Optimizing Image Quality When Evaluating Blood Flow at Doppler US: A Tutorial

04 Nov.,2022

 

doppler ultrasound machine

Doppler US is an essential component of nearly all diagnostic US procedures. In this era of increased awareness of the effects of ionizing radiation and the side effects of iodine- and gadolinium-based contrast agents, Doppler US is poised to play an even bigger role in medical imaging. It is safe, cost-effective, portable, and highly accurate when performed by an experienced operator. The sensitivities and specificities of Doppler US for detecting blood flow and determining the direction and velocity of blood flow in various organs and vascular systems have increased dramatically in the past decade. With use of advanced flow techniques that are available for use with most modern equipment, US can provide vascular information that is comparable to or even more accurate than that obtained with other cross-sectional and interventional modalities. However, there remains concern that US (including newer more advanced flow-evaluating techniques) will not be used to its full potential owing to dependence on operator skill and expertise. Thorough understanding of image optimization techniques and expanded knowledge of the physical principles, instrumentation, application, advantages, and limitations of this modality are of utmost importance. The authors provide a simple practical guide for optimizing images for vascular flow detection by reviewing various cases and focusing on the parameters that should be optimized.

SA-CME LEARNING OBJECTIVES

After completing this journal-based SA-CME activity, participants will be able to:

  • ■ Explain the basic principles of Doppler US.

  • ■ Optimize Doppler US images to improve flow detection and avoid misinterpretation of findings.

  • ■ Use a practice-based approach to problem solving in Doppler US image optimization.

Introduction

Since its inception approximately 40 years ago, Doppler US has become an essential component of nearly all diagnostic US examinations. Despite the apparent benefits of this modality in the diagnosis of various diseases when it is performed by a knowledgeable and experienced operator, including cost-effectiveness, portability, safety, and high accuracy, Doppler US has overall been relatively underutilized in the field of radiology. The reasons for choosing more expensive and invasive modalities such as CT angiography, MR angiography, and conventional angiography are multifactorial but include the significant dependence on operator experience and the limited training that radiologists receive to master Doppler US skills, specifically with regard to image acquisition, parameter optimization, and findings interpretation.

In this article, by using a case-based approach, we describe elements that are essential for image optimization and provide a guide for avoiding common mistakes related to inadequate adjustment of Doppler parameters. We also briefly review Doppler-related artifacts, with emphasis on their cause and clinical relevance, and offer strategies for avoiding or, in some instances, enhancing them.

Basic US Principles

Doppler Frequency Shift

Medical Doppler US is based on the principle that a transmitted frequency from an ultrasound transducer undergoes a change when it becomes reflected by moving objects such as red blood cells (RBCs), and this change in frequency (ie, frequency shift) is proportional to the velocity of the moving object (Fig 1). Thus, the Doppler frequency shift, Δfd, is the difference between the received and transmitted frequencies, and it can be calculated by using the following formula (1):where c is the speed of sound, cos is cosine, v is the flow velocity, θ is the angle between the direction of blood flow and the axis of the ultrasound beam, ft is the transmitted frequency, and fr is the received frequency. Therefore, the change in Doppler frequency shift is proportional to the flow velocity.

Figure 1. Schematic diagram illustrates the Doppler frequency shift (Δf) principle and the acquisition of US data. The transducer transmits the ultrasound pulse at a transmitted frequency (ft) that is reflected from the moving RBCs within a vessel at a received frequency (fr). The difference between the transmitted frequency and the received frequency is the Doppler frequency shift. The angle between the direction of blood flow and the direction of the ultrasound beam is the Doppler angle (θ). The Doppler frequency shift equation enables one to calculate the velocity of moving blood. The amount of Doppler shift depends on the frequency of the transducer, frequency of the reflected echo, velocity of the moving reflector (v), angle between the transmitted beam and the flow direction (ie, Doppler angle), and velocity of sound in the reflecting medium (c). Note: The display of parameters on the image display is vendor specific, and one should refer to a vendor manual to ensure correct interpretation and identification of the settings.

Download as PowerPointOpen in Image Viewer

The received sound frequency can be higher or lower than the transmitted sound frequency, depending on whether the RBCs are moving toward the transducer (higher) or away from it (lower). With use of the Doppler frequency shift equation, the US system provides quantitative and directional estimates of flowing blood (Fig 1). In addition, this technique allows the creation of color-coded flow maps of the vasculature (2). Related terminology and definitions are provided in a glossary (Table 1).

Table 1: Glossary of Terms

Doppler US Modes

Three primary acquisition modes are used in Doppler US imaging: color Doppler imaging (CDI), spectral Doppler imaging (SDI), and power Doppler imaging (PDI). These modes are related, and each has advantages and limitations (Table 2). Only the mode that is active during the US examination can undergo optimization.

Table 2: Pros and Cons of Doppler US Modes

CDI provides information about the presence or absence of flow, mean flow velocity, and direction of flow within the selected color box (1). Note that no absolute value of velocity can be determined from color Doppler images. Flow-related information is superimposed on a gray-scale US image; this allows the examiner to know where the flow signals originate. Therefore, CDI has dual capability: for evaluation of tissue characteristics (on the gray-scale image) and assessment of vascular hemodynamics. With capability to display mean velocities, a CDI system can facilitate a quick assessment of blood flow in an area of interest and enables the detection of normal and abnormal flow patterns, which can be interrogated in more detail later at spectral Doppler analysis.

SDI differs from CDI in that information is not obtained from the entire color box but rather from a specified gate window, a 2–4-mm–wide sample volume, that typically is placed in the center of a vessel or in tissue identified at CDI. Similar to CDI, SDI yields information regarding flow direction. The main advantage of using SDI is that it enables one to determine the absolute velocity and the location (depth) from which the velocity is originating. Thus, it enables quantitative assessment of the velocities of blood flow within a sample volume and allows the examiner to analyze different features of the flow patterns—namely, the pulsatility and phasicity of vascular flow (Fig 2).

Figure 2. Spectral Doppler waveform of a common carotid artery (CCA). Only the spectral waveform is included in the image. Information regarding the waveform is obtained from a small gate placed in the center of the CCA. The waveform indicates the flow direction, velocities at a specific time, and amplitude within the sample volume. The velocities are proportional to the frequency shift, and each point on the waveform corresponds to a specific velocity. The slope of the curve represents acceleration, and inflection points correspond to changes in acceleration. An appropriate Doppler angle (≤60°) is essential. Spectral Doppler US yields information regarding the flow direction, absolute velocity of the moving blood, phasicity, flow hemodynamics, arterial resistance, and pulsatility. Inv = inversion.

Download as PowerPointOpen in Image Viewer

Analysis of the spectral tracing facilitates assessment of specific blood flow parameters, including acceleration or deceleration rate, resistive index, and peak systolic and end-diastolic velocities. These characteristics are of critical importance in the diagnosis of various vascular and organ diseases.

In contrast to CDI and SDI, which are based on the pulse-echo principle, PDI involves the use of a significantly different method for signal analysis. It does not yield absolute values of velocity; rather, it depicts the strength or amplitude of the Doppler signals detected from each location within a selected area. With PDI, a uniform background color is assigned to noise, and, thus, the dynamic range of the Doppler signal is extended. For this reason, PDI can reveal relatively low velocities and is more sensitive to low flow states than are CDI and SDI.

Another advantage of PDI is that it can enable a global assessment of organ and tissue perfusion (2,3). Conventional PDI does not yield information about flow direction; however, more advanced applications of this mode, such as directional PDI, allow determination of the flow direction. Some of the newer techniques for evaluating microvascular flow (discussed later) are based on fundamental PDI principles.

Image Optimization

Image optimization plays a crucial role in any successful Doppler US evaluation. A lack of understanding of image optimization is at the heart of inadequate US evaluations and finding misinterpretations. The Doppler parameters should be adjusted throughout the US examination, which is performed in real time during the evaluation of different areas of interest. The principles of Doppler US image optimization are universal to any organ, vessel, or other structure and can be applied to any part of the body or any vascular bed. The findings of poor or no flow, excessive flow signal, and/or abnormal appearance of waveforms should always cause the examiner to question whether the obtained findings are attributable to a true pathologic process or an artifact associated with suboptimal image optimization. In the sections that follow, we provide general guidelines for troubleshooting the most commonly encountered pitfalls of Doppler US and thus allow the examiner to have high confidence in the obtained information.

Poor Flow Detection

The main advantage of using Doppler imaging is the ability to detect vascular flow within a vessel, organ, or other structure of interest. When flow is not identified, it is important to be able to differentiate the true absence of vascular flow from suboptimal image optimization. A relatively simple algorithm can be used to quickly check all important parameters and ensure appropriate image settings, thus increasing confidence regarding the presence or absence of flow (Table 3) (Movie 1). Readers can refer to the appropriate vendor manual for specific lists of imaging parameters. A number of factors can cause poor flow detection at CDI and SDI.

Table 3: Troubleshooting Poor Visualization of Vascular Flow

Get the Flash Player to see this video.

Movie 1 Locations of different parameters on the color Doppler image display. CDI video of the proximal right main renal artery (RT MRA PROX) demonstrates different gray-scale and color Doppler settings (arrows) that can be determined from the image display. In the right upper corner, the chosen presettings for evaluation of the abdominal vasculature (ABD VASC) are displayed, with the transducer choice and frame rate listed below the application. In the left upper corner, the output power selected for the Doppler US evaluation is displayed. Note that the output energy differs, depending on the application for which it is used. For example, much less power is applied for an obstetric application. The range of the velocity scale can be determined from the color bar (top right). The frame rate is listed under the transducer choice, and the two-dimensional (2D) settings include gain and dynamic rate. CDI settings include gain, expressed as the saturation percentage, and wall filter (WF). PDI settings, although not shown in this video, also include gain and wall filter. The focal zone and depth (arrows) also are displayed on the image. CF = color flow.

Download Original Video (55.5 MB)

Poor or No Vascular Flow Detection

Color Gain Is Set Too Low.—The inability to detect flow at CDI can be due to low gain settings; a common example is apparent poor filling of the carotid artery lumen due to low gain settings. The gain defines the amount of amplification of echoes from the receiver. When the color gain setting is suboptimal, the returned signals are not sufficiently amplified and an otherwise patent vessel may not entirely fill with color (Movie 2). If the color gain setting is extremely low, the vessel lumen may not fill with color at all and the findings will be misinterpreted as absence of flow (1,4).

Get the Flash Player to see this video.

Movie 2 Poor flow detection in the left CCA due to suboptimal (low) color gain settings. The color gain indicates the amount of amplification of signals from the receiver. When the returned signals are not sufficiently amplified, the vessel of interest is not entirely filled with color. To improve blood flow detection, the gain should be increased to the approximately 50% saturation level. Note that the velocity scale and wall filter settings are well optimized. No specific percentage of gain is recommended; rather, the optimization is set on a case-by-case basis.

Download Original Video (28.2 MB)

To enhance the color information, it is essential to keep the gain setting on “gray-scale low” while performing CDI, as the color data will be suppressed if the gray-scale gain is too high. To optimize the gain setting, the gain should be increased so that the color on the image display becomes visible outside of the boundaries of the vessel wall. A subsequent decrease in gain until all of the color specks outside of the vascular wall are removed will ensure proper adjustment of this setting (Movie 3).

Get the Flash Player to see this video.

Movie 3 Optimization of the color Doppler gain. When the gain is set too high, color specks will be seen in and adjacent to the vessel lumen. This is caused by the amplification of low frequency shifts that are received from the minimal motion in the surrounding tissues. To optimize the images, the color gain should be decreased so that no specks are seen in the adjacent soft tissues.

Download Original Video (21.6 MB)

Wall Filter Is Set Too High.—At CDI, incomplete filling of a vessel can be seen with a high wall filter setting (Fig 3). Use of a high filter setting at SDI will result in the loss of diastolic flow in an arterial waveform or a floating appearance of the waveform above the baseline. This is seen most commonly while imaging the portal vein (5–7). In this case, the small frequency shifts corresponding to low velocities will be eliminated and not displayed on the CDI or SDI scan (Fig 4). Changing the filter setting to low will ensure the display of low-amplitude velocities and normalize the waveform appearance.

Figure 3. Doppler US image shows poor flow detection due to suboptimal wall filter setting. Selecting the medium wall filter setting, in this case 140 Hz (circled), when performing Doppler US assessment of the portal venous system causes the low-amplitude signals to be filtered out. As a result, the images will display no color filling of the inferior vena cava (IVC) lumen and poor color filling of the main portal vein (MPV). Also, flow appears to be absent in the hepatic veins (HV). The hepatic vasculature should always be evaluated with low wall filter settings. The higher the filter setting, the wider the color baseline band on the color bars (arrows). Here and in all subsequent figures and movies showing Doppler image displays, CF = color flow, FR = frequency, 2D = two dimensional, WF = wall filter.

Download as PowerPointOpen in Image Viewer

Figure 4a. Loss of slow vascular flow–related information due to a high wall filter setting. (a) Longitudinal SDI scan of the proximal abdominal aorta (Ao) shows nearly complete loss of diastolic flow at the high wall filter setting (180 Hz) (circled). Arrow points to the spectral waveform. (b) SDI scan of the portal vein obtained at a high wall filter setting (110 Hz) (circled) shows the lower-amplitude velocities filtered out, resulting in loss of spectral information immediately above the baseline. The spectral waveform (arrow) appears to float above the baseline. Optimizing the wall filter to a lower setting will ensure the display of lower-amplitude velocities. PW = pulsed-wave Doppler mode.

Download as PowerPointOpen in Image Viewer

Figure 4b. Loss of slow vascular flow–related information due to a high wall filter setting. (a) Longitudinal SDI scan of the proximal abdominal aorta (Ao) shows nearly complete loss of diastolic flow at the high wall filter setting (180 Hz) (circled). Arrow points to the spectral waveform. (b) SDI scan of the portal vein obtained at a high wall filter setting (110 Hz) (circled) shows the lower-amplitude velocities filtered out, resulting in loss of spectral information immediately above the baseline. The spectral waveform (arrow) appears to float above the baseline. Optimizing the wall filter to a lower setting will ensure the display of lower-amplitude velocities. PW = pulsed-wave Doppler mode.

Download as PowerPointOpen in Image Viewer

Filter settings are usually preset by the manufacturer, and a maximal, high, medium, or minimal filter setting can be applied to SDI, CDI, and PDI. Wall filters selectively filter out all frequency shifts that fall below a selected threshold, with the intent of eliminating the lowest Doppler shifts that usually result from vessel wall motion and motion in the surrounding solid tissues. These shifts are referred to as noise, clutter, or motion artifacts and are characterized by a low frequency and a high intensity and/or high amplitude (5,8). However, the US machine cannot distinguish between low-frequency Doppler shifts originating from slow-moving blood and those originating from tissue movement. Consequently, both of these low-frequency shifts will be removed when a high filter setting is selected (Fig 3) (9,10). To avoid a loss of signal related to slow flow, filter settings should be kept at the lowest possible setting.

With advances in modern equipment, failure to optimize the wall filter is less of an issue, as these units have an “auto-scan” control function that automatically adjusts settings (including wall filter) according to the selected application (ie, portal vein or abdominal aorta assessment).

Velocity Scale Is Set Too High.—Slow flow (ie, low PSV) may not be detected at CDI or SDI if the selected range of velocities (ie, velocity scale) is too high. For example, poor color filling of the carotid artery lumen may be seen if the selected velocity scale is too high (Fig 5). A high velocity scale signifies a high range of velocities depicted on the color Doppler or spectral Doppler scan and is usually selected when high-amplitude velocities are expected in the examined vessel or structure, as in cases of stenotic areas in a vessel or arteriovenous fistulas. With a high velocity scale setting, emphasis is placed on correctly displaying mean and absolute velocities in an area of increased flow. Low-amplitude velocities will not be displayed in this setting, as they will be filtered out by the US unit. This is because the unit will consider the low-amplitude velocities to be originating from the surrounding soft tissues and thus representative of noise.

Figure 5. Poor flow detection due to a suboptimally wide velocity scale. Sagittal CDI scan of the left common carotid artery (L CCA) obtained on a wide velocity scale (0–86 cm/sec) shows only partial filling of the artery lumen with color. This is due in part to the elimination of low frequency shifts, with automatic engagement of high filter settings. Changes in the range of velocities are linked to the PRF (yellow rectangles). Thus, the PRF increases when velocities increase, and it decreases when velocities decrease.

Download as PowerPointOpen in Image Viewer

In most Doppler US systems, the PRF or velocity scale control and the wall filter control are linked. For example, when the velocity scale or PRF is increased, the wall filter also will increase to the high setting, and, therefore, a higher range of relatively low frequency shifts will be filtered out and not displayed on the color or spectral image. To improve flow visualization, the examiner should decrease the velocity scale range or PRF (Movie 4). Understanding this phenomenon is essential during attempts to detect flow in, for example, a vein suspected of having a bland or tumor thrombus or in a suspicious mass. In such instances, the velocity scale must be decreased to as low as necessary (∼5–10 PRF) to have confidence that the flow within a lesion is truly absent (Movie 5).

Get the Flash Player to see this video.

Movie 4 Optimization of the color Doppler velocity scale. CDI cine clip of the carotid artery demonstrates laminar flow with an optimal velocity scale (0–28 cm/sec). When the range of velocities is increased to 86 cm/sec, the lumen of the carotid artery is only partially filled with color. When the range of velocities is decreased to 3.4 cm/sec, aliasing occurs, manifesting as a color mosaic. Note that changes in the range of velocities are linked to changes in PRFs.

Download Original Video (26.9 MB)Get the Flash Player to see this video.

Movie 5 Image optimization for evaluation of suspicious lesions and masses. CDI cine clips acquired through the pelvis in a 64-year-old man with metastatic testicular cancer demonstrate no significant vascular flow when the velocity scale and PRF are set high (55 cm/sec) and the gain is suboptimal (set at 42%). With a decrease in the velocity scale to 33 cm/sec, subtle vascularity becomes apparent in the mass. When the velocity scale is decreased to 7 cm/sec and the gain is increased to 56% of saturation, an ample amount of neovascular flow is detected. This can be critical in the distinction between neoplastic and nonneoplastic entities.

Download Original Video (22.0 MB)

Angle of Insonation Is at a 90° Angle to the Direction of Flow.—An artifactual absence of blood flow can be seen when the angle of insonation of the ultrasound beam to the vessel is at 90°. This artifact is most commonly seen at a turn of a very tortuous vessel or at a branching point of a vessel such as an intrahepatic portal vein (Fig 6). This phenomenon can be explained by using the Doppler frequency shift equation (Fig 1), which infers that a frequency shift is directly proportional to the cosine of the Doppler angle (θ). According to this equation, when a reflector object (eg, RBC) is moving parallel to the source of the sound (transducer), the frequency change is nearly proportional to the velocity of the blood cells, with the Doppler shift at its maximum (Fig 7). When the sound waves and blood cells are not moving in parallel directions, the equation must be modified to account for less Doppler shift. The maximal Doppler shift occurs at an angle (θ) of 0°. Therefore, the best signal and best spectral image are obtained when the direction of flow is parallel to the ultrasound beam. When the direction of flow is perpendicular to the transducer, the frequency shift is zero, because the cosign of a 90° angle is equal to zero, which means no flow (Fig 6). Thus, a Doppler angle of 90° does not display flow on the CDI or SDI waveform.

Figure 6. No vascular flow because the ultrasound beam is perpendicular (90° angle of insonation) to the direction of blood flow in the main portal vein (MPV). Color Doppler scan of the liver shows a lack of color filling in a portion of the anterior branch (arrow), resembling partial vein thrombosis. This is due to the angle of insonation of the ultrasound beam being approximately 90° to this part of the vessel (diagram at right). Changing the position of the transducer to reduce the angle will help to avoid misinterpretation. Cor = coronal.

Download as PowerPointOpen in Image Viewer

Figure 7. Diagram illustrates the relationship between Doppler angle and velocity. The maximal Doppler shift (v = max) occurs at a 0° angle when the ultrasound beam is parallel to the flow. The Doppler shift varies as a function of the cosign of the Doppler angle (θ) of the transmitted pulse and the axis of the vessel, for a fixed blood velocity (v). For angles greater than 0°, the measured Doppler shift is less by a factor cosine θ, and the velocity estimates are compensated by 1/cos θ. If the motion is perpendicular to the ultrasound beam, no Doppler signal is obtained (v = 0). An average Doppler shift (v = ave) occurs at an angle of less than 60°.

Download as PowerPointOpen in Image Viewer

When an absence of flow is encountered, the examiner should always confirm that the angle of insonation is less than 90° to the vessel of interest. Selecting a different acoustic window may ensure a change in the angle of insonation and thus result in improved flow detection. It should be noted that the Doppler signal in the gray-scale mode is strongest at a 90° angle, mainly because of the better axial resolution. Therefore, there is a trade-off between gray-scale US image optimization and Doppler US image optimization.

Inappropriate Transducer Selection and Deep Positioning of the Color Box

The inability to detect flow can also be related to a suboptimal choice of transducer. The capability of the US machine to detect flow decreases with increasing depth of the vessel of interest, as the ultrasound beam attenuates with increasing distance to a target (1). The rate of attenuation, or attenuation coefficient, depends on the medium (tissue through which the ultrasound beam propagates) and the US frequency. The sound waves emitted from a higher-frequency transducer attenuate faster because they have a higher attenuation coefficient. An inappropriate choice of transducer, with an inadequate Doppler frequency, will result in the inability to detect flow in a deeper positioned vessel (Fig 8). This can be avoided by switching to a lower-frequency transducer that allows deeper penetration and has a higher sensitivity threshold for the detection of signals (11).

Figure 8. Inability to detect blood flow due to inappropriate transducer selection in a 92-year-old patient who experienced multiple transient ischemic attacks. CDI scan obtained for evaluation of the right internal carotid artery (ICA) shows the ICA to be tortuous and deep in the neck, resulting in significant attenuation of the ultrasound beam generated by a high-frequency (12–7-MHz) transducer. Signal loss occurs, and there is no color filling in the deeply positioned segment of the vessel (dashed line, ★). In addition, significant tortuosity of the ICA results in areas of absent color flow when the angle of insonation is 90° to the vessel. Flow in the ICA proximal (PROX) and distal (DIST) to the tortuous segment implies that there is no actual obstruction to flow in the vessel. Switching to a lower-frequency transducer may help to eliminate this artifact and improve the visualization of color in the vessel lumen. The aliasing in the distal ICA is due to a low velocity scale setting. Normal waveforms and velocities were obtained proximal and distal to the tortuous segment, further reducing suspicion of vascular obstruction (not shown).

Download as PowerPointOpen in Image Viewer

In contrast, higher-frequency transducers are more sensitive for the detection of slow flow because they generate higher frequency shifts and therefore an increased amplitude of the returning echoes, resulting in better blood flow detection (12). Therefore, superficial vascular structures such as peripheral arteries and veins (eg, subclavian or femoral) or superficially located organs such as the testes and subcutaneous tissues are best examined by using 12–7-MHz linear transducers. In some cases, a trial use of several transducers may result in the best compromise between penetration and signal strength. Progress in US technology has led to great advances in transducer capabilities, with some of the newer models equipped with a wide range of frequencies (17–3 MHz).

Deep positioning of the color box also has a negative effect on the detection of flow, as the ultrasound beam becomes more attenuated as the depth increases; the returning signals are weaker and may not be displayed on the image. To improve blood flow detection, the depth of imaging should be kept to a minimum. Occasionally, selecting a different acoustic window may help to decrease the distance to a target vessel or organ, resulting in shallower placement of the color box and improved flow detection (Fig 9). The size of the color box also plays a role in flow detection. A smaller color box allows higher frame rates, which increase the PRF and improve the sensitivity to flow.

Figure 9a. Poor blood flow visualization with increased distance to the target. The color box is positioned too deep. (a) CDI scan through the left kidney (LK) obtained by using a left lateral decubital approach shows no flow in the abdominal aorta (AO). Flow is present in the left kidney, which is positioned closer to the transducer. The distance from the skin to the aorta is 10 cm (bracket). (b) On a color Doppler scan obtained by using a different approach, by way of the epigastrium, aortic flow is readily detectable; the distance to the aorta is 4 cm. With increased depth, the image frame rate is decreased, and, thus, the maximal allowable PRF is decreased.

Download as PowerPointOpen in Image Viewer

Figure 9b. Poor blood flow visualization with increased distance to the target. The color box is positioned too deep. (a) CDI scan through the left kidney (LK) obtained by using a left lateral decubital approach shows no flow in the abdominal aorta (AO). Flow is present in the left kidney, which is positioned closer to the transducer. The distance from the skin to the aorta is 10 cm (bracket). (b) On a color Doppler scan obtained by using a different approach, by way of the epigastrium, aortic flow is readily detectable; the distance to the aorta is 4 cm. With increased depth, the image frame rate is decreased, and, thus, the maximal allowable PRF is decreased.

Download as PowerPointOpen in Image Viewer

Poor or No Detection of Flow at SDI

Absence of a spectral waveform or very faint visualization of the waveform also may be a product of poor image optimization. As with CDI settings, an increase in spectral gain, a decrease in the velocity scale, and an adjustment of the angle of insonation to less than 90° can lead to improved flow visualization at SDI (Fig 10) (Movies 6, 7). Selecting a different acoustic window to decrease the distance to the target vessel also can be helpful for flow detection.

Get the Flash Player to see this video.

Movie 6 Optimization of the spectral Doppler velocity scale. When the range of the velocity scale is very wide, up to 500 cm/sec, the obtained waveforms become much smaller in relation to the confines of the graph, making it difficult to accurately determine absolute velocities. When the spectral velocity scale is decreased to below 45 cm/sec, aliasing occurs, because the obtained velocity measurements are higher than the upper limit value on the selected scale. Note that the highest velocities are represented below the baseline. When the velocity scale is optimized to the 0–160 cm/sec range, the obtained waveforms fit well within the confines of the graph.

Download Original Video (29.9 MB)Get the Flash Player to see this video.

Movie 7 Optimization of the spectral Doppler gain. The gain should be adjusted so that the ideal signal-to-noise ratio, manifesting as minimal background noise and a clear spectral window, is reached. Increasing the spectral Doppler gain increases the brightness of the spectrum on the screen and maximizes visualization of the spectral tracing. When the gain setting is too high, the velocity envelope is degraded and a significant amount of noise artifact is produced in the background of the spectral graph; this can result in the appearance of spectral broadening. When the gain is set too low, the returning signal is not well amplified and the displayed graph is poorly seen. When the gain is set at zero, no waveform is produced, falsely suggesting absent flow.

Download Original Video (40.8 MB)

Figure 10. Poor blood flow detection due to a low gain setting on the SDI unit. Sagittal SDI scan of the CCA shows faint waveforms obtained when the gain setting is very low. When the gain is set too low (0% gain [circled], insert image), the returning signal is not well amplified and the displayed tracing is poorly seen. When the gain is set at zero, a waveform may not be seen, falsely indicating absence of flow. To improve flow visualization, the gain has to be increased so that an ideal signal-to-noise ratio, which facilitates minimal background noise and a defined spectral trace, can be reached. The gain setting for SDI can be adjusted only when the SDI mode is active.

Download as PowerPointOpen in Image Viewer

New Platforms for Detection of Vascular Flow

Doppler techniques, including CDI, SDI, and PDI, are established tools that are used in daily clinical practice to characterize tumor and organ vascularity and confirm vascular patency. These tools enable excellent quantitative and qualitative assessment of the vascular bed. However, the evaluation of microvessels and vascular beds with slow flow is limited with use of these techniques.

A number of innovative vascular US technologies focused on the detection of blood flow, specifically slow flow in microscopic vessels and in diseased vessels, organs, or other structures, have emerged during the past decade. These technologies include but are not limited to B-flow imaging, microflow imaging, and superb microvascular imaging. Some of these newer applications are based on power Doppler US fundamentals (microflow and superb microvascular flow), while others are entirely different. For example, B-flow imaging is a form of gray-scale scanning (13–22). Moreover, contrast-enhanced US is gaining more acceptance and being utilized more often in different US applications and on a larger scale by both academic and private radiology practices (23).

All of these newer techniques were designed with a common goal: to make US, as compared with other cross-sectional and interventional modalities, such as CT angiography, MR angiography, and conventional angiography, as useful or more useful for the detection of vascular flow and vessel continuity (Figs 11, 12) (24). The main intent in using these technologies is to dramatically remove clutter while maintaining very high frame rates. This aids in removing motion artifact (ie, clutter) while attempting to visualize low-velocity blood flow.

Figure 11a. Advanced imaging techniques for evaluating slow or microvascular flow. (a) B-flow US image of the left proximal (Prox) ICA obtained for evaluation of ulcerated plaque (P) and focal dissection shows slow flow in the ulceration of the soft plaque (short arrow). The focal dissection flap (long arrow) is more apparent at B-flow imaging than at Doppler and gray-scale imaging. CDI scan (top left insert) shows no flow within the plaque, as well as blooming artifact over the dissection flap. IJV = internal jugular vein. (b) Dual-imaging gray-scale (left and right) and monochrome superb microvascular flow (color overlay, right) images of a kidney obtained for evaluation of renal parenchymal vascularity show fine microvascular branching vessels in the renal cortex and crisp main hilar renal vasculature, resembling angiographic findings. The direction of flow and absolute flow velocity cannot be determined with this technique. (Figure 11b courtesy of Pat Washko, BS, RVT, RDMS, Rex Hospital, Raleigh, NC.)

Download as PowerPointOpen in Image Viewer

Figure 11b. Advanced imaging techniques for evaluating slow or microvascular flow. (a) B-flow US image of the left proximal (Prox) ICA obtained for evaluation of ulcerated plaque (P) and focal dissection shows slow flow in the ulceration of the soft plaque (short arrow). The focal dissection flap (long arrow) is more apparent at B-flow imaging than at Doppler and gray-scale imaging. CDI scan (top left insert) shows no flow within the plaque, as well as blooming artifact over the dissection flap. IJV = internal jugular vein. (b) Dual-imaging gray-scale (left and right) and monochrome superb microvascular flow (color overlay, right) images of a kidney obtained for evaluation of renal parenchymal vascularity show fine microvascular branching vessels in the renal cortex and crisp main hilar renal vasculature, resembling angiographic findings. The direction of flow and absolute flow velocity cannot be determined with this technique. (Figure 11b courtesy of Pat Washko, BS, RVT, RDMS, Rex Hospital, Raleigh, NC.)

Download as PowerPointOpen in Image Viewer

Figure 12a. Advanced techniques for evaluation of microvascularity and slow flow. (a) At microflow imaging (MFI) of the right kidney (RK), the microvascular branching pattern of the renal vasculature is overlayed (right) on a gray-scale image. Subtraction monochromatic images, without gray-scale superimposition, also could be obtained to show only an angiogram-like appearance of the vasculature. (b) Color Doppler (A), gray-scale (B), and contrast-enhanced (C) US images of the right (RT) kidney show diffuse arterial enhancement of a hypoechoic mid-pole (MID) renal mass (M) (arrow in C). The mass exhibited indeterminate characteristics, no significant enhancement at other imaging modalities (contrast-enhanced CT and gadolinium-enhanced MRI), and no significant vascular flow at CDI (arrows in A). Renal mass protocol (contrast-enhanced CT) and MR images are not shown.

Download as PowerPointOpen in Image Viewer

Figure 12b. Advanced techniques for evaluation of microvascularity and slow flow. (a) At microflow imaging (MFI) of the right kidney (RK), the microvascular branching pattern of the renal vasculature is overlayed (right) on a gray-scale image. Subtraction monochromatic images, without gray-scale superimposition, also could be obtained to show only an angiogram-like appearance of the vasculature. (b) Color Doppler (A), gray-scale (B), and contrast-enhanced (C) US images of the right (RT) kidney show diffuse arterial enhancement of a hypoechoic mid-pole (MID) renal mass (M) (arrow in C). The mass exhibited indeterminate characteristics, no significant enhancement at other imaging modalities (contrast-enhanced CT and gadolinium-enhanced MRI), and no significant vascular flow at CDI (arrows in A). Renal mass protocol (contrast-enhanced CT) and MR images are not shown.

Download as PowerPointOpen in Image Viewer

These technologies also may involve the use of subtraction capabilities, whereby only the vasculature is displayed while the remaining background signal is eliminated (Table 4). The ability to detect vascular flow with use of these newer techniques is very impressive (Figs 11, 12) (Movie 8).

Table 4: New Platforms for Vascular Flow Detection

Get the Flash Player to see this video.

Movie 8 Ulcerated ICA plaque at B-flow imaging. B-flow imaging of the left ICA depicts intermittent blood flow within a small soft plaque ulceration. This was not seen at CDI owing to blooming artifact and low sensitivity to slow flow.

Download Original Video (18.8 MB)

Excessive Color Flow

Several factors can cause the appearance of excessive color flow on CDI scans. A few such factors that examiners must keep in mind while performing Doppler US examinations are described in the sections that follow.

Color Doppler Gain Set Too High

US images may appear to be filled with color if the gain is set too high (Fig 13). With high gain settings, even minimal motion, including background noise, is assigned a certain color, filling the entire color box with color specks and color inside and outside the vascular wall (Movies 7, 9, 10). To optimize the image, the color gain setting should be decreased until color noise is eliminated (Movie 3). No specific percentage of saturation is recommended; rather, the optimization is applied on a case-by-case basis.

Get the Flash Player to see this video.

Movie 9 Excessive color flow with high gain settings. CDI cine clip of the left kidney demonstrates excessive color flow within and adjacent to the left renal hilum when the color gain is at its maximum (95%). There is poor discrimination between noise and true flow in the renal vessels, as the color fills not only vascular structures but also surrounding soft tissues. Decreasing the color gain facilitates minimization of aliasing and better assessment of the vascular structures.

Download Original Video (20.3 MB)Get the Flash Player to see this video.

Movie 10 Excessive color flow with high gain settings. CDI cine clip of the left ovary (LO) demonstrates excessive color flow in the left adnexa, with a change in color gain to the maximal level (95%). No meaningful information is obtained, as the color fills not only the vascular structures but also the surrounding soft tissues. By decreasing the color gain, aliasing is minimized and the vascular structures are better assessed. FT = fallopian tube.

Download Original Video (21.0 MB)

Figure 13. Overestimation of vascular flow due to high color gain setting. Sagittal CDI scans of the left CCA show color specks in and adjacent to the vessel lumen (arrows). The gain is set at 84% of saturation (circled). The flow overestimation is caused by the amplification of low frequency shifts received from the minimal motion in the surrounding tissues. To improve the image, the color gain should be decreased until no color specks are seen outside of the vessel walls. No specific percentage of gain is recommended; rather, the optimization is applied on a case-by-case basis.

Download as PowerPointOpen in Image Viewer

Related Artifacts

Blooming Artifact.—Blooming artifact, usually seen with high gain settings, is a phenomenon in which the color spreads, or “bleeds,” outside of the margins of a blood vessel, overriding the gray-scale information. PDI scans are free of blooming artifact, because with this imaging mode, the signal is not amplified; rather, the strength and amplitude of the signal are detected and imaged. In daily clinical practice, blooming artifact may result in the inability to detect a dissection flap, an atherosclerotic ulcerated plaque, or even a partial vascular thrombus owing to obscuration of the findings by color blooming (Fig 14). Decreasing the gain settings, adjusting the velocity scale, and/or using PDI may help to prevent this artifact.

Figure 14a. Blooming artifact and excessive blood flow at CDI. (a) Gray-scale image of the porta hepatis shows a nonocclusive thrombus (arrow) in the main portal vein (MPV). (b) Color Doppler scan of the same region shows color that “bleeds” over the vessel wall, masking the thrombus (arrow) and mimicking the appearance of a fully patent main portal vein. Blooming artifact can be prevented by lowering the color gain settings and/or using PDI, at which the received signal is not amplified.

Download as PowerPointOpen in Image Viewer

Figure 14b. Blooming artifact and excessive blood flow at CDI. (a) Gray-scale image of the porta hepatis shows a nonocclusive thrombus (arrow) in the main portal vein (MPV). (b) Color Doppler scan of the same region shows color that “bleeds” over the vessel wall, masking the thrombus (arrow) and mimicking the appearance of a fully patent main portal vein. Blooming artifact can be prevented by lowering the color gain settings and/or using PDI, at which the received signal is not amplified.

Download as PowerPointOpen in Image Viewer

Flash Artifact.—Excessive color flow may be seen on an image when color flow is applied to any significant motion detected by the US unit (Movie 11). The basis of this phenomenon, which is known as flash artifact, is that the Doppler effect is used to measure a change in the frequency that is generated by motion of any moving reflector, not just RBCs, including the frequency shifts generated by motion of the diaphragm, bowel, or detector itself (such as a transducer probe). Thus, bowel peristalsis, cardiac activity, and transmitted pulsations from great vessels may generate color flash artifact (Movies 11, 12). Flash artifact is commonly associated with PDI and can be minimized by acquiring images when the transducer is still or the patient is holding his or her breath, or by positioning the color box away from peristalsing bowel. If flash artifact is suspected, SDI can be performed to confirm that the detected flow is not real, as only vascular flow exhibits the expected cardiac flow dynamics.

Get the Flash Player to see this video.

Movie 11 Flash artifact due to overestimation of flow. CDI evaluation of the liver demonstrates filling of the color box with a burst of colors, which represents flash artifact caused by motion of the diaphragm during inspiration. PDI is the most susceptible to flash artifact owing to the low frame rate of image acquisition (not shown).

Download Original Video (18.3 MB)Get the Flash Player to see this video.

Movie 12 Flash artifact due to bowel peristalsis. Longitudinal CDI scan of the abdominal aorta demonstrates a sudden surge of color, or flash artifact, caused by peristalsis in the adjacent bowel. The examiner should recognize the artifact and be sure not to record it, as it may simulate findings associated with other vascular disease.

Download Original Video (17.6 MB)

Perivascular Tissue Vibration.—Excessive flow signal may also be detected in perivascular structures at a site of significant stenosis or arteriovenous fistula (Movie 13). This phenomenon is known as tissue bruit artifact, color bruit artifact, or visible bruit artifact and is seen in the CDI mode. Color bruit artifact results from low-level frequency shifts that are generated when high-velocity jets induce vibrations in the tissues surrounding an area of stenosis or fistula (Fig 15) (3). The recognition of tissue bruit is essential for the accurate detection of areas of flow disturbance at vascular US. Adjacent vessels should be thoroughly interrogated for the presence of stenosis or arteriovenous fistulas with use of SDI. Tissue bruit artifact can alert the examiner to the possible presence of an abnormality and thus should be distinguished from flash artifact, which requires image optimization for correct interpretation of findings.

Get the Flash Player to see this video.

Movie 13 Recurrent distal femoral artery stent stenosis and perivascular tissue vibration. CDI cine clip through a right superficial femoral artery (SFA) stent demonstrates aliasing within the distal portion of the stent, manifested by a mosaic of colors within the stenotic area. Laminar flow is seen in the more proximal portion of the stent. Also note the visible color bruit artifact (arrow) in the adjacent soft tissues that is due to tissue vibration at the site of turbulent flow. Spectral Doppler interrogation of the area of flow abnormality revealed markedly elevated PSVs, consistent with significant stenosis.

Download Original Video (29.2 MB)

Figure 15. Overestimation of flow at CDI (top) and SDI (bottom) due to perivascular bruit artifact (ie, visible bruit artifact). SDI of a stenotic celiac artery (CA) shows high-signal-intensity noise at baseline during systole (red arrow, red outline), with associated spectral broadening and an elevated PSV of 342 cm/sec. Similar color bruit artifact and aliasing (yellow arrows) are seen at the level of stenosis on the CDI portion of the display. At CDI and SDI, perivascular bruit artifact is most conspicuous during systole and less prominent during diastole. This artifact can be reduced by increasing the PRF and/or the wall filter. Perivascular bruit artifact is an important diagnostic tool in the identification of arteriovenous fistulas, anastomotic sites, and arterial stenoses. Ao = abdominal aorta.

Download as PowerPointOpen in Image Viewer

Aliasing

Intact vasculature depicted on a properly optimized US image usually demonstrates a laminar flow pattern with homogeneous color distribution in the vessel(s). A disturbed flow pattern, such as that in an area of stenosis, or the lack of proper image optimization can result in a multicolored heterogeneous color display, which is referred to as aliasing. When aliasing occurs owing to improper image optimization, a few Doppler parameters might require adjustment. These parameters include velocity scale, color baseline, depth and size of the color box and/or sample volume gate, transducer frequency, and angle of insonation (Fig 16). For a better understanding of why certain parameters require adjustment, a brief summary of the underlying physical principles may be helpful.

Figure 16. Diagram illustrates CDI and SDI parameters that may need adjusting to address aliasing. AVF = arteriovenous fistula.

Download as PowerPointOpen in Image Viewer

Aliasing can occur in the CDI and SDI modes, in which the Doppler signal is formed by using multiple pulse-echo sequences. In a pulse-echo system, sampling of the Doppler signal occurs at a specific rate, or PRF. When the sampling rate of the Doppler signal is sufficient, blood flow in a vessel is assigned an appropriate color coding that reflects its velocity and direction. To accurately reproduce a wavelength of a specific Doppler frequency, a minimum of two samples per cycle (wavelength) is required to unambiguously determine the Doppler signal frequency, fs, and convert it to a corresponding velocity. Therefore, the sampling rate, or PRF, must be at least twice the Doppler signal frequency (Δfsmax = PRF/2 or PRF = 2fs), where Δfsmax is the maximal signal frequency shift. This is known as the Nyquist limit or Nyquist criterion.

If the sampling rate is lower than the Nyquist rate, aliasing occurs (Fig 17). This phenomenon is encountered when high Doppler shift frequencies are undersampled (at low PRF), with the result being signals that are inaccurately reconstructed at a lower frequency (longer wavelength). The assumption is that blood will move away from the transducer if the reflected frequency is lower than the transmitted frequency, and, thus, the Doppler shift will be assigned a negative value (or blue color on color Doppler scans) (Fig 18). This explains why the aliased signals are assigned an opposite color at CDI and an opposite velocity direction and amplitude at SDI. In other words, aliased Doppler signals become “wrapped around” the color bar and velocity scale and thus appear in a reversed flow direction (Fig 19).

Figure 17. Aliasing due to suboptimal image optimization. CDI scan of the abdomen shows a supranumerary left kidney (LK) ectopically positioned in the left lower abdomen. A mosaic of colors (arrows) is seen in the renal hila and four main renal arteries supplying the anomalous kidney, compatible with aliasing. This is due to the very low velocity scale range selected at the time of imaging to improve the detection of the four main renal arteries. Also note the aliasing in the distal abdominal aorta (Ao). The gain and wall filter settings (circled) are well optimized. Aliasing occurs because the peak velocities in the aorta and renal arteries are much higher than the selected velocity scale.

Download as PowerPointOpen in Image Viewer

Figure 18. Diagrams illustrate sampling of a Doppler signal. At a high PRF (top), the original echo signal (solid blue line) gets sampled multiple times per cycle (ie, per wavelength [λ]). The original wave is reconstructed by using information obtained during each sampling (blue dots). After the dots are connected, the reconstructed waveform displays the same frequency (ie, same wavelength). No aliasing occurs. At a low PRF (middle), the original echo signal gets sampled less than two times per cycle (ie, per wavelength), resulting in a reconstructed waveform that is much lower in frequency (long wavelength) (connected red dots). This results in a lower-frequency aliased signal owing to inadequate sampling. At a PRF of the Nyquist limit (PRF/2) (bottom), the initial signal will be sampled two times per cycle, resulting in a reconstructed waveform of the same frequency as the original echoes. No aliasing occurs.

Download as PowerPointOpen in Image Viewer

Figure 19. Aliasing at the level of stenosis. SDI scan of the celiac artery (CA) at the level of stenosis shows aliasing in the color portion of the display (white arrow). Because the color velocity scale and PRF are not well optimized, high velocities are assigned the wrong color and direction. In this case, aliasing is a helpful sign, enabling quick assessment of the area of potential stenosis and placement of the spectral gate for evaluation of the absolute PSV at the focal stenosis. The spectral waveform portion of the display (bottom) fits well within the confines of the graph, without aliasing (yellow arrow), owing to downward placement of the spectral baseline. The PSV at the level of the stenosis is 253 cm/sec. Ao = abdominal aorta.

Download as PowerPointOpen in Image Viewer

The presence of aliasing should not be confused with a true change in the direction of blood flow. At imaging, a true change in the direction of flow is indicated by a black line (or bar) between the red and blue colors, indicating that the direction change is genuine (Movie 14). An absent demarcating black line with intermixed red and blue colors is an example of color aliasing (Fig 19). To eliminate aliasing error, the velocity scale must be adjusted to a higher range or PRF (Movie 15) (2,4,6).

Get the Flash Player to see this video.

Movie 14 True change in flow direction. CDI video of the splenic vein at the level of the pancreas demonstrates a black line or bar (arrow) between the red and blue colors, representing a true change in the direction of flow and indicating that the direction change is genuine.

Download Original Video (12.3 MB)Get the Flash Player to see this video.

Movie 15 Aliasing at low velocity scale. Sagittal CDI video of the left CCA shows that using a low velocity range results in the development of mosaic colors due to aliasing. Increasing the PRF and velocity scale eliminates aliasing.

Download Original Video (14.6 MB)

A few additional techniques can be used to avoid aliasing. These techniques include downward or upward adjustment of the color or spectral baseline, which allows a greater range of sampled frequencies to be assigned positive (downward adjustment) or negative (upward adjustment) velocities (Movies 16, 17). Also, the color box size and/or sample volume depth can be reduced to allow an increase in the PRF and a higher rate of sampling. Choosing a lower-frequency transducer or larger Doppler angles can minimize the Doppler shift frequency and thus prevent aliasing (2).

Get the Flash Player to see this video.

Movie 16 Optimization of the color Doppler baseline. With adjustment of the baseline downward, the positive velocities are emphasized and aliasing is seen in the vein. With adjustment of the baseline upward, the negative velocities are emphasized and aliasing is seen in the artery. When the baseline is positioned in the middle of the color bar, the positive and negative velocities are equally emphasized and laminar flow is seen in the artery and vein.

Download Original Video (20.2 MB)Get the Flash Player to see this video.

Movie 17 Optimization of the spectral Doppler baseline. When the baseline is moved upward, the higher velocities will be wrapped around and displayed below the baseline. When the baseline is moved downward, the negative velocities will be wrapped around and displayed at the top of the spectral display. With optimal settings of the spectral baseline, the waveform is accurately displayed on the graph.

Download Original Video (17.7 MB)

When aliasing is associated with true blood flow disturbance at CDI, it allows the operator to quickly localize the area of highest velocity and thus aids in identifying areas of stenosis or high-velocity flow abnormalities, such as arteriovenous fistulas, the neck of pseudoaneurysms, and tortuous vessels (Fig 19) (Movies 18–20). The identification of aliasing at CDI facilitates placement of the spectral gate for quantitative assessment of velocity. Therefore, the experienced examiner always pays close attention to focal areas of aliasing to aid in accurately diagnosing areas of abnormal flow (Fig 19).

Get the Flash Player to see this video.

Movie 18 Aliasing at the level of vascular shunt placement. CDI cine clip shows a middle hepatic vein (MHV)–to–right hepatic vein (RHV) shunt in a 19-year-old man with a liver transplant (TPL). An abnormal direct communication between the right and middle hepatic veins is seen, with retrograde flow in the branch of the right hepatic vein that shunts blood directly from the middle hepatic vein into the main right hepatic vein. Aliasing is noted where the middle hepatic vein takes a sharp turn at the periphery of the liver (arrow). Optimized parameters enable one to recognize the direction of abnormal flow in the vein and render the correct diagnosis. Cor = coronal.

Download Original Video (29.7 MB)Get the Flash Player to see this video.

Movie 19 Aliasing at the level of a left common femoral artery pseudoaneurysm. CDI cine clip (top) depicts aliasing in the neck of the pseudoaneurysm (arrow). SDI (bottom) depicts the characteristic “to-and-fro,” or bidirectional, flow in the neck of the pseudoaneurysm (PSA). Note the large partially thrombosed pseudoaneurysmal sac. LT = left, TRV = transverse.

Download Original Video (14.2 MB)Get the Flash Player to see this video.

Movie 20 Aliasing in a tortuous right ICA. CDI of the right ICA depicts aliasing in areas where the blood moves at the highest velocity, manifesting as a color mosaic. Aliasing (arrows) is noted throughout the artery during diastole and at sharp turns of the vessel during the systolic portion of the cardiac cycle. Increasing the velocity scale (or PRF) and/or decreasing the gain can help to prevent this artifact.

Download Original Video (26.0 MB)

Flow Direction Ambiguity

Although CDI, SDI, and directional PDI enable the examiner to determine the direction of flow, several scenarios may prevent him or her from doing so. The aliasing just described is one such scenario. In addition, positioning the color and spectral baselines too high or too low, as well as the inability to determine mean and absolute velocities, in an examined vascular bed also may result in flow direction ambiguity (Fig 20).

Figure 20a. Directional ambiguity with poor optimization of the color and spectral baselines. (a) Sagittal CDI scan of the CCA and internal jugular vein (IJV) shows the effect of altering the positioning of the color baseline. When the baseline is shifted downward from the midportion of the color bar (circled), the positive frequency shifts (ie, velocities) are emphasized (bracket); the larger color bar at far right is a magnification of the smaller color bar. Thus, the CCA is red; however, aliasing is seen in the internal jugular vein. No direction of blood flow can be determined from the image, and no mean velocity can be estimated in the internal jugular vein. A similar color assignment is seen in the vessel adjacent to the ICA (arrow). With adjustment of the color baseline upward (not shown), the emphasis can be changed to negative velocities. When the position of the baseline is changed, the color velocity range also is changed, from 28 cm/sec when the color baseline is in the midline (not shown) to +49 cm/sec (rectangles) with downward adjustment of a color baseline and to +7 cm/sec when the color baseline is adjusted upward (not shown). (b) Sagittal SDI scan of the right brachial artery (R BA) shows the effect of altered positioning of the spectral baseline. The spectral baseline divides the spectral display into positive and negative Doppler shifts, with a Doppler shift of zero at the baseline itself (yellow line). The frequency shifts above the baseline represent blood velocities moving toward the transducer, and the shifts below the baseline represent blood moving away from the transducer. When the baseline is moved upward, the higher velocities will be “wrapped around” and displayed at the bottom under the baseline (arrow), consistent with aliasing. The negative velocity will be emphasized (bracket). No absolute velocities or flow direction can be calculated with this suboptimal baseline placement. A similar effect is produced when the baseline is moved downward, resulting in emphasis on positive velocities and resultant wraparound of negative velocities that will be displayed at the top of the spectral display (not shown). Adjustment of the color and spectral Doppler baselines is useful for avoiding aliasing when high PSVs are detected in a vessel and the velocity scale is already optimized to its maximum. The baselines can be adjusted to allow higher velocities to be seen in the correct flow direction and thus prevent wraparound of the data (ie, aliasing).

Download as PowerPointOpen in Image Viewer

Figure 20b. Directional ambiguity with poor optimization of the color and spectral baselines. (a) Sagittal CDI scan of the CCA and internal jugular vein (IJV) shows the effect of altering the positioning of the color baseline. When the baseline is shifted downward from the midportion of the color bar (circled), the positive frequency shifts (ie, velocities) are emphasized (bracket); the larger color bar at far right is a magnification of the smaller color bar. Thus, the CCA is red; however, aliasing is seen in the internal jugular vein. No direction of blood flow can be determined from the image, and no mean velocity can be estimated in the internal jugular vein. A similar color assignment is seen in the vessel adjacent to the ICA (arrow). With adjustment of the color baseline upward (not shown), the emphasis can be changed to negative velocities. When the position of the baseline is changed, the color velocity range also is changed, from 28 cm/sec when the color baseline is in the midline (not shown) to +49 cm/sec (rectangles) with downward adjustment of a color baseline and to +7 cm/sec when the color baseline is adjusted upward (not shown). (b) Sagittal SDI scan of the right brachial artery (R BA) shows the effect of altered positioning of the spectral baseline. The spectral baseline divides the spectral display into positive and negative Doppler shifts, with a Doppler shift of zero at the baseline itself (yellow line). The frequency shifts above the baseline represent blood velocities moving toward the transducer, and the shifts below the baseline represent blood moving away from the transducer. When the baseline is moved upward, the higher velocities will be “wrapped around” and displayed at the bottom under the baseline (arrow), consistent with aliasing. The negative velocity will be emphasized (bracket). No absolute velocities or flow direction can be calculated with this suboptimal baseline placement. A similar effect is produced when the baseline is moved downward, resulting in emphasis on positive velocities and resultant wraparound of negative velocities that will be displayed at the top of the spectral display (not shown). Adjustment of the color and spectral Doppler baselines is useful for avoiding aliasing when high PSVs are detected in a vessel and the velocity scale is already optimized to its maximum. The baselines can be adjusted to allow higher velocities to be seen in the correct flow direction and thus prevent wraparound of the data (ie, aliasing).

Download as PowerPointOpen in Image Viewer

Another cause of directional ambiguity is flow inversion due to variable positioning of the transducer relative to the vessel of interest. For example, if the portal vein at its bifurcation is examined by using the epigastric approach, the direction of blood flow in this vein will be perceived as traveling away from the transducer, and, thus, the blood flow will be assigned a blue color at CDI. This mistake can be rectified by scanning the patient by using a right intercostal approach (Movie 21).

Get the Flash Player to see this video.

Movie 21 Color flow inversion. When the main portal vein (MPV) is scanned by using a subcostal or epigastric approach, as in this case, the direction of the blood flow in the vein is seen flowing away from the transducer, and, therefore, a blue color is assigned to the main portal vein. This could be misinterpreted as hepatofugal flow, which is commonly seen with portal hypertension. Note that the color bar is not inverted. When the images are obtained by using the intercostal approach, the direction of the blood flow is toward the transducer, and, thus, the vascular flow is correctly assigned a red color, indicating hepatopetal flow. AO = abdominal aorta.

Download Original Video (28.5 MB)

Other factors that might lead to difficulty interpreting the flow direction correctly include inversion of the color bar and/or spectral graph, altered steering of the color box, and a flipped transducer. The inversion control enables the operator to invert the Doppler display, with flow away from the transducer displayed above the baseline and flow toward the transducer displayed below the baseline (Fig 21). Activation of the inversion setting can be recognized by the display of negative velocity values above the baseline. Care must be taken to note the direction of flow on the Doppler displays so that the misinterpretation of reversed flow can be avoided. In addition, the color box should be steered in a way that emphasizes anterograde flow toward the transducer, and the transducer notch should be facing the head or right side of the patient.

Figure 21. Flow inversion. SDI scan of the liver shows inversion of the color bar and the resultant incorrect interpretation (ie, color assignment) of the blood flow in the main portal vein (MPV) as going away from the liver (arrow). The spectral graph is not inverted and correctly displays the blood flow going toward the liver. Color and spectral graphs can be inverted. It is imperative to pay attention to the color bar and spectral graph values to ensure that the scale is not erroneously inverted and avoid misinterpreting the finding.

Download as PowerPointOpen in Image Viewer

Related Artifacts

Cross talk.—Directional ambiguity, or cross talk, refers to an artifact seen at SDI and characterized by a display of waveforms with nearly equal amplitude above and below the baseline in a mirror image pattern (25) (Fig 22). This artifact is generated when the interrogating beam intersects the vessel at an approximately 90° angle (26). At such angles, the ability to discriminate between forward and reversed flow is impaired, and, thus, the Doppler signal is displayed in both directions (12). This artifact is also seen when interrogating small vessels, particularly those that travel in and out of the plane of the ultrasound beam within a sample volume or when the gain is increased (Movie 22) (27). By interfering with the determination of flow direction, this artifact may also adversely affect velocity measurements.

Get the Flash Player to see this video.

Movie 22 Cross talk artifact with high gain settings. SDI of the CCA at the optimal gain setting demonstrates a low-resistance waveform that is seen above the baseline only. Increasing the gain causes a display of waveforms with nearly equal amplitude above and below the baseline in a mirror image pattern. The waveform below the baseline is fainter than the original one. To prevent this artifact, the gain should be decreased or the angle of insonation should be smaller.

Download Original Video (17.3 MB)

Figure 22. Directional ambiguity or cross talk artifact on SDI scan of the left portal vein. The interrogating beam within a sample gate is almost perpendicular (pink arrow) to the direction of flow in the vessel (yellow arrow, right diagram); this results in simultaneous display of the velocities, with nearly the same amplitude but in opposite directions above and below the baseline (yellow line) in the mirror image pattern. This artifact interferes with the determination of flow direction. It can be addressed by decreasing the gain. The pulsatile waveform in the left portal vein is a sign of portal hypertension and hepatic congestion.

Download as PowerPointOpen in Image Viewer

Cross talk differs from bidirectional flow caused by the simultaneous display of velocities with the same amplitude but in opposite directions above and below the baseline. Bidirectional flow is characterized by a display of velocities in one direction during systole and in the opposite direction during diastole (Movie 23). It is sometimes attributable to a change in intra-abdominal pressure during inspiration and expiration, which results in highly pulsatile flow, or to different factors in various pathologic processes. For example, bidirectional flow may be seen in the renal arteries in the setting of acute renal vein thrombosis; in the hepatic veins in the setting of right heart failure or tricuspid regurgitation; and in the neck of a pseudoaneurysm (Movie 19). In addition, the observation of helical flow on a CDI scan (for example, at the bifurcation of a portal vein or carotid artery) may cause confusion regarding the directionality of flow. Helical flow is predominantly unidirectional at SDI and is due to spiraling of blood at the level of vessel widening.

Get the Flash Player to see this video.

Movie 23 Pulsatile or bidirectional flow in the main portal vein. CDI cine clip demonstrates pulsatile bidirectional flow in the main portal vein and its branches in a patient with right heart failure. Hepatopetal flow is observed during systole, and hepatofugal flow is seen during diastole. Cor = coronal.

Download Original Video (19.8 MB)

Spectral Broadening

When the blood flow is laminar, the spectral waveforms demonstrate a clear window under the spectral envelope, signifying that the majority of RBCs are moving at a similar velocity and in a uniform fashion (Fig 2). In the setting of flow disturbance (as a result of RBC collisions and the generation of eddy currents), a wide range of velocities “fill in” the spectral envelope. This phenomenon is referred to as spectral broadening (Fig 23). Although in clinical practice spectral broadening is considered a marker for stenosis and correlates with vascular disease processes, spectral broadening is also commonly seen in routine studies and may be related to a number of technical issues in the absence of significant vascular obstruction. Accurate identification and correction of the suboptimal settings that result in spectral broadening can decrease the number of misinterpretations and misdiagnoses. The main technical causes of spectral broadening are briefly discussed below.

Figure 23. Sagittal SDI scan shows spectral broadening caused by the selection of a high gain. When the gain setting is too high (100%, circled), it degrades the velocity envelope on the spectral display and produces a significant amount of noise artifact in the background of the spectral display (artifactual fill-in of the spectral waveform), resulting in spectral broadening (arrow). The velocity range also is elevated when the gain setting is too high (★), with the PSV increasing from 120 cm/sec (optimized gain, not shown) to 160 cm/sec on the image.

Download as PowerPointOpen in Image Viewer

Spurious spectral broadening may result from very high gain settings, large Doppler angles, a large sample volume gate (>3.5 mm), and/or the placement of a sample volume adjacent to the wall of a vessel (Figs 22, 23) (Movie 24). When gain settings are too high, overamplification of Doppler shifts may result in thickening of the velocity tracing and thus spectral broadening (Fig 23) (Movie 7). A clue to recognizing this phenomenon is the finding of substantial background noise in addition to spectral broadening (26).

Get the Flash Player to see this video.

Movie 24 Spectral broadening artifact. SDI of the carotid artery demonstrates a low-resistance waveform with a clear window under the spectral envelope. The spectral gate is centered within the artery lumen. When the spectral gate is moved close to the vessel wall, the waveform pattern is changed, and as a result, the spectral window becomes completely filled, resulting in spectral broadening. This is due to the presence of a wide range of velocities at the vascular wall as a result of friction and turbulent flow. Placement of the spectral gate back in the center of the vessel results in normalization of the obtained waveforms.

Download Original Video (13.2 MB)

Very large Doppler angles cause significant error in the estimation of flow velocities and, in turn, spectral broadening (Fig 22). Decreasing Doppler angles to 45°–60° aids in correcting the waveform appearance. In addition, if the angle is too small—for example, smaller than 20°, refraction and critical angle interactions can cause problems with accurate velocity estimation and waveform appearance (28).

With a large sample volume, Doppler shifts may be obtained from surrounding vessels and result in contamination of the obtained data and thus spectral broadening. Decreasing the size of the sample volume will prevent the mixing of data from different sources. By convention, a small sample gate size of 2–3 mm is considered optimal for minimizing unwanted data from structures in the vicinity of the examined vessel and thus decreases spectral broadening (29,30).

When a sample gate is located too close to the vascular wall, a wide range of velocities, due to local turbulent effects and wall friction, will be analyzed and plotted on the graph (31). A similar phenomenon is observed with small-caliber vessels, especially vertebral arteries, owing to sampling of the entire vessel, including the periphery. Placing the sample volume in the center of the vessel will ensure that only the most representative velocities in a vessel are displayed (Movie 24).

Detection of Vascular Flow Where Flow Is Absent

There are instances in which flow is depicted where none is present. Mirror image artifact, beam width artifact, pseudoflow, edge artifact, flash and tissue bruit artifacts, and twinkling artifact belong to this artifact category (Table 5).

Table 5: Artifacts and Artifact-Troubleshooting Techniques

Duplication or Mirror Image Artifact

Doppler mirror image artifact occurs when the reflected echoes are redirected by a strong specular reflector (ie, “mirror”) and eventually return to the transducer (Fig 24). To distinguish a mirror image artifact from the true target, one should remember that a strong reflector, or mirror, will always be present along a straight line between the transducer and the artifact, and both the target and the artifact will be equally distant from the mirror (32,33).

Figure 24. Drawing illustrates the generation of a vascular mirror image artifact, whereby flow is detected where none is present. Some of the transmitted ultrasound beams from the transducer (thinner [left] blue arrow) generate echoes from a hepatic vessel (thinner [left] yellow arrow), and part of the beam (short blue arrow) continues to a strong reflector (diaphragm), where a very strong echo is produced (short yellow arrow). These echoes travel from the diaphragm to the vessel, producing another set of echoes that are now directed back to the diaphragm. These echoes are then re-reflected from the diaphragm to the transducer (thicker and longer [right] yellow arrow). Returning echoes from the diaphragm to the transducer (thicker and longer [right] yellow arrow) are also amplified by the echoes generated by the ultrasound beam directly traveling to the diaphragm (thicker and longer [right] blue arrow]). The back-and-forth travel of the second echo set from a hepatic vessel produces echoes that return to the transducer at a later time. The machine assumes that the returning echoes are coming from a different, deeper vessel and displays the vessel at a deeper location (mirror image of the vessel) beyond the diaphragm. The true hepatic vessel and its mirror image are the same distance from the diaphragm. The mirror image lies along the line of the transmitted beam, perpendicular to the diaphragm. Mirror image artifacts can be reduced by changing the scanning angle.

Download as PowerPointOpen in Image Viewer

Duplication images, or mirror image artifacts, seen at vascular US are usually more complicated. The velocity, direction of flow, and signal strength in the artifactual (ie, mirrored) vessel are not always comparable or identical to those in the true vessel (Movie 25) (5). Alternating the angle of insonation to the strong reflector by changing the orientation of the transducer, adjusting the Doppler angle during SDI, and/or steering the color box during CDI may help in distinguishing a duplication artifact from a true vessel in ambiguous cases (5).

Get the Flash Player to see this video.

Movie 25 Vascular mirror image artifact. Longitudinal CDI of the liver depicts hepatopetal flow in the main portal vein (MPV) and hepatofugal flow in the hepatic veins (HV) and inferior vena cava (IVC). An additional area of color flow (arrows) erroneously displayed on the other side of the diaphragm in the region of the right lower lobe of the lung represents a vascular mirror image artifact. To prevent this artifact, the examiner can change the acoustic window so that the high reflector is not in the path of the ultrasound beam.

Download Original Video (26.1 MB)

Twinkling Artifact

Twinkling artifact is characterized by the display of flow at CDI when none is present. It is most commonly seen in association with renal or biliary stones (Fig 25). Twinkling artifact is especially valuable in US because it can aid in the diagnosis of various calcified lesions such as urinary, renal, and biliary stones; gallbladder adenomyomatosis; calcifications in chronic pancreatitis; encrustations on indwelling catheters; vascular parenchymal calcifications; and other lesions (Movie 26). To improve the detection of twinkling artifact, the focal zone should be positioned deeper than the area of interest, the color-write priority threshold should be set to the high setting (Table 1), and the gray-scale gain and PRF should be set to the lower settings.

Get the Flash Player to see this video.

Movie 26 Twinkling artifact produced behind air-filled loops of bowel. CDI of the middle region of the abdomen demonstrates a mosaic of colors at the nondependent portion of the air-filled bowel.

Download Original Video (9.6 MB)

Figure 25. Twinkling artifact. Longitudinal CDI scan of the left kidney (LK) shows rapidly alternating multicolored flow-mimicking signals (white arrow) that are deep to a nonobstructive echogenic stone seen at gray-scale US (not shown). The focal zone (pink arrow) is deep to the stone, leading to generation of the artifact.

Download as PowerPointOpen in Image Viewer

Pseudoflow

Pseudoflow is characterized by the presence of color flow in a fluid medium other than flowing blood (eg, motion in ascites, amniotic fluid, ureteral jets, gallbladder sludge) (25). This artifact is caused by the motion of particles in a fluid medium and can be seen in the CDI and PDI modes. Flow is detected for as long as the fluid motion continues (Movie 27). Spectral Doppler analysis can aid in preventing misinterpretations, as the waveform pattern will lack the flow characteristics of the arterial or venous system (25).

Get the Flash Player to see this video.

Movie 27 Pseudoflow artifact generated by a right ureteral jet. CDI of the urinary bladder demonstrates a color jet emanating from the right ureterovesical junction. Flow is detected for as long as the fluid motion (ie, urine flow) continues during the ureteral jet. This artifact is particularly important in the diagnosis of complete or partial ureterovesical junction obstruction, with no detectable ureteral jet observed in cases of complete obstruction and a weak prolonged continuous ureteral jet seen in cases of partial ureteral obstruction. A small right ureterocele also is seen.

Download Original Video (15.8 MB)

Edge Artifact

When the Doppler signal is generated at the surface of a very strong and echogenic reflector such as a calculus or cortical bone, or at the surface of an iatrogenic static structure such as a Foley balloon or surgical catheter, the US machine perceives the reflected signal to be flow. This reflected signal is referred to as edge artifact. Edge artifact has a characteristic appearance on the CDI display; it is seen as a continuous rim of color along the edge of the structure (Fig 26) (25). The absence of a vessel at the site of color flow at gray-scale imaging and the presence of a straight line instead of a waveform at SDI are helpful clues for recognition of this artifact. To eliminate this artifact, the PRF should be increased or the high wall filter settings can be manually optimized.

Figure 26. Edge artifact. CDI scan of the urinary bladder (B) shows a decompressed bladder around a Foley catheter. A continuous rim of variable colors (arrow) is seen along the surface of the Foley balloon. Increasing the PRF may help to prevent this artifact.

Download as PowerPointOpen in Image Viewer

Beam Width Artifact

Beam width artifact is encountered when a highly reflective object, located outside of the imaging plane but within the widest portion of the US beam (deep to the focal zone and beyond the actual width of the transducer), is depicted within the imaging plane. This artifact can be seen at Doppler imaging, where it is related to the incorrect depiction of flow in the imaging plane where no flow is present. This depiction error is due to contamination from the sampling of structures adjacent to a vessel (34). It can be recognized when the blood flow pattern does not correspond to the vessel sampled (ie, a venous waveform depicted when an artery is interrogated) (Fig 27). This artifact can be prevented by adjusting the focal zone so that it is positioned at the level of the vessel and by decreasing the size of the sample volume.

Figure 27. Bandwidth artifact in a 59-year-old patient with treatment-resistant hypertension. Transverse SDI scan of the left kidney (LK) shows superimposed arterial (red arrow) and venous (blue arrow) waveforms. Although the spectral sample gate (white arrow) is positioned at the level of the interlobar vein, the adjacent interlobar artery also is sampled, producing arterial waveforms that are placed on the same spectral display. This “contamination” is due to the three-dimensional nature of the sample volume, in which structures that are not seen on the two-dimensional color image become sampled and are displayed on the spectral image. This artifact can be prevented by decreasing the sample gate size.

Download as PowerPointOpen in Image Viewer

Spuriously Elevated Peak Velocity

While performing SDI, the examiner must always maintain an angle of correction of 45°–60° within the sample volume to avoid introducing significant error in the calculation of the PSV. At this range, there is a nearly linear relationship between velocity and Doppler shifts. At angles higher than 60°, minor errors in angle accuracy can result in large errors in velocity estimations (25,30). Ultrasound beam steering may lead to an improved Doppler angle.

Conclusion

With increasing technologic advances and innovations, Doppler US has become the primary tool for detecting and characterizing blood flow in clinical practice. In experienced and knowledgeable hands, this technique can have a pivotal role in the diagnosis and monitoring of vascular conditions. A thorough understanding of the strengths and weaknesses of available Doppler modes and the techniques used for image optimization and artifact recognition and avoidance can enable US operators to confidently use Doppler US to its full potential. This understanding may also ensure the widespread growth of this modality in this era of increased awareness regarding radiation dose and cost containment.

Disclosures of Conflicts of Interest.— J.S.P.Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: research grant from GE Medical Systems. Other activities: disclosed no relevant relationships.

Acknowledgments

The authors thank Mark Saba for assistance with the illustrations; Henry Douglas for help with the images; Lei Wang for help with video recording; Alexandria Brackett for assistance with literature searches; Melody Polio, Tricia Haggerty, Victoria Clifford, and Rachel Katz for help obtaining images; Mary Jo Smallwood, Cindy Rapp, and Andrew Hatch for help with education activities regarding the new US vascular platforms; and Kevin M. Johnson, MD, for insights regarding the content of this article.

Recipient of a Certificate of Merit award for an education exhibit at the 2017 RSNA Annual Meeting.

For this journal-based SA-CME activity, the author J.S.P. has provided disclosures all other authors, the editor, and the reviewers have disclosed no relevant relationships.

disclosed no relevant relationships.research grant from GE Medical Systems.disclosed no relevant relationships.