Real-Time Monitoring of High Intensity Focused Ultrasound Lesion Formation with Combined Acoustic Force Elastography and Acousto-Optic Imaging. Andrew Draudt, Puxiang Lai, Ronald A. Roy, Todd Murray, Robin O. Cleveland Department of Mechanical Engineering, Boston University, Boston, MA 02215 Abstract: High Intensity Focused Ultrasound (HIFU) is a non-invasive method by which ultrasound can be used to thermally ablate tissue (tumors, for example). However, real-time imaging of the tissue heating and lesion formation process is challenging and is a major barrier to widespread clinical use of HIFU. Here we report on an approach to monitor HIFU lesions by means of changes in both optical and mechanical properties. The optical properties were detected by means of an acousto-optic interaction between the HIFU and a diffuse optical wave. The stiffness was detected by acoustically forcing the tissue with the HIFU beam and using a high frequency ultrasound probe to detect the motion. By simultaneously combining multiple modalities we anticipate developing a robust approach to tracking HIFU lesion formation in real time.

Motivation and State of the Art

Technical Approach

Clinical applications of HIFU therapy employ MRI to image the ultrasound-induced lesions as they are being produced. However, it’s expense and space requirements inhibit the adoption of HIFU as a viable therapy for cancer. Imaging lesion formation using ultrasound would be preferable, but lesions do not show detectable contrast with standard ultrasound imaging. Other techniques investigated include acoustic forcing elastography (1,2), and attenuation imaging (3). However, since no single approach has seen clinical success, our goal is to combine mechanical strain imaging with acousto-optic imaging to obtain a robust multi-property imaging system.

1. Acoustic Radiation Force Elastography. The HIFU transducer is used to simultaneously create the lesion and apply a 39Hz oscillatory radiation force to the lesion area, resulting in local tissue motion in the lesion region. Tissue necrosis causes a marked change in stiffness, which results in a change in this tissue motion. Motion is monitored using a pulsed diagnostic transducer and “speckle tracking” correlation software. See refs. 1, 2. 2. Acousto-Optic Imaging. Use the significant advances made in our department in the area of Acousto-Optic Imaging (AOI) to image HIFU lesions, which exhibit optical contrast. Real-time results are obtained by using a lock-in amplifier tuned to the 39Hz signal used in the forcing modulation described above.

Acousto-Optical Imaging Results

Experimental Setup

Motion is induced in tissue via acoustic forcing. An intense sound field exerts a pushing force at the focus. In our case, the beam is amplitude-modulated at 39Hz, causing the tissue to oscillate. Measuring the magnitude of this motion yields the local tissue stiffness (Hooke’s law), which will change when a lesion is formed. Tissue motion is measured by “speckle tracking”. Ultrasound imaging A-lines like the one shown at right consist of echoes from the scattering particles in the tissue. If there are local movements in the tissue, the echoes will shift over successive time frames. Crosscorrelation between segments of successive A-lines indicate how much each segment has moved.

1.1 MHz HIFU transducer , with concentric 7.5 MHz pulse transducer)

BK 12 MHz linear array probe

To laser

Below right are measured results, and the output of a finite element (Comsol Inc.) model of the entire process. Lens 3

Focal zone

Interrogation transducer

Tissue HIFU transducer

-6

Water tank 2.5

2 amp(m)

Measured sinusoidal amplitude taken before, during, and after lesion formation. Chicken breast exposed to 7MPa pk pressure Hifu and interrogated before(blue) and after(green) lesion formation.

x 10

1.5

Tissue or Phantom

1

Optical signal beam

Schematic showing lesion zone. 0.5

0

0.005

0.01

4

0.015 0.02 depth(m)

0.025

0.03

1000

1050 1100 time samples

1150

1200

4

x 10 1 0

3.5

x 10

3

Conclusion

2.5

amp(m)

volts

Finite element model of same conditions above, accounting for thermal lesion formation and acoustic forcing. Absorption and stiffness are functions of cumulated thermal exposure.

0 -1 950

volts

Setup used for making HIFU lesions in tissue samples, and detecting their formation using ultrasonic elastography and acousto-optic imaging.

-6

1

2

•The first real-time detection of a HIFU lesion using acousto-optic imaging has been accomplished.

1.5

1

1 -1 950

1000

1050 1100 time samples

1150 0.5

Regions of two successive A-lines, showing how a segment has shifted in time.

0

0

0.005

0.01

Problem: Nonlinearity

0.015 0.02 depth(m)

0.025

0.03

0.035

• Acoustically-forced tissue motion has been observed, and a major source of error in this method has been discovered and solved. FEM models reveal that the displacement is governed my many conflicting processes, revealing the limits of this method for lesion monitoring.

6

4 2 0

4 2 0

phase=90 phase=270 Average of phase 90 and 270

8 6

Continuously flipping the phase relationship of the imaging pulse with respect to the HIFU wave enabled the cancellation of the nonlinear effect by averaging the displacement calculated for each phase.

-2

-2

10

dis plac em ent (m ic rons )

6 pressure (MPa)

pressure (MP a)

8

4 2 0 -2 -4 -6 -8

-4

-4 -1

-0.5

0 0.5 time (microseconds)

1

-1

-0.5

Acoustic density fluctuations cause changes in optical index of refraction, absorption, and scattering. i Diff Diffuse photons h that h pass through the HIFU focus will thus have their velocities changed, causing small changes in the output optical wavefront. These photons are detected by mixing the light with a reference beam within a photo-refractive crystal (PRC). The crystal turns the interference pattern into a grating, which diffracts the beam into a photodetector. Acoustically-induced changes in the wavefront from the sample don’t match the original interference pattern, and won’t be diffracted by the PRC. The optical detector level thus decreases. A higher optical absorption in the focal region, as in the case when a lesion forms, will weaken the above acousto-optic effect, decreasing the drop in signal. In order to enable real-time detection of lesion formation, the signal-to-noise ratio is boosted by taking advantage of the f i modulation forcing d l i employed l d iin the elastography portion of this work. By bursting the HIFU at 39Hz, the resulting switched photodetector signal is fed into a lock-in amplifier tuned to the modulation frequency.

frequency reference

Signal generator (burst mode)

output

Lock-in amplifier

waveform Figure courtesy of Lei Sui.

Representative photos of lesion made with 8MPa, 30Hz modulated HIFU beam. 35

30

25

Lesion absorbs or scatters light, reducing creation of tagged photons, so l k i signal lock-in i ld drops.

20

15

10

5

Hifu turned on at time=0.

0

0

5

10

15

20

25 time(s)

30

35

40

45

50

Output of lock-in, showing lesion formation in real time

Solution: Phase Averaging

The speed of sound in a fluid changes with pressure. Thus, the speed at which the imaging pulse travels depends on whether it travels in a high pressure (below left) or low pressure (below right) part of the forcing HIFU wave. Since displacement is calculated from the arrival time of the echoes of the pulse from the tissue, a change in the speed of the pulse will result in a false displacement. 8

To photodetector

0.035

x 10

Monitoring lesion formation has been a long time technological goal of researchers , doctors, and makers of ultrasound equipment. Our research will aid them in this task. Since we are the first to show that lesions can be imaged with AOI, commercialization is possible. The field appears to be open.

lo ock-in output(mV)

Acoustic Force Elastography Results

Technology Transfer

0 0.5 time (microseconds)

Figure depicting an interrogation pulse “riding” on a high pressure crest(left), and low pressure trough(right) of the forcing HIFU wave. Due to nonlinearity, the pulse can arrive earlier or later than it should.

1

-10

0

5

10 15 20 axial distance into tissue(mm)

25

30

Tissue displacement vs depth for two values of phase of interrogation pulse with respect to HIFU wave. Dashed lines show the result of averaging the two together. True values for displacement are then found by subtracting thermally-induced displacement error.

References:

Publications Acknowledging NSF Support:

1. B.Fahey, K.Nightingale, D.Stutz, G. Trahey. Acoustic Radiation Force Impulse Imaging of Thermally and Chemically-Induced Lesions in Soft Tissues: Preliminary ex-vivo results. Ultrasound in Med. & Biol., Vol. 30, No. 3, pp. 321–328, (2004) 2. Maleke C, Konofagou E. Harmonic motion imaging for focused ultrasound (HMIFU): a fully integrated technique for sonication and monitoring of thermal ablation in tissues. Phys. Med. Biol. 53 1773– 1793 3 Anand A, Kaczkowski P. Monitoring formation of high intensity focused ultrasound (HIFU) induced lesions using backscattered ultrasound ARLO 5(3) 88-94 (2004)

1. A. Draudt, P. Lai, R. A. Roy, T. W. Murray, and R. O. Cleveland, Detection of HIFU lesions in Excised Tissue Using Acousto-Optic Imaging International Symposium on Therapeutic Ultrasound 2008, (Minneapolis, MN, 2008) 2. A. Draudt, P. Lai, R. A. Roy, T. W. Murray, and R. O. Cleveland, A Multi-modal Approach to Monitoring HIFU Lesions. Presentation, ASME congress, 2008 3. P. Lai, R. A. Roy, and T. W. Murray, Sensing the optical properties of diffusive media by acoustooptic pressure contrast imaging SPIE 7177, 71771G

This work was supported in part by Gordon-CenSSIS, the Bernard M. Gordon Center for Subsurface Sensing and Imaging Systems, under the Engineering Research Centers Program of the National Science Foundation (Award Number EEC-9986821).

Pi Contact Information: Prof. Robin Cleveland Mechanical Engineering Dept. Boston University, Boston, MA, 02215 Phone: 617-353-7767 Email: [email protected]

Fundamental Science

R2 R1

R3

Validating TestBEDs Bio-Med

S1

S2 S3

Enviro-Civil

S4

S5