Page 1 
Previous  1 of 1  Next 

small (250x250 max)
medium (500x500 max)
Large
Extra Large
large ( > 500x500)
Full Resolution
All (PDF)

This page
All

Estimating Tissue Particle Sizes Using Ultrasound Breanna Swan and Dr. Anthony Gerig Physics Department, Viterbo University 900 Viterbo Drive, La Crosse, WI 54601 Email: bswan03013@viterbo.edu Methods and Materials Phantoms. Gelatinbased tissuelike structures were created mixing Knox’s Flavorless Gelatin, water, and Liquid Germal Plus. Beads were mechanically sorted then assessed using ProgRes CapturePro 2.6 and iSolution Lite. The mass density of the beads was calculated so the amount of glass beads needed could be determined. Speed of Sound and Attenuation. Narrow and broadband transducers of either 5 or 7 MHz sent pulses through the phantom. Matlab interpreted the data and generated the speed of sound and attenuation coefficient of the phantom. Backscatter Coefficient. Narrowband transducers of either 5 or 7 MHz sent a pulse to phantoms containing only one size of beads. Data was collected from the received echoes that bounced off of the glass beads and the backscatter coefficient was calculated using Matlab. Adjustments were made to the code involving the hanning window, attenuation value, and sample length. Sizing beads. The backscatter coefficient was compared to theoretical graphs to determine the size of the scattering particles. After determining the sizing code was accurate, the process was repeated with phantoms containing different concentrations of large beads. Results The shape and amplitude of the backscatter coefficient graph were compared to the theoretical graph. The amplitude of the experimental data was lower than the theoretical by a factor of 10. (Figure 1) The results of the particlesizing code were interpreted using different statistical methods. Three boxplots were created to illustrate the results; small beads, large beads (Figure 2), and phantoms containing mixed beads (Figure 3). Figure 2. In this figure, the four different boxplots represent four phantoms of 10g/mm3 mass density of large glass beads (90106nm). Two of the data were collected with 5 MHz transducers and two with 7.5 MHz transducers. The statistics used were the mean, median, maximum, and medians of quartile 1 and quartile 3. Abstract In medical ultrasound, sound waves are transmitted within tissue and the amplitude of the resulting echo signal is processed to generate an image. This project focuses on extracting the frequency content of the echo signal, called backscatter, and using that information to estimate sizes of the particles that generated the echo signal. Echo signals produced by phantoms, or gelatinbased structures infused with glass beads to generate tissuelike scattering, were recorded and processed using programs written in Matlab to calculate backscatter coefficients and estimate glass bead sizes. These values were then compared against the known backscatter coefficients and sizes for the glass beads. The accuracy of the sizepredictions was first examined using broadband and narrowband transducers of 5 and 7.5 MHz with phantoms containing only one size glass bead. Next, the accuracy was tested using phantoms of large, 90106nm, and small, 6375nm, beads. The size predictions were proven accurate as the results showed that even if the concentration of beads within the phantom varied, the predicted sizes remained within the correct size range. Similarly, as the ratio of large beads to small beads increased in the phantoms containing mixed beads, the predicted size also increased. The accuracy of the results allow for future work to be done such as the comparison the mixed size results against theoretical models. Introduction In the 1943 and 1957, two scientists made great strides in the field of ultrasound. Doctor Karl Theodore Dussik of Austria published the first paper on medical ultrasonics based on transmission ultrasound investigation of the brain while Professor Ian Donald of Scotland developed practical applications for ultrasound. (Kline) The accomplishments of these two men opened a new window in the world of medicine. Today, medical ultrasound is used by physicians and researchers to study the structure of function of tissues in situ, without having to perform surgery. Most medical ultrasound techniques create images based on the intensity of the echo signal. Instead, in this study, the backscatter coefficient, or the differential scattering cross section per unit volume as a function of frequency, will be measured and interpreted. The ultrasonic backscatter coefficient is useful in characterizing tissues, especially when it can be estimated with high accuracy and precision. Faran’s scattering theory is useful in evaluating the accuracy of the backscatter coefficient. Faran’s work predicts scattering of plane waves from a single, rigid spherical scatterer within an inviscid medium. (Anderson et al) The size of the scatterers within a tissue is a key factor in determining the backscatter coefficient dependent on frequency. Biological tissues often have scatterers of varying sizes which make it challenging to estimate particle sizes. One goal of this experiment is to correctly predict the backscatter coefficient of phantoms, or gelatinbased structures infused with glass beads with tissuelike values for speed of sound and attenuations, specifically 1540 m/s and 0.5 dB/MHz/cm, with glass beads of both the same size and varying sizes. A second goal of this experiment is to accurately estimate the size of the glass beads within the phantom. Size estimation accuracy is largely dependent on the distribution of bead sizes within the phantom. This experiment will explore the effects on size estimates when the concentration and size of glass beads vary within a phantom. Figure 3. In this figure, the four box plots represent the predicted particle sizes of four phantoms of increasing concentrations of small:large beads (80:20, 60:40, 40:60, 20:80). The data was interpreted using the Faran Model. Discussion One of the goals of this project was to determine why the amplitude of the backscatter coefficient graphs was lower than the theoretical graph. Unfortunately this problem was not solved. Three factors within the backscatter coefficient code, attenuation, length sample, and hanning window, were studied as to what effects it had on the backscatter coefficient value. Varying these three factors did not lead to a solution to the amplitude problem. Fortunately, the backscatter coefficient amplitude problem did not need to be fixed before experimenting with sizing the beads. First, the bead sizing code was determined accurate when analyzing the predicted sizes of bead in a phantom containing only one size. Figure 2 shows that even with phantoms of varying concentrations of beads, the code accurately predicts the sizes of beads to be within the correct size range (90106nm). The hypothesis for the results of the mixed beads phantoms was that the predicted bead size would gradually increase as the concentration of large beads increased. As seen by Figure 3, this hypothesis was proven correct. The predicted size gradually increases as the concentration of large beads increases. Continued work is needed to determine the reason for the low backscatter coefficient amplitude. One proposal is to obtain phantoms from UWMadison which already have calculated backscatter coefficients and test the backscatter coefficient using the procedures of this experiment. If the backscatter coefficient amplitude increases with UWMadison’s phantoms, there is an issue with the procedure of making phantoms. The problem could be with uniformly distributing the glass beads when creating the phantom. If, for example, the beads settle near the edges of the phantom, it would explain the low amplitude of the backscatter coefficient since the center of the phantom would have a lower density of beads. On the other hand, if the backscatter coefficient amplitude of UWMadison’s phantoms has the same low amplitude, there is a problem with the measurements. References Anderson et. al (2010). Interlaboratory Comparison of Backscatter Coefficient Estimates for TissueMimicking Phantoms. Ultrasonic Imaging, 32, 4864. Insana, M. F. (1996). Ultrasonic Imaging of Microscopic Structures in Living Organs. International Review of Experimental Pathology, 36, 7392. Kline, J. P. (2011). Ultrasonic Guidance in Anesthesia. AANA Journal, 79(3), 209217. Acknowledgments Viterbo University Dr. Anthony Gerig In this figure, the colored lines represent the backscatter coefficient of phantoms containing 10g/mm3 of large glass beads (90106nm). The data collected was split into ten parts and had ten different graphs. The data was collected using 5 MHz transducers. The top red line represents the theoretical data that would be expected with the same parameters. The bottom red line is the theoretical data reduced by a factor of ten. Figure 1.
Object Description
Title of Work  Estimating Tissue Particle Sizes Using Ultrasound 
Author  Breanna P. Swan 
Scholarship Group  Seven Rivers Undergraduate Research Symposium 
Summary of Work  In medical ultrasound, sound waves are transmitted within tissue and the amplitude of the resulting echo signal is processed to generate an image. This project focuses on extracting the frequency content of the echo signal, called backscatter, and using that information to estimate sizes of the particles that generated the echo signal. Echo signals produced by phantoms, or gelatinbased structures infused with glass beads to generate tissuelike scattering, were recorded and processed using programs written in Matlab to calculate backscatter coefficients and estimate glass bead sizes. These values were then compared against the known backscatter coefficients and sizes for the glass beads. The accuracy of the sizepredictions was first examined using broadband and narrowband transducers of 5 and 7.5 MHz with phantoms containing only one size glass bead. Next, the accuracy was tested using phantoms of large, 90106nm, and small, 6375nm, beads. The size predictions were proven accurate as the results showed that even if the concentration of beads within the phantom varied, the predicted sizes remained within the correct size range. Similarly, as the ratio of large beads to small beads increased in the phantoms containing mixed beads, the predicted size also increased. The accuracy of the results allow for future work to be done such as the comparison the mixed size results against theoretical models. 
Advisor  Anthony L. Gerig 
Date of Student Graduation  2013 
Date of Work  20111111 
Degree Name  Bachelor of Science 
Major/Program  Mathematical Physics 
School  School of Natural Sciences and Mathematics 
Type of Material 
Text Image 
File Name  2011_7r_swa.pdf 
Material Format  PDF/A 
Usage Rights  This item cannot be copied, reproduced, or transmitted in any form, by any means, without the express written permission of the author. 
Description
Title of Work  Page 1 
Full Text Search  Estimating Tissue Particle Sizes Using Ultrasound Breanna Swan and Dr. Anthony Gerig Physics Department, Viterbo University 900 Viterbo Drive, La Crosse, WI 54601 Email: bswan03013@viterbo.edu Methods and Materials Phantoms. Gelatinbased tissuelike structures were created mixing Knox’s Flavorless Gelatin, water, and Liquid Germal Plus. Beads were mechanically sorted then assessed using ProgRes CapturePro 2.6 and iSolution Lite. The mass density of the beads was calculated so the amount of glass beads needed could be determined. Speed of Sound and Attenuation. Narrow and broadband transducers of either 5 or 7 MHz sent pulses through the phantom. Matlab interpreted the data and generated the speed of sound and attenuation coefficient of the phantom. Backscatter Coefficient. Narrowband transducers of either 5 or 7 MHz sent a pulse to phantoms containing only one size of beads. Data was collected from the received echoes that bounced off of the glass beads and the backscatter coefficient was calculated using Matlab. Adjustments were made to the code involving the hanning window, attenuation value, and sample length. Sizing beads. The backscatter coefficient was compared to theoretical graphs to determine the size of the scattering particles. After determining the sizing code was accurate, the process was repeated with phantoms containing different concentrations of large beads. Results The shape and amplitude of the backscatter coefficient graph were compared to the theoretical graph. The amplitude of the experimental data was lower than the theoretical by a factor of 10. (Figure 1) The results of the particlesizing code were interpreted using different statistical methods. Three boxplots were created to illustrate the results; small beads, large beads (Figure 2), and phantoms containing mixed beads (Figure 3). Figure 2. In this figure, the four different boxplots represent four phantoms of 10g/mm3 mass density of large glass beads (90106nm). Two of the data were collected with 5 MHz transducers and two with 7.5 MHz transducers. The statistics used were the mean, median, maximum, and medians of quartile 1 and quartile 3. Abstract In medical ultrasound, sound waves are transmitted within tissue and the amplitude of the resulting echo signal is processed to generate an image. This project focuses on extracting the frequency content of the echo signal, called backscatter, and using that information to estimate sizes of the particles that generated the echo signal. Echo signals produced by phantoms, or gelatinbased structures infused with glass beads to generate tissuelike scattering, were recorded and processed using programs written in Matlab to calculate backscatter coefficients and estimate glass bead sizes. These values were then compared against the known backscatter coefficients and sizes for the glass beads. The accuracy of the sizepredictions was first examined using broadband and narrowband transducers of 5 and 7.5 MHz with phantoms containing only one size glass bead. Next, the accuracy was tested using phantoms of large, 90106nm, and small, 6375nm, beads. The size predictions were proven accurate as the results showed that even if the concentration of beads within the phantom varied, the predicted sizes remained within the correct size range. Similarly, as the ratio of large beads to small beads increased in the phantoms containing mixed beads, the predicted size also increased. The accuracy of the results allow for future work to be done such as the comparison the mixed size results against theoretical models. Introduction In the 1943 and 1957, two scientists made great strides in the field of ultrasound. Doctor Karl Theodore Dussik of Austria published the first paper on medical ultrasonics based on transmission ultrasound investigation of the brain while Professor Ian Donald of Scotland developed practical applications for ultrasound. (Kline) The accomplishments of these two men opened a new window in the world of medicine. Today, medical ultrasound is used by physicians and researchers to study the structure of function of tissues in situ, without having to perform surgery. Most medical ultrasound techniques create images based on the intensity of the echo signal. Instead, in this study, the backscatter coefficient, or the differential scattering cross section per unit volume as a function of frequency, will be measured and interpreted. The ultrasonic backscatter coefficient is useful in characterizing tissues, especially when it can be estimated with high accuracy and precision. Faran’s scattering theory is useful in evaluating the accuracy of the backscatter coefficient. Faran’s work predicts scattering of plane waves from a single, rigid spherical scatterer within an inviscid medium. (Anderson et al) The size of the scatterers within a tissue is a key factor in determining the backscatter coefficient dependent on frequency. Biological tissues often have scatterers of varying sizes which make it challenging to estimate particle sizes. One goal of this experiment is to correctly predict the backscatter coefficient of phantoms, or gelatinbased structures infused with glass beads with tissuelike values for speed of sound and attenuations, specifically 1540 m/s and 0.5 dB/MHz/cm, with glass beads of both the same size and varying sizes. A second goal of this experiment is to accurately estimate the size of the glass beads within the phantom. Size estimation accuracy is largely dependent on the distribution of bead sizes within the phantom. This experiment will explore the effects on size estimates when the concentration and size of glass beads vary within a phantom. Figure 3. In this figure, the four box plots represent the predicted particle sizes of four phantoms of increasing concentrations of small:large beads (80:20, 60:40, 40:60, 20:80). The data was interpreted using the Faran Model. Discussion One of the goals of this project was to determine why the amplitude of the backscatter coefficient graphs was lower than the theoretical graph. Unfortunately this problem was not solved. Three factors within the backscatter coefficient code, attenuation, length sample, and hanning window, were studied as to what effects it had on the backscatter coefficient value. Varying these three factors did not lead to a solution to the amplitude problem. Fortunately, the backscatter coefficient amplitude problem did not need to be fixed before experimenting with sizing the beads. First, the bead sizing code was determined accurate when analyzing the predicted sizes of bead in a phantom containing only one size. Figure 2 shows that even with phantoms of varying concentrations of beads, the code accurately predicts the sizes of beads to be within the correct size range (90106nm). The hypothesis for the results of the mixed beads phantoms was that the predicted bead size would gradually increase as the concentration of large beads increased. As seen by Figure 3, this hypothesis was proven correct. The predicted size gradually increases as the concentration of large beads increases. Continued work is needed to determine the reason for the low backscatter coefficient amplitude. One proposal is to obtain phantoms from UWMadison which already have calculated backscatter coefficients and test the backscatter coefficient using the procedures of this experiment. If the backscatter coefficient amplitude increases with UWMadison’s phantoms, there is an issue with the procedure of making phantoms. The problem could be with uniformly distributing the glass beads when creating the phantom. If, for example, the beads settle near the edges of the phantom, it would explain the low amplitude of the backscatter coefficient since the center of the phantom would have a lower density of beads. On the other hand, if the backscatter coefficient amplitude of UWMadison’s phantoms has the same low amplitude, there is a problem with the measurements. References Anderson et. al (2010). Interlaboratory Comparison of Backscatter Coefficient Estimates for TissueMimicking Phantoms. Ultrasonic Imaging, 32, 4864. Insana, M. F. (1996). Ultrasonic Imaging of Microscopic Structures in Living Organs. International Review of Experimental Pathology, 36, 7392. Kline, J. P. (2011). Ultrasonic Guidance in Anesthesia. AANA Journal, 79(3), 209217. Acknowledgments Viterbo University Dr. Anthony Gerig In this figure, the colored lines represent the backscatter coefficient of phantoms containing 10g/mm3 of large glass beads (90106nm). The data collected was split into ten parts and had ten different graphs. The data was collected using 5 MHz transducers. The top red line represents the theoretical data that would be expected with the same parameters. The bottom red line is the theoretical data reduced by a factor of ten. Figure 1. 