Occurrence of adverse events after magnetic resonance-guided high-intensity focused ultrasound (MR-HIFU) therapy in symptomatic uterine fibroids—a retrospective case-control study

Abstract Objectives Our study aims at the comprehensive analysis of adverse events (AEs) in patients with symptomatic uterine fibroids (UFs) who underwent magnetic resonance-guided high-intensity focused ultrasound (MR-HIFU) in the last 6 years in one of the major Polish centers performing this type of therapy. Methods The presented retrospective case-control study was conducted in the Department of Obstetrics and Gynecology, Pro-Familia Hospital, Rzeszów in cooperation with the Second Department of Obstetrics and Gynecology, Center of Postgraduate Medical Education, Warsaw. The study enrolled 372 women with symptomatic UFs who underwent MR-HIFU and reported AEs during or after the procedure. The occurrence of particular AEs was analyzed. Statistical comparison of two cohorts (patients with and without AEs) was conducted based on epidemiological factors, UF characteristics, fat layer thickness, the presence of abdominal scars and technical parameters of the procedure. Results The overall mean occurrence rate of AEs was 8.9% (n = 33). No major AEs were reported. The only statistically significant risk factor of AEs was the treatment of type II UFs according to Funaki (OR 2.12, CI 95%, p = 0.043). Other investigated factors did not have a statistically significant influence on AE occurrence. Abdominal pain was the most common AE. Conclusion Our data showed that MR-HIFU seemed to be a safe procedure. The AE rate after the treatment is relatively low. According to the obtained data it seems that the occurrence of AEs does not depend on the technical parameters of the procedure and the volume, position and location of UFs. Further prospective, randomized studies and with long follow-up are necessary to confirm the final conclusions.


Introduction
Uterine fibroids (UFs), also known as leiomyomas, are benign monoclonal tumors of the genital tract [1][2][3].According to available studies, their pathogenesis is complex and not fully elucidated.Genetics and sex steroids play an essential role in UF development [3].
UFs are the most common uterine neoplasms diagnosed in up to 70% of women before the onset of menopause [1].Despite a high incidence in the general population, only 25-50% of women report clinically relevant symptoms.These include heavy, prolonged menstrual bleeding leading to iron-deficiency anemia.Bulk symptoms, pelvic discomfort and urine incontinence are also commonly reported [3].UFs also have a great impact on pregnancy complications including fertility disorders, miscarriage or preterm delivery [4].
According to a systematic review by Stewart et al., the main risk factors of UF occurrence are black race, a family history of UFs, hypertension, perimenopausal status, time since last birth, soy milk and food additive consumption [1].Obesity was also suggested to have an association with UF occurrence [5,6].
There are a number of modalities in UF therapy.As only patients with symptomatic UFs will require treatment, it should be adjusted to clinical manifestations and the patient's expectations as much as possible [7].Currently, treatment methods can be divided into conservative ones including pharmacological therapy (mostly concentrated on symptom relief) such as non-steroidal anti-inflammatory drugs, oral contraceptives, progestins and GnRH agonists/antagonists.All these modalities either have a limited success rate or are associated with adverse drug reactions which are not acceptable for some groups of patients [4,8].Conversely, surgical methods, such as myomectomy or hysterectomy, resolve the cause of the symptoms related to UFs.These operative methods might be related to a number of shortand long-term adverse events (AEs) and long recovery time [4,7,9,10].According to a systematic review published by Soliman et al. the total costs (direct and indirect) connected with the diagnosis or surgery of UFs ranged from 11,717 to 25,023 US dollars per patient annually [11].
Recently, invasive radiology has given an opportunity to patients to use new, less invasive methods.Uterine artery embolization (UAE) and magnetic resonance-or ultrasoundguided focused ultrasound (MRgFUS/USgFUS) may be considered as modalities with an acceptable effect on symptom relief and a good safety profile compared to classic surgical methods, especially hysterectomy [7,12].MR-HIFU/MRgFUS, being the only completely noninvasive technique conducted under highly precise control of MRI, is a particularly beneficial modality for patients who would like to avoid surgery and general anesthesia.Although there are no recommendations for women who would like to conceive in the future, the results of current studies are encouraging in this area, even in comparison with other minimally-invasive methods, such as UAE [13].The results of a recent meta-analysis by Akhatova et al. comparing obstetric outcomes after minimally invasive treatment options in UF therapy revealed a similar live birth rate after all procedures with a significantly lower miscarriage rate after focused ultrasound therapy [14].According to currently available data, UAE and myomectomy seem to generate similar costs.However, MR-HIFU is expected to be even cheaper due to faster recovery.Research is still ongoing to confirm the above [15,16].
MR-HIFU/MRgFUS is a noninvasive ablative technique, which triggers the necrosis of the targeted UF tissue through a concentrated ultrasound beam.Magnetic resonance imaging (MRI) enables the treatment planning and temperature control of the treated area in real time.As already mentioned, MRgFUS is a relatively new therapeutic option in case of UFs.Available data concerning the safety profile support the hypothesis that complications after this type of treatment occur rather rarely, especially in case of severe AEs [17,18].A meta-analysis by Verpalen et al. dated from 2019 and our recent systematic review showed that due to the lack of a clear definition of AE and reporting bias in analyzed studies, it is difficult to compare and statistically analyze the results [19,20].It only shows that there is still no consensus among specialists using this method of treatment on this very important issue.Since our center has a large patient population in the method and indications in question, we decided to present the data and experience we have to other centers.
Our study aims at the comprehensive analysis of AEs in patients undergoing MRgFUS in the last 6 years in our reference center.

Material and methods
A group of 372 women suffering from UFs underwent treatment with MR-HIFU between April 2015 and May 2022 at the Department of Obstetrics and Gynecology of Pro-Familia Hospital in Rzesz ow (Poland) cooperating with the second Department of Obstetrics and Gynecology.In this retrospective case-control study, we focused on patients who reported AEs after or during treatment.Some of the treated patients had also participated in our previous study [21].
The research met the principles of the Declaration of Helsinki and received an approval of the Bioethics Committee (approval no 6/2022 from 12 January 2022) at the Center of Postgraduate Medical Education, Warsaw, Poland.Informed written consent regarding the use of medical and epidemiological data for scientific purposes was obtained from all the patients.Data were retrospectively retrieved from medical records.
The inclusion criteria were: the diagnosis of UFs, symptoms correlated with the existing UF/UFs (abnormal menstrual bleeding, anemia, bulk symptoms, painful intercourse, miscarriage or preterm delivery in the past, infertility (patients whose cause of infertility was other than UF had been excluded till 2018)).In the years 2015-2018, due to the scientific grant, the inclusion criteria were more strict.Only women till the age of 43 were qualified.After that time, the cohort was extended by women aged until 50 and those in whom UFs were not the causes of infertility.Some patients received pharmacological agents (oxytocin, misoprostol with diclofenac) directly before or during the treatment.The implementation of this medication was a part of the protocol in previous research and aimed to improve the outcomes of treatment.The qualification criteria, procedure description and outcomes were analyzed in previous studies [22,23].Those women were also included in our analysis regarding AEs.
The exclusion criteria for all the patients were: large UFs (>13 cm), UFs on the back wall of the uterus with direct contact with the rectum, multiple UFs (>2), pedunculated UFs, a history of operations due to UFs (relative), asymptomatic UFs, contraindications to MRI.
Medical interview was the first step of the qualification.Patients were also asked to complete the present authors' questionnaire regarding the symptoms of UFs.On clinical examination, we performed gynecological evaluation with cervical smear, the assessment of the adnexa, uterus and UFs with special attention paid to their mobility, tenderness and location.Transvaginal ultrasound, which was the next step, was performed by experienced sonographers.The assessment included the following features: the number, size, structure, location and position of UFs in the uterine cavity, the flow index (FI), the evaluation of the vascular index (VI) and the evaluation of the flow-vascular index.
The next step involved the performance of the MRI of the pelvic area.UFs were described according to Funaki et al. [24,25].The Funaki classification divides UFs based on signal intensity in T2-weighted MR images (T2WI) in relation to the intensity of skeletal muscles and myometrium.Type I-low intensity of the signal (comparable to that of skeletal muscles), type II-intensity lower than that of the myometrium, but higher than that of skeletal muscles, type IIIintensity equal or higher than that of the myometrium.The best therapeutic effect is achieved only in types I and II, so only those UFs were qualified for the treatment [24,25].Another purpose of the MRI assessment of the pelvic area was to establish the location and position of UFs in the uterine cavity with reference to the adjacent organs such as the intestines, bladder and rectum in order to enhance the safety of the procedure.Sonalleve integrated with Ingenia 3 T MRI scanner (Philips, Amsterdam, the Netherlands) was the MRgFUS system used for the qualification and treatment.All patients who were included in the study underwent an MRI scan with the administration of intravenous gadolinium contrast (Gadovist; Bayer Schering Pharma [Leverkusen, Germany]) with T1-and T2-weighted scans performed before and after receiving the contrast.At the end of the procedure, the patients were given an additional dose of gadolinium to establish the non-perfused volume ratio (NPVR) of UFs, which was the measurement of therapy success.
Thereafter, two follow-up visits after 3 and 6 months were planned.Gynecological examination and transvaginal ultrasound were performed during those visits.The patients received questionnaires concerning the quality of life and were asked about the occurrence of any AEs in this period.Due to the lack of appropriate standardized questionnaires for AE reporting in studies we decided to assess the AEs by asking the patients directly during the follow-up appointment and in the next step we classified those AEs with the Society of Interventional Radiology (SIR) classification of AEs scale to make the results more objective [26,19].This classification system sorts AEs into six groups: (1) Class A: no therapy, no consequence; (2) Class B: nominal therapy, no consequence; (3) Class C: require therapy, minor hospitalization (<48 h); (4) Class D: required major therapy, unplanned increase in the level of care, prolonged hospitalization (>48 h); ( 5) Class E: permanent adverse sequelae; (6) Class F: death.According to this scale, Classes A and B are described as minor complications; Class C, Class D, Class E and Class F are categorized as major complications.We measured the level of pain directly after the procedure with the Numerical Rating Scale (NRS) where 0 means no pain and 10 is the highest level of pain in life [27].
We compared patients with and without AEs after or during the treatment.Additionally, we analyzed epidemiological factors (age, body mass index (BMI), birth rate), characteristics of UFs (position, location, number, volume and Funaki type of the UFs), flexion of the uterus, distance between skin surface and the UF, the thickness of the subcutaneous tissue, presence of abdominal scars, use of misoprostol and oxytocin during the treatment and the technical parameters of the procedure (sonication time, maximal power and maximal temperature).All imaging characteristics were collected during MRI qualification before treatment.
Statistical analysis was conducted and in order to assess the impact of the studied factors on the occurrence of AEs, a generalized linear logistic regression model was used.Descriptive statistics of all variables was conducted with the Student's t and v 2 test, depending on the character of a variable.The results are presented as odds ratios (OR) with a 95% confidence interval (CI) in the univariate and multivariate analysis of selected variables.A p-value lower than 0.05 was considered as statistically significant.

Results
In the analyzed period, we treated 372 women with MR-HIFU due to symptomatic UFs.AEs were found in 33 patients (8.9%) after or during the treatment.No major AEs according to the SIR classification occurred in the analyzed group [26].Apart from UFs, three patients from AE cohort had a history of hypothyroidism, one had gallbladder polyps, one suffered from asthma and one reported an unidentified allergy.The remaining patients in this cohort reported no chronic diseases.The mean age of patients with AEs was 34.76 (±5.05; range 24-43 years).The mean BMI in the same group was 23.94 (±4.69; range 18.29-39.3).The mean birth rate before treatment was 0.29 (±0.53; range 0-2) in AE patients.The mean NPVR in AE group was 73% (±23%; range 15-100%).The mean volume of treated UFs in AE cohort was 84.14 mL (±86.21;range 288.3-8.76).All the above-mentioned results were compared to those of patients who did not report any AEs during or after the treatment.The results are presented in Table 1.There were no statistically significant differences in either group.
Regarding UF characteristics, the specific position, location of the UF and flexion of the uterus were not associated with a significantly increased risk of AE occurrence.The proportion between the UFs on the posterior and anterior wall was similar in both groups.The subserosal, submucosal or intramural location of the UF and the flexion of the uterus was also similarly prevalent in AE and AE-free patients.Interestingly, in the group of AE patients, multiple UFs were treated significantly less frequently than in AE-free cohort (30 vs. 55.5%,OR 0.34, CI 95%, p ¼ 0.013).However, the multivariate analysis did not confirm the statistical significance (OR 0.45, CI 95%, p ¼ 0.09).It is difficult to define its conclusive cause.We suppose this observation may result from a more rigorous selection of patients with multiple UFs than with a solitary lesion.As a large group of patients with multiple UFs did not meet the eligibility criteria, the procedure could be conducted in selected patients with specific UF characteristics.Nevertheless, this hypothesis needs to be evaluated in further research.
In the cohort of AE patients, 42.4% (n ¼ 14) of UFs were classified as type I, and 57.6% (n ¼ 19) as type II according to Funaki et al. [24].Funaki II UFs were treated more often in AE group than in AE-free patients (57.65% vs. 39.1%,respectively).The risk of AE occurrence was more than two-fold higher during the treatment of Funaki II versus Funaki I (OR 2.216, CI 95%, p ¼ 0.043).The multivariate analysis confirmed the findings (OR 2.6, CI 95%, p ¼ 0.04).
As regards the group of 33 patients with AEs, the mean sonication time was 111.09 min, the mean treatment time was 154.39 min, and the mean maximal sonication power was 183.13 W.There were no statistically significant differences regarding technical parameters between the groups.All above mentioned data are presented in Tables 1-4.
Interestingly, the occurrence of several AEs in the same patient was reported much more commonly than the occurrence of single AEs.Out of the whole AE group only 27.3% (n ¼ 9) reported a single AE, and the remaining 72.7% (n ¼ 24) experienced multiple AEs after the procedure.
Abdominal pain was the most commonly reported AE.It occurred in 52% (n ¼ 17) of women.The assessment with the NRS scale was available in only 73% (n ¼ 24) of patients during the procedure.The mean admitted level of pain during the treatment according to the NRS scale was 7.8.Fever and malaise were the second most common AEs.They both occurred in 21% (n ¼ 7) of patients.Hematuria and other urological problems occurred in 12% (n ¼ 4) of cases.Gastrointestinal AEs, such as diarrhea, nausea or flatulence, were reported in 9% (n ¼ 3) of patients.The group of patients who experienced AEs is presented in detail in Table 5.

Discussion
MR-HIFU/MRgFUS is quite a new, but exciting option in UF therapy.It was first accepted by the US Food and Drug Administration in 2004 in this indication [7].Since then, it has started to be a treatment opportunity in clinical use, not only as a part of research protocols.Currently, this method is  present in selected recommendations of scientific societies for the treatment of UFs, for example, the Society of Obstetricians and Gynecologists of Canada (SOGC) [7].
Although the method is still not widely available, numerous studies concerning the safety of the procedure were conducted [17,18,[28][29][30][31].Our recent systematic review showed that the method was rather safe, but the incidence of AEs varies a lot among the analyzed studies [19].Due to the lack of unification in reporting, the precise establishment of the occurrence of particular AEs is difficult.Our data from this retrospective analysis support the hypothesis that the overall incidence of AEs is relatively low.
Regarding the results of our study, the only characteristic of the UFs that had a relevant impact on AE occurrence was related to the treatment of Funaki II lesions.The risk of AEs was more than two-fold higher if the UF was classified as Funaki II versus Funaki I.This observation may result from a slightly decreased potential for thermoablation in this type of UFs compared to Funaki I.It might be due to a difference in the structure and biophysical parameters.According to current studies, signal intensity in T2WI is positively correlated to the number of uterine smooth muscle cells and negatively with the presence of fibroid tissue [32,33].Cellrich fibroids are much more difficult to treat compared to those with more fibers [34].Heterogenicity and higher signal intensity in T2WI are independent factors correlated with poorer treatment effects due to higher vascularization and difficulty in energy deposition [33].In order to reach the same therapeutic effect, more energy may need to be applied over a longer period.Although numerous authors demonstrated similar outcomes in the treatment of Funaki I and II UFs, some studies showed significantly lower NPV and higher reintervention rates in Funaki II versus Funaki I UFs [17,25,30,35].Nevertheless, the available data do not allow the comparison of the results in this aspect.Further research in this area is necessary to establish the clinical impact on patient selection for this type of UF therapy.
Another important issue resulting from our research is the impact of the thickness of the subcutaneous tissue and the presence of abdominal scars.Although some authors considered these factors as potentially increasing the incidence of AEs, especially skin burns [33], our study showed there was no statistically significant association between those two factors and AE occurrence.Mindjuk et al. showed that the thickness of the subcutaneous tissue and the presence of abdominal scars did not have a statistically significant impact on NPVR [35].However, the results obtained by Inbar et al. revealed significantly decreased NPVR in patients with abdominal scars compared to those without scars [36].Nevertheless, the authors did not note any serious AEs in both groups.The results of our study are somehow coherent with available data [35,36].As primarily the presence of abdominal scars was treated as a contraindication to MR-HIFU/MRgFUS, the results showed that more confidence in the described method may be needed in this group of patients.The individualized and careful assessment of each case may increase the eligibility of the procedure, while maintaining the same safety profile.Still, more prospective studies on larger groups of patients are needed to establish the final recommendations in this area.
It is worth comparing the occurrence of AEs in MRgFUS/MR-HIFU therapy and the incidence of those in other treatment methods of UFs.Although there are numerous studies comparing different treatment methods of UFs, data regarding AEs after or during the treatment are limited.Two studies published in 2003 and 2009 compared the safety of MRgFUS and hysterectomy [12,37].The overall incidence of AEs was significantly lower in MRgFUS than in hysterectomy group in both studies.These findings were also confirmed by a more current research in a meta-analysis by Tsai et al. published in 2021 comparing HIFU and conventional surgical methods [38].The meta-analysis consisted of 10 studies (4217 patients), but only 2 of them regarded MR-HIFU/MRgFUS.This fact may be somehow limiting, but in our opinion it does not seem to have a great impact on the final conclusions.
In case of UAE and MRgFUS, some authors also tried to compare the safety of both procedures.In 2017 Bernard et al. conducted a randomized controlled trial comparing MRgFUS to UAE [39].AEs constituted one of the outcomes of the study.The occurrence was similar and statistically insignificant in both groups.However, the authors stated that due to a limited number of participants, the assessment of safety parameters could be underpowered.A similar study by Ikink et al. revealed a significantly lower incidence of AEs in MR-HIFU/MRgFUS group [40].Those two studies were comprised in two meta-analyses comparing MR-HIFU/MRgFUS and UAE [41,42].The differences were not statistically significant, but a small sample size was the limitation of the studies.
The available data comparing AE occurrence after MR-HIFU/MRgFUS and after minimally-invasive surgical methods are limited.The majority of authors concentrated on US-HIFU/USgFUS.As the method shares some similarities with MR-HIFU/MRgFUS, some concerns about AEs could be extrapolated on the method presented in our article.Two recent meta-analyses comparing the safety profile of ablative treatment methods of UFs to surgical modalities (myomectomy and hysterectomy) revealed a significantly lower incidence of AEs, especially major ones, after ablative techniques compared to surgical methods.Skin burns were the only AEs, which occurred more often after ablative treatment [38,43].One of above-mentioned meta-analyses compared solely myomectomy to USgFUS [43].The incidence of major AEs in ablative group was significantly lower compared to myomectomy group.The limitation of this study is the inclusion and comparison of only major AEs.As the majority of AEs are minor ones in case of ablative techniques, it is difficult to compare the overall incidence.Further research should be conducted to compare the safety profile of surgical methods, especially minimally-invasive ones, such as laparoscopic myomectomy, with MRgFUS in UF therapy.
According to available literature, pain (mainly located in the lower abdomen) was the most commonly described AE in case of MRgFUS therapy in UFs.It was reported in 22 out of 43 studies analyzed in our systematic review [19].Our current research seems to support this hypothesis, as pain was reported by 17 out of 33 patients (51.5%).In our manuscript, we assessed the pain level with the NRS scale.It is difficult to compare the results to other similar papers as most studies lack information concerning any scale used to describe pain.It seems to be reasonable to use the same method to describe the intensity of pain in patients who reported this type of AE during or after the treatment.It would facilitate conducting a meta-analysis on larger groups of patients and draw conclusions concerning factors that may have influence on the pain level in certain groups of patients.
The majority of studies used the SIR scale to classify AEs, so we also decided to use it.According to the SIR classification, no major AEs (class C-F) occurred during or after the treatment in our center.The most of cases (87.9%) were classified as A (no therapy, no consequences), and the remaining ones (12.1%) were assessed as class B (nominal therapy, no consequence; includes overnight admission for observation only).
It is also worth considering the influence of the experience of medical professionals conducting MRgFUS therapy on the occurrence of AEs.This type of treatment needs very close cooperation and understanding between the radiologist and gynecologist.Adequate patient selection and correct treatment techniques are crucial for the effectiveness and safety of the procedure.As already mentioned, the ablative therapy of UFs is quite a new modality, widely available since 2004 [7].Some studies concerning the impact of experience on treatment outcomes have been published since that time.Japanese authors compared two groups of patients treated in the years 2003-2005 and 2005-2006 [44].They reported significantly fewer skin burns in the second group.In another study, Browne et al. compared two groups of patients treated in 2005-2009 and 2013-2019 with two generations of MRgFUS systems.The occurrence of AEs, especially serious ones, was significantly lower in the second group.Although technical improvement could greatly impact the outcomes, the authors claimed that AEs might largely be predicted through appropriate and rigorous screening.Our data seem to support the hypothesis, while the majority of AEs occurred in early years.
In our opinion, our study has a limitation which is related to the use of a non-standardized questionnaire regarding the types of AEs.All the AEs reported during and after the procedure were self-reported by the patients, mostly orally, directly after or during the treatment or at a follow-up appointment.We collected the answers from the patients and classified them according to the SIR classification to make the results more objective.We do not suppose that it could have a great impact on the outcomes of the study.However, the diversity of data might make it difficult to compare them to other studies.Although the SIR classification may somehow facilitate the comparison of the results, the great majority of AEs after MR-HIFU are classified as A or B based on this classification, and the only assessed aspect of AEs is their severity.This limitation was also mentioned in our systematic review concerning AEs after MRgFUS [19].The researchers did not use any standardization in AE reporting in most of the studies.In our viewpoint, the establishment of standardized questionnaires by experts is a key issue that would help researchers to statistically analyze the results and draw legible conclusions and researchers focusing on the area should be encouraged to do so.

Conclusions
Our study is the first available analysis of the risk factors of AE occurrence in patients treated with MR-HIFU/MRgFUS in UF therapy.According to our data, MR-HIFU/MRgFUS seems to be a relatively safe procedure.The occurrence of AEs after treatment is fairly low.We did not register any serious AEs.The occurrence of AEs is not dependent on the reached NPVR, volume, location or the position of the UF.Fat layer thickness and the presence of abdominal scars also do not increase the risk of AEs.The only factor that had a significant impact on AE occurrence the was treatment of Funaki II fibroids.Although our results confirm the safety of the procedure, more prospective, randomized studies on larger cohorts need to be conducted in this area.Increasing the awareness of the clinicians about the potential risks of the treatment would certainly help them plan the most patient-tailored therapy.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The study was funded by the Center of Postgraduate Medical Education, Warsaw, Poland, grant number 501-1-022-26-23.

Table 1 .
Comparison between patients with and without AEs.

Table 2 .
Univariate analysis for AE occurrence.

Table 3 .
Multivariate analysis for AE occurrence.

Table 4 .
Technical data of the procedure in patients with AEs.

Table 5 .
Characteristics of patients who reported AEs after treatment.