Simultaneous Passive Acoustic Mapping and Magnetic Resonance Thermometry for Monitoring of Cavitation-Enhanced Tumor Ablation in Rabbits Using Focused Ultrasound and Phase-Shift Nanoemulsions

Abstract

Thermal ablation of solid tumors via focused ultrasound (FUS) is a non-invasive image-guided alternative to conventional surgical resection. However, the usefulness of the technique is limited in vascularized organs because of convection of heat, resulting in long sonication times and unpredictable thermal lesion formation. Acoustic cavitation has been found to enhance heating but requires use of exogenous nuclei and sufficient acoustic monitoring. In this study, we employed phase-shift nanoemulsions (PSNEs) to promote cavitation and incorporated passive acoustic mapping (PAM) alongside conventional magnetic resonance imaging (MRI) thermometry within the bore of a clinical MRI scanner. Simultaneous PAM and MRI thermometry were performed in an in vivo rabbit tumor model, with and without PSNE to promote cavitation. Vaporization and cavitation of the nanoemulsion could be detected using PAM, which led to accelerated heating, monitored with MRI thermometry. The maximum heating assessed from MRI was well correlated with the integrated acoustic emissions, illustrating cavitation-enhanced heating. Examination of tissue revealed thermal lesions that were larger in the presence of PSNE, in agreement with the thermometry data. Using fixed exposure conditions over 94 sonications in multiple animals revealed an increase in the mean amplitude of acoustic emissions and resulting temperature rise, but with significant variability between sonications, further illustrating the need for real-time monitoring. The results indicate the utility of combined PAM and MRI for monitoring of tumor ablation and provide further evidence for the ability of PSNEs to promote cavitation-enhanced lesioning.

Keywords: Cavitation; Focused ultrasound; Magnetic resonance imaging; Passive acoustic mapping; Phase-shift nanoemulsion; Rabbit; Thermal ablation; Tumor ablation; VX2.

Figure 1:

Setup for dual mode imaging of focused ultrasound (FUS) ablation. (a) Experimental setup (not to scale). A 1.5MHz single element FUS transducer and ultrasound (US) imaging array were aligned and attached to a 3-axis positioning system. A tumor-bearing rabbit was coupled to the transducer using water. A receive-only surface coil was used for MR imaging. (b) Electrical setup. A two-step waveform generated by two arbitrary waveform generators (AWG) was combined, amplified and delivered to the FUS transducer. The US array was connected to a research ultrasound platform (Verasonics) via an extension cable and MR penetration panel. (c) Example images showing (left) T2-weighted MRI for treatment planning: the FUS transducer, tumor and approximate focus position are shown; (mid) MRI phase image showing focal heating; (right) B-mode US image: the outline of the thigh and approximate focus position are shown. Scale bar = 1cm. (d) Timing diagram. A trigger signal from the MR scanner was used to synchronize B-mode, passive acoustic mapping (PAM) and FUS pulses. A new MR image was obtained every 2.6s. The routine shown was repeated for the duration of each treatment.

Figure 2:

Effect of phase-shift nanoemulsion (PSNE) on passive acoustic mapping as a function of electrical power. (a) Acoustic maps showing the summation of acoustic emissions with and without PSNE. The imaging array is located at the top of the images and focused ultrasound propagation is out of the page. (b) Peak amplitude from summed acoustic maps for 3 trials of each condition (n=3 repeats × 3 powers × 2 animals). (c) Averaged spectra from raw ultrasound data with and without PSNE at 32W. The frequency axis is normalized with respect to the FUS transmission frequency of 1.5MHz.

Figure 3:

Effect of phase-shift nanoemulsion (PSNE) on passive acoustic mapping over time (32W electrical power). (a) Acoustic maps showing the summation of acoustic emissions for subsets of the data with and without PSNE. The imaging array is located at the top of the images and focused ultrasound propagation is out of the page. The control maps are shown with two different color scales for comparison. (b) Top: peak amplitude and Mid: focal index over time for 3 repeats of each condition (n=3 repeats × 2 animals); Bottom: comparison of the two metrics over all of the time points.

Figure 4:

Effect of phase-shift nanoemulsion (PSNE) on MRI-observed temperature rise as a function of electrical power. (a) Temperature maps about the focus from sonications with and without PSNE at three power levels. Maps at the time point of maximum heating are shown.Focused ultrasound propagation is from bottom to top. Scale bar = 1cm. (b) Top: Maximum temperature rise; Bottom: width (FWHM = full-width half-maximum) of the heated region exceeding 50°C for 3 trials of each condition (n=3 repeats × 3 powers × 2 animals).

Figure 5:

Effect of phase-shift nanoemulsion (PSNE) on MRI-observed temperature rise over time (32W electrical power). (a) Temperature maps about the focus at selected time points for sonications with/without PSNE. Focused ultrasound propagation is from bottom to top. Scale bar = 1cm. (b) Maximum temperature rise in all temperature maps for 3 trials of each condition (n=3 repeats × 2 animals).

Figure 6:

Correlation between MRI and passive acoustic mapping (PAM) metrics. Maximum temperature rise from MRI vs. (a) peak amplitude from PAM at each time point, (b) cumulative peak amplitude from PAM up to each time point, (c) focal index from PAM at each time point. Data from 266 time points are included (n=3 repeats × 3 powers × 2 animals × c. 15 time points per treatment). The PAM and MRI data were matched based upon the time stamps of each recording; PAM data were averaged over the time between adjacent MR images. ρ=Pearson’s correlation coefficient.

Figure 7:

No phase-shift nanoemulsion (control): histological section of the thigh muscle of a rabbit with a VX2 tumor after treatment. The lesion has an appearance typical for thermal lesions. A well-demarcated area of tumor (green arrows) has three zones of tissue injury: a thermally coagulated central core (a), surrounding ring of edematous tissue (b), and adjacent tumor region (c) with some signs of thermal damage, and small tumor cell foci adjacent to the necrotic area. Tumor cells in the central core are shrunken with condensed nuclei. Beyond the thermal lesion (d), the intact tumor exhibits a relatively solid pattern. The tumor cells are round or oval, are rich in cytoplasm and show many mitotic figures (M). Dotted areas (a-c) are enlarged on right to show detail; detail image (d) of normal tumor was taken from a region away from any thermal lesions. Scale bar = 1mm for main image, 100μm for detail images.

Figure 8:

With phase-shift nanoemulsion (PSNE): histological section of the thigh muscle of a rabbit with a VX2 tumor after treatment. The lesion has an appearance typical for thermal lesions. Note the large areas of coagulative necrosis (a) in the central lesion core, the cavities produced by vaporization of the nanoemulsion (b), multiple hemorrhages resulted from inertial cavitation (c), hemorrhagic necrotic areas (d), and small tumor foci adjacent to the central lesion core (e). Dotted areas (a-e) are enlarged on right to show detail. Scale bar = 1mm for main image, 100μm for detail images.

Figure 9:

Cumulative amplitude of acoustic emissions and maximum heating at 32W electrical power (n=3 animals with PSNE, 3 control). (a) Peak amplitude from cumulative passive acoustic maps (PAM) over duration of each sonication. (b) Peak temperature rise at time of maximum heating from each sonication. ‘Control’ indicates sonications without phase-shift nanoemulsion (PSNE) (n=54 sonications), ‘PSNE’ indicates sonications with PSNE (n=52 sonications), ‘PSNE + Cav’ indicates the subset of PSNE sonications in which acoustic emissions exceeded those in the corresponding control animal (n=19 sonications). Stars indicate statistical significance assessed using 1-way ANOVA followed by Tukey-Kramer multiple comparison testing (***=p<0.001, **=p<0.01, n.s.=not significant).