversal in their common carotid arteries. The area of reverse flow was located near the vessel wall and present only briefly in late systole or early diastole. This p h e n o m e n o n is a consequence of the flow deceleration that occurs uniformly across the carotid lumen, slowing the higher velocities in the center stream and reversing the lower velocities within the boundary layer adjacent to the vessel wall. The presence of flow separation near the origin of the external carotid artery represents another part of the complex secondary flow patterns in the carotid bifurcation. Although the clinical value of color-flow imaging has not been established, it offers several potential advantages over conventional duplex systems. The color-flow image facilitates identification of the carotid bifurcation branches. This could be particularly helpful when vascular abnormalities such as kinks, coils, or occluded segments are present. Once the carotid branches have been imaged, any localized areas with increased flow velocities should be apparent as lighter shades of red or white. These areas can then be evaluated in detail with pulsed Doppler and spectral waveform analysis. Thus, it is possible that color-flow imaging will reduce the need for evaluating flow patterns at multiple sites with a single pulsed Doppler sample volume. This may shorten the time required for an examination and make it easier to identify flow disturbances that are limited to short arterial segments. Visualization of the interface between the flowing blood and the arterial wall might also enable direct measurement of lumen size from the B-mode image. One of the sources for error in color-flow imaging is the dependence of Doppler frequency shift on the angle between the ultrasound beam and the direction of blood flow. If it is assumed that the angle is constant, changes in this angle along the course of an imaged artery will result in different color assignments, even if the flow velocity does not change. This situation may occur with tortuous vessels and at origins of arterial branches. A new version of thi: instrument allows the examiner to correct for Doppler angle in the color-flow image. Different color-flow imaging systems are certain to utilize different color schemes for displaying flow velocities. Thus, direct comparisons between images from various instruments will be difficult. While duplex scanning currently has a sensitivity of 99%, the specificity is only 84% when compared to contrast arteriography (Roederer et al., 1982). Distin-
476
Ultrasound in Medicine and Biology
guishing between normal carotid arteries and those with minimal disease is particularly difficult and remains one of the major sources of variability in duplex scanning (Kohler et al., 1985). The presence of flow separation in the carotid bulb has been associated with a normal appearance of this segment on arteriography. When the bulb is filled with atherosclerotic plaque, flow separation does not occur at that site. Areas of flow reversal or separation are easily recognized in a color-flow image by the presence of red and blue shades in adjacent sections of the vessel lumen; the blue separation zone in the carotid bulb characteristically changes size and shape during the cardiac cycle, while the red portion of the stream indicates continuous forward flow. Since these features are readily apparent on the color-flow image, this method should improve the specificity of duplex scanning. Work is currently in progress to define specific criteria for interpretation of color-flow images. Preliminary experience indicates that spectral waveform features such as spectral broadening are associated with distinct changes in color patterns. Since this instrument also functions as a conventional duplex scanner, the previously established criteria for interpreting spectral waveforms can continue to be used for diagnostic purposes if center stream pulsed Doppler signals are obtained at an angle of 60 degrees. It is hoped that the addition of color-flow imaging will enhance the diagnostic value of noninvasive testing and make the details of carotid flow dynamics available to those technologists and physi-
August 1987, Volume 13, Number 8
cians who have been reluctant to rely on spectral waveform analysis alone. REFERENCES Bharadvaj B. K., Mabon R. F. and Giddens D. P. (1982) Steady flow in a model of the human carotid bifurcation. Part I--Flow visualization. ,L Biomechanics 15, 349-362. Fell G., Phillips D. J., Chikos P. M., Harley J. D., Thiele B. L. and Strandness D. E.. Jr. (1981) Ultrasonic duplex scanning for disease of the carotid artery. Circulation 64, 1191-1195. Kohler T., Langlois Y., Roederer G. O., Phillips D. J., Beach K. W., Primozich J., Lawrence R., Nicholls S. C. and Strandness D. E., Jr. (1985) Sources of variability in carotid duplex examination: a prospective study. Ultrasound in Med. & Biol. 11, 571-576. Ku D. N. and Giddens D. P. (1983) Pulsatile flow in a model carotid bifurcation. Arteriosclerosis 3, 31-39. Ku D. N.. Giddens D. P., Phillips D. J. and Strandness D. E., Jr. (1985) Hemodynamics of the normal human carotid bifurcation: in vitro and in vivo studies. Ultrasound in Med. & Biol. 11, 13-26. Langlois Y., Roederer G. O., Chan A., Phillips D. J., Beach K. W., Martin D., Chikos P. M. and Strandness D. E., Jr. (1983) Evaluating carotid artery disease: the concordance between pulsed Doppler/spectrum analysis and angiography. Ultrasound in Meal. & Biol. 9, 51-63. LoGerfo F. W.. Nowak M. D. and Quist W. C. (1985) Structural details of boundary layer separation in a model human carotid bifurcation under steady and pulsatile flow conditions. J. Vase. Surg. 2, 263-269. Phillips D. J., Greene F. M., Langlois Y., Roederer G. O. and Strandness D. E., Jr. (1983) Flow velocity patterns in the carotid bifurcations of young, presumed normal subjects. Ultrasound in Med.& Biol. 9, 39-49. Roederer G. O., Langlois Y. E., Chan A., Primozich J., Lawrence R. J., Chikos P. M. and Strandness D. E., Jr. (1982) Ultrasonic duplex scanning of extracranial carotid arteries: improved accuracy using new features from the common carotid artery. J. Cardiovasc. Ultrason. 1,373-379. Strandness D. E., Jr. (1986) Ultrasound in the study of atherosclerosis. Ultrasound in Med. & Biol. 12, 453-464.
0301-5629/87 $3.00 + .00 © 1987 Pergamon Journals Ltd.
OOriginal Contribution NONINVASIVE A S S E S S M E N T OF N O R M A L CAROTID BIFURCATION H E M O D Y N A M I C S WITH COLOR-FLOW U L T R A S O U N D IMAGING R . E U G E N E ZIERLER, D A V I D J. PHILLIPS, K I R K W . BEACH, JEAN F. P R I M O Z I C H and D. E. STRANDNESS, JR. Division of Vascular Surgery, Department of Surgery RF-25, University of Washington School of Medicine, Seattle, WA 98195, U.S.A. (Received in first and final form 24 March 1987) Abstract--The combination of a B-mode imaging system and a single range-gate pulsed Doppler flow velocity detector (duplex scanner) has become the standard noninvasive method for assessing the extracranial carotid artery. However, a significant limitation of this approach is the sniall area of vessel lumen that can be evaluated at any one time. This report describes a new duplex instrument that displays blood flow as colors superimposed on a real-time B-mode image. Returning echoes from a linear array of transducers are continuously processed for amplitude and phase. Changes in phase are produced by tissue motion and are used to calculate Doppler shift frequency. This results in a color assignment: red and blue indicate direction of flow with respect to the ultrasound beam, and lighter shades represent higher velocities. The carotid bifurcations of 10 normal subjects were studied. Changes in flow velocities across the arterial lumen were clearly visualized as varying shades of red or blue during the cardiac cycle. A region of flow separation was observed in all proximal internal carotids as a blue area located along the outer wall of the bulb. Thus, it is possible to detect the localized flow patterns that characterize normal carotid arteries. Other advantages of color-flow imaging include the ability to rapidly identify the carotid bifurcation branches and any associated anatomic variations.
Key Words: Duplex scanning, Carotid artery, Flow separation, Pulsed Doppler, Spectrum analysis, Linear array.
1985). When duplex scanning was initially used to evaluate carotid arteries the specificity of the test, or the ability to identify normal vessels, was very low (Fell et al., 1981). This was because the normal complex flow disturbances in the carotid bulb were being interpreted as abnormal. It was not until the work of Phillips et al. (1983) that the nature of normal carotid bifurcation flow was fully appreciated. Recognition of these complex normal flow patterns in the carotid bulb and increased reliance on the c o m m o n carotid flow pattern improved the specificity to 84% (Roederer et al., 1982); however, many who use duplex scanning have difficulty identifying these complicated three-dimensional flow patterns. Most duplex instruments employ a pulsed Doppler with a single range-gate that can be positioned at any site in the B-mode image. While this facilitates selective sampling of flow patterns, the region of the vessel lumen that can be assessed at any one time is relatively small. Thus, flow disturbances that are limited to a short arterial segment or a discrete section of the flow stream may be overlooked. This report describes a new duplex approach that presents a real-time color
INTRODUCTION
In the past decade, ultrasonic duplex scanning, which c o m b i n e s real-time B - m o d e imaging and pulsed Doppler flow velocity detection with spectral waveform analysis, has become the standard method for noninvasive evaluation of carotid bifurcation disease (Strandness, 1986). This combination of methods has been valuable for both clinical diagnosis and research. A d v a n c e s in u l t r a s o u n d h a r d w a r e and Doppler signal processing techniques have continuously improved the diagnostic accuracy of duplex scanning when compared to contrast arteriography (Langlois et al., 1983). Experience with the duplex approach has also provided a better understanding of the complex nature of flow in the carotid bulb. For example, it is now recognized that a region of boundary layer separation is present in normal carotid bifurcations (Ku et al., Address correspondenceto: R. EugeneZierler, M.D., Department of Surgery (112), Seattle VA Medical Center, 1660 South Columbian Way, Seattle, Washington 98108. 471
472
Ultrasound in Medicine and Biology
image of blood flow within a two-dimensional Bmode image of tissue. Since this method displays a component of the speed and direction of blood flow in color, it is possible to demonstrate many of the complex flow patterns in the carotid bifurcation with a degree of spatial and temporal detail not available by other noninvasive or invasive methods. In order to establish a standard examination technique, 10 subjects with normal carotid bifurcations were studied. METHODS
Instrumentation Examinations were performed with a real-time B-mode and color Doppler duplex scanner, t The scan head for this instrument contains a linear array of ultrasound transducers approximately 3 cm long that operate at 7.5 MHz for both B-mode imaging and Doppler functions. This 3-cm configuration permits easy access to the cervical carotid arteries and provides a rectangular scanning field 3 cm wide and either 4 cm or 6.7 cm deep. Electronic phasing of the array elements produces dynamic continuous focus over the scanning field to provide both improved lateral resolution and smaller Doppler sample volumes at all depths in the image. The system permits rapid visualization of arterial and venous flow patterns by superimposing a real-time color image showing regions of blood flow on a two-dimensional B-mode tissue image. This combination is made possible by high speed, electronic array processing of ultrasound signals. Returning echoes from the scan field are continuously processed for amplitude and phase. Changes in phase are used to calculate a Doppler shift frequency that provides information on the speed and direction of blood flow at a particular site in tissue. Echoes from stationary or very slow moving tissue interfaces are used to create the B-mode image. The real-time m onoc hr om e B-mode image is generated in the conventional way by amplitude detection of the received echoes. For each transmitted ultrasound pulse, the echo amplitude is sampled once every 0.26 us to give an axial depth resolution of approximately 0.6 mm at 40 dB. A gray-scale value related to this amplitude is then used to construct the tissue image. The color-flow image is produced by analyzing the phase changes between echoes from each ultrasound scan line. To generate one scan line in the color image, a series of several pulse-echo cycles are t AngioDynograph 1 Vascular Imaging System, Quantum Medical Systems, Inc., Issaquah, Washington.
August 1987, Volume 13, Number 8
required. For each depth along the scan line, the phase change of the echo from pulse to pulse is measured. The changes in phase between successive echoes from the same depth in tissue are then used to calculate a Doppler frequency shift. The axial depth resolution of the Doppler system is the same as that for the B-mode image. When a Doppler shift is detected, a color is assigned to the corresponding depth along the scan line. The hue and intensity of the color are determined by the direction and magnitude of the Doppler frequency shift. Shades of red and blue are used to show the direction of flow approaching or receding from the transducer; however, the relationship between the red or blue color and a particular flow direction can be specified by the examiner. The video display of the instrument has been set up so the subject's head is on the left and the primary direction of arterial flow is from fight to left. For the examinations described in this report, red generally represents flow away from the heart and blue represents flow toward the heart. Increases in Doppler shift frequency produce discrete changes in color desaturation for both directions of flow: lighter shades indicate higher flow velocities with the maximum velocities shown as white. The B-mode and Doppler information within the scan field is updated 18 times per second for the shortest field of view. This results in a real-time display of the flow patterns in the twodimensional tissue image. In addition to the composite B-mode and colorflow image, a single range-gate pulsed Doppler sample volume can be used to evaluate the flow pattern at selected arterial sites. This Doppler signal is processed by a real-time spectral waveform analyzer and displayed with time on the horizontal axis, frequency or velocity on the vertical axis, and amplitude represented by a gray-scale. The angle between the ultrasound beam and the direction of blood flow can be determined for each selected sample site by placement of a cursor and used to scale the velocity according to the Doppler equation. In these calculations it is generally assumed that blood flow is parallel to the vessel walls. For the examinations reported here, this Doppler angle correction was only available for signals obtained with the single sample volume. The instrument has now been modified to allow Doppler angle adjustment for the color-flow image.
Examination technique The carotid bifurcations of 10 presumed normal subjects were examined. Subjects ranged in age from 23 to 41 and were without signs or symptoms of cerebrovascular disease. An 18 degree wedge-shaped standoff was attached to the transducer array to as-
Normal carotid bifurcation • R. E. ZIERLERel al.
473
Center Stream
Systole
Adjacent to Wall
Diastole
Fig. 1. Common carotid artery color-flow images in systole and diastole with pulsed Doppler spectral waveforms obtained in the center stream and adjacent to the arterial wall. Center stream flow is forward throughout the cardiac cycle; a brief period of flow reversal is seen in diastole near the arterial wall. Arrows indicate location of Doppler sample volume (circle) and corresponding spectral waveform features.
Fig. 2. Color-flow image of a carotid bifurcation at peak systole. A separation zone (blue) is present along the outer wall of the bulb. Forward flow (red and white) is seen along the flow divider. The Doppler sample volume (circle) is positioned near the flow divider, and the corresponding spectral waveform shows forward flow throughout the cardiac cycle.
Fig. 3. Color-flow image similar to that shown in Fig. 2, but with Doppler sample volume (circle) positioned in the separation zone. Spectral waveform shows forward and reverse flow during systole with minimal flow during diastole.
474
Ultrasound in Medicineand Biology
August 1987, Volume 13, Number 8
Fig. 4. Color-flow image of a carotid bifurcation in diastole. Separation zone now appears predominantly black because the very low flow velocities do not register in the color image.
Fig. 5. Color-flow image of a carotid bifurcation showing flow separation (blue or black) at the origins of the internal and external carotid arteries.
sure a favorable Doppler angle for vessels parallel to the skin surface. The scan head was coupled to the skin with acoustic gel. Scans were performed with the subjects supine and the head rotated away from the side being examined. The carotid system was imaged in the longitudinal plane from the proximal c o m m o n carotid artery just above the clavicle to the distal internal carotid above the angle of the mandible. Particular care was taken to visualize the flow pattern in the carotid bulb or the proximal segment of the internal carotid artery. The external carotid and its flow pattern were also identified. Each examination was recorded on videotape and 35-mm slide film. A complete examination of both carotid bifurcations required approximately 30 minutes.
tern was characterized by a more uniform distribution of colors indicating a blunt or plug-like flow profile. A feature of flow in the c o m m o n carotid artery that is not widely recognized is the presence o f a small, transient region of reverse flow near the vessel wall occurring in late systole or early diastole. This was observed bilaterally in five of the 10 subjects as a narrow blue area along the vessel wall, usually in the proximal segment of the c o m m o n carotid. Figure 1 illustrates these normal c o m m o n carotid color-flow patterns. Typical examples of systolic and diastolic flow patterns in the proximal internal carotid artery are shown in Figs. 2, 3 and 4. The normal internal carotid origin is characterized by a bulb of variable size (Bharadvaj et al., 1982). A zone of boundary layer separation was present in the proximal internal carotid arteries of all l0 subjects. This zone was located along the outer wall of the bulb on the side opposite the flow divider, as predicted by model studies (Ku and Giddens, 1983; LoGerfo et al., 1985) and clinical pulsed Doppler studies (Phillips et al., 1983; Ku et al., 1985). The length and width of this separation zone fluctuated during the cardiac cycle but were largest around peak systole. Figures 2 and 3 show the separa-
RESULTS The common, internal, and external carotid arteries were readily identified in all subjects. During systole the flow velocity in the c o m m o n carotid was higher in the center of the stream than near the vessel wall, as indicated by the predominance of white in the center of the vessel lumen. The diastolic flow pat-
Normal carotid bifurcation • R. E. ZIERLERet al. tion zone during systole as an oval-shaped blue area. Forward flow along the inner wall of the bulb adjacent to the flow divider was clearly visualized by shades of red. During diastole a dark space often appeared in the bulb between the forward m o v i n g stream and the outer wall, indicating extremely slow or stationary flow in the separation zone (Fig. 4). These zones of flow separation were limited to the proximal internal carotid or bulb segment and were not observed in the distal internal carotid artery. A smaller zone of flow separation was noted near the origins of six external carotid arteries in four subjects (Fig. 5). The single range-gate pulsed Doppler was used to evaluate spectral waveforms at selected sites within the color-flow image. Center stream spectral waveforms were normal according to the criteria based on a 5 M H z transmitting frequency and a 60 degree Doppler angle described by Fell et al. (1981), with peak velocities less than 120 cm/sec and minimal spectral broadening. Typical c o m m o n carotid artery spectral waveforms are shown in Fig. 1. In the center stream o f the c o m m o n carotid, forward flow is present t h r o u g h o u t the cardiac cycle, and the Doppler shift frequencies or velocities remain above the baseline. However, a brief period of reverse flow in diastole is indicated by the presence of frequencies below the baseline in waveforms taken near the vessel wall. Spectral waveforms from the separation zone in the bulb were characterized by a prominent period of flow reversal during systole (Fig. 3), while spectral waveforms obtained near the flow divider showed forward flow throughout systole and diastole (Fig. 2).
DISCUSSION The distinctive feature of color-flow ultrasound imaging is the ability to display in real-time one vector c o m p o n e n t o f the flow velocities within a two-dimensional tissue image. This composite image facilitates visualization of the complex flow patterns at the carotid bifurcation. Although it is possible to characterize these flow patterns with a conventional duplex system using a single range-gate pulsed Doppler, such an approach requires more time and experience on the part of the examiner. With the color-flow instrument, the area o f boundary layer separation near the outer wall of the normal carotid bulb described by Phillips et al. (1983) is easily recognized. Both the spatial and temporal characteristics of the velocity distribution in this area can now be visualized in real-time.
475
In addition to the area of flow separation in the bulb, color-flow imaging revealed several other important features of normal carotid flow. Half of the subjects studied showed flow reversal in their common carotid arteries. The area of reverse flow was located near the vessel wall and present only briefly in late systole or early diastole. This p h e n o m e n o n is a consequence of the flow deceleration that occurs uniformly across the carotid lumen, slowing the higher velocities in the center stream and reversing the lower velocities within the boundary layer adjacent to the vessel wall. The presence of flow separation near the origin of the external carotid artery represents another part of the complex secondary flow patterns in the carotid bifurcation. Although the clinical value of color-flow imaging has not been established, it offers several potential advantages over conventional duplex systems. The color-flow image facilitates identification of the carotid bifurcation branches. This could be particularly helpful when vascular abnormalities such as kinks, coils, or occluded segments are present. Once the carotid branches have been imaged, any localized areas with increased flow velocities should be apparent as lighter shades of red or white. These areas can then be evaluated in detail with pulsed Doppler and spectral waveform analysis. Thus, it is possible that color-flow imaging will reduce the need for evaluating flow patterns at multiple sites with a single pulsed Doppler sample volume. This may shorten the time required for an examination and make it easier to identify flow disturbances that are limited to short arterial segments. Visualization of the interface between the flowing blood and the arterial wall might also enable direct measurement of lumen size from the B-mode image. One of the sources for error in color-flow imaging is the dependence of Doppler frequency shift on the angle between the ultrasound beam and the direction of blood flow. If it is assumed that the angle is constant, changes in this angle along the course of an imaged artery will result in different color assignments, even if the flow velocity does not change. This situation may occur with tortuous vessels and at origins of arterial branches. A new version of thi: instrument allows the examiner to correct for Doppler angle in the color-flow image. Different color-flow imaging systems are certain to utilize different color schemes for displaying flow velocities. Thus, direct comparisons between images from various instruments will be difficult. While duplex scanning currently has a sensitivity of 99%, the specificity is only 84% when compared to contrast arteriography (Roederer et al., 1982). Distin-
476
Ultrasound in Medicine and Biology
guishing between normal carotid arteries and those with minimal disease is particularly difficult and remains one of the major sources of variability in duplex scanning (Kohler et al., 1985). The presence of flow separation in the carotid bulb has been associated with a normal appearance of this segment on arteriography. When the bulb is filled with atherosclerotic plaque, flow separation does not occur at that site. Areas of flow reversal or separation are easily recognized in a color-flow image by the presence of red and blue shades in adjacent sections of the vessel lumen; the blue separation zone in the carotid bulb characteristically changes size and shape during the cardiac cycle, while the red portion of the stream indicates continuous forward flow. Since these features are readily apparent on the color-flow image, this method should improve the specificity of duplex scanning. Work is currently in progress to define specific criteria for interpretation of color-flow images. Preliminary experience indicates that spectral waveform features such as spectral broadening are associated with distinct changes in color patterns. Since this instrument also functions as a conventional duplex scanner, the previously established criteria for interpreting spectral waveforms can continue to be used for diagnostic purposes if center stream pulsed Doppler signals are obtained at an angle of 60 degrees. It is hoped that the addition of color-flow imaging will enhance the diagnostic value of noninvasive testing and make the details of carotid flow dynamics available to those technologists and physi-
August 1987, Volume 13, Number 8
cians who have been reluctant to rely on spectral waveform analysis alone. REFERENCES Bharadvaj B. K., Mabon R. F. and Giddens D. P. (1982) Steady flow in a model of the human carotid bifurcation. Part I--Flow visualization. ,L Biomechanics 15, 349-362. Fell G., Phillips D. J., Chikos P. M., Harley J. D., Thiele B. L. and Strandness D. E.. Jr. (1981) Ultrasonic duplex scanning for disease of the carotid artery. Circulation 64, 1191-1195. Kohler T., Langlois Y., Roederer G. O., Phillips D. J., Beach K. W., Primozich J., Lawrence R., Nicholls S. C. and Strandness D. E., Jr. (1985) Sources of variability in carotid duplex examination: a prospective study. Ultrasound in Med. & Biol. 11, 571-576. Ku D. N. and Giddens D. P. (1983) Pulsatile flow in a model carotid bifurcation. Arteriosclerosis 3, 31-39. Ku D. N.. Giddens D. P., Phillips D. J. and Strandness D. E., Jr. (1985) Hemodynamics of the normal human carotid bifurcation: in vitro and in vivo studies. Ultrasound in Med. & Biol. 11, 13-26. Langlois Y., Roederer G. O., Chan A., Phillips D. J., Beach K. W., Martin D., Chikos P. M. and Strandness D. E., Jr. (1983) Evaluating carotid artery disease: the concordance between pulsed Doppler/spectrum analysis and angiography. Ultrasound in Meal. & Biol. 9, 51-63. LoGerfo F. W.. Nowak M. D. and Quist W. C. (1985) Structural details of boundary layer separation in a model human carotid bifurcation under steady and pulsatile flow conditions. J. Vase. Surg. 2, 263-269. Phillips D. J., Greene F. M., Langlois Y., Roederer G. O. and Strandness D. E., Jr. (1983) Flow velocity patterns in the carotid bifurcations of young, presumed normal subjects. Ultrasound in Med.& Biol. 9, 39-49. Roederer G. O., Langlois Y. E., Chan A., Primozich J., Lawrence R. J., Chikos P. M. and Strandness D. E., Jr. (1982) Ultrasonic duplex scanning of extracranial carotid arteries: improved accuracy using new features from the common carotid artery. J. Cardiovasc. Ultrason. 1,373-379. Strandness D. E., Jr. (1986) Ultrasound in the study of atherosclerosis. Ultrasound in Med. & Biol. 12, 453-464.