Acquisition of Ultrasonic Echoes for use in Ultrasonic Thermometry

Thomas Simpson

Supervised by Drs. R. Martin Arthur and Jason W. Trobaugh
Department of Electrical and Systems Engineering
Washington University in St. Louis
Spring 2005

Background

Hypothermia is becoming a supplement to radiation in modern cancer treatment.  In heat treatments the temperature of the tissue is elevated to 41C-45C.  An accurate measure, to within 0.5C, of the temperature of a volume of tissue is needed to monitor the hypothermia of the treatment region.   Most of the current measurement techniques are invasive.  Invasive temperature measurements are limited due to the small number of measurement sites and the resulting lack of information necessary to describe the temperature distribution of the heated tissue.  Ultimately a more complete rendering of the temperature distribution is desired.1,2  

Studies performed at Washington University in St. Louis by Dr. Arthur and coworkers suggest that temperature estimation in tissue can be done noninvasively using the energy of backscattered ultrasound.  In initial studies, experiments were carried out using an ultrasonic transducer with center frequency of 7.5 MHz.1,2  It is hypothesized that moving to higher frequency transducers will provide more detail on the temperature dependence of scatterers.

Backscattered Ultrasound System

Figure 1 shows the basic diagram of the backscattered ultrasound system.  The tissue sample is placed on a rubber mat which has the same speed of sound and density of water to minimize the reflections due to the water rubber interface.  The insulated tank in which the sample is located is filled with degassed water.  The temperature of the tank is controlled by a circulating heater.  The transducer is activated using a pulse circuit.  Echoes from the tissue are digitized using a digital oscilloscope.  The data are then captured by the computer.  Once the data are recorded from one tissue site, the stepper motor moves the transducer to the next site, and another echo is recorded.  After the tissue scan is complete, the temperature is changed and the capture process is repeated. 

Figure 1.  Experimental Setup (from Reference 1)

As previously stated, experiments have been performed with a 7.5 MHz transducer.  At this center frequency the echoes were sampled at 50 MHz.  When the center frequency of the transducer is increased, the sampling frequency also needs to increase.  This necessitates an oscilloscope with a faster sampling rate than the scope used in previous experiments. 

Project Description

The objective of this study was to integrate and automate the system shown in Figure 1 using Matlab as the controlling software. 

The original software was written in BASIC by Dr. Arthur.  The new software was written in Matlab and follows the same structure as the original program.  The first task in accomplishing the integration is to initialize the Tektronix TDS220 oscilloscope from Matlab.  Initializing the scope consists of setting the time base and initial gain for each channel, setting the scope to trigger off the excitation pulse on the EXT channel, setting the sampling mode, and setting the delay from the excitation pulse to the received echo signal.  Within the oscilloscope window there are 2500 samples, so the sampling rate is set by changing the time setting.  The following table summarizes some sampling rates for the specified time base.  These rates were determined by querying the oscilloscope to determine the sample spacing at each time base setting. 

Time base (μs)

Sample Spacing (seconds)

Sampling Frequency (Hz)

0.01

4*10-11

25*109

0.025

1*10-11

10*109

0.05

2*10-10

5*109

0.1

4*10-10

2.5*109

0.25

1*10-9

1*109

0.5

2*10-9

500*106

1

4*10-9

250*106

2.5

1*10-8

100*106

5

2*10-8

50*106

The sampling rate of the scope is 1 GHz, but in the table above the sampling frequency is higher than 1 GHz for some of the time base settings.  This is because the oscilloscope upsamples the data to obtain 2500 points, which gives the higher sampling frequencies listed in the above table.  Because the sampling rate is based on the time base setting, the center frequency of the transducer determines the correct setting.  Thus, a switch was implemented in the scope initialization routine to set the sampling rate to 50 MHz and 250 MHz for transducer center frequencies less than 10 MHz and greater than 10 MHz, respectively. 

Data were captured from the oscilloscope via a GPIB interface.  Oscilloscope gain was adjusted under computer control.  This was accomplished by querying the scope to obtain the data and checking to see if the maximum value was between half and full scale.  If it was then the gain was correct.  If the maximum value is larger than full scale the gain was reduced one setting and the data was recaptured to check the maximum value and adjusted accordingly.  If the maximum value is less than half scale then the gain is incremented one setting, and the data was checked again to make sure the maximum falls between full and half scale.  Once the correct gain setting was determined, the data was stored.  Figure 2 shows a picture of me with a plot of a captured data from the oscilloscope.  The plot shown in Figure 2 is the echo seen from a coffee mug shown in Figure 3. 

Figure 2.  Tom Simpson with ultrasonic echo captured using the instrumentation shown in Figure 1.

Once the software could effectively communicate with the scope, motor control was added.  This was done using the Matlab file posmotorfn.m written by Dr. Arthur as modified by Suvimol Sangkatumvong, which controls the motor movement.  This m-file was used by collecting an echo signal at a particular site, then moving the transducer with the stepper motor to the next site and collecting another echo.  This is done over the length of the tissue then the transducer is moved back to its original position.

 

Figure 3.  Stepper motor and transducer during acquisition of the signal in Figure2 and at the top of the page.

The temperature control was also easy to integrate using the m-file ChangeTemp.m originally written by William L. Straube and modified by Thomas Simpson.  ChangeTemp.m controls a ThermoHaake P170 Heating Circulator, which changed the temperature of the water bath to the specified value based on feedback from a temperature probe inserted into the tissue. 

With all of the parts of the backscattered ultrasound system in working fashion automation was the final step.  This was simple to implement using a for-loop to cycle through temperatures 37C to 50C by 0.5C steps.  The algorithm is: 1) initialize the scope; 2) set the temperature; 3) scan the tissue; 4) repeat 1-3 for the each temperature of interest.  

The project was completed successfully.  The backscattered ultrasound system is in working order, and data have been collected from pork tissue with a 7.5 MHz transducer over the temperature range 37C to 50C with 0.5C steps.  The next step is to test the system with a 50 MHz transducer, which requires a high-frequency pulse circuit and receiver now under development.

References:

1 R.M.Arthur, W.L. Straube, J.D. Starman, E.G. Moros, "Noninvasive temperature estimation based on the energy of backscattered ultrasound." Med. Phys 30(6), pp. 1021-1028, June 2003.
2 R.M. Arthur, J.W. Trobaugh, W.L. Straube, E.G. Moros, "Temperature Dependence of Ultrasonic Backscattered Energy in Motion-Compensated Images." IEEE Transactions on UFFC, October 2005.

Edited by R. Martin Arthur, 9 September 2005.