Noninvasive thermal ablation of uterine leiomyomas is possible with extracorporeal magnetic resonance (MR) imaging guided focused ultrasound surgery.

Transcription

1 Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at Fiona M. Fennessy, MD, PhD Clare M. Tempany, MD Nathan J. McDannold, PhD Minna J. So, MD Gina Hesley, MD Bobbie Gostout, MD Hyun S. Kim, MD George A. Holland, MD Dennis A. Sarti, MD Kullervo Hynynen, PhD Ferenc A. Jolesz, MD Elizabeth A. Stewart, MD 1 From the Departments of Radiology (F.M.F., C.M.T., N.J.M., M.J.S., K.H., F.A.J.) and Obstetrics and Gynecology (E.A.S.), Harvard Medical School/Brigham and Women s Hospital, 75 Francis St, Boston, MA 02115; Departments of Radiology (G.H.) and Obstetrics and Gynecology (B.G.), Mayo Clinic, Rochester, Minn; Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, Md (H.S.K.); Department of Radiology, Lahey Clinic, Burlington, Mass (G.A.H.); and Radnet Management, Los Angeles, Calif (D.A.S.). From the 2005 RSNA Annual Meeting. Received February 10, 2006; revision requested April 12; revision received July 17; accepted August 23; final version accepted October 4. Supported by InSightec, Haifa, Israel. Supported by National Institutes of Health grants 5R25CA (F.A.J., M.J.S.), 5P01CA (F.A.J.), and 1U41RR A2 (F.A.J.). F.M.F. supported by GE AUR Radiology Research Academic Fellowship. Address correspondence to F.M.F. ( ffennessy@partners.org). Uterine Leiomyomas: MR Imaging guided Focused Ultrasound Surgery Results of Different Treatment Protocols 1 Purpose: Materials and Methods: To prospectively assess patient response (after 12 months) to magnetic resonance (MR) imaging guided focused ultrasound surgery in treatment of uterine leiomyomas by using two treatment protocols. This prospective clinical trial was approved by institutional review boards and was HIPAA compliant. After giving informed consent, patients with symptomatic leiomyomas were consecutively enrolled and treated at one of five U.S. centers by using an original or a modified protocol. Outcomes were assessed with the symptom severity score (SSS) obtained at baseline and 3, 6, and 12 months after treatment. Adverse events (AEs) were recorded. Statistical analysis included Student t test, Fisher exact test, analysis of covariance, Spearman correlation, and logistic regression. Results: One hundred sixty patients had a mean SSS of (standard deviation) at baseline, which decreased to at 3 months (P.001) and to at 6 months (P.001) and was at 12 months (P.001). Ninety-six patients (mean age, 46.0 years 4.6) were treated with an original protocol, and 64 (mean age, 45.9 years 3.9) were treated with a modified protocol. Patients in the modified group had a significantly greater SSS decrease at 3 months (P.037) than those in the original group, and 73% of those in the original group and 91% of those in the modified group reported a significant decrease in SSS (of 10 points or greater) at 12 months. No serious AEs were recorded. Fewer AEs were reported in the modified group than in the original group (25% vs 13% reporting no event). Of evaluable patients, fewer in the modified group chose alternative treatment (28%) than in the original group (37%). Conclusion: MR imaging guided results in symptomatic improvement, sustained to 12 months after treatment. Treatment with a modified protocol results in greater clinical effectiveness and fewer AEs. RSNA, 2007 Clinical trial registration no. IDE G Protocol UF005 ORIGINAL RESEARCH VASCULAR AND INTERVENTIONAL RADIOLOGY RSNA, 2007 Radiology: Volume 243: Number 3 June

2 Noninvasive thermal ablation of uterine leiomyomas is possible with extracorporeal magnetic resonance (MR) imaging guided focused ultrasound surgery. The potential surgical application of was demonstrated more than 6 decades ago (1). However, the clinical use of focused ultrasound surgery as a therapy has been hampered by the difficulty in precisely targeting the ultrasound beam and obtaining feedback about thermal damage. MR imaging has excellent anatomic resolution and high sensitivity for tumor depiction and therefore offers accurate planning of the tissue to be targeted. MR imaging can quantify temperature changes by exploiting the temperature dependence of the water proton resonant frequency (2), which allows targeting of the beam during subthreshold ultrasound exposures (3) and online estimation of the ablated volume (4,5). Phase imaging is used to estimate the temperature-dependent proton resonant frequency shift with a fast spoiled gradient-recalled-echo sequence (6). This temperature dependence has been shown to be linear above the coagulation threshold (7 9). Therefore, temperature-sensitive MR images obtained before, during, and after a focused ultrasound sonication can be used to monitor tissue temperature elevations and prevent unwanted tissue damage with depiction of slight temperature elevations. By using interactive target segmentation at MR imaging, an outlined volume can be noninvasively treated and monitored with online MR temperature control. Advances in Knowledge We demonstrated a significant correlation (P.007 with original guidelines, P.039 with modified guidelines) between nonperfused volume at treatment and fibroid symptom score decrease. Less restrictive treatment guidelines resulted in a greater early symptom decrease, fewer adverse events reported, and a decrease in number of patients seeking alternative treatment. As a noninvasive technique for treatment of uterine leiomyomas, MR imaging guided has been shown to result in focal welldefined areas of protein denaturation and coagulative necrosis while sparing overlying and surrounding tissues (10,11). This technique provides precise targeting of the beam and has been demonstrated to be well tolerated, feasible, and safe (10,11). Results of initial studies have demonstrated substantial fibroid-related symptom reduction at 6 months, with more than 70% of women reporting improvement in their symptoms (12,13), which was sustained in 51% of women at 12 months (13). The purpose of our study was to prospectively assess patient response (after 12 months) to MR imaging guided focused ultrasound surgery in the treatment of uterine leiomyomas by using two different treatment protocols. Materials and Methods InSightec (Haifa, Israel) funded and provided the equipment for the clinical trial. Authors who were not consultants for InSightec had control over inclusion of any data and information submitted for publication. This was a prospective phase III multicenter clinical trial, which was approved by all local institutional review boards. All patients gave informed consent after the nature of the MR imaging guided procedure and our study was explained to them. All patients treated were part of a continued access study, which was an extension of a pivotal trial (to determine treatment effectiveness and to monitor for complications) of MR imaging guided in the treatment of uterine fibroids, once preliminary evidence from the pivotal study showed this technique to be likely effective and without important safety concerns. In this study, some patients were treated according to the original treatment guidelines used in the pivotal trial, and some were treated according to modified guidelines. This study was Health Insurance Portability and Accountability Act compliant. Patients Patients seeking treatment for symptomatic leiomyomas at one of five U.S. clinical centers (in Boston, Mass; Rochester, Minn; Baltimore, Md; Los Angeles, Calif; Burlington, Mass) were considered. The eligibility criteria for enrollment in this study have been previously described (12,13) and included adult premenopausal women with no future childbearing plans. Pregnant women, postmenopausal women, and those with MR imaging contraindications were excluded. The anterior abdominal wall of women who had previously undergone abdominal surgery was examined for scars, and if these were extensive or thought to be in the path of the ultrasound beam, these patients were excluded because of increased risk of skin burns (14). Clinical Selection Suitable candidates were asked to complete an eight-item section of a uterine fibroid symptom and quality of life (UFSQOL) questionnaire (15) to determine a symptom severity score (SSS). Components were scored on a fivepoint Likert scale, with responses ranging from not at all to a very great deal. An SSS score of 21 of 40 possible points, as a reflection of substantial fibroid-related symptoms at enrollment, was required for entry into the clinical Published online before print /radiol Radiology 2007; 243: Abbreviations: AE adverse event NPV nonperfused volume SSS symptom severity score UFSQOL uterine fibroid symptom and quality of life Author contributions: Guarantor of integrity of entire study, F.M.F.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, F.M.F., C.M.T., E.A.S.; clinical studies, all authors; statistical analysis, F.M.F., C.M.T., E.A.S.; and manuscript editing, F.M.F., C.M.T., N.J.M., G.H., H.S.K., K.H., F.A.J., E.A.S. Clinical trial registration no. IDE G Protocol UF005 See Materials and Methods for pertinent disclosures. 886 Radiology: Volume 243: Number 3 June 2007

3 trial. Age, body mass index, hormonal (pre- or postmenopausal) status, and the presence of chronic medical disease were recorded for each patient. These demographics were compared between patients treated according to original guidelines and those treated according to modified guidelines. Pretreatment Imaging After documentation of a negative pregnancy test result, all patients were evaluated with MR imaging performed by using a standard imaging protocol. Imaging was performed with a 1.5-T MR system (Signa; GE Healthcare, Milwaukee, Wis) with an LX computer platform with version 8.3 or higher software, by using a receive-only pelvic coil (USA Instruments, Aurora, Ohio). Initial three-plane localizer images with a large field of view were obtained first. Standard T2-weighted fast spinecho images, T1-weighted spin-echo images, and multiphase fat-suppressed T1- weighted spoiled gradient-recalled-echo images were obtained before and after intravenous injection of gadopentetate dimeglumine (0.1 mmol per kilogram of body weight) (Magnevist; Berlex Laboratories, Wayne, NJ). Parameters for the T1- and T2-weighted and spoiled gradient-recalled-echo MR images have been previously described (10). All MR images were obtained at 1.5 T by using the same protocol. MR images were analyzed to determine the number, size, and location of uterine leiomyomas, to evaluate for contrast enhancement within the leiomyomas, and to exclude other causes of reported symptoms such as adenomyosis. Patients without MR evidence of fibroids or with extensive abdominal wall scarring were not deemed eligible. Image analysis of the ultrasound beam path was performed in all patients to evaluate for possible obstructions to treatment, such as bowel loops anterior to the uterus at the level of the uterine fibroid. Pretreatment MR images were also used to select leiomyoma(s) to be treated. If there were multiple leiomyomas, clinical symptoms and accessibility of target leiomyomas were reviewed in detail and target leiomyoma(s) were selected. We attempted to select the leiomyoma that was most likely to explain the dominant symptom. Treatment Protocols Consecutive patients were treated in the continued access study from April 2003 to December Prior to April 30, 2004, they were treated according to original treatment guidelines, and after April 30, 2004, they were treated according to modified treatment guidelines (Table 1), subsequent to Food and Drug Administration satisfaction with results in those treated by using original guidelines and Food and Drug Administration approval of less restrictive treatment guidelines. The goal of treatment was to deliver therapeutic sonications to as large an area as possible according to treatment guidelines to induce a subsequent large area of nonperfusion. In patients treated according to modified guidelines, it was recorded if a second treatment was deemed necessary, and treatment was performed within guidelines. Equipment Sonications were performed by using an MR imaging compatible focused ultrasound system (ExAblate 2000; In- Sightec). This system is based on an earlier clinical design (7,16). It consists of a focused piezoelectric phased-array transducer (208 elements; frequency, MHz), a computer-controlled positioning system, and a multichannel radiofrequency amplifier system. All of these components are integrated with a standard 1.5-T clinical MR imaging unit (Signa; GE Healthcare). The array is located within a specially designed table in a water bath (5,10). Imaging is performed with a custom receive-only pelvic coil (USA Instruments). The MR imaging compatible focused ultrasound system automatically prescribes the temperature-sensitive imaging sequence performed with the MR imaging unit. Patient Preparation All patients were instructed to shave the hair of the anterior abdominal wall from the umbilicus to the superior aspect of the pubic bone and were asked not to use creams on the skin to allow maximal skin coupling with the ultrasound beam. Women arrived at the radiology department in a fasting state and with a companion to escort them home. On arrival, a urine-based pregnancy test was performed, and an intravenous line and Foley catheter were inserted. Informed Table 1 Comparison between Treatment Guidelines for MR Imaging guided Focused Ultrasound Surgery of Uterine Fibroids Parameter Original Treatment Guidelines Modified Treatment Guidelines Treatment volume of fibroid Limited to 33%, independent of fibroid location Limited to 50% of all fibroids, except submucosal (33%) Treatment volume of fibroid 100 cm 3 per fibroid or 150 cm 3 if more than 150 cm 3 if one or more fibroid treated one fibroid treated Maximum treatment time (min) Distance from edge of sonication to 1.5 No restriction endometrium (cm) Distance from edge of sonication to serosal surface (cm) Distance to fibroid capsule 0.5 cm on serosal surface No restriction Second treatment Not allowed Yes, if within 14-day period Radiology: Volume 243: Number 3 June

4 consent was obtained for use of intravenous conscious sedation (fentanyl citrate and midazolam hydrochloride), per hospital guidelines. Patient preparation and monitoring during treatment was performed (10 13). A registered nurse was in the room with the patient at all times. The patient could press an emergency stop button during the treatment, which allowed her to terminate any single sonication should she be required to do so. Radiologists (F.M.F., C.M.T., M.J.S., G.H., G.A.H., D.A.S.) supervised the procedures and anesthesia administration. The patient was in the magnet in the prone position, and the anterior abdominal wall was placed over the transducercontaining water tank. The pelvic coil was mounted on the water tank (5,10). The relationship between the fibroid and the transducer was checked by using threeplane localizer images with a large field of view, and the patient was repositioned if necessary. The pelvis was imaged by using T2-weighted fast spin-echo images in transverse, coronal, and sagittal planes, which were transferred to the system workstation. Treatment The volume to be ablated was delineated by one of seven radiologists (F.M.F., C.M.T., M.J.S., G.H., H.S.K., G.A.H., D.A.S.) on a selected coronal T2-weighted image by using the system software. The years of experience of these radiologists with MR procedures ranged from 3 to 15 years. Equally spaced individual sonications were then placed such that the entire target volume was covered. The system software allowed display of all the individual sonications superimposed on all tissues through which it would pass, in three orthogonal planes. The beam path was reviewed, and care was taken to ensure safe passage to avoid any contact with bowel loops and remain a safe distance from the sacrum. If necessary, the transducer beam could be tilted up to 20 to provide a suitable pathway. The sonication spot size could be altered in power, frequency, diameter, and length to ablate the outlined volume in as few sonications as possible. Locations of the planned sonications could be changed if any patient motion occurred during the procedure. Once planning was complete, lowenergy sonications were delivered to confirm correct anatomic location (5,10). The power was then increased until the temperature-sensitive images displayed a therapeutic focus. Sonications were typically all delivered in the same coronal plane. Sonication duration was typically seconds, followed by seconds of cooling time. The patient gave constant feedback regarding any symptoms after each sonication. This, in addition to image-based feedback, was routinely used to modify the sonication parameters so that each sonication achieved a sufficient thermal dose without causing discomfort. Thermometric results provided immediate measurement of thermal dose, and repeated sonications could be performed as necessary. Thermal Map Acquisition Phase-subtraction fast spoiled gradientrecalled-echo imaging was used to monitor temperature elevations and construct temperature maps (2) during treatment sonications. Each sonication was monitored with MR imaging, which was timed to coincide with the deposition of thermal energy (5,10). The first image was triggered 5 seconds before the sonication to create a baseline image. During the heating and early cooling phase, multiple images were created every few seconds in a single imaging plane. The MR imaging unit was programmed to construct the magnitude, real, and imaginary images for each imaging time point, which were used to calculate phase subtraction between two imaging time points (6). If MR thermometric results indicated that the thermal dose delivered was insufficient or that temperatures were too low, the power was increased. The sonication frequency was altered as deemed necessary, according to patient anatomy and/or reported pain during sonications. Posttreatment Follow-up Immediately after completion of treatment, MR imaging with administration of intravenous gadopentetate dimeglumine was performed to assess the therapeutic effect, or nonperfused volume (NPV). Patients were evaluated and released from the medical center and returned within 1 week for clinical evaluation. At 3 months after the procedure, a phone interview was conducted with patients. At 6 and 12 months after the procedure, patients returned for clinical evaluation (E.A.S.) and underwent repeat MR imaging examination. At all three time points (3, 6, and 12 months), the effect of treatment on leiomyomarelated symptoms was assessed by completion of a self-assessment UFSQOL questionnaire. The 3-month questionnaire was completed at home, and the patient called in her answers to the clinic by telephone. The 6- and 12- month questionnaires were completed at the clinic. If data for a time point were acquired outside a protocol-specified allowance, the patient was considered nonevaluable and their data for that specific time point were not used for this study. Patients did, however, remain in the study for data collection at future time points. There was a 2-week window allowance for the 3-month follow-up, a 3-week window allowance for the 6-month follow-up, and a 6-week window allowance for the 12-month follow-up. The number of patients in the study population who were followed up to all time points was recorded. In addition, the numbers of patients whose data were obtained but were obtained outside the window allowance for follow-up, who were overdue or not yet due for follow-up, who were lost to follow-up, and who had chosen alternative treatment for their uterine fibroids at each specific time point were recorded. The primary end point of the study was improvement in SSS, according to the UFSQOL questionnaire, from baseline. SSSs were obtained and transformed to create a summed comparable score (15). The SSS assesses menorrhagia and bulk-related symptoms on a single 100-point scale, with higher scores indicating worse symptoms. In the initial validation UFSQOL questionnaire, patients with uterine fibroids had a mean score of 44.0, while a reference control population had a mean score of 22.5 (15). The SSS will, however, vary 888 Radiology: Volume 243: Number 3 June 2007

5 depending on the severity of fibroids in the population in question. SSS at baseline and at 3, 6, and 12 months after treatment was obtained. The rate of improvement in SSS from baseline to 3, 6, and 12 months after treatment was calculated and compared between patients treated according to original and those treated according to modified guidelines by using analysis of covariance. In addition, the percentage of patients with a 10-point improvement in the transformed SSS was calculated. A 10-point improvement is consistent with a clinical improvement and was the primary end point of the pivotal study of the evaluation of MR imaging guided focused ultrasound surgery on uterine fibroids (12,13). The odds of a 10-point improvement in SSS in those with a greater NPV ( 30% vs 30%) were calculated. At each time point, all adverse events (AEs) were documented. Serious AEs were defined according to Standard Code of Federal Regulation definitions for use of an investigational product (17). AEs were reported to the institutional review board and the Food and Drug Administration as required. AEs were classified as not important or important (lasting more than 14 days). Important AEs were classified according to severity (mild, moderate, or severe). MR Data Analyses MR images from all five medical centers were sent to a core laboratory, where baseline total fibroid volume load (ie, total volume of all fibroids per patient at Table 2 baseline) and NPV were measured. Fibroid volume load and NPV were outlined by an independent MR radiologist (with 20 years of experience in genitourinary MR imaging) in a core laboratory. The fibroid volume load was calculated on standard T2-weighted fast spinecho images. NPV immediately after treatment was calculated on coronal spoiled gradient-recalled-echo images after administration of gadopentetate dimeglumine and was defined as the area of nonenhancement within the uterine fibroid. Commercially available software (ProVision; Algotec, Rochester, NY) was used to calculate volume. The outlines of the leiomyomas and nonenhancing treated areas were contoured with electronic calipers to compute the area per section. The volume was computed by multiplying the sum of the measured areas by the distance between the centers of two consecutive images. NPV as a percentage of fibroid load at treatment was also calculated. Statistical Analysis Baseline age, body mass index, and total fibroid volume load were compared by using a two-sample t test assuming unequal variance. Race (white vs nonwhite), hormonal status, and history of chronic medical disease were compared by using the Fisher exact test. The number of patients who did not report any AEs in each treatment group was also compared by using the Fisher exact test. The transformed SSS was expressed as mean standard deviation. Student one-sample t tests were used for statistical comparisons of mean SSS at baseline and that reported at 3, 6, and 12 months for both protocol groups. Analysis of covariance was used to compare the mean change in SSS at 3, 6, and 12 months between patients treated according to the original protocol and those treated according to the modified protocol. Linear regression analysis and Spearman correlation were used to evaluate for a correlation between SSS change and treatment NPV. Software (SAS; SAS Institute, Cary, NC) was used for all data analysis. A P value of less than.05 was considered to indicate a significant difference. Results Patients One hundred sixty patients (mean age, 46.0 years 4.3; range, years) were enrolled and treated in this prospective study. One hundred forty-four (90.0%) of 160 treated women were evaluable at 6 months, and 76 (47.5%) were evaluable at 12 months (Table 2). Some patients sought alternative treatment; others were lost to or were overdue for follow-up at 6 and 12 months (Table 3). Ninety-six patients (mean age, 46.0 years 4.6; range, years) were consecutively treated according to the original treatment protocol, and 64 patients (mean age, 45.9 years 3.9; range, years) were treated according to the modified treatment protocol. Five and nine patients treated according to original guidelines were lost to follow-up at 6 and 12 months, respectively. Eight and 23 patients treated according to original guidelines sought alternative Number of Patients in Study at Different Time Points Time Point Total Study Population No. of Patients No. of Treated Patients with Original Guidelines Baseline months months months Treated with Modified Guidelines Table 3 Disposition of Patients at Different Time Points Reason 6 Months 12 Months Sought alternative treatment 8 30 Lost to follow-up 6 10 Overdue or not due for follow-up 2 27 Outside window allowance for follow-up 0 17 Completed follow-up and data available Note. Data are numbers of patients. Numbers reported at 12 months include those accounted for at 6 months. Data obtained outside the window allowance for a specific time point were omitted for the purposes of this study, but the patient continued on to the next data time point. Radiology: Volume 243: Number 3 June

6 treatment at 6 and 12 months, respectively. When expressed as a percentage of those patients whose outcome was known (and assuming those lost to follow-up also had a negative outcome), 12.9% (13 of 101) of patients at 6 months and 37% (32 of 87) at 12 months had sought alternative treatment after treatment according to the original protocol. Of those treated according to the modified protocol, one patient was lost to follow-up at both 6 and 12 months, and zero and seven patients had sought alternative treatment at 6 and 12 months, respectively. Therefore, assuming treatment failure in those lost to follow-up, 2% (one of 57) of patients had sought alternative treatment at 6 months, and 28% (eight of 29) of patients had sought alternative treatment at 12 months. Twenty-four patients treated according to modified guidelines underwent a second MR imaging guided treatment within a 14-day period. There was no significant difference in age (P.84), body mass index (P.12), race (P.59), hormonal status (P.999), or underlying chronic medical disease status (P.8) between the two treatment groups (Table 4). There was no significant difference (P.1)intotal fibroid volume load between those treated according to original guidelines (385.0 cm [mean standard deviation]; range, cm 3 ) and those treated according to the modified protocol (498.4 cm ; range, cm 3 ). Clinical Outcomes In the overall study population (Table 5), 79.2% of evaluable patients had a 10- point or greater symptom improvement at 3 months after treatment, which was sustained in 79.2% of patients at 6 months and in 78% of patients at 12 months. For those treated by using original treatment guidelines, results of clinical follow-up showed that 74% of patients at 6 months and 73% at 12 months had a greater than 10-point improvement in SSS. We found that, in patients treated according to the modified protocol, 88% reported a 10-point or greater symptom improvement at 6 months and 91% had significant symptom improvement at 12 months. In the overall study population (Fig 1), the mean UFSQOL SSS at baseline was This decreased to at 3 months and to at 6 months and was at 12 months. This decrease in score was significant at all time points up to 12 months after treatment (P.001). Both the original and modified treatment guidelines independently resulted in a significant decrease in SSS, which was sustained at 12 months. The rate of improvement was maximal in the first 3 months after treatment (23.8-point decrease in patients treated according to original protocol, 30.9-point decrease in those treated according to modified treatment guidelines) (Fig 2). This rate was accelerated in the modified group, because analysis of covariance comparison between the two groups of patients (with the change in SSS from baseline as the dependent variable) showed a significantly greater symptom score decrease (P.037). By 6 and 12 months after treatment, this rate of response was no longer significantly different between the two groups. Treatment Volumes The NPV was calculated in both groups and was found to be 59.4 cm (range, cm 3 ) in the original protocol group and cm (range, cm 3 ) in the modified protocol group. When the NPV was calculated as a percentage of the total fibroid volume load, it was 16.65% 16.1 (n 88) in the original protocol group and significantly increased to 25.79% 21.8 (n 44) in the modified protocol group (P.001, two-tailed t test). When correlated with the symptom score change Table 4 Patient Demographics at MR Imaging guided Focused Ultrasound Surgery of Uterine Fibroids Baseline Demographic All Patients (n 160) Patients Treated with Original Guidelines (n 96) Patients Treated with Modified Guidelines (n 64) P Value* Age (y) (35, 58) (35, 58) (38, 53).84 Body mass index (kg/m 2 ) (18.4, 39.5) (18.4, 38.4) (19.1, 39.5).12 Race (%).59 Asian 1.3 (2/160) 1 (1/96) 2 (1/64) African American 5.6 (9/160) 7 (7/96) 3 (2/64) White 90 (144/160) 89 (85/96) 92 (59/64) Other 3.1 (5/160) 3 (3/96) 3 (2/64) Hormonal status 95.6 (153/160) 96 (92/96) 95 (61/64).999 Percentage of those with history of chronic medical disease 29.3 (47/160) 27 (26/96) 33 (21/64).80 * Demographics in those patients treated according to original guidelines were compared with those treated according to modified guidelines by using a two-sample t test assuming unequal variances (age, body mass index) and Fisher exact test (race, hormonal status, and history of chronic medical disease). P value for race is for white versus nonwhite. Data are means standard deviations, with minimum and maximum values in parentheses. Data in parentheses are numbers used to calculate percentages. Data are percentages of premenopausal women, with numbers used to calculate percentages in parentheses. Numbers used to calculate percentages are in parentheses. 890 Radiology: Volume 243: Number 3 June 2007

7 from baseline to 6 months, this correlation was (P.007, Spearman correlation) for the original treatment guidelines and a moderate (P.039, Spearman correlation) for modified treatment guidelines. Linear regression analysis of the correlation slope showed a decrease from 0.28 to 0.36 after expansion of treatment guidelines, consistent with increasing predictive power of the nonperfused treatment volume. However, with multiple regression analyses, this difference was not significant (P.7). The odds of a 10-point improvement in SSS at 12 months after treatment in patients treated according to the modified protocol was 2.8 in those with an NPV of 30% or greater compared with those with a NPV of less than 30% (P.038). Safety: AEs No serious AEs were reported. A total of 290 AEs was recorded (mean, 1.8 AEs per patient) in the entire study population. In the original treatment group, 13% of patients reported no AE, which increased to 25% in the modified treatment group. This difference was not significant (P.06). In the original treatment group, two important AEs (ie, ones that did not resolve without sequelae within 14 days) were reported. One patient reported paresthesia at the site of the intravenous cannula that resolved within 6 weeks. A second patient reported mild sonication-related leg pain, which resolved within 2 days, but was nonetheless considered important for Food and Drug Administration reporting. No important events were reported in the modified treatment group. All other AEs were not important, the most common being pain or discomfort related either to position within the magnet or uterine discomfort due to sonications; these AEs occurred in 54% and 47% of patients treated according to the original and modified protocols, respectively (Table 6). Table 5 Ten-Point Improvement in SSS after MR Imaging guided Focused Ultrasound Surgery of Uterine Fibroids Study Population and Time Point* Figure 1 Time-Point Change Ten-Point Improvement (%) All patients 3 months (n 149) 3 months from baseline 79.2 (118/149) 6 months (n 144) 6 months from baseline 79.2 (114/144) 12 months (n 76) 12 months from baseline 78 (59/76) Those treated according to original guidelines 3 months (n 91) 3 months from baseline 76 (69/91) 6 months (n 88) 6 months from baseline 74 (65/88) 12 months (n 55) 12 months from baseline 72 (40/55) Those treated according to modified guidelines 3 months (n 58) 3 months from baseline 85 (49/58) 6 months (n 56) 6 months from baseline 88 (49/56) 12 months (n 21) 12 months from baseline 91 (19/21) * n Is number of patients treated according to guidelines at time point indicated. Data in parentheses are numbers used to calculate percentages of patients showing 10-point improvement. Figure 1: Bar graph demonstrates UFSQOL transformed SSSs at baseline, 3, 6, and 12 months after MR imaging guided in patients in continued access study (CAS) and in those subdivided into treatment according to original and modified guidelines. Mean scores at 3, 6, and 12 months were compared with baseline score with Student t test. There was a significant decrease in mean score at all time points. Discussion To our knowledge, our study results are the first to demonstrate that improvement in fibroid-related symptoms can be sustained 12 months after MR imaging guided. Improvements in technique have lead to greater clinical response and safety and fewer AEs. Our study results demonstrate that there was a significant correlation between the percentage of NPV at treatment and the UFSQOL SSS change Radiology: Volume 243: Number 3 June

8 Figure 2 from baseline to 6 months after therapy. Although this correlation was moderate and the time period of follow-up was relatively short, it is likely that there is an association between treatment effectiveness and devascularization, as has been suggested with uterine artery embolization (18). Furthermore, expanded treatment with a modified MR imaging guided focused ultrasound surgery protocol resulted in a significantly greater early reduction in symptom score at 3 months after treatment. In the total population evaluable in this continued access study, 79.2% and 78% of patients treated reported a 10- point or greater improvement in SSS at 6 and 12 months after treatment, respectively. This result demonstrates a sustained response in a far greater number of patients in comparison to that in the pivotal study (13), in which only Figure 2: Line graph demonstrates UFSQOL SSS decrease from baseline to 3, 6, and 12 months after treatment according to original and modified guidelines. At 3, 6, and 12 months, respectively, n was 91, 88, and 55 in original group and 58, 56, and 21 in modified group. Score decrease was compared between the two treatment profiles at each time point by using analysis of covariance. Mean score decrease was obtained by subtracting least-squares mean change from baseline at 3, 6, and 12 months. Change in least-squares mean was significantly greater in modified treatment group at 3 (P.037) months after treatment. Table 6 Common Not Important AEs AE Patients Treated with Original Guidelines (n 96) Pain or discomfort 52 (54) 30 (47) Abdominal cramping, pain, or tenderness 18 (19) 11 (17) Sonication-related pain 23 (24) 16 (25) Positional back pain 11 (12) 6 (9) Gynecologic 29 (30) 9 (14) Vaginal discharge 9 (9) 2 (3) Heavy menses 8 (8) 3 (5) Gastrointestinal 25 (26) 12 (19) Nausea or vomiting 21 (22) 5 (8) Patients Treated with Modified Guidelines (n 64) Note. Data are number of patients, with percentages in parentheses. One patient may have experienced more than one AE. 51% of patients were reported to have a significant clinical response at 12 months. The overall significant decrease in mean SSS from 62.1 at baseline to 32.7 at 12 months after treatment also confirms the durability of response. We do, however, recognize that longer-term follow-up and evaluation are indicated. These processes are currently underway, and data collection at 24 and 36 months is in progress. Optimization of treatment protocol by expansion of treatment guidelines resulted in a far greater number of evaluable patients (91%) reporting a significant clinical response at 12 months after treatment. Symptom reduction occurred early, with a significant decrease seen at 3 months after treatment. A significantly greater reduction in symptom score at 3 months was seen in patients treated according to the modified treatment protocol than in those treated according to the original protocol. While this difference in symptom reduction was not significantly greater at 6 and 12 months, this may reflect the smaller number of patients in each group at the later time points, and continued follow-up is warranted. The apparent trend is for those treated according to the modified protocol to maintain a greater symptom reduction. Assuming treatment failure in those lost to follow-up, there was a smaller percentage of patients seeking alternative treatment at 6 and 12 months after MR imaging guided focused ultrasound surgery performed according to the modified protocol than after MR imaging guided performed according to the original protocol. Although the number of women followed up to 12 months is small (particularly in the modified protocol group) and longer-term follow-up is essential, these findings point toward a trend for better clinical outcome and increased patient satisfaction when larger volumes of tissue are ablated. This conclusion is further supported by the relationship between SSS change at 6 months and NPV after treatment. The correlation is stronger in patients with larger volumes treated according to the modified protocol. The implication of this finding is that greater clinical 892 Radiology: Volume 243: Number 3 June 2007

9 symptom reduction is likely achieved with maximal NPV. Therefore, we believe that it is an important goal to treat as much tissue as feasible and as safely as possible in each woman. We recognize that the decrease in SSS 12 months after uterine artery embolization is greater (from to 19.23) (19) than that after MR imaging guided. However, treatment guideline limitations with MR imaging guided focused ultrasound surgery (and subsequent relatively small NPV as a percentage of total fibroid volume load) should be taken into account when the difference in the decrease in SSS with MR imaging guided is being weighed against that with uterine artery embolization. Future technical improvements in MR imaging guided focused ultrasound surgery, which would allow for a greater NPV, may provide greater SSS decrease. While the modified guidelines may explain our results, there may be factors such as experience, familiarity with the device, and better patient preparation and education that play a role. The number of AEs reported in the entire population was 290 (mean, 1.8 AE per patient). This number is decreased when compared with that in the pivotal trial, in which 271 AEs were recorded in 109 treated patients (mean, 2.5 AEs per patient) (20). Improved patient selection, enhanced explanation of the procedure to the patient, continuous communication with the patient and the treating physician during the procedure, and optimization of fibroid targeting and thermometry all likely contributed to the reduction of AEs. An important limitation of this study was that this was a single treatmentarm trial. Patients were not randomly assigned into treatment with MR imaging guided or accepted alternatives, such as hysterectomy or uterine artery embolization; therefore, direct comparison with these accepted forms of therapy cannot be done. We also do not have a no-treatment control arm to confirm lack of symptom change in those not receiving treatment. Although we are continuing to follow up patients to 12 months and beyond to 24 and 36 months, another important limitation was the small number of patients currently evaluable at 12 months and the reporting of outcomes in symptom relief on the basis of data in the evaluable population. Longer-term follow-up and use of an intent-to-treat population to report sustainability of response to treatment and the number of patients seeking alternative treatments is required. In summary, in this study, we demonstrate sustained symptom improvement at 12 months, confirm the reproducibility of MR imaging guided focused ultrasound surgery for uterine leiomyomas, and demonstrate that treatment according to a modified protocol results in a greater clinical response and a decrease in the number of patients seeking alternative treatment and the number of AEs reported. Acknowledgments: The authors thank Louise Greenberg, Patty Devine, RT, Kevin Lanctot, RT, Sharon Price, RN, and Xiangtao Yin, PhD. References 1. Lynn JG, Zwemer RL, Chick AJ, Miller AE. A new method for the generation and use of focused ultrasound in experimental biology. J Gen Physiol 1942;26: Ishihara Y, Calderon A, Watanabe H, et al. A precise and fast temperature mapping using water proton chemical shift. Magn Reson Med 1995;34(6): Hynynen K, Vykhodtseva NI, Chung AH, Sorentino V, Colucci V, Jolesz FA. Thermal effects of focused ultrasound on the brain: determination with MR imaging. Radiology 1997; 204(1): Chung AH, Jolesz FA, Hynynen K. Thermal dosimetry of a focused ultrasound beam in vivo by magnetic resonance imaging. Med Phys 1999; 26(9): McDannold N, Tempany CM, Fennessy FM, et al. Uterine leiomyomas: MR imaging based thermometry and thermal dosimetry during focused ultrasound thermal ablation. Radiology 2006;240(1): Chung AH, Hynynen K, Colucci V, Oshio K, Cline HE, Jolesz FA. Optimization of spoiled gradient-echo phase imaging for in vivo localization of focused ultrasound beam. Magn Reson Med 1996;36(5): Hynynen K, Freund WR, Cline HE, et al. A clinical, noninvasive, MR imaging monitored ultrasound surgery method. RadioGraphics 1996; 16(1): Kuroda K, Chung AH, Hynynen K, Jolesz FA. Calibration of water proton chemical shift with temperature for noninvasive temperature imaging during. J Magn Reson Imaging 1998;8(1): Peters RD, Hinks RS, Henkelman RM. Ex vivo tissue-type independence in proton-resonance frequency shift MR thermometry. Magn Reson Med 1998;40(3): Tempany CM, Stewart EA, McDannold N, Quade BJ, Jolesz FA, Hynynen K. MR imaging guided of uterine leiomyomas: a feasibility study. Radiology 2003; 226(3): Stewart EA, Gedroyc WM, Tempany CM, et al. Focused ultrasound treatment of uterine fibroid tumors: safety and feasibility of a noninvasive thermoablative technique. Am J Obstet Gynecol 2003;189(1): Hindley J, Gedroyc WM, Regan L, et al. MRI guidance of focused ultrasound therapy of uterine leiomyomas: early results. AJR Am J Roentgenol 2004;183(6): Stewart EA, Rabinovici J, Tempany CM, et al. Clinical outcomes of for the treatment of uterine fibroids. Fertil Steril 2006;85(1): Leon-Villapalos J, Kaniorou-Larai M, Dziewulski P. Full thickness abdominal burn following magnetic resonance guided focused ultrasound therapy. Burns 2005;31(8): Spies JB, Coyne K, Guaou Guaou N, Boyle D, Skyrnarz-Murphy K, Gonzalves SM. The UFS- QOL, a new disease-specific symptom and health-related quality of life questionnaire for leiomyomata. Obstet Gynecol 2002;99(2): Cline HE, Hynynen K, Watkins RD, et al. Focused US system for MR imaging guided tumor ablation. Radiology 1995;194(3): Standard Code of Federal Regulation definitions (21 CFR IND safety reports). U.S. Food and Drug Administration Web site Pelage JP, Guaou NG, Jha RC, Ascher SM, Spies JB. Uterine fibroid tumors: long-term MR imaging outcome after embolization. Radiology 2004; 230(3): Spies JB, Myers ER, Worthington-Kirsch R, et al. The FIBROID Registry: symptom and qualityof-life status 1 year after therapy. Obstet Gynecol 2005;106(6): Exablate device labeling. U.S. Food and Drug Administration Web site. /cdrh/pdf4/p html. Accessed October 22, Radiology: Volume 243: Number 3 June