There were 12 leiomyomata treated in 11 patients. In the 11th patient treated, two leiomyomata were treated during one MR-guided HIFU treatment session. Average treatment session duration, defined as from the time the patient was positioned on the HIFU table until transport to the recovery room, was 3.67 hours ± 0.7 (range, 2.67–5.55 h). Preplanning time, defined as the time from when the patient was placed on the table to the first survey scan, was 13.2 minutes ± 5.4 (range, 8 –28 min). The planning time, defined as the time from the first survey scan to delivery of the first sonication, was 104.4 minutes ± 49.2 (range, 48 –241 min). The sonication time, defined as the time from the first sonication to the last sonication, was 54.6 minutes ± 18.0 (range, 23– 86 min). The posttreatment duration, defined as the time from performance of the last sonication to transferring the patient off the HIFU table, was 34.2 minutes ± 12.0 (range, 12– 60 min). The duration of time for transfer of the patient to the recovery area, subsequent to removal off the HIFU tabletop, was 13.2 minutes ± 9.6 (range, 1–16 min).
Of 11 patients, 9 underwent hysterectomy, with a mean time of 9 days between MR-guided HIFU and hysterectomy (range, 3–20 d); 2 patients decided against having hysterectomy citing that the postoperative recovery would interfere with their work schedule and withdrew from the study before surgery. Nontarget thermal injury was not observed on contrast-enhanced MR images obtained after treatment, during surgical inspection at hysterectomy, or at gross pathology assessment. At gross pathology assessment, all leiomyomata showed confluent treatment in the expected location. In eight of nine cases, confluent treatment in the expected location was confirmed at histopathologic assessment with hematoxylin and eosin staining. In one of these eight cases, one patient had an additional 0.8-cm focus of hemorrhagic necrosis within uterine myometrium adjacent to a 2-cm submucosal leiomyoma that moved during treatment. This movement was observed on MR imaging during treatment; concurrently, the patient experienced pain and pressed the emergency button, interrupting treatment after 24.4 seconds of sonication. A ninth patient underwent MR-guided HIFU of a small area within an 11-cm leiomyoma containing extensive myxomatous degeneration. Although identified by needle localization on gross pathology, the area treated by HIFU at histopathology was difficult to identify, being in close proximity to the area of preexisting necrosis. delineates ablation volumes determined by MR imaging assessment using the contour method, 30 CEM43 and 240 CEM43 thermal dose estimates, and histopathologic findings for each patient. The histopathologically confirmed mean ablation volume was 6.92 cm3 ± 10.7 with a mean targeted leiomyoma volume of 246.7 cm3 ± 375.8. These intentionally small volumes of treatment were in keeping with the study goal, which was to treat enough tissue to allow comparisons between MR imaging findings and pathology.
Comparison of Ablation Volumes Determined by Magnetic Resonance (MR) Imaging Assessment, Thermal Dose Estimates, and Specific Histopathologic Findings Observed after MR Imaging–guided High-Intensity–focused Ultrasound Therapy
Complications during and after HIFU were described using the Society of Interventional Radiology Classification System for Complications by outcome (25
). During and after HIFU, no major complications were reported. Of the minor complications reported, transient abdominal cramping (n = 9) or leg pain (n = 2) were the most frequently reported. These expected minor adverse events were self-limited, resolving by 72 hours in all cases. Specifically, no skin burns or nerve injuries were observed. One patient with chronic constipation had blood in the stool in the recovery area after HIFU treatment. This adverse event was reported to be a relatively common occurrence for this patient, and was not attributed to HIFU treatment. Another patient who developed transient suprapubic pain and urinary frequency 1 week after HIFU treatment had negative urinalysis and culture; physical assessment did not reveal any findings. The patient was prescribed ibuprofen to relieve the pain symptoms. The beam path was not near the symptomatic area, and these symptoms were attributed to Foley catheterization.
MR-guided HIFU Spatial Accuracy
MR-guided HIFU spatial targeting was assessed by comparing intended and actual treated locations and calculating misregistration values. Test sonications were discarded from misregistration analyses because they were used to calibrate spatial targeting. Mean misregistration perpendicular to the ultrasound beam axis was 0.8 mm ± 1.2 (range, −2.6 to +3.7 mm) from the planned location in the feet-head direction and 0.1 mm ± 1.0 (range, −4.9 to +1.7 mm) from the planned location in the left-right direction. Misregistration along the beam axis was −0.7 mm ± 3.1 from the planned location (range, −8.5 to +6.2 mm).
Comparison of MR-guided HIFU Treatment Volumes by MR Imaging Assessment, Thermal Dose Estimates, and Histopathology
Mean contour and ellipsoid measured volumes were equivalent and in agreement by Bland-Altman analysis (P > .05) enabling subsequent comparisons to be performed using only contour volumes. Mean contour MR imaging ablation volumes were not different from histopathology ablation volumes (P = .5348) and were in agreement with histopathology ablation volumes by Bland-Altman analysis (P = .5345).
Paired comparisons between contour MR imaging ablation volumes and 30 CEM43 thermal dose estimates indicated no significant difference between the mean values of these data (P
= .5510). Contour MR imaging ablation volumes and 30 CEM43 values showed agreement by Bland-Altman analysis (P
= .5500), showing small magnitudes of differences within tight thresholds (; available online at www.jvir.org
). However, contour MR imaging volumes were significantly different from the smaller 240 CEM43 thermal dose estimates (P
= .03) and did not show agreement with the smaller 240 CEM43 thermal dose estimates (P
= .0328) (; available online at www.jvir.org
), suggesting that 30 CEM43 may provide a more accurate description of the treatment volume. Comparisons between histopathology ablation volumes and the 30 CEM43 and 240 CEM43 thermal dose estimates indicated no difference between the mean values of these data (P
= .93 and P
= .33). Agreement between histopathology ablation volumes and the 30 CEM43 and 240 CEM43 thermal dose estimates was observed via Bland-Altman analysis (P
> .05) and could be visually noted by the tight thresholds and small mean differences with minimal data falling outside the thresholds (; available online at www.jvir.org
). Considered together, the MR imaging, thermal dose, and histopathology measurements showed agreement (intraclass correlation coefficient = 0.6996).
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 (more ...)