| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2002 ARRS Executive Council Award |
1 All authors: Department of Radiology, University of Michigan Hospitals, 1500 E. Medical Center Dr., Ann Arbor, MI 48109-0030.
Received April 5, 2002;
accepted after revision June 24, 2002.
Presented at the annual meeting of the American Roentgen Ray Society,
Atlanta, April—May 2002.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Struvite (ammonium magnesium phosphate) stone fragments, wire mesh, and a flat surface were scanned in a water bath with a sonography scanner using a high-frequency linear array probe fixed in a ring clamp. Pulse repetition frequency, color-write priority, gray-scale gain, and spectral Doppler gain were varied. Color and spectral Doppler modes were used.
RESULTS. Twinkling artifact and spectral broadening were seen most intensely behind struvite stone fragments, and both were seen more strongly behind wire mesh with greater surface roughness than behind wire mesh with less surface roughness or a flat surface. The appearance of the twinkling artifact is highly dependent on machine settings. System noise measured on a flat surface generates a band-limited Doppler shift on spectral displays with a mean frequency shift of 0 Hz and a mean (± SD) absolute fluctuation of 86 ± 10 Hz over a pulse repetition frequency range of 1250-10,000 Hz. Rough surfaces increase the spectral bandwidth.
CONCLUSION. The appearance of the twinkling artifact is highly dependent on machine settings and is likely generated by a narrow-band, intrinsic machine noise called phase (or clock) jitter. Surface roughness secondarily broadens the noise spectrum. With a strongly reflecting, rough surface such as a renal stone, the high amplitude, broadband signal appears as random motion in color Doppler sonography. Understanding of the twinkling artifact may result in better use of its clinical appearance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Since its initial description, the twinkling artifact has been reported mainly in association with nephrolithiasis [2, 3]. The incidence of twinkling artifact with certain compositions of stones (calcium phosphate and calcium oxalate dihydrate) [3] and the occurrence of the artifact in different anecdotal cases [4, 5] have been reported in only a handful of studies.
Present knowledge of twinkling artifact is limited, and the artifact has not been well characterized. Its cause has only been speculated. Rahmouni et al. [1] proposed that the artifact is generated by rough surfaces with multiple reflectors splitting the incident beam into a complex beam pattern. This was thought to create an increased pulse duration of the received radiofrequency signal that was then interpreted as movement [1]. However, roughness alone is not enough to produce this artifact, and although Rahmouni et al. briefly commented that some fluctuation is needed in the incidental beam, they did not describe how this fluctuation occurs. Little is known about the twinkling artifact, and its characterization with various machine settings has not been investigated. Our goal was to further understanding of the twinkling artifact by systematic evaluation. In addition, we propose a new cause of the artifact. Understanding of the twinkling artifact may result in better use of its clinical appearance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Fragments from a struvite stone were embedded in a form of modeling adhesive (Paper Tak; Ace Hardware, Oak Brook, IL) to ensure that the stone did not move during scanning (Fig. 1). This procedure also ensured that the object was not affected by ultrasound radiation force. Stone fragments were analyzed using color Doppler sonography. Three random positions on various stone fragments were scanned. The color Doppler box was chosen to encompass the entire stone as well as all artifact seen behind the stone. Color-write priority, pulse repetition frequency, and gray-scale gain were varied systematically. The wall filter was set at "low" and color gain was set at 85%, which are both standard default settings. Doppler frequency was 6 MHz, the standard default setting for the transducer configuration. Color-write priority was serially increased around the center value of zero from a minimum value (-8) to a maximum value (+8) for a total possible color-write priority setting of 17. At each increment of color-write priority, the pulse repetition frequency values of 700, 1000, and 1500 Hz were recorded. We chose these pulse repetition frequency values because they are standard defaults for color Doppler sonography. Gray-scale gain on the Philips HDI 5000 machine is controlled by a knob with 42 possible positions and a set of slide pods for the time-gain compensator at each depth position. To ensure consistency, we set the time-gain compensator at all depth positions at maximum so that only the rotating knob controlled gray-scale gain. We could use this setting because our targets were total reflectors, and there was essentially no overlying attenuation in the water bath. The gray-scale gain knob was turned three clicks between any gray-scale gain measurements. Thus we made measurements at a total of 15 different gain settings during any measurement sequence in which gain was the independent variable.
|
When collecting color Doppler data for analysis, we filled a 117-frame cine loop at each gain setting. To ensure that there was no selection bias in the images chosen for analysis at each gain setting, we always analyzed images 20, 40, 60, 80, and 100 of each cine loop for a total of five different images at each setting.
Each image was processed using software developed in Matlab (Mathworks, Natick, MA). For the region of interest in the color Doppler box, the total number of color pixels was computed. The mean and SD were computed over the five images for each setting. Plots were generated of the number of color pixels at each stone site as a function of color-write priority and as a function of gray-scale gain. A different plot was generated for each of the three pulse repetition frequency values for each of the three different stone positions or nine total plots. A linear regression line was then calculated for each plot to show if a relationship existed between the magnitude of the twinkling artifact and the gray-scale gain and color-write priority. The linear regressions were then compared by determining if there was overlap of the 95% confidence intervals (CIs) of their slopes. If there was no overlap, the regressions were considered different.
Stone fragments were then analyzed using spectral Doppler sonography. Three separate areas were again scanned. Interrogation points were chosen in those areas in which twinkling artifact was present as seen on color Doppler sonography. Spectral Doppler-gain settings were systematically increased and recorded at 59%, 65%, and 70%. These settings were selected as a general range of settings that showed the broadest range of spectral waveforms before saturation. The machine does not allow settings to be selected at 60%; therefore, a 59% setting was used at this level. Higher gains frequently tended to saturate, and lower gains often produced too little signal. Spectra were sampled at the "slow" scrolling rate, which filled the spectral window in 10 sec, amounting to 518 individual spectral lines for each 10-sec sample. Unfortunately, spectral saturation could occur at higher spectral Doppler gain settings in an unpredictable fashion because of the strong reflections from the stone's surface. When waveforms saturated during acquisition before 10 sec, the acquisition was shortened to the length of time allowed by the machine just before saturation. Statistically, this shortening did not matter because we still had literally hundreds of independent, randomly distributed data samples to analyze. At each spectral Doppler gain setting, pulse repetition frequency settings were varied systematically at 1250, 1500, 2500, 3731, 5000, 6250, 8333, and 10,000 Hz. The lowest pulse repetition frequency setting allowed in spectral Doppler mode by the machine was 1250 Hz, which was sequentially increased in increments as allowed by machine settings. Images were acquired with the scale units in frequency mode.
The spectral images were digitized and processed using Matlab. For each image, the spectra were averaged over time, and then the mean absolute frequency and SD corresponding to the positive or negative frequency bands were computed. The mean overall frequency shift was not used because it was zero. Three independent experts voted on the area that showed the most spectrally broadened waveforms. If disagreements occurred, they were resolved by consensus. The area that was thought to represent the most spectrally broadened waveforms was then analyzed using linear regression and was compared with a flat surface and wire mesh.
Next, we scanned a flat surface in a water bath using the same sonography machine and fixed probe. The bottom of the water bath, which was made of plastic, was used as the flat surface. Machine settings were systematically varied, and color Doppler images and spectral images were obtained. On spectral images, Doppler gain settings were systematically increased and recorded at 59%, 65%, and 70%. At each Doppler gain setting, pulse repetition frequency settings were varied from 1250 to 10,000 Hz, like those for the stones. The spectral images were also digitized and processed the same as the stone data, using regression analysis.
To simulate different surface roughnesses, stainless steel wire mesh of varying wire diameter and mesh density (mesh 40, wire 0.01 inch [0.25 mm], mesh 60, wire 0.0075 inch [0.19 mm], and mesh 60, wire 0.0045 inch [0.11 mm]; [The Cleveland Wire Cloth & Manufacturing, Cleveland, OH]) were affixed to a flat metal plate securely with four paper clips and scanned at three different positions. (Mesh refers to the number of holes per inch. Wire corresponds to the thickness of the wire surrounding the holes.) Initial gray-scale scanning confirmed that the mesh was flat on the metal plate without gaps at the scanning site. Evaluation was performed using spectral imaging. Interrogation points were chosen in those areas that showed twinkling artifact on color Doppler sonography. To control for color-write priority when detecting twinkling, we kept color-write priority constant at zero during these acquisitions. Spectral Doppler gains were systematically increased and recorded at 59%, 65%, and 70%, as before. At each spectral Doppler gain, acquisitions were obtained at pulse repetition frequency values ranging from 1250 to 10,000 Hz. Spectral images were obtained with these systematic variations in machine settings and processed to determine the absolute mean frequency and SD of the noise spectrum either above or below the baseline just as for stones. The SD and absolute mean values would be representative of spectral broadening as a function of pulse repetition frequency. Again, the overall mean for the entire noise spectrum was zero. Three independent experts voted on the sampled area that showed the most spectrally broadened waveforms, and any disagreements were resolved by consensus. The area that was believed to contain the most spectrally broadened waveforms was then analyzed and compared with the flat surface and the most broadened spectrum from the stone on the basis of a linear regression line drawn through the data.
To increase the apparent roughness further, we clipped two wire meshes together (mesh 40, wire 0.01 inch and mesh 60, wire 0.0075 inch) and then secured the wire meshes to a flat metal plate and scanned at three independent positions. Positions scanned with the two meshes on top of each other were chosen for the closest contact with the affixed flat metal plate surface. Evaluation was performed using spectral imaging, and interrogation points were chosen in those areas that showed twinkling artifact on color Doppler sonography. Images were obtained exactly the same way as with the single meshes described previously.
Spectra from the stone fragments and mesh were then plotted against pulse repetition frequency. These data were compared with spectra of a flat surface by plotting linear regressions through the spectral data for each surface and determining the slopes of these lines with their 95% confidence limits (CLs). Differences in slope between the regression lines were taken as differences in spectral broadening and hence roughness of the surfaces. Slopes were considered different if their 95% CLs did not overlap.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
These effects are shown for pulse repetition frequencies 700, 1000, and 1500 Hz in Table 1. The measurements are made at the maximum (+8), middle (0), and minimum (-8) color-write priorities. Each cell contains the slope of the best-fit regression line through the number of color Doppler pixels representing the twinkling artifact as a function of gray-scale gain at a given pulse repetition frequency and color-write priority, as seen in Figure. 3. The 95% CIs around the slope are shown. The appearance of twinkling artifact varies with all the measured parameters. Notice, for example, that the number of color pixels in the twinkling artifact increases with increasing gray-scale gain at a color-write priority of +8, but decreases with increasing gray-scale gain at -8. The dependence of color gain and color-write priority has a lesser effect at a priority of zero. At a color-write priority of zero and a pulse repetition frequency of 700 Hz, the 95% CI of the regression line includes zero, suggesting no gray-scale gain dependence. Spectral Doppler imaging of the stone fragments showed a strong broadband signal that appeared to alias, on the basis of filling of the entire frequency band at a given pulse repetition frequency. This aliasing was especially seen at lower pulse repetition frequency values (1250-5000 Hz). At higher pulse repetition frequency values (5000-10,000 Hz), the broadband signal was less constant, depending on the intensity of the twinkling artifact but still could randomly appear to alias (Fig. 4). In particularly strong areas of twinkling artifact, spectral broadening at all pulse repetition frequency values was present (Fig. 4).
|
Flat surfaces did not show any color Doppler twinkling artifact. Spectral Doppler imaging of the flat surface showed only narrow-band signal with consistent absolute values above and below the baseline at all pulse repetition frequency values and no broadband signal (Fig. 5). The mean absolute value of the noise from the flat surface was 86 ± 10 Hz using the absolute means of the positive or negative noise bands over the sampled pulse repetition frequency range (Fig. 6). The slope of the narrow-band signal did not vary with pulse repetition frequency or Doppler gain to at least a factor of approximately 2 x 10-3 kHz/Hz, although the intensity increased with Doppler-gain settings. In other words, the slope was essentially zero to the limits of our measurements and was statistically different from the rough surfaces (Table 2).
|
|
|
|
Twinkling artifact was seen in wire meshes affixed to a flat metal surface especially with mesh 40, wire 0.010 inch (Fig. 7) and less so with mesh 60, wire 0.0075 inch. Mesh 60, wire 0.0045 inch did not show any twinkling artifact. When mesh 40, wire 0.010 inch was superimposed on mesh 60, wire 0.0075 inch, a slightly more intense twinkling artifact was present as well as further broadening of the spectral wave-forms. Subjectively, the twinkling artifact and spectral broadening were not as intense on wire mesh as on stone fragments, however.
|
Spectral Doppler sonography of the wire meshes that showed twinkling artifact on color Doppler sonography revealed a broadband signal at lower pulse repetition frequency values (1250-2500 Hz) (Fig. 8). At higher pulse repetition frequency values, the broadband signal was not definitely seen, and only a narrowband signal was seen. However, there was clearly a trend toward more broadening with rougher surfaces, with the stone fragments showing the most broadening and the flat surface showing the least broadening on the basis of no overlap of the 95% CIs of the slopes of the regression lines through the curves of the different surfaces (Table 2). This trend was further manifested by the increasing absolute values of the slopes of the regression lines with roughness (Table 2 and Fig. 9).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The appearance of color Doppler signal is affected by several variables including color-write priority, gray-scale gain, and pulse repetition frequency, among other things, in addition to the motion of the object scanned with respect to the transducer. Conventional Doppler evaluations reflect the velocity of the object based on a frequency shift in the transmitted signal due to motion of the target. Therefore, underlying motion could potentially influence the appearance of the twinkling artifact. However, in our investigation, the transducer was clamped to a stationary ring stand, and the objects we scanned were stationary. This procedure ensured that outside motion did not affect our results and that our results were restricted to machine effects alone.
We found that the appearance of the twinkling artifact is highly dependent on machine settings and is also probably dependent on the equipment used. We could clearly show the trends, but some effects were highly nonlinear and somewhat surprising (Table 1). For instance, one would expect that color-write priority and gray-scale gain should be interrelated. In all sonography machines, color-write priority decides whether to write a pixel as color or as gray scale. If the gray-scale signal is above some threshold, the pixel is written as gray. If it is below the threshold, it is written as color. Hence, the higher the gray-scale signal, the less color written in an image. We would, therefore, expect that the color-write priority would be directly related to the twinkling artifact, and gray-scale gains would be inversely related to the twinkling artifact. That is, the artifact should increase in prominence with increasing color-write priority, but it should become less prominent with increasing gray-scale gain. This outcome however, did not seem to be the case in general. As the gray-scale gain increased, the degree of visible artifact varied with color-write priority and gray-scale gain. At low color-write priority, color-write priority and gray-scale gain were inversely related as would be expected—that is, low gain and high priority gave the most artifact. At a middle range priority, they were close to independent—that is, the amount of color seemed more independent of gain. At high gain, the amount of artifact was directly related to write priority, an outcome that was not expected.
The cause for this unusual relationship is difficult to explain except that the color-write priority is a manufacturer-designed feature, and the influence of the color-write priority on the appearance of the artifact can vary depending on what priorities and contingencies a given manufacturer uses in its design. However, it is fair to say that the appearance of the artifact is highly dependent on the gray-scale gain and color-write priority settings (Fig. 2).
In addition, even given our systematic sampling, we believe that it is likely that changing one parameter actually changed several different functions in the machine. However, we used a clinical machine because this issue is a clinical one. We did not design and test our own sonography scanner, in which we could control every facet of the imaging. Thus although we tested a range of parameters in our investigation, it will likely be difficult to extrapolate our quantitative results directly to other machines or even different transducers on the same machines. For example, the Doppler output probably decreased automatically with increasing pulse repetition frequency in our machine (Schwartz G, personal communication). Hence, although we tried to control one variable at a time, other variables may have changed without our knowledge during our experiments.
The plots comparing artificial surfaces of progressively increasing roughness and stones show that increasing roughness broadens the Doppler spectrum (Fig. 9). The smoother the surface, the lower the absolute value of the slope of the regression lines through Doppler frequency as a function of pulse repetition frequency (Table 2). None of the regression line slopes overlap at the 95% CLs. We can therefore conclude that the slopes were different; hence, the spectral broadening is a consequence of the roughness of the surface. The smooth surface had essentially a near-zero slope, approximately 2 x 10-3 units in the scale we used (kHz/pulse repetition frequency in units of Hz). Thus, the amount of broadening in the smooth surface was nearly independent of pulse repetition frequency.
The differences in slopes of the regression lines require some explanation (Table 2). Clearly, the higher the pulse repetition frequency, the larger the broadening that can be displayed, where the broadening by our measures ranges from 0 Hz to either the upper or lower aliasing frequency. The aliasing frequency increases with increasing pulse repetition frequency, so more broadening can be shown. Thus, a regression line through spectra produced from a smooth structure with little broadening would not alias and, therefore, would not be influenced by pulse repetition frequency. Non-aliased waveforms at low pulse repetition frequencies will clearly not alias at higher pulse repetition frequencies. However, as spectra get broader and broader, they will have more and more pulse repetition frequency dependence by aliasing at progressively higher pulse repetition frequencies. The regression lines would start with a negative slope when aliasing at low pulse repetition frequencies but then would turn positive as the broadening occurs at the highest frequencies as well. We saw this pattern.
Despite the experimental variations in the color Doppler sonographic appearance of the artifact, the qualitative findings still hold, and the relationships of the multiple parameters and the Doppler spectra that we have studied to the twinkling artifact still exist. The primary findings are that there is an intrinsic, band-limited noise in Doppler sonography that can be seen when scanning even flat surfaces (Fig. 5). This noise can be seen near the baseline when scanning rough surfaces as well (Fig. 8). When a rough surface is scanned, the noise spectrum broadens. These two findings are consistent with what we observed in the literature, given the artifact signal that has been seen along the bony surfaces of the calvarium on transcranial Doppler sonography [5] or highly reflective hemangiomas in the liver [6] and the random, broad-banded color signal typical of the twinkling artifact behind stones [1, 2].
The narrow-band artifacts are likely generated by intrinsic machine noise caused by what is referred to as phase (or clock) jitter. Phase jitter is caused by the slight random time fluctuations in the digital clock that synchronizes the firings of the ultrasound pulse transmissions. Because rates of change in phase represent Doppler shifts in color and spectral Doppler signals, if there is a fluctuation in the clock sampling of a signal, there will be a perceived Doppler shift even if there is no true motion. This fluctuation must be small, or the noise would dominate the true Doppler signal generated by moving targets. In our experiments, the mean absolute error was 86 Hz (Fig. 6). Given the Doppler carrier frequency of 6 MHz, a fractional error of 86 Hz out of 6 MHz is 1.4 x 10-5 or 14 parts per million (ppm). The cycle period for a 6-MHz transducer is 1.7 x 10-7 sec. With an error of 14 ppm, the clock error corresponding to 86 Hz is 2.4 x 10-12 sec or 2.4 psec. This is obviously a small interval. However, this slight error would explain the narrow-band spectral waveform that was seen on flat surfaces. The variations of this error frequency were largely independent of pulse repetition frequency, as one would expect (Fig. 6), and in many cases, this small frequency shift will not be visible as a color Doppler artifact because of the wall filter. However, because the artifact signal is many orders of magnitude greater in amplitude than standard Doppler motion produced by moving blood, this signal could "break through" the wall filter in some circumstances and produce apparent flow along large reflecting objects such as renal stones or flat bones even with relatively high wall filters. Notice that the error can be as high as 150 Hz or higher at different times and pulse repetition frequency values (Fig. 6). Thus, unless the wall filter was greater than these extreme values, this noise would be definitely visible and may be particularly noticeable in power Doppler sonography, in which the imager is optimized to show flow [7], although signal should be present in standard color Doppler sonography as well.
So how does this narrow-band signal error become a broadband "twinkling artifact"? As suggested by Rahmouni et al. [1], some fluctuation of the sound field must occur to generate the apparent Doppler shift. Clearly, we now have the fluctuation with phase jitter. However, this fluctuation is independent of the complexity of the field, and it is much less broadband than the standard twinkling artifact. We propose that the twinkling artifact is probably generated through slight variations in path length of the transmitted and reflected ultrasound. Focusing of an ultrasound array transducer depends on precise phasing among the elements in the scanhead array. Even minor phase errors could produce fluctuations in the ultrasound field generated by a transducer. These fluctuations could easily alter the effective beam path slightly from firing to firing, although the beam will be pointed in the correct direction on average. The same types of errors only increase on reception of the reflected wave in which the same phasing issues also occur.
A crude estimate of the degree of directional variations one might expect
with a 2.4-psec phase jitter may be obtained as follows: the standard equation
of the size of the focal spot for a sonography scanner is 1.22 
F#,
where is the wavelength and F# is focal length divided by the
diameter of the aperture. Although this is the equation for a circular
aperture, it is similar for a linear array. Multiplying 2.4 psec times the
speed of sound in tissue, 1540 m/sec, we can estimate the variational
displacement across the assumed flat plane in the focus with perfect phasing.
This variation is approximately 3.7 x 10-6 mm or 3.7 nm.
Assuming a focal length of 1 and a wavelength of 250 µm for the 6-MHz
transducer, we calculated a focal plane diameter of approximately 305 µm.
Using these numbers to calculate the tangent of the angle between the two
phase fronts, one with perfect phasing and one tipped by a 2.4-psec phase
jitter, we obtained an angle of roughly 0.0007°.
Slight variations such as the one shown will generally have no visible
effect in a continuous medium. However, if there is a strong reflector with a
rough surface, these slight variations in beam direction could be magnified to
produce apparent aliased Doppler shifts
(Fig. 10). Multiple
reverberations would further magnify this effect by projecting the artifact
below the reflecting surface. Although the two vertical arrows are slightly
angularly displaced in the figure, the sonography machine thinks they are
along identical paths. If the difference in distance, 
L, in the paths
from the transducer to the rough surface is one half of a wavelength, the
reflected signal would appear as aliased motion in a Doppler mode. For a 6-MHz
transducer, this path length is only approximately 125 µm. Multiple
reflections from each transmission would be compared and produce multiple
different randomly distributed positive and negative Doppler shifts—that
is, the twinkling artifact in color Doppler sonography.
|
Our experiments were limited by the inability to control all machine settings separately as described previously. Additionally, we did not use an optical table, so there could have theoretically been external motion transmitted into the imaging system. We doubt that this is significant because of the consistency of our results, which we would have expected to vary from day to day and instant to instant if external sources were influencing the results. In addition, the amount of motion restraint we used is at least as much as has been used by others [2, 3]. Finally, the only really narrow-band signals we saw in the Doppler spectra were the phase-jitter artifacts. The broadband, roughness-influenced signals were either aliased or not visible at each pulse repetition frequency. This finding could be due to our relatively coarse pulse repetition frequency sampling. We were limited to the pulse repetition frequency values assigned by the machine, or a nonlinear effect could still influence the Doppler spectra from roughness. This issue will require further investigation.
In conclusion, the appearance of the twinkling artifact is highly machine- and setting-dependent. The underlying cause is likely a narrow bandwidth noise introduced by phase (or clock) jitter in the Doppler circuitry of the sonography scanner. The effect of roughness appears to be only secondary and serves to broaden the spectrum. The effect is most commonly seen in urinary stones, and the rougher the surface, the greater the twinkling artifact. The usefulness of the twinkling artifact remains to be investigated. However, increased understanding of the twinkling artifact may allow better evaluation and decision making when the twinkling artifact occurs in the clinical setting. In the future, this understanding could possibly lead to a means of measuring surface roughness.
Acknowledgments
We thank Gary Schwartz of Philips Ultrasound and Matthew O'Donnell of the
University of Michigan for their insightful comments and Seema Sonnad for her
help with statistics.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Andrulli, A. Turrin, M. C. Bigi, P. Ravani, A. Trinchieri, and F. Locatelli Colour Doppler twinkling in kidney stones: artefact or sign? NDT Plus, October 13, 2009; (2009) sfp145v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. O'Flynn and P. S. Sidhu The Sonographic Twinkling Artifact in Testicular Calcification J. Ultrasound Med., April 1, 2009; 28(4): 515 - 517. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Tchelepi and P. W. Ralls Color Comet-Tail Artifact: Clinical Applications Am. J. Roentgenol., January 1, 2009; 192(1): 11 - 18. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Suh, J. J. Cronan, and J. M. Monchik Primary Hyperparathyroidism: Is There an Increased Prevalence of Renal Stone Disease? Am. J. Roentgenol., September 1, 2008; 191(3): 908 - 911. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Brunese, A. Romeo, S. Iorio, G. Napolitano, S. Fucili, B. Biondi, G. Vallone, and A. Sodano A New Marker for Diagnosis of Thyroid Papillary Cancer: B-Flow Twinkling Sign J. Ultrasound Med., August 1, 2008; 27(8): 1187 - 1194. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Viduetsky Unexpected Doppler Waveform Patterns in the Lower Extremity Arteries Journal of Diagnostic Medical Sonography, September 1, 2007; 23(5): 293 - 297. [Abstract] [PDF]
|
|
|
|
|
|
S. J. Jeon and S. E. Yoon Color Doppler Twinkling Artifact in Hepatic Bile Duct Hamartomas (von Meyenburg Complexes). J. Ultrasound Med., March 1, 2006; 25(3): 399 - 402. [Full Text] [PDF]
|
|
|
|
|
|
A. Kamaya, S. Abate, B. Nan, I. Grover, R. S. Adler, D. Jamadar, and J. M. Rubin Characterization of a Linear Streak Artifact With Pulse Inversion Tissue Harmonics in Musculoskeletal Sonography J. Ultrasound Med., December 1, 2004; 23(12): 1597 - 1605. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Roubidoux, M. S. Sabel, J. E. Bailey, C. G. Kleer, K. A. Klein, and M. A. Helvie Small (<2.0-cm) Breast Cancers: Mammographic and US Findings at US-guided Cryoablation--Initial Experience Radiology, December 1, 2004; 233(3): 857 - 867. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Mitchell, M. A. Pozniak, J. Zagzebski, and M. Ledwidge Twinkling Artifact Related to Intravesicular Suture J. Ultrasound Med., December 1, 2003; 22(12): 1409 - 1411. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), 40 and 60 wire meshes attached to each other (
), and
flat surface (+). Each symbol represents mean absolute spectral frequency
above or below baseline at each pulse repetition frequency at Doppler gain of
65% for given structure. Error bars correspond to SDs. Different slopes of
regression lines reflect different levels of spectral broadening as function
of roughness. This clearly shows that stone had most broadening.
Dot—dashes represent regression line for stone, dots alone represent
regresion line for 40 + 60 mesh pair, and solid line represents regression
line for flat surface.
L is difference in path length. If path from transducer to rough
surface varies one-half wavelength between firings, reflected signals would
appear as aliased motion in spectral Doppler mode.