6651

Physics in Medicine & Biology

Effect of hydrodynamic cavitation in the tissue erosion by pulsed high-intensity focused ultrasound (pHIFU)

Yufeng Zhou1 and Xiaobin Wilson Gao

School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore

E-mail: yfzhou@ntu.edu.sg

Received 31 March 2016, revised 8 July 2016Accepted for publication 1 August 2016Published 19 August 2016

AbstractHigh-intensity focused ultrasound (HIFU) is emerging as an effective therapeutic modality in clinics. Besides the thermal ablation, tissue disintegration is also possible because of the interaction between the distorted HIFU bursts and either bubble cloud or boiling bubble. Hydrodynamic cavitation is another type of cavitation and has been employed widely in industry, but its role in mechanical erosion to tissue is not clearly known. In this study, the bubble dynamics immediately after the termination of HIFU exposure in the transparent gel phantom was captured by high-speed photography, from which the bubble displacement towards the transducer and the changes of bubble size was quantitatively determined. The characteristics of hydrodynamic cavitation due to the release of the acoustic radiation force and relaxation of compressed surrounding medium were found to associate with the number of pulses delivered and HIFU parameters (i.e. pulse duration and pulse repetition frequency). Because of the initial big bubble (~1 mm), large bubble expansion (up to 1.76 folds), and quick bubble motion (up to ~1 m s−1) hydrodynamic cavitation is significant after HIFU exposure and may lead to mechanical erosion. The shielding effect of residual tiny bubbles would reduce the acoustic energy delivered to the pre-existing bubble at the focus and, subsequently, the hydrodynamic cavitation effect. Tadpole shape of mechanical erosion in ex vivo porcine kidney samples was similar to the contour of bubble dynamics in the gel. Liquefied tissue was observed to emit towards the transducer through the punctured tissue after HIFU exposure in the sonography. In summary,

Y Zhou and X W Gao

Printed in the UK

6651

PMB

© 2016 Institute of Physics and Engineering in Medicine

2016

61

Phys. Med. Biol.

PMB

0031-9155

10.1088/0031-9155/61/18/6651

Paper

18

6651

6667

Physics in Medicine & Biology

Institute of Physics and Engineering in Medicine

IOP

1 Author to whom any correspondence should be addressed.School of Mechanical and Aerospace Engineering, Nanyang Technological University, 50 Nanyang Ave. 639798,

Singapore.

0031-9155/16/186651+17$33.00 © 2016 Institute of Physics and Engineering in Medicine Printed in the UK

Phys. Med. Biol. 61 (2016) 6651–6667 doi:10.1088/0031-9155/61/18/6651

6652

the release of HIFU exposure-induced hydrodynamic cavitation produces significant bubble expansion and motion, which may be another important mechanism of tissue erosion. Understanding its mechanism and optimizing the outcome would broaden and enhance HIFU applications.

Keywords: high-intensity focused ultrasound (HIFU), tissue erosion, hydrodynamic cavitation, high-speed imaging

(Some figures may appear in colour only in the online journal)

Introduction

High-intensity focused ultrasound (HIFU) is emerging as an effective and noninvasive ther-apeutic procedure (ter Haar 2001). High-amplitude ultrasound energy from the transducer can be focused through the skin and overlying tissue to a tumor or cancer extracorporeally. Absorption of acoustic energy in the focal region leads to high and rapid temperature elevation in the tissue and formation of irreversible necrosis within seconds or even sub-seconds. The boundary between HIFU-induced protein denaturalization and intervening normal tissue is around 200 µm, which suggests accurate deposition of acoustic energy in the ablation. HIFU treatment is attractive in clinics because of its noninvasiveness, effective tissue ablation, few associated complications, and faster recovery for patients, and a large number of clinical trials have been successfully carried out (Zhou 2011).

While HIFU thermal ablation is the dominant interaction at low acoustic intensities, high intensities can introduce other bioeffects. The large tensile wave of HIFU pulse can induce a cloud of bubbles whose rapid growth and violent collapse mediates the tissue destruction, such as the hemorrhagic lesion over the spleen, intestines, and peritoneum and subcapsular hematomas due to vascular rupture (Crum and Hansen 1982, Zhong et al 2001, Matlaga et al 2008). Nonlinear acoustic effects on finite-amplitude bursts result in the formation of a shock front when propagating towards the focus. The abrupt pressure change creates significant mechanical stress in the tissue and growth and coalescence of microcracks by the cyclic stress (Lokhandwalla and Sturtevant 2000). The waveform distortion results in the generation of high order harmonics. Because the absorbed acoustic energy is proportional to the frequency, temperature could rise to 100 °C in milliseconds during HIFU sonication to produce large boiling bubbles (Bailey et al 2003). Recent research has shown that the presence of shocks and bubbles can break up or mechanically erode soft tissues to tiny debris in histotripsy and boiling histotripsy (Maxwell et al 2012). The fragmentation effect, liquefying the tissue rather than thermally destroying it for easily discharging through the natural body orifices or being reabsorbed by surrounding tissue, is similar to extracorporeal shockwave lithotripsy (ESWL) for kidney stones. Both histotripsy techniques have been demonstrated in animal studies, such as the treatment of malignant tumors, benign prostatic hyperplasia, deep vein thrombosis, and congenital heart defects (Roberts et al 2006, Xu et al 2010, Khokhlova et al 2011). However, mechanisms of tissue erosion generated by HIFU pulses still have not been fully understood.

Hydrodynamic cavitation is another type of cavitation, and is generated by the pressure gradients in a stream of variable speed in which nucleuses of bubble migrate (Shestakov 2005). It has different characteristics of pressure and time scale of bubble cavitation in com-parison to acoustic cavitation (Arrojo and Benito 2008). The hydrodynamic cavitation causes the bubble collapse and produces shock waves, micro-jets, and micro-streaming. As a result, the material in proximity to the collapsing bubble will be eroded. The damage starts as pitting and may gradually become large (Young 1989). Microscale hydrodynamic cavitation through

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6653

a 147 µm-probe at 9790 kPa pressure has been successfully used for kidney stone erosion at a rate of 0.31 mg min−1 (Perk et al 2012). In addition, hydrodynamic force could disintegrate covalent bond of long polymers and yeast cells when the pressure difference across the drop exceeds their surface tension (Doulah et al 1975, Save et al 1994).

In our previous study, mechanical erosion in the shape of a ‘tadpole’ at the center and ther-mal necrosis on the boundary were produced in the gel phantom as well as in ex vivo porcine kidney using HIFU pulses with pulse duration of 20 ms that is shorter than the time for boiling and pulse repetition frequency (PRF) of 1 Hz (Zhou and Gao 2013). The erosion is mainly in the pre-focal region towards HIFU transducer. In this study, the bubble dynamics immediately after the termination of HIFU exposure was investigated using high-speed photography, from which the role of hydrodynamic cavitation in the mechanical erosion of gel phantom was illustrated. It is found that the bubble will expand and move towards the transducer in 1–2 ms and then shrink. The characteristics of hydrodynamic cavitation (i.e. the maximum bubble displacement, the initial bubble size, and the maximum bubble size) were derived with the progress of HIFU exposure. HIFU termination-induced hydrodynamic cavitation depends on the HIFU parameters (varied pulse duration of 10–30 ms and PRF of 0.5–2 Hz), the number of pulses delivered, and the characteristics of the pre-existing lesion (i.e. distribution of tiny bubbles around the mechani-cal erosion). Mechanical erosion produced in ex vivo porcine kidney had similar tadpole shape of bubble contour found in the phantom. Emission of disintegrated and liquefied tissue through the punctured site on the kidney surface towards the transducer was observed in B-mode sonogra-phy after HIFU exposure. It suggests that hydrodynamic cavitation after the termination of HIFU exposure may be another mechanism for the bubble dynamics and subsequent tissue erosion.

Material and methods

HIFU transducer

An annular focused HIFU transducer (H-102, outer diameter = 69.94 mm, inner diame-ter = 22.0 mm, F = 62.64 mm, Sonic Concepts, Woodinville, WA, USA) working at its third harmonic frequency (3.3 MHz) was used in this study (see figure 1). The HIFU transducer was immersed in the degassed and deionized water (O2 < 4 mg l−1, T = 25 °C, measured by DO700, Extech Instrument, Waltham, MA, USA) of a Lucite tank (L × W × H = 70 × 50 × 30 cm) and driven by sinusoidal bursts produced by a function generator (AF3021B, Tektronics, Beaverton, OR, USA) together with a 55 dB power amplifier (A150, ENI, Rochester, NY, USA). An acoustic absorber was put on the opposite wall of the testing tank to prevent the ultrasound reflection. The HIFU transducer was attached to a three-axis positioning system (XSlide, Velmex, Bloomfield, NY, USA) to align its focus to the desired position (i.e. 2 cm from the sample surface). A LabView program (National Instruments, Austin, TX, USA) was written to control the pulse delivery. Pulse duration varied from 10 to 30 ms, and 200 HIFU pulses were delivered to the gel phantom at varied pulse repetition frequency (PRF) from 0.5 Hz to 2 Hz.

Gel phantom

Polyacrylamide was used to prepare optically transparent gel phantoms, which have the acoustic impedance of 1.6 MRayls, the speed of sound of 1544 m s−1, and the acoustic attenuation of 0.15 dB/cm/MHz (Lafon et al 2005). The liquid mixture of gel constituents was degassed for about 1 h in a descant chamber (420100000, Scienceware, Pequannock, NJ, USA) with a vacuum pump (VTE8, Thomas, Sheboygan, WI, USA) at a pressure of 150 mbars, then poured into a mold for polymerization at room temperature within 1 min

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6654

of adding N,N,N′,N′-tetra-methylethylene/diamine (TEMED, Sigma-Aldrich, St. Louis, MO, USA). No bovine serum albumin (BSA) was added in order to avoid the optical opaque due to protein denaturalization at high temperature.

High-speed imaging system

Bubble dynamics in the tissue-mimicking gel phantom were observed using a high-speed camera (S-PRI plus, AOS Technologies AG, Dättwil, Switzerland), working at 1000 frames per second in a resolution of 640 by 480 pixels. An optical lens (focal length from 12.5 to 53 mm, f/1.8D, Nikon, Tokyo, Japan) and a set of the close-up lens (Nikon) were mounted on the camera to achieve an appropriately magnified view of bubble cavitation. The resolution of the established imaging system was calibrated by photographing a ruler and determined to be 0.04 mm/pixel. The camera was triggered by a transistor-transistor logic (TTL) pulse from a digital delay generator (DG535, Stanford Research Systems Inc, Sunnyvale, CA, USA) to synchronize the photography with HIFU exposure. To capture bubble dynamics in the whole progress of HIFU exposure, the camera started and ended 10 ms before and at least 50 ms after each HIFU pulse, respectively. A 1000 W white light (Model H10, Hi-Tech Electronics, Singapore) was used as back-illumination in the photography. Under such setup, the gel phantom and bubble appeared as transparent background and dark region, respectively. The individual frame could be extracted from the captured videos using software (AOS imaging Studio V3, AOS Technologies AG) companied with the high-speed camera.

Characterization of bubble dynamics

In the extracted images the bubble displacement, Δd, and change of bubble size, D(t), after the termination of HIFU exposure could be determined quantitatively in digital image processing

Figure 1. The schematic diagram of experimental setup.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6655

software (Photoshop, Adobe System, San Jose, CA, USA) as shown in figure 2. Here bub-ble width was used because of asymmetry of the bubble in the axial direction. Such a work was repeated until the disappearance or collapse of this bubble. Three parameters were used to characterize the hydrodynamic cavitation: the bubble maximum displacement towards the source (HIFU transducer), initial bubble size, and the maximum bubble expansion size according to the release of sonication.

Ex vivo study

In ex vivo experiment, fresh porcine kidney purchased from a local slaughterhouse (Primary Industries Pte Ltd, Singapore) was immersed in phosphate-buffered saline (PBS) solution, degassed for at least 30 min, and used within 4 h of harvest for HIFU ablation. Kidney sample was fixed in front of gel phantom (see figure 8(a)), and HIFU focus was aligned under the guidance of an ultrasound imaging system (T3000, Terason, Burlington, MA, USA) with a linear array probe (8L2, Terason). B-mode sonography during HIFU ablation was recorded and then stored in the control PC. After the ablation, porcine kidney samples were cut, and the generated lesion was photographed by a digital camera (PowerShot SX230 HS, Canon, Tokyo, Japan).

Statistical analysis

At each experimental condition, at least five data were collected. ANOVA was performed in SPSS® Statistics (IBM Software, Somers, NY, USA) to determine the statistical difference between two testing groups that was fixed at the confidence value of p < 0.05.

Results

Hydrodynamic bubble cavitation

High-speed camera recorded the bubble dynamics after the termination of HIFU exposure up to 200 pulses. Representative images of hydrodynamic cavitation (pulse duration of 20 ms and PRF of 1 Hz) were shown in figure 3(a), and the corresponding changes to bubble size

Figure 2. The method of calculating bubble displacement, Δd, initial size at the termination of ultrasound exposure, D0, and expansion size, D(t), from high-speed photography taken (a) at and (b) after the termination of ultrasound exposure.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6656

in figure 3(b). It is found that 1–2 ms after the HIFU exposure the bubble moved towards the transducer, which may be caused by the sudden decrease of pressure or applied acoustic radiation force to it and the recovery of compressed material towards its initial position, and expanded to its maximum size. The velocity of bubble motion was around 800–1000 mm s−1. Meanwhile, a plenty of small bubbles remained in the postfocal region after the jet emission (Zhou and Gao 2013), and the bubble jet induced by HIFU exposure (i.e. Δt = 0 ms, N = 20) retreated. Afterwards, the bubble shrank exponentially in size. Bubble dissolution took mostly more than 100 ms due to the high viscosity and limited liquid content of gel phantom. The

Figure 3. (a) Representative high-speed images of bubble dynamics in the gel phantom using HIFU pulses with pulse duration of 20 ms and pulse repetition frequency of 1 Hz at the delivered 1, 20, 50, 100, and 200 pulses and retarded time of 0, 1, 2, 5, 10, 20, and 45 ms, and (b) the corresponding changes of bubble size.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6657

first bubble induced by HIFU was quite small (~0.5 mm in diameter), and the bubble expan-sion by hydrodynamic cavitation was insignificant (increasing by 1.14 folds). At N = 20 and N = 50, the bubble expansion was clear (increasing by 1.76 and 1.56 folds, respectively). With more HIFU pulses delivered (N = 100–200) the bubble expansion and shrinkage became less significantly again although the initial bubble size was large. In addition, the asymmetry of bubble dynamics (lateral versus axial direction) became significant. The characteristics of hydrodynamic cavitation after the termination of HIFU exposure are summarized and listed in table 1.

Effect of pulse duration

Keeping the PRF as 1 Hz while varying the pulse duration from 10 ms to 30 ms results in significant changes (see figure 4). With the increase of pulse duration, there were more tiny scattered bubbles in the tail of lesion, and the asymmetric expansion and shrinkage of the initially larger bubble were more significant (more retaining in the axial direction than that in the lateral one). The bubble displacement increased monotonically with the number of pulses delivered at the pulse duration of 10 ms, but became a peak distribution for longer pulses with the peak moving towards the beginning of exposure (see figure 5(a)). The initial bubble sizes induced by HIFU pulse with the pulse duration of 20 ms and 30 ms were quite similar at N = 20–50, but much larger than those of 10 ms. Afterwards, 20 ms-bursts generated larger initial bubbles, which may be due to the shielding effect of many scattered tiny bubbles in the prefocal region induced by 30 ms-bursts (see figure 5(b)). The changes of bubble size (Dmax − D0) by 10 ms and 30 ms bursts gradually increased and decreased during the HIFU exposure, respectively. For 20 ms burst, the trend had a peak at N = 50.

Effect of PRF

The bubble dynamics due to hydrodynamic cavitation at PRF of 0.5 Hz and 2 Hz are shown in figure 6, and the corresponding bubble displacement and size are shown in figure 7. Using the low PRF (i.e. 0.5 Hz) the scattered tiny bubbles in the tail of lesion were reduced because of longer pulse interval time for bubble dissolution, and the changes of bubble size had small variation from 0.19 ± 0.05 mm to 0.28 ± 0.09 mm with the progress of HIFU exposure. In comparison, the characteristics of hydrodynamic cavitation induced by 20 ms HIFU pulses at PRF of 2 Hz were similar to those of 30 ms ones at PRF of 1 Hz. Although the initial bubble size reached its maximum at N = 20, the changes of bubble size was minimal (0.21 ± 0.11 mm). Afterwards, the initial bubble size gradually decreased to 1.16 ± 0.11 mm at N = 200. The bubble displacements induced at PRF of either 0.5 Hz or 2 Hz were overall smaller than that at PRF of 1 Hz with significant differences (p < 0.05).

Ex vivo study

The lesion in porcine kidney induced by 200 HIFU pulses at the pulse duration of 20 ms and PRF of 1 Hz is shown in figure 8(b). Two types of lesion were found clearly: thermal lesion due to the temperature elevation and mechanical erosion due to both acoustic and hydrody-namic cavitation (Zhou and Gao 2013). Round tail of mechanical erosion had similar shape as hydrodynamic cavitation-induced bubble found in the gel phantom using high-speed images. If the HIFU focus was close to the surface of the kidney sample, fragmented renal tissues were found retreated into the coupling degassed water through the punctured site on the tissue surface (see figure 9).

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6658

Tab

le 1

. C

ompa

riso

n of

hyd

rody

nam

ic c

avita

tion

char

acte

rist

ics

at th

e te

rmin

atio

n of

HIF

U p

ulse

s at

the

puls

e re

petit

ion

freq

uenc

y of

0.5

Hz,

1

Hz,

and

2 H

z an

d va

ried

pul

se d

urat

ion

of 1

0 m

s, 2

0 m

s, a

nd 3

0 m

s.

Puls

e

dura

tion

(ms)

Puls

e re

petit

ion

fr

eque

ncy

(Hz)

Max

(Δd)

(m

m)

Nm

ax(Δ

d)

Max

(D0)

(m

mx)

()

ND

max

0

Max

(Dm

ax)

(m

m)

()

ND

max

max

Max

(Dm

ax −

D0)

(m

m)

()

−N

DD

max

max

0

101

0.67

± 0

.18

200

0.55

± 0

.215

00.

84 ±

0.1

215

00.

32 ±

0.1

220

020

10.

96 ±

0.1

415

01.

4 ±

0.1

920

01.

68 ±

0.1

150

0.38

± 0

.08

5030

10.

99 ±

0.0

850

1.18

± 0

.22

150

1.52

± 0

.09

150

0.55

± 0

.15

2020

0.5

0.74

± 0

.04

501.

18 ±

0.1

620

01.

56 ±

0.1

120

00.

29 ±

0.1

120

020

20.

81 ±

0.1

650

1.06

± 0

.33

201.

31 ±

0.0

950

0.41

± 0

.17

100

Max

(Δd)

. max

imum

bub

ble

mot

ion,

Nm

ax(Δ

d): t

he n

umbe

r of

pul

ses

deliv

ered

for

the

max

imum

bub

ble

mot

ion,

max

(D0)

: max

imum

initi

al b

ubbl

e si

ze, N

max

(Δd)

: the

num

ber

of p

ulse

s de

liver

ed f

or th

e m

axim

um in

itial

bub

ble

size

, max

(Dm

ax):

max

imum

bub

ble

expa

nsio

n si

ze, N

Dm

axm

ax(

): th

e nu

mbe

r of

pul

ses

deliv

ered

for

the

max

imum

bub

ble

expa

nsio

n si

ze,

max

(Dm

ax −

D0)

: max

imum

cha

nge

of b

ubbl

e si

ze, N

DD

max

max

0(

)−

: the

num

ber

of p

ulse

s de

liver

ed f

or th

e m

axim

um c

hang

e of

bub

ble

size

.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6659

Discussion

HIFU could not only generate irreversible coagulation in the tissue thermally but also dis-integrate or liquefy tissue mechanically with the presence of bubble cloud or large boiling bubble. It is thought mostly that the mechanism of HIFU-induced tissue fragmentation is because of the interaction of distorted ultrasound bursts (i.e. shock front) with bubbles and, subsequently, violent bubble collapse in acoustic cavitation. In this study, bubble dynamics in the transparent gel phantom immediately after the termination of HIFU pulse were captured using high-speed photography. After ultrasound exposure, HIFU-induced bubble in the focal region moved towards the transducer and expanded within 1–2 ms, and then reduced its size gradually. The effects of pulse duration and pulse repetition frequency of HIFU burst on the maximum bubble displacement and change of bubble size were investigated. Although it is

Figure 4. (a) Representative high-speed images of bubble dynamics in the gel phantom using HIFU pulses with pulse duration of 10 ms and 30 ms and pulse repetition frequency of 1 Hz at the delivered 1, 20, 50, and 100 pulses and retarded time of 0, 1, 2, 5, 10, 20, and 45 ms, and (b) the corresponding changes of bubble size (solid line: 10 ms, dash line: 30 ms).

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6660

highly reasonable to extrapolate that the hydrodynamic scenario observed in the transparent gel phantom would also occur in ex vivo and in vivo tissue, direct evidence is not available now. In our ex vivo study, the puncture on the surface of kidney sample occurred after a certain

Figure 5. The effect of varied pulse duration from 10 ms to 30 ms on (a) the bubble displacement and (b) bubble size (D0: initial bubble size, Dmax: the maximum bubble size) after HIFU exposure at the pulse repetition frequency of 1 Hz.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6661

number of pulses delivered. Afterwards, fragmented and liquefied tissue emitted towards the HIFU transducer. The mechanical erosion in the tissue generally has the shape of tadpole, more erosion in the prefocal region, which is similar to the bubble motion, expansion, and shrink-age observed in the phantom. However, due to the limited frame rate of the most sonography systems (30–40 fps), the hydrodynamic cavitation in the tissue was not found here, but may be observed using compounded planar-wave transmissions at ultrafast frame rate (>1000 fps) (Montaldo et al 2009).

The hydrodynamic cavitation occurred towards HIFU transducer because of the pres-sure release and recovery of compressed material after the termination of HIFU exposure. Exponential time constants for bubble displacement by the acoustic radiation force are

Figure 6. (a) Representative high-speed images of bubble dynamics in the gel phantom using HIFU pulses with pulse duration of 20 ms and pulse repetition frequency of (a) 0.5 Hz and (b) 2 Hz at the delivered 5,10, 20, 50, 100, and 200 pulses and retarded time of 0, 1, 2, 5, 10, 20, and 45 ms.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6662

independent of the bubble radius and inversely proportional to the Young’s modulus of the surrounding medium if treating the bubble as a point-like target (Erpelding et  al 2005). The maximum bubble displacement is expressed as x Ir cE2max /= , where I is the acoustic

Figure 7. The effect of varied pulse repetition frequency from 0.5 Hz to 2 Hz on (a) the bubble displacement and (b) bubble size (D0: initial bubble size, Dmax: the maximum bubble size) after HIFU exposure with the pulse duration of 20 ms.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6663

intensity, r is the bubble radius, c is the speed of sound, and E is the Young’s modulus, under the assumptions of planar incident wave, completely acoustic reflection by the bubble, and ideal dynamics in the homogenous medium. However, there are several discrepancies in compariso n to our cases. First, the acoustic radiation force of HIFU applied to a large bub-ble is different from that to a small bubble (on the order of µm), F I r c2 2/π= . The arbitrary incident beam can be described as a sum of planar waves by employing angular spectrum decomposition (Sapozhnikov and Bailey 2013). Then a superposition of the scattered fields from all angular spectrum components (each planar incident wave) in the far field is used to derive the radiation stress tensor. Thus, radiation force is a second order quantity and more complicated especially for the radially asymmetric bubble. Second, bubble shape changes during displacement. Third, the radius-dependent drag coefficient and inertial terms are also ignored in the Voigt model, but should be accounted for large bubbles. So no good correlation was found between the bubble displacement and expansion with the initial bubble size and temporal-average acoustic intensity (R2 < 0.5, data not included).

Acoustic and hydrodynamic cavitation have inherent differences. Acoustic cavitation is produced by sound waves due to pressure variation while hydrodynamic cavitation is produced by pressure gradients in a stream of variable speed (Arrojo and Benito 2008). The pressure to cause acoustic cavitation is high, e.g. about 100 atm in pure water (Moholkar et al 1999) while the corresponding value for hydrodynamic cavitation is much lower, about 1–8 atm

Figure 8. (a) Representative photo of setup for ex vivo experiment and (b) the erosion generation inside the porcine kidney.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6664

(Jyoti and Pandit 2001). Another significant difference between them is the time scale of processes. Hydrodynamic cavitation process works from 1 to 50 ms which is three orders longer than 1–50 µs for acoustic cavitation. The experimental results illustrate the duration of bubble dynamics after HIFU exposure is on the order of ms. Even without the flow of

Figure 9. Representative B-mode ultrasound images of porcine kidney (a) before HIFU exposure and (b)–(d) after termination of HIFU exposure after 50 pulses (triangle: hyperechoes in the sonography due to the HIFU-generated bubble, arrow: emission of fragmented tissues into coupling degassed water, and (e) the photos of erosion on the surface of kidney.

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6665

liquefied material the dominant mechanism is the large pressure and stress gradient applied to the bubble so that the cavitation found here should be categorized as hydrodynamic cavita-tion. Although acoustic cavitation has been studied extensively over the past decades, there is hardly any industrial chemical processing due to the diverse expertise required (e.g. material science, acoustics, chemical engineering) for scaling up the lab setup. In comparison, hydro-dynamic cavitation is a cheaper and easier alternative and causes the fluid close to the bubble to move rapidly (e.g. microstreaming or microjet) (Moholkar and Pandit 1997). The micro-streaming causes large localized shear stress up to 600–800 dyn cm−2 and results in physi-cal damage to the materials (Williams et al 1970, Williams 1971). Hydrodynamic cavitation could lyse fatty oils (Pandit and Joshi 1993), disrupt cells (Save et al 1997), and polymerize/depolymerize aqueous polymeric solution (Chivate and Pandit 1993) that has been proved more energy-efficient compared to its counterpart.

Gel phantom studies showed that 20 ms HIFU burst at PRF of 1 Hz could achieve the maxi-mum bubble displacement and expansion among the all HIFU parameters tested, which cor-relates well with the largest mechanical erosion in ex vivo tissue in our previous study (Zhou and Gao 2013). The maximum bubble expansion depends on the magnitude of acoustic radia-tion force applied to the bubble, the bubble induced by HIFU pulse, and the medium proper-ties (i.e. viscoelastic and intact tissue, or liquefied tissue debris, or coagulated tissue with increased stiffness). Small bubble at the early stage of HIFU exposure at low power or short pulse duration in the gel phantom has a small bubble displacement and expansion. Although the bubble size becomes larger at the late stage and liquefied phantom has less damping and resistant effect on bubble motion, the presence of tiny scattered bubbles in the prefocal region would reduce the equivalent radiation force applied to the bubble even at long pulse duration. HIFU bursts with pulse duration of 30 ms at PRF of 1 Hz were found to produce the larger thermal lesion in ex vivo tissue but less mechanical erosion in comparison to that of 20 ms HIFU bursts (Zhou and Gao 2013). Pulse interval time at high PRF (i.e. 2 Hz) may not be long enough for complete bubble dissolution. However, the undissolved bubble at the focus may be beneficial for bubble growth by the following HIFU pulse. Thus, low PRF (i.e. 0.5 Hz) also results in small initial bubble for less hydrodynamic bubble expansion. Altogether, opti-mal HIFU parameters for effective and maximum hydrodynamic cavitation could produce large tissue erosion. Furthermore, transmitting a secondary ultrasound pulse to bubble at its collapse stage may enhance the collapse strength between two primary HIFU bursts, which has already been used in ESWL to increase the stone comminution by 60–80% (Xi and Zhong 2000, Zhou et al 2004). The amplitude and duration of the acoustic pulse should be optimized because of the different characteristics of hydrodynamic cavitation in comparison to those of acoustic cavitation.

Conclusions

The hydrodynamic cavitation (bubble displacement towards the transducer and change of bubble size on the order of ms) after the termination of HIFU exposure illustration using high-speed photography in this study may be another important mechanism in mechanical tissue disintegration using HIFU technology. A good agreement was found between with the find-ings of gel phantom and ex vivo tissue experiment in which the larger mechanical erosion is located in the pre-focal region. In comparison to the acoustic cavitation, the characteristics of hydrodynamic cavitation are significantly different. Large initial bubble size (~1 mm), bubble displacement (1 mm), and significant bubble expansion (by 1.7 folds) result in large shear stress applied to the surrounding medium for tissue disintegration in a large volume. Pulse

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6666

duration, PRF of HIFU exposure, and the number of pulses delivered have an influence of hydrodynamic cavitation. Optimal utilization of such a mechanism may lead to enhancement in the safety and efficiency of tissue disintegration and improved possibility of clinical trans-lation. However, the role of cavitation in the mechanical erosion still remains incompletely understood. Deep insights are required in the physical investigation and medical applications.

Acknowledgment

This study was financially supported by the proof-of-concept (POC) grant of National Research Funding (NRF), Singapore, 2014NRF-POC001-039. The authors would like to thank Miss Siyu Carine Tan in the help of carrying out the experiment.

References

Arrojo  S and Benito  Y 2008 A theoretical study of hydrodynamic cavitation Ultrason. Sonochem. 15 203–11

Bailey  M, Khokhlova  V, Sapozhnikov  O, Kargl  S and Crum  L 2003 Physical mechanisms of the therapeutic effect of ultrasound (a review) Acoust. Phys. 49 369–88

Chivate M and Pandit A 1993 Effect of hydrodynamic & sonic cavitation on aqueous polymeric solutions Indian Chem. Eng. 35 52–7

Crum L and Hansen G 1982 Growth of air bubbles in tissue by rectified diffusion Phys. Med. Biol. 27 413–7

Doulah M, Hammond T and Brookman J 1975 A hydrodynamic mechanism for the disintegration of saccharomyces cerevisiae in an industrial homogenizer Biotechnol. Bioeng. 17 845–58

Erpelding T N, Hollman K W and O’Donnell M 2005 Bubble-based acoustic radiation force elasticity imaging IEEE Trans. Ultrason. Ferroelectr. Freq. Control 52 971–9

Jyoti K and Pandit A B 2001 Water disinfection by acoustic and hydrodynamic cavitation Biochem. Eng. J. 7 201–12

Khokhlova  T  D et  al 2011 In vivo tissue emulsification using millisecond boiling induced by high intensity focused ultrasound J. Acoust. Soc. Am. 129 2477

Lafon C, Zderic V, Noble M L, Yuen J C, Kaczkowski P J, Sapozhnikov O A, Chavrier F, Crum L A and Vaezy S 2005 Gel phantom for use in high-intensity focused ultrasound dosimetry Ultrasound Med. Biol. 31 1383–9

Lokhandwalla M and Sturtevant B 2000 Fracture mechanics model of stone comminution in ESWL and implications for tissue damage Phys. Med. Biol. 45 1923–40

Matlaga B R, McAteer J A, Connors B A, Handa R K, Evan A P, Williams J C, Lingeman J E and Willis  L  R 2008 Potential for cavitation-mediated tissue damage in shockwave lithotripsy J. Endourol. 22 121–6

Maxwell A, Sapozhnikov O, Bailey M, Crum L, Zhen X, Fowlkes B, Cain C and Khokhlova V 2012 Disintegration of tissue using high intensity focused ultrasound: two approaches that utilize shock waves Acoust. Today 8 24–37

Moholkar V, Kumar P S and Pandit A 1999 Hydrodynamic cavitation for sonochemical effects Ultrason. Sonochem. 6 53–65

Moholkar V S and Pandit A B 1997 Bubble behavior in hydrodynamic cavitation: effect of turbulence AIChE J. 43 1641–8

Montaldo G, Tanter M, Bercoff J, Benech N and Fink M 2009 Coherent plane-wave compounding for very high frame rate ultrasonography and transient elastography IEEE Trans. Ultrason. Ferroelectr. Freq. Control 56 489–506

Pandit A and Joshi J 1993 Hydrolysis of fatty oils: effect of cavitation Chem. Eng. Sci. 48 3440–2Perk O Y, Şeşen M, Gozuacik D and Koşar A 2012 Kidney stone erosion by micro scale hydrodynamic

cavitation and consequent kidney stone treatment Ann. Biomed. Eng. 40 1895–902Roberts W W, Hall T L, Ives K, Wolf J S, Fowlkes J B and Cain C A 2006 Pulsed cavitational ultrasound:

a noninvasive technology for controlled tissue ablation (histotripsy) in the rabbit kidney J. Urol. 175 734–8

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651

6667

Sapozhnikov O A and Bailey M R 2013 Radiation force of an arbitrary acoustic beam on an elastic sphere in a fluid J. Acoust. Soc. Am. 133 661–76

Save S, Pandit A and Joshi J 1994 Microbial cell disruption: role of cavitation Chem. Eng. J. Biochem. Eng. J. 55 B67–72

Save  S, Pandit  A and Joshi  J 1997 Use of hydrodynamic cavitation for large scale microbial cell disruption Food Bioprod. Process. 75 41–9

Shestakov S 2005 Mathematical model of hydrodynamic cavitation Sixteenth Session Russ. Acoust. Soc. 2 71–3

ter Haar G 2001 Acoustic surgery Phys. Today 54 29–34Williams  A 1971 Disorganization and disruption of mammalian and amoeboid cells by acoustic

microstreaming J. Acoust. Soc. Am. 50 99–100Williams  A, Hughes  D and Nyborg  W 1970 Hemolysis near a transversely oscillating wire Science

169 871–3Xi  X and Zhong  P 2000 Improvement of stone fragmentation during shock-wave lithotripsy using

a combined EH/PEAA shock-wave generator—in vitro experiments Ultrasound Med. Biol. 26 457–67

Xu Z, Owens G, Gordon D, Cain C and Ludomirsky A 2010 Noninvasive creation of an atrial septal defect by histotripsy in a canine model Circulation 121 742–9

Young F 1989 Cavitation (New York: McGraw-Hill)Zhong P, Zhou Y and Zhu S 2001 Dynamics of bubble oscillation in constrained media and mechanisms

of vessel rupture in SWL Ultrasound Med. Biol. 27 119–34Zhou Y-F 2011 High intensity focused ultrasound in clinical tumor ablation World J. Clin. Oncol. 2 8–27Zhou Y, Cocks F H, Preminger G M and Zhong P 2004 Innovations in shock wave lithotripsy technology:

updates in experimental studies J. Urol. 172 1892–8Zhou Y and Gao X W 2013 Variations of bubble cavitation and temperature elevation during lesion

formation by high-intensity focused ultrasound J. Acoust. Soc. Am. 134 1683–94

Y Zhou and X W Gao Phys. Med. Biol. 61 (2016) 6651