Pulse-echo Methods for Determination of
Broadband Ultrasonic Attenuation
to Image Temperature in Tissue
Dr. R. Martin Arthur
Department of Electrical and Systems Engineering
Washington University in St. Louis
When treating cancer patients with radiation or chemotherapy, raising the temperature of the cancerous tissue increases the effectiveness of the treatment. To do this while also keeping healthy tissue at normal body temperature, it is important to have a real-time temperature mapping of the tissue. We investigated the ability to generate broadband ultrasonic attenuation measurements from a single piston transducer that could be useful for temperature imaging based on attenuation change with temperature. Attenuation was estimated as a function of frequency with focused-piston transducers with center frequencies of 5, 7.5, and 10 MHz.
The goal of this project was to determine whether or not a programmable pulser/receiver designed and built by Dr. W. David Richard in Computer Science and Engineering could provide a wider bandwidth echo signal than a commercially available unit, a 5900 Panametrics (Waltham, MA) pulser/receiver. The utility of the broader band signal source was then applied to pulse-echo measurement of attenuation in a tissue phantom constructed in Dr. Arthur's laboratory.
As shown in Figure 1, the experimental setup consisted of a metal ball bearing inside a container full of water. Initially an Olympus (Waltham, MA) 7.5 MHz transducer was attached to a PC driven Newport Universal Motion Controller (UMC) (Irvine, CA) and was partially submerged in the water directly above the ball bearing.
Each pulser sent an electric pulse to the transducer, which changed the electrical pulse into an ultrasonic pulse directed toward the ball bearing. The echo from the ball bearing was measured by the receiver attached to the transducer, after which it was amplified and then sent to a 14-bit, 200 MHz Gage (Lockport, IL) digital oscilloscope in the PC.
Figure 1: Experiment setup used to compare ultrasonic pulsers.
Two pulser/receivers were compared: 1) A 5900 Panametrics unit and 2) A unit designed and built by Dr. W. David Richard (wdr). The latter unit was tested on two different settings (10 MHz and 50 MHz) intended to maximize acoustic output at those frequencies. They were first tested on the metal ball bearing setup shown in Figure 1 as a basis for comparison with the metal plate, which is part of the apparatus used to measure tissue or tissue phantoms. The ball bearing ensured the echo signal was parallel to the axis of the transducer, where as the metal plate could be slightly tilted and could cause a saw tooth effect in the frequency spectra of the echo, as seen in Figure 7. Next both pulser/receivers were compared using a metal plate as reflector in place of the ball bearing shown in Figure 1. In the future, temperature-dependent measurements in tissue will be referred to signals from the metal plate.
To choose the better pulser/receiver, we compared the Fourier Spectra of the echoes. We wanted as wide a frequency band as possible centered above 20 dB down from the maximum of the signal, which was the apparent signal-to-noise limit. This frequency band was centered around the transducer's center frequency. As shown in Figures 2 and 3, the wdr pulser/receiver on the 50 MHz setting had the widest frequency band in echoes from the metal plate. Also, the wdr pulser/receiver can apply a variable gain to the echo from 0 to 48 dB which is helpful in maximizing the signal-to-noise ratio.
Figure 2: Frequency domain comparison of pulses off of metal ball bearing.
Figure 3: Frequency domain comparison of pulses off of metal plate.
Next, echoes were measured from the metal plate with and without a 1.5 cm think graphite in gelatin phantom to determine the attenuation caused by the phantom as shown in Figure 4. Specifically, an echo was captured through the carbon phantom at the middle of the plate and directly off the plate to the east and west of the phantom. The experiment was run with three different transducers with center frequencies at 5 MHz, 7.5MHz, and 10 MHz. Results are shown in Figures 5, 6, and 7. The cleanest data (from the 7.5 MHz transducer), shows attenuation in the phantom to be about 0.75 dB/cm/MHz, which matches value from 5 MHz to 10 MHz reported by Madsen and coworkers1.
Figure 4. Experimental setup to measure attenuation of a
tissue phantom. The transducer is at the end of the column
over the specimen fixed in place by a 4 x 6 cm lucite form.
Figure 5: Frequency domain attenuation of ultrasound through phantom with 7.5 MHz transducer.
Figure 6: Frequency domain attenuation of ultrasound through phantom with 10 MHz transducer.
Figure 7: Frequency domain attenuation of ultrasound through phantom with 5 MHz transducer.
Run the same experiment with phantoms and tissue specimens at temperatures from 37 to 45oC. Our expectation is that the results will show that attenuation can be described by a single-pole model with a relaxation frequency that varies monotonically with temperature2.
1Ernest L. Madsen, James A. Zagzebski, Richard A. Banjavie, and Ronald E. Jutila, "Tissue mimicking materials for ultrasound phantoms" Med. Phys. 5, 391 (1978)
2R. Martin Arthur and K.V. Gurumurthy, "A single pole model for the propagation of ultrasound in soft tissue" The Journal of the Acoustical Society of America, Volume 77, Issue 4, April 1985, pp.1589-1597