Magnetic resonance imaging-guided volumetric ablation of symptomatic leiomyomata: correlation of imaging with histology

Abstract

Purpose: To describe the preliminary safety and accuracy of a magnetic resonance (MR) imaging-guided high-intensity-focused ultrasound (HIFU) system employing new technical developments, including ablation control via volumetric thermal feedback, for the treatment of uterine leiomyomata with histopathologic correlation.

Materials and methods: In this phase I clinical trial, 11 women underwent MR-guided HIFU ablation (Sonalleve 1.5T; Philips Medical Systems, Vantaa, Finland), followed by hysterectomy within 30 days. Adverse events, imaging findings, and pathologic confirmation of ablation were assessed. The relationship between MR imaging findings, thermal dose estimates, and pathology and HIFU spatial accuracy were assessed using Bland-Altman analyses and intraclass correlations.

Results: There were 12 leiomyomata treated. No serious adverse events were observed. Two subjects decided against having hysterectomy and withdrew from the study before surgery. Of 11 women, 9 underwent hysterectomy; all leiomyomata demonstrated treatment in the expected location. A mean ablation volume of 6.92 cm(3) ± 10.7 was observed at histopathologic examination. No significant differences between MR imaging nonperfused volumes, thermal dose estimates, and histopathology ablation volumes were observed (P > .05). Mean misregistration values perpendicular to the ultrasound beam axis were 0.8 mm ± 1.2 in feet-head direction and 0.1 mm ± 1.0 in and left-right direction and -0.7 mm ± 3.1 along the axis.

Conclusions: Safe, accurate ablation of uterine leiomyomata was achieved with an MR-guided HIFU system with novel treatment monitoring capabilities, including ablation control via volumetric thermal feedback.

Trial registration: ClinicalTrials.gov NCT00837161.

Conflict of interest statement

None of the other authors have identified a conflict of interest.

Copyright © 2012 SIR. Published by Elsevier Inc. All rights reserved.

Figures

Figure 1

MR-guided HIFU volumetric ablation. (a) Schematic of HIFU transducer and beam, applying focused acoustic energy in concentric circles within a treatment cell. (b) Treatment cells 4 mm, 8 mm, or 12 mm in diameter, with a ratio of cell diameter to length of approximately 1:2.5.

Figure 2

Intraprocedural MR-guided HIFU monitoring and MR imaging and histopathologic findings after leiomyoma ablation with MR-guided HIFU. (a– d) Graphic user interface displays multiplanar three-dimensional T2-weighted imaging and overlaid temperature maps (a– b) and overlaid thermal dose estimates (c– d) during sonication of an anterior intramural leiomyoma within the body of the uterus. Accumulated thermal dose information in the treated volume is displayed at the end of each sonication as a thermal dose estimate. These thermal doses are reported in CEM43, with 30 CEM43 (beige polygon, c– d) corresponding to onset of tissue alteration and 240 CEM43 (white polygon, c– d) representing predicted territory of complete necrosis. Both 30 CEM43 and 240 CEM43 thermal dose estimates are updated after each sonication. (e–f) Sagittal (e) and coronal (f) contrast-enhanced MR images after HIFU show nonenhancing treated region (black arrows). (g) Bivalved gross uterine specimen shows hemorrhagic necrosis in the area of treatment (white arrow). (h–j) Low-magnification (4×) histologic images of margin (h), high-magnification (10×) images of margin (i), and high-magnification images of the center of the ablation zone (j) confirm necrosis (asterisk) and narrow zone of transition (white arrows) between viable and necrotic HIFU-treated tissue. (Available in color online at www.jvir.org)

Figure 3

Comparison of mean temperature (T) elevations as a function of time during MR-guided HIFU heating of leiomyomata with 8-mm (a) and 12-mm (b) treatment cells (gray lines) versus feedback cells (black lines). Feedback cells refer to sonications performed with thermal feedback, for which exposure duration is determined by the system, and sonication duration is adjusted according to measured mean cell temperature. Treatment cells are sonications performed without thermal feedback for which sonication duration is fixed and linked with cell size. (a) An 8-mm treatment cell (mean temperature displayed as gray line) is set to deliver energy for a fixed duration of 27 seconds, despite ablative temperatures (60°– 62°C) not being achieved at the end of sonication (black arrow). Exposure duration for the 8-mm feedback cell is controlled by an algorithm monitoring mean temperature (black line) and continues heating until ablative temperatures are successfully achieved. (b) A 12-mm treatment cell (mean temperature displayed as gray line) delivers energy for a fixed time, exceeding the desired ablative temperature range after cessation of therapy (black arrow), unnecessarily lengthening the treatment time. Exposure duration of the 12-mm feedback cell is adjusted based on the mean temperature achieved (black line) and does not exceed the desired ablative temperature range during the course of therapy.

Appendix 2

Therapy planning software enables delineation of the planned treatment volume (PTV) and individual sonications during treatment planning. Each sonication cell (either treatment cell or feedback cell) may be assigned its own acoustic power setting, sonication frequency, and cell diameter. (a) Individual sonications are delineated, and cumulative volume of individual planned sonications, known as a treatment cluster, is compared with PTV to determine the percentage of target volume to be ablated (in this example, 28%). Regulatory stipulations for this trial permitted ablation of 30% of target fibroid volume. (b) Three-dimensional T2-weighted magnetic resonance (MR) imaging allows for multiplanar assessment and treatment planning, with PTV (red ellipse) and planned sonications (green ellipses) superpositioned on anatomic images. After populating PTV with intended sonication cells, the radiologist verifies on all planes to confirm that cells remain within the PTV outline.

Appendix 5

Bland-Altman analyses assessing pair-wise agreements between magnetic resonance (MR) imaging measurements via the contour method, 30 cumulative equivalent minutes at 43°C (CEM43) (a) and 240 CEM43 (b) thermal dose estimates, and between histopathology volumes and 30 CEM43 (c) and 240 CEM43 (d) thermal dose estimates. Bland-Altman analyses show the mean of the results of the two methods being compared on the x-axis, and their absolute difference on the y-axis. Good agreement between the two methods being compared was defined as (i) a mean difference close to zero (and lack of statistical significance on the t-test), (ii) ≥ 95% of the data falling within the threshold bars (representing 2 standard deviations of the mean difference, with tighter thresholds indicating more consistency in measurements), and (iii) almost no extreme outliers falling outside 3.5 standard deviations from the mean difference.