Using high frequency ultrasound envelope statistics to determine scatterer number density in dilute cell solutions
Tunis AS, Baddour RE, Czarnota GJ, Giles A, Worthington AE, Sherar MD, Kolios MC
High frequency ultrasound (HFUS) has the potential to monitor structural changes occurring in tissues during cell death. Other than spectroscopic methods, which requires the acquisition of RF data, backscatter envelope statistics can also be used. It has previously been shown in tissue-mimicking phantoms and simulations that the envelope statistics of ultrasound backscatter are affected by changes in the scatterer properties. At higher frequencies the wavelength begins to approach the size of cells and cellular components. We hypothesize that at this scale the envelope statistics of backscattered ultrasound become more sensitive to structural changes within cells. To investigate the relation between the envelope statistics and scatterer properties in a biological system, experiments were performed with cells in dilute solutions. The physical meaning of the fit parameters was evaluated by investigating HFUS backscatter envelope statistics from suspensions of various concentrations of two different cell lines of different sizes.
The two cells lines used were AML cells (diameter ~11 µm) and PC3 cells (diameter ~23 µm). A volume of each type of cells was measured and then diluted with a corresponding volume of phosphate buffered saline (PBS) solution in a sample holder to form suspensions with various volume concentrations of cells, ranging from 0.1% to 1.6% cells per volume. Data were acquired using a commercial HFUS scanner (VisualSonics VS-40B) with a focused transducer (f# 2.35, focal length 20 mm, centre frequency 20 MHz, bandwidth 100%). B-scan images and radio-frequency (RF) data were recorded from approximately 200 independent locations separated by at least one beamwidth (250 mm) within the solution. The B-scan images were recorded in real-time and the RF data were acquired over a period of approximately 2-3 minutes. Using custom software written in Java, RF data were extracted from a homogeneous region-of-interest within the stored data. The maximum likelihood estimation technique was implemented in Matlab to obtain the fit parameter values of the Generalized Gamma (GG) function, one of the functions that can be used to describe the envelope statistics.
As demonstrated in theoretical work by Shankar¹, the a parameter of the GG function depends on the effective scatterer cross-section and the ratio of c/v depends on the effective scatterer number density. Experimental results demonstrated that the a parameter was higher for the larger PC-3 cells than the AML cells (0.5 ±0.1 vs. 0.09 ±0.02). Furthermore, the ratio c/v increased with cell number density, with similar values for both cell lines at equal number density independent of cell size. These results confirm the interpretation of the parameters proposed in the theoretical predictions of Shankar¹ and extend the work to the biological model of sparse solutions of cells. This technique demonstrates that HFUS signal statistics can be used to measure cell properties.
¹ Shankar PM. Ultrasonic tissue characterization using a generalized Nakagami model. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control. 2001, 48(6).