A tissue phantom for visualization and measurement of ultrasound-induced cavitation damage

Abstract

Many ultrasound studies involve the use of tissue-mimicking materials to research phenomena in vitro and predict in vivo bioeffects. We have developed a tissue phantom to study cavitation-induced damage to tissue. The phantom consists of red blood cells suspended in an agarose hydrogel. The acoustic and mechanical properties of the gel phantom were found to be similar to soft tissue properties. The phantom's response to cavitation was evaluated using histotripsy. Histotripsy causes breakdown of tissue structures by the generation of controlled cavitation using short, focused, high-intensity ultrasound pulses. Histotripsy lesions were generated in the phantom and kidney tissue using a spherically focused 1-MHz transducer generating 15 cycle pulses, at a pulse repetition frequency of 100 Hz with a peak negative pressure of 14 MPa. Damage appeared clearly as increased optical transparency of the phantom due to rupture of individual red blood cells. The morphology of lesions generated in the phantom was very similar to that generated in kidney tissue at both macroscopic and cellular levels. Additionally, lesions in the phantom could be visualized as hypoechoic regions on a B-mode ultrasound image, similar to histotripsy lesions in tissue. High-speed imaging of the optically transparent phantom was used to show that damage coincides with the presence of cavitation. These results indicate that the phantom can accurately mimic the response of soft tissue to cavitation and provide a useful tool for studying damage induced by acoustic cavitation.

Figure 1

High speed imaging apparatus used to acquire images prior to, during, and after phantom insonation. The RBC layer of the gel was positioned within the ultrasound focus. The high speed camera was positioned imaging the plane of RBCs, and a flash lamp was used to backlight the phantom. A digital camera was also used to take high-quality images before and after insonation.

Figure 2

A photograph of the cell phantom in a polycarbonate holder. The phantom consists of 3 layers of agarose, with the middle layer containing 5% red blood cells. The insert shows the visual appearance of the red blood cell layer during examination under a microscope.

Figure 3

Images of (a) histology of a lesion generated in ex-vivo kidney using 1500 pulses in 3 adjacent focal spots with 2 mm spacing and (b) a lesion generated in the cell phantom using identical parameters. Ultrasound (US) propagation is from left to right. Lesions in kidney appear as a more homogeneous region due to cell structure disruption, while lesions in the phantom show increased transparency. Magnified images of the boundary between treated (T) and untreated (UT) regions of the lesions in Figure 3a and Figure 3b are also shown in 3c and 3d, respectively. The disrupted zone of the phantom shows virtually no red blood cells remaining, while the region only 50μm from the lesions shows similar appearance to untreated phantom.

Figure 4

Eight lesions generated in a cell phantom by applying ultrasound perpendicular to the RBC layer. Each lesion is an accumulation of cavitation damage caused by applying 1500 pulses to a single focal spot. The phantom shows a completely disrupted region within the center of the focal zone, and small microlesions surrounding this area.

Figure 5

Histogram of collateral microlesions’ diameters recorded in the 8 focal spots shown in Figure 4 (not including the central lesion). A local peak is observed at about 150 μm diameter, which corresponds well with single bubbles observed on high speed images of the phantom.

Figure 6

Comparison of photographs and B-Mode images of lesions. (a) shows a photograph of kidney lesion histology and (b) is a lesion in the cell phantom. Histotripsy ultrasound (US) propagation is from left to right. The B-Mode image has a hypoechoic appearance similar to that seen for histotripsy lesions in ex-vivo kidney tissue (c). B-Mode image of the phantom lesion is shown in (d).

Figure 7

Images captured by the high-speed camera during a single ultrasound pulse. Ultrasound propagation is from left to right. The pulse arrives on the left at 0 μs and passes through the focus between 0–20 μs. The bubble cloud generated during the pulse then collapses 110 μs after ultrasound exposure.

Figure 8

Examples of a bubble cloud and the corresponding lesion generated after ultrasound exposure. Ultrasound (US) propagation is from left to right. (a) High speed image of the bubble cloud 5 μs after ultrasound pulse. (b) Subtraction image of the lesion generated by the single pulse. All lesions matched well in shape and dimensions with the corresponding bubble clouds. Lesions due to individual bubbles surrounding the focus are also evident.

Figure 9

High speed images of the cell phantom under 20× magnification. Prior to insonation, individual cells are observed in the agarose. After cavitation, the red blood cells are still visible as inhomogeneous regions (t = 300 μs). When another image is captured after 50 seconds, no cells are apparent in the area.

Figure 10

Ultrasound backscatter coefficient (BSC) vs. blood hematocrit in agarose at 10 MHz. The peak occurs at 13% in this case. A 5% phantom yields has a BSC of 0.7 compared with the peak.

Figure 11

Image of a lesion generated from 3 single ultrasound pulses (a) and overlapping images of 3 bubble clouds taken from binary images (b). Ultrasound propagation was from left to right. Lesion damage matches the shape and location of the bubble clouds, although appears darker near the edges of the cloud.