Hiromi Kumon md phd, Department of Urology, Okayama University Graduate School, 2-5-1 Shikata, Okayama 700-8558, Japan. E-mail: email@example.com
Abstract Background: We have developed velocity-flow urodynamics using Doppler sonography based on the hypothesis that microbubbles formed in the urethra are responsible for Doppler signals. In order to confirm this hypothesis derived from Bernoulli's principle, we investigated the simultaneous detection of cavitation noise and Doppler signals in an experimental system.
Methods: An experimental circuit was built in which a stenosis was created using a glass or silicon tube with tap water used as the sample fluid. Doppler signals, pressure before and after the stenosis, flow rate, flow velocity and cavitation noise were measured. Direct detection of cavitation with a high-speed charged-coupled device (CCD) camera was conducted in the glass tube. The relationship between cross-sectional area and flow velocity in terms of the detection of Doppler signals was analyzed in the silicon tube study.
Results: In the glass tube study, a high-speed CCD camera clearly detected masses of microbubbles associated with cavitation. The range of flow rates creating cavitation completely corresponded with those producing Doppler signals detected by ultrasonography. A similar correlation was observed in the silicon tube study, which showed that a low flow velocity of 41.5 cm/sec through a stenosis with a cross-sectional area of 20 mm2 created Doppler signals at a flow rate of 8.3 mL/sec.
Conclusion: The results of the present study confirmed that microbubbles created in flowing urine are responsible for Doppler signals. Measurement of velocity-flow urodynamics has great potential to become a non-invasive and reliable alternative to conventional pressure- flow urodynamic studies.
It has been previously reported that Doppler ultrasonography could visualize urine flow in the human urethra, resulting in the establishment of a method of assessing velocity-flow urodynamics using Doppler ultrasonography.1–4 It was believed to be impossible to measure intraurethral urine flow velocity because urine does not contain particles that can reflect ultrasound signals.5 However, our initial experiment using an artificial urethra showed that Doppler signals could be detected, even though the fluid did not contain any particles.1 Based on this observation, we have employed the transperineal approach to establish velocity-flow parameters for the diagnosis of bladder outlet obstruction (BOO) in men.2–4 According to Bernoulli's principle, a decrease in pressure would be caused by accelerated urine flow from the bladder into the urethra and we hypothesized that this would lead to the creation of microbubbles associated with cavitation that could reflect Doppler signals.1,6 In order to confirm this hypothesis, we investigated the occurrence of cavitation and Doppler signals by using a glass or silicon tube as an experimental urethra. We attempted the simultaneous detection of microbubbles due to cavitation with a charged-coupled device (CCD) camera and Doppler signals with an ultrasound scanner. Cavitation noise and cavitation hysteresis were also examined.
Glass tube study
An experimental circuit was built as shown in Figure 1a. A transparent glass tube was placed in the circuit and a pump was used to circulate a sample fluid through the circuit from a tank. Tap water was used as the sample fluid and was allowed to stand in the tank with the top open for at least 10 h before starting the experiment. A glass tube with a narrow segment (inside diameter, 15.0 mm; inside diameter at the stenosis, 2.25 mm) was used to examine cavitation (Fig. 1b). The pressure before and after the stenosis, the flow rate and temperature of the fluid as well as the noise arising from cavitation were measured. Development of cavitation within the tube was investigated by using a high-speed (CCD) camera (CCN-2712Y, Ytechdesign Co., Diepikon, Switzerland) and a Doppler ultrasound machine with a 3.75-MHz scanner (SSA-340, Toshiba Medical Co., Tokyo, Japan). The scanner was placed vertically on the polyvinyl tube of the circuit slightly distal to the glass tube. The recording site was wrapped with an ultrasound-transmitting material (Stand-off pad, Kitecko, Tokyo, Japan) and the angle for the measurement of flow velocity was set at 45 degrees. The flow rate was measured with an impeller-type flowmeter (TUR-15 L, Japan Flow-cell Co., Tokyo, Japan), which had a range of 1.0–16.0 L/min with an accuracy of ±3%. A microphone was mounted on the glass tube to record the noise arising from cavitation on a video deck simultaneously with the ultrasonographic recording. Spectral analysis of the recorded noise was performed using a fast Fourier transformation (FFT) analyzer (AD-3524, A&D Co., Tokyo, Japan).
Silicon tube study
Instead of the glass tube, a 350-mm silicon tube (inside diameter of 6 mm and wall thickness of 1 mm, Fuji Systems Co., Shizuoka, Japan) was used to create experimental conditions that were closer to the physiological state (Fig. 2a). A Hoffman pinch clamp with sandwich boards 10 mm wide was used to compress the tube in order to create a stenosis. The cross-sectional area of the stenosis was controlled by the height of the tube and calculated mathematically (Fig. 2b). The flow rate was set by using an impeller-type flowmeter (TUR-10S, Japan Flow-cell Co., Tokyo, Japan), which had a range of 0.2–3.0 L/min and an accuracy of ±3% on the full scale. Four flow rates (0.5, 1.0, 1.5, and 2.0 L/min; 8.3, 17.0, 25.0, and 33.0 mL/sec) were employed to observe Doppler signals and measure cavitation noise via the auditory output for pulsed Doppler ultrasound. These flow rates are relevant to ordinary urine flow rates.
Glass tube study
Ultrasonographic images obtained at various flow rates are shown in Figure 3. When the flow rate was increased from 3.6 to 4.9 L/min, clear Doppler signals were suddenly detected from a flow rate of 4.4 L/min. Conversely, when the flow rate was decreased from 4.9 to 3.6 L/min, Doppler signals were no longer detected at flow rates below 4.4 L/min, demonstrating cavitation hysteresis. When the flow rate was fixed so that clear Doppler signals were detected, no signals were obtained upstream of the narrowing in the tube.
Using a high-speed CCD camera, pictures of cavitation were also obtained at flow rates of 4.4 L/min or higher, when the rate was increased from 3.6 to 4.9 L/min (Fig. 4). After cavitation began to occur, it was detected continuously and masses of microbubbles were seen to expand and disperse downstream. The range of flow rates producing cavitation detected by the CCD camera completely corresponded with the flow rates producing Doppler signals.
The results of FFT analysis of the cavitation noise are shown in Figure 5. When the flow rate was increased gradually, the noise level showed a sudden increase at 4.4 L/min. Conversely, the noise disappeared at 3.6 L/min when the flow rate was decreased gradually. This high frequency noise was thought to be created by the occurrence and disappearance of cavitation and it showed peaks at three common frequencies (5.5, 8.4 and 13.5 kHz) at all flow rates tested. The size of the cavitation bubbles was calculated to be 0.6, 0.4, and 0.2 mm at the respective frequencies according to the model described by Rayleigh.7 Thus, this acoustic analysis also confirmed that the Doppler signals were closely related to the occurrence of cavitation.
The flow pressure study demonstrated that a decrease of at least 138 kPa across the pressure gradient (downstream vs upstream pressure) was needed for the onset of cavitation when the flow rate was increased gradually at 22°C. When the flow rate was decreased gradually at 22°C, the pressure gradient required for the maintenance of cavitation became smaller and the difference between the two states was 5.9 kPa, clearly demonstrating the existence of cavitation hysteresis. The flow velocity measured by ultrasonography correlated extremely well with that obtained using the flowmeter at various temperatures, including 22°C, 27°C and 37°C, and the maximum difference between the two methods was ±2% (data not shown).
Silicon tube study
Figure 6 demonstrates the results of FFT analysis of cavitation noise at a flow rate of 17 mL/ s. Although there was no change in the high frequency range, more noise was detected in the low frequency range at higher velocities (92–225 cm/sec) compared with low velocity (59 cm/sec). In the flow pressure study performed at a flow rate of 17 mL/sec, a 2.4-kPa decrease of the pressure gradient (downstream vs upstream pressure) was noted at a flow velocity of 92 cm/sec. The decrease tended to show a plateau around 5.7 kPa at a flow velocity of 155 cm/sec. Similar changes associated with the detection of Doppler signals were also observed at the other flow rates studied (data not shown). Detection of Doppler signals was also analyzed while the pinch clamp was closed or opened gradually. The cross-sectional area of the stenosis was calculated from the height of the tube and the average flow velocity within the stenotic portion was also calculated at flow rates of 8.3, 17.0, 25.0 and 33 mL/sec. The detection of Doppler signals was plotted on curves representing the relationship between the calculated cross-sectional area of the tube and the flow velocity at each flow rate (Fig. 7). A higher flow velocity was associated with more constant Doppler signals at each flow rate and signals could even be detected when the cross-sectional area was 28 mm2 (no clamp). Doppler signals could also be detected at a lower flow velocity when this was combined with lower flow rates. For example, a flow velocity of 41.5 cm/sec with a cross-sectional area of 20 mm2 led to the detection of Doppler signals at a flow rate of 8.3 mL/sec.
We have developed a method of assessing velocity-flow urodynamics using Doppler ultrasonography and have successfully applied it to the clinical diagnosis of bladder outlet obstruction.1–4 According to Bernoulli's principle, an increase in flow velocity produces a decrease of pressure that leads to the formation of microbubbles in the urine.6 Our initial study using an experimental urethra suggested that the decrease of pressure at a high velocity caused dissolved gas in the urine to form microbubbles.1 However, Wang et al. suggested that the specific gravity is an important factor in detecting urine flow by Doppler ultrasonography.8 Since various factors might influence the detection of Doppler signals in the urethra, it is important to confirm whether the creation of microbubbles via a basic hydrodynamic phenomenon is the principal mechanism underlying these signals in order to expand the clinical application of ultrasound urodynamics.
In the present study, the simultaneous detection of cavitation and Doppler signals was investigated. Our experiment using a glass tube demonstrated that microbubbles specific to cavitation could be clearly detected with a CCD camera and corresponded with the detection of Doppler signals by ultrasonography. Due mainly to the rigidity of the glass tube, cavitation occurred at a far higher flow rate (4.4 L/min; 73 mL/sec) than the range during actual voiding and a gradient of at least 138 kPa (downstream vs upstream pressure) was needed when the flow rate was increased gradually at 22°C. Although these data cannot be applied directly to the human urethra, the existence of cavitation hysteresis was clearly demonstrated. Our previous study showed that it was sometimes difficult to obtain a Doppler signal in the proximal prostatic urethra, especially in patients with an enlarged prostate, although a clear signal was obtainable in the distal prostatic urethra.3 This observation suggests that the accumulation of microbubbles during flow through the urethra leads to clear Doppler signals in the distal prostatic urethra and membranous urethra, because microbubbles that develop inside the urethra would tend to persist.
In the silicon tube study, cavitation developed at a flow rate closer to that in the human urethra during actual voiding.9 A flow rate of 8.3 mL/sec and a flow velocity of 41.5 cm/sec with a cross-sectional area of 20 mm2 are parameters frequently observed in clinical Doppler ultrasounds.3 Although we assumed that the cross-sectional shape of the silicon tube was oval, it would actually have become peanut-shaped when the pinch clamp was closed gradually.10 The presence of stenosis with a complicated shape would produce an inhomogeneous flow that would contribute to cavitation, and the above-mentioned flow velocity should represent the mean velocity of complex flow through the stenosis. Liquid in a tube shows laminar or turbulent flow and the type of flow is determined by a Reynolds number (Re = VD ρ/µ where V represents flow velocity; D, inside diameter; ρ, specific gravity; and µ, viscosity). A high Reynolds number (Re > 2100) means turbulent flow while a low Reynolds number (Re < 2000) means laminar flow.11 Due primarily to a low viscosity, urine and water are likely to show turbulent flow which increases fluctuation of the regional flow velocity and flow pressure. Interactions between such factors inside a collapsible tube may explain why a smaller pressure decrease was required to create Doppler signals in the silicon tube study.
It is difficult to identify the precise factors or conditions that may cause cavitation in the human urethra from the present experimental results. Nevertheless, these results clearly indicated that the creation of microbubbles in a flowing liquid (model urine) was responsible for the detection of Doppler signals. In clinical practice, information on the Doppler shift frequencies at different flow velocities is converted to color-coded data and the results for each 9 pixels (1 mm2 of the sample volume) are averaged to obtain the mean flow velocity within the region of interest set in the urethra.2–4 This averaging process is helpful for reducing measurement artifacts when obtaining velocity-flow parameters.
Although various local factors may influence the creation of microbubbles in the urethra, we have confirmed that microbubbles derived from a basic hydrodynamic phenomenon are the principal reason for the detection of Doppler signals in flowing urine.
This work was supported by a Grant-in-Aid for Scientific Research (B) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (project number 13557135).