Does Body Position Affect Ultrasound Measurements of Bladder Wall Elasticity?

Abstract

Ultrasound Bladder Vibrometry (UBV) parameters have been shown by previous studies to strongly correlate with Urodynamic Studies (UDS) measurements. Just like UDS, UBV can be performed in supine and sitting positions. The objective of this study is to compare UBV parameters obtained for the two different positions using statistical methods. We recruited 8 volunteers with healthy bladders for this purpose. The elasticity, group velocity squared and thickness of the bladder were the UBV parameters of interest, and their values were recorded at different bladder volumes for each volunteer. The results presented here indicate that the measurements made for the two positions are in agreement using the Bland Altman method and a parameter q, which compares the values at each bladder volume for each volunteer. UBV parameters were also repeatable for measurements recorded in the supine and sitting positions.

INTRODUCTION

The urinary bladder is responsible for storing urine and controls urination. Bladder activity is regulated by the central nervous system (Fowler et al. 2008). A normal bladder is compliant and expands to accommodate urine to a certain amount at low pressure, thereby controlling one’s frequency of urination. However, neurological disorders such as spinal cord injuries can affect the compliance of a bladder (Vilensky et al. 2004; Wyndaele et al. 2011) by increasing its rigidity due to the excessive growth of bladder fibrotic tissue (Nguyen et al. 1996). Physiological abnormalities like sleep apnea which causes bladder oxidative stress can also contribute to detrusor instability and bladder non-compliance (Witthaus et al. 2015). Changes in the compliance of the bladder happen gradually over months (Johnson and Singh 2013) and thus assessing bladder performance in patients at risk of developing bladder non compliance and then providing the appropriate medication is an important process.

Currently, Urodynamic Studies (UDS) is the accepted gold standard for the assessment of bladder compliance in the clinic. UDS comprises two parts: the filling and storage phase and the voiding phase (Nitti 2005). One of the tests under the filling and storage phase of UDS is the cystometric test. This test monitors how pressure builds up in the bladder and tests the amount of urine the bladder can hold before one feels the need to urinate. A cystometric test is performed by inserting a catheter into an empty bladder, which is filled up gradually with water using a mechanical pump. Two pressure sensors are used; one is placed in the bladder and the other in the vagina or rectum. The bladder pressure and the abdominal pressure are recorded and the difference between the two measurements is used to calculate the detrusor pressure (de Jong and Klijn 2009, Nitti 2005; Winters et al. 2012). Subsequently, bladder compliance can be determined. Although this method is accurate, it is invasive, uncomfortable, and expensive. There is also the risk of urinary tract infections. A non-invasive, low cost and objective method is therefore desirable.

Ultrasound Bladder Vibrometry (UBV), as introduced in (Nenadic et al. 2013; Nenadic et al. 2016) provides a non-invasive method to measure bladder wall elasticity and indirectly assess bladder compliance. In this method, the bladder is modeled as a viscoelastic shell filled with an incompressible fluid, and acoustic radiation force (ARF) is used to excite transient waves in the bladder tissue. The motion of the tissue is then tracked using high-speed imaging and Doppler based speckle tracking algorithm (Bayat et al. 2018; Nenadic et al. 2016). The Lamb wave model is used in UBV to retrieve parameters using the least square fitting method and wave dispersion analysis along the bladder wall (Nenadic et al. 2016). The parameters reported in UBV are the bladder wall’s modulus of elasticity and the lamb wave group velocity. Previous studies have shown a strong correlation between UBV and UDS values, thus making this method a non-invasive alternative to UDS (Bayat et al. 2018; Bayat et al. 2019; Nenadic et al. 2016).

An important factor in the development of UBV as a new measurement tool is to assess measurement repeatability. One potential source of measurement variability is the patient position. Similar to the UDS, UBV measurements may be performed with the patient in a supine or sitting position. Arunkalaivanan et al. (2004) studied the effect of body posture on UDS parameters related to voiding sensations and their results showed a significant difference between the supine and sitting positions. Therefore, a logical question is if the patient’s position influences the UBV measurements and, if so, then how the position must be considered to obtain repeatable results. This question has motivated us to investigate the potential influence of patient position in the UBV.

In this study, we aim to analyze and compare UBV measurements obtained in supine and sitting positions in a group of healthy volunteers. The elasticity, group velocity, and bladder thickness were recorded, and results were analyzed using statistical tools to determine the influence of the subject position on the UBV parameters.

MATERIALS AND METHOD

Ultrasound Bladder Vibrometry

The principle of UBV operates on the idea that an acoustic radiation force beam is generated by an ultrasound transducer on the bladder wall, which excites anti-symmetric lamb waves (Nenadic et al. 2016). The motion of the lamb waves in the bladder wall is tracked by the receive transducer. If a wall thickness of 2h is considered, the lamb wave dispersion can be described as

$$4k_L^3\mathrm{\beta }\tanh (\beta h)=k_S^4+(k_S^2-k_L^2)\tan (k_{L}h)$$

where $\beta =\sqrt{k_L^2-k_S^2}$, k_L = ω/c is the Lamb wave number, ω is the angular frequency, c is the frequency-dependent Lamb wave phase velocity, $k_S=\mathrm{\omega }\sqrt{\mathrm{\rho }/\mathrm{\mu }}$ is the shear wavenumber, ρ is the tissue density, and μ = c² / ρ is the bladder wall shear modulus.

The steps used to estimate the UBV parameter values, that is, group velocity and elasticity, are outlined in (Bayat et al. 2018; Nenadic et al. 2016). The thickness of the bladder was estimated from the B-mode image. Fig. 1 illustrates the experimental setup used for the study.

Fig. 1:

Experimental setup in (a) supine and (b) sitting positions

Volunteer Population:

This prospective pilot study was the Health Insurance Portability and Accountability Act (HIPAA) compliant, and it was approved by the Mayo Clinic institutional review board (IRB Application # 17-002139). A signed written informed consent with permission for publication was obtained from each enrolled volunteer prior to the study. Eight healthy volunteers with no history of bladder diseases were recruited for this study, three of which were females and the rest males. Volunteers were between the ages 21 and 72 years old. The mean BMI was 26.59±3.60. We ensured volunteers were comfortable throughout the procedure and were given the opportunity to opt-out of the study at any point should they feel uncomfortable.

UBV data collection and processing:

Volunteers consumed about 1 liter of fluid (tea or water) and were asked to avoid voiding 2 hours before the start of the study to allow the natural filling of the bladder. Before any UBV measurements were recorded, the bladder volume was measured with the BladderScan device (Verathon, Bothell, WA). Volunteers then voided approximately 50-100mL of urine at a time until all urine was voided. UBV measurements for both positions were recorded after each incremental voiding. The voided volume was noted, and the bladder volume was measured at the time of each UBV measurement. Verasonics (Kirkland, WA), a programmable research ultrasound machine, and a linear array transducer (L7-4, Philips Healthcare, Andover, MA) with a center frequency of 5MHz were used for this study. UBV measurements were acquired using the UBV concepts explained in (Bayat et al. 2018; Nenadic et al. 2016). All ultrasound images were taken in the transverse plane. MATLAB (2016) (Mathworks Inc., Natick, MA) was used to process the data.

RESULTS AND DISCUSSION

The results obtained from the volunteers were analyzed to understand the relationship between the UBV values in supine and sitting positions. Different bladder volumes were recorded as urine was voided with the first acquisition corresponding to the full bladder and the higher acquisitions corresponding to lower volumes. At each bladder volume, 2 or more UBV measurements were recorded and averaged. The volunteers showed a variation in the total bladder volumes due to the difference in the rate of natural voiding in different people. The bladder thickness increases as more urine is voided. This is expected as the bladder wall stretches as more fluid fills the bladder. When all urine is voided, however, the true thickness of the bladder becomes more visible. The typical range for bladder thickness in an adult is between 2.9mm and 4.4mm (Hakenberg et al. 2000). In this study, the average bladder thickness recorded when all urine was voided in the supine position was around 3.6mm, and that of sitting was 3.1mm. The group velocity squared, which was obtained by squaring the measured group velocity and elasticity measurements, generally decreased gradually as more urine was voided, a characteristic of a compliant bladder.

Bland Altman Analysis

The Bland Altman plot, which plots the difference between measurements for supine and sitting positions against the mean of the 2 measurements, was used to statistically analyze the agreement between the two methods of measurements (Fig. 2). The thickness measurements recorded in the supine position on average were higher than the sitting position measurements shifting the mean to 0.12. The limits of agreement, however, show a quite narrow band (−0.59 – 0.82), which captures about 93% of the data. With the exception of 2 measurements, all other recordings for thickness were between the lower and upper limits of agreement. Contrary to what was observed for the thickness, GV² measurements recorded in the sitting position were higher than that of supine, thereby shifting the bias (mean) to a value of −0.5. The majority of the data plotted here lie very close to the mean line, which is an indication of a consistently small difference between the two measurements for different volunteers. The supine position is again noticed to produce higher values for elasticity than sitting with a mean difference of 0.2. However, this value is clinically small, and therefore the values recorded for supine and sitting positions can be considered to show no significant difference. Outliers noticed in Fig. 2 could be due to experimental errors. The MedCalc software version 19.2 (MedCalc, Seoul, Republic of Korea) was used to plot the Bland Altman plots. A good agreement between the two methods is seen in all three plots.

Fig. 2.

a: Bland Altman plot for thickness. The different colors represent different volunteers and the number of acquisitions obtained from each volunteer corresponds to the number of times a specific legend appears

b: Bland Altman plot for group velocity squared. The different colors represent different volunteers and the number of acquisitions obtained from each volunteer corresponds to the number of times a specific legend appears

c: Bland Altman plot for elasticity. The different colors represent different volunteers and the number of acquisitions obtained from each volunteer corresponds to the number of times a specific legend appears

The Bland Altman plot presented earlier put together all the measurements acquired from all volunteers across all volumes, and therefore gives an overall estimate of the agreement between the measurements in the two positions of interest; supine and sitting. We present a parameter q, which compares a specific UBV parameter at each bladder volume for each volunteer. This parameter is defined by the equation, $q=\frac{{\sigma }_S+{\sigma }_{SU}}{avg\left|S-SU\right|};$ where σ is the standard deviation of the total number of measurements at a specific bladder volume, S represents the value for supine, and SU represents the value for sitting up. A value ≥ 0.5 indicates less variation and, thus, more similarity between the two measurements. In addition to this, an exception is made where the exact values were recorded for both supine and sitting methods leading to a division by zero using the parameter q formula. Only one such case was present in our study and was categorized as a total agreement between the two measurements. The q values for the bladder thickness measurement showed about 57% agreement between the two measurements for supine and sitting positions and a 43% disagreement.

For grouped velocity squared, half of the volunteers showed a 100% agreement between the measurements recorded in the two positions using the q parameter. Two of the remaining 4 volunteers showed 80% agreement; one showed an approximate of 67% agreement, and the last volunteer had a 50% agreement between the supine and sitting measurements. Elasticity showed a total agreement between the measurements obtained for the two positions for 6 volunteers using their q values. One volunteer out of the remaining 2 showed an 86% agreement, and the other showed about 60% agreement between measurements for supine and sitting positions.

The overall percentage of agreement for all volunteers using the q values are 57%, 84%, and 90% respectively for thickness, group velocity squared and elasticity, respectively. This is consistent with the results from the Bland Altman plot as we notice a widely scattered plot in the plot for thickness than the other UBV parameters.

Repeatability Analysis of UBV measurements

The coefficient of variation calculated as the ratio of the standard deviation to the mean of the measurements, mathematically expressed as $\mathit{CV}=\sigma ∕\stackrel{‒}{\mathrm{X}}$ was used to estimate the repeatability of the UBV measurements. The CV was computed across all volunteers and the result is displayed in Fig. 3 using the mean (a) and the median (b). A higher CV indicates less repeatability in the measurements and vice versa. Thickness measurements for supine and sitting positions differed by about 0.003 in both Figs. 3(a) and (b). Between Fig. 3(a) and (b), however, we see a difference of 0.0001 between the two statistical measurements. A difference of 0.04 and 0.08 is also observed in the GV² and elasticity measurements, respectively in Fig. 3(a) and (b). Within the measurements in the mean plot (Fig. 3(a)), a small difference (~0.012) is seen in GV² and a smaller difference of about 0.008 is noticed in Fig. 3(b) for the same parameter. The elasticity plot in Fig. 3(a) also shows a difference of about 0.007 and Fig. 3(b) produces a relatively small difference of 0.004 between supine and sitting measurements. Generally, across the two plots, supine produces higher values for thickness, and the opposite is true for the GV² measurements. A rather conflicting situation is seen in the plots for elasticity, where in Fig. 3(a), sitting position produces higher CV values, and Fig. 3(b) produces lower values compared to the supine position measurements. This could be due to the outliners present in the data set, thereby affecting the overall average.

Fig. 3:

CV for UBV parameters using mean (a) and median (b)

One limitation of this study is the number of volunteers involved. To arrive at a much clearer conclusion, further studies will need to be conducted with a greater number of volunteers.

CONCLUSIONS

In this paper, we statistically compared the UBV parameters measured for two positions: supine and sitting. The two positions produced results that were in agreement for the majority of the data presented for all UBV parameters. Both the Bland Altman approach, which looks at the overall performance of the two positions across all volunteers, and the q parameter, which compares a specific UBV parameter value at each bladder volume for each volunteer, all indicate a good agreement between the supine and sitting measurements. Repeatable results are seen for all UBV measurements as well.