Histotripsy-Induced Bactericidal Activity Correlates to Size of Cavitation Cloud In Vitro

Abstract

Large abscesses are walled-off collections of pus and bacteria that often do not respond to antibiotic therapy. Standard of care involves percutaneous placement of indwelling catheter(s) for drainage, a long and uncomfortable process with high rehospitalization rates. The long-term goal of this work is to develop therapeutic ultrasound approaches to eradicate bacteria within abscesses as a noninvasive therapeutic alternative. Inertial cavitation induced by short pulses of focused ultrasound (histotripsy) is known to generate lethal mechanical damage in bacteria. Prior studies with Escherichia coli (E. coli) in suspension demonstrated that bactericidal effects increase with increasing peak negative amplitude, treatment time, and duty cycle. The current study investigated correlates of bactericidal activity with histotripsy cavitation cloud size. Histotripsy was applied to E. coli suspensions in 10-mL sample vials at 810 kHz, 1.2 MHz, or 3.25 MHz for 40 min. The cavitation activity in the sample vials was separately observed with high-speed photography. The cavitation cloud area was quantified from those images. A linear relationship was observed between bacterial inactivation and cavitation cloud size ( ), regardless of the acoustic parameters (specifically frequency, pulse duration, and power) used to produce the cloud.Index Terms- Abscess, bacterial inactivation, bactericidal activity, cavitation, high intensity focused ultrasound (HIFU), histotripsy, therapeutic ultrasound.

Figure 1.

Experimental setups for cavitation cloud imaging at three different frequencies. (a) Studies at 0.81 and 1.2 MHz were conducted with a Sonic Concepts transducer (model H-161) mounted to a cylindrical water bath degassed to < 20% saturation. The sample vial was held with magnets that kept the vial aligned with the transducer. (b) Studies at 3.25 MHz were performed with a custom-built transducer immersed in a large degassed (< 20%) water bath. Alignment of the transducer and sample vial were performed manually, as described in the text. For either setup, cavitation was generated at the focus, 15 mm above the vial’s bottom, approximately at the center of the vial. The same LED light source back-illuminated the cavitation cloud. Additional details are described in the text

Figure 2.

Focal pressure waveforms at the three frequencies were obtained from hydrophone measurements in degassed water (excluding the sample vial) and compared with numerical simulations: (a) Frequency = 0.81 MHz, Acoustic power = 926 W (b) Frequency = 1.20 MHz, Acoustic power = 268 W (c) Frequency = 3.25 MHz, Acoustic power = 87 W. At 3.25 MHz hydrophone measurements could not be performed due to the cavitation occurring at the hydrophone tip. These waveforms correspond to the lowest output acoustic power levels (just above the consistent cavitation threshold) at that frequency.

Figure 3.

Image processing of grayscale images to obtain bubble cloud area. a) Representative grayscale image at 0.81 MHz, 24.6 μs pulse duration, 926 W acoustic power. The arrow is in the direction of the histotripsy pulse. b) Conversion to binary image based on a binary threshold. c) The outline of the bubble clouds (black) in the binary images was highlighted and area was calculated based on the number of pixels. Note that the dark region at the bottom of Fig. 3b which corresponds to the vial membrane is not included in the area calculation

Figure 4.

Bubble cloud grayscale images in the growth medium (3% Tryptic Soy Broth) at the three frequencies just above the consistent cavitation threshold. The asterisk indicates geometric focus. a) Frequency = 0.81 MHz, Acoustic power = 926 W b) Frequency = 1.20 MHz, pulse duration 20 μs, Acoustic power = 268 W c) Frequency = 3.25 MHz, Acoustic power = 87 W. A decrease in dimensions with increasing frequency is observed.

Figure 5.

Bubble cloud grayscale images in the growth medium (3% Tryptic Soy Broth) at 1.20 MHz, 20 μs with increasing acoustic powers. Geometric focus is marked with an asterisk for reference. With increasing power, the bubble cloud expands pre-focally.

Figure 6.

Bubble cloud area dependence on peak acoustic power for the three frequencies and two pulse durations. Error bars correspond to standard deviation of the bubble cloud area across histotripsy pulses. The increase in the bubble cloud size is due to an increase in acoustic power and independent of the transducer frequency or pulse length. Points marked with a star indicate the protocols used for inactivation experiments

Figure 7.

E. coli log reduction dependence on bubble cloud area. Seven pulse protocols marked with a star in Fig.6 and also listed in Table II were applied to 10 mL E. coli suspensions for 40 minutes. A strong linear dependence of log reduction on the bubble cloud area was observed (r = 0.99, p < 0.0001, n=4). Vertical error bars correspond to standard deviation of log reduction, while horizontal error bars correspond to the standard deviation of the bubble cloud area from Fig.6.