Long-term outcome and risk factors of reintervention after high intensity focused ultrasound ablation for uterine fibroids: a systematic review and meta-analysis

Abstract

Objectives

To quantify the reintervention rate and analyze the risk factors for reintervention after high-intensity focused ultrasound (HIFU) ablation of uterine fibroids.

Methods

Eighteen studies were selected from the seven databases. A meta-analysis was applied to synthesize the reintervention rates for fibroids across various follow-up durations. Subgroup-analysis was conducted based on the year of surgery, sample size, guide methods, and non-perfusion volume ratio (NPVR). Signal intensity of T2-weighted imaging (T2WI) was independently evaluated for reintervention risk.

Results

The study enrolled 5216 patients with fibroids treated with HIFU. There were 3247, 1239, 1762, and 2535 women reaching reintervention rates of 1% (95% confidence interval (CI): 1-1), 7% (95% CI: 4-11), 19% (95% CI: 11-27), and 29% (95% CI: 14-44) at 12, 24, 36, and 60-month after HIFU. The reintervention rates of patients treated with US-guided HIFU (USgHIFU) were significantly lower than those of patients treated with MR-guided focused ultrasound surgery (MRgFUS). When the NPVR of fibroids was over 50%, the reintervention rates at 12, 36 and 60-month after HIFU were 1% (95% CI: 0.3–2), 5% (95% CI: 3–8), and 15% (95% CI: 9–20). The reintervention risk for hyper-intensity fibroids on T2WI was 3.45 times higher (95% CI: 2.7–4.39) for hypo-/iso-intensity fibroids.

Conclusion

This meta-analysis showed that the overall reintervention rates after HIFU were acceptable and provided consultative suggestions regarding treatment alternatives for patients with fibroids. Subgroup-analysis revealed that USgHIFU, NPVR ≥ 50%, and hypo-/iso-intensity of fibroids on T2WI were significant factors in reducing reintervention.

Systematic Review Registration

PROSPERO, CRD42023456094

Keywords:

• Uterine fibroid
• high-intensity focused ultrasound
• reintervention rate
• meta-analysis

Introduction

Uterine fibroids, also known as leiomyomas, are highly prevalent tumors that commonly occur in the female genital tract. The highest prevalence is 70% in reproductive-age women, and 30% of patients with uterine fibroids are symptomatic. Uterine fibroids may cause menstrual disorders, abnormal vaginal bleeding, pelvic compression, or infertility [1,2]. Various treatment options are available for patients diagnosed with uterine fibroids. These include the administration of medications, surgical approaches (myomectomy and hysterectomy), and non-surgical approaches (uterine artery embolization (UAE) and high-intensity focused ultrasound (HIFU) ablation). The treatment choice depends on the patient’s age, desire for fertility, and fibroid characteristics. Myomectomy is considered the most effective surgical procedure for individuals who desire to maintain fertility. However, it is important to note that the recurrence rate of fibroids increases significantly over time [3]. Uterine artery embolization (UAE) is also used in the management of uterine fibroids; however, the adverse effects of this treatment on ovarian function have significantly limited its clinical application [4].

In 2002, a pioneering noninvasive treatment called High-Intensity Focused Ultrasound (HIFU) was introduced for the treatment of uterine fibroids [5]. In 2004, the United States Food and Drug Administration (FDA) approved magnetic resonance imaging-guided focused ultrasound surgery (MRgFUS) as a viable treatment option for uterine fibroids. Numerous studies have conclusively demonstrated the high level of safety and efficacy inherent in utilizing HIFU for the treatment of uterine fibroids [6–8]. A prospective study with 2411 subjects from 20 centers compared HIFU treatment with myomectomy and hysterectomy, and the results showed that patients treated with HIFU had a similar long-term therapeutic outcome and lower postoperative morbidity than patients treated with surgical approaches [9]. Recently, a systematic review of minimally invasive approaches to uterine fibroids showed that HIFU significantly reduced uterine fibroid-related symptoms and improved the quality of life (QoL) of patients [10].

However, long-term follow-up results from different studies have varied. The FIRSTT study included 43 patients in the group treated with MRgFUS and 40 patients treated with UAE and showed that patients treated with MRgFUS had a higher reintervention rate within 24-month than patients treated with UAE [11]. However, Hu et al. retrospectively analyzed 412 patients with uterine fibroids who were treated with ultrasonic-guided HIFU (USgHIFU), and the results showed that the reintervention rate at 24 months was 3.15% [12]. The re-intervention rate is a vital factor for consulting different therapeutic methods for both doctors and patients. Because of the lack of consistent reintervention results, guidelines from either the American College of Obstetricians and Gynecologists (ACOG) or the Society of Obstetricians and Gynecologists of Canada (SOGC) have not yet recommended HIFU as a preferred treatment method and population indication for uterine fibroids [13,14].

Therefore, we conducted a systematic meta-analysis to determine the reintervention rate of uterine fibroids after HIFU and analyze the influencing risk factors for reintervention, aiming to recommend preferred treatment strategies and tailor suitable populations for HIFU.

Materials and methods

Research design

The meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, which provided a standardized framework for the design, implementation, and reporting of the study [15]. Literature search and screening, quality evaluation, and data extraction and analysis were conducted according to guidelines and regulations.

Information sources

Original articles were identified by searching PubMed, Embase, Web of Science, Cochrane, Scopus, Chinese National Knowledge Infrastructure (CNKI), and Wanfang databases from January 2002 to August 2023.

Search strategy

The keywords and terms used in the search included uterine fibroids/leiomyomas and HIFU ablation. The search strategies are listed in Supplementary 1.

Eligibility criteria

Articles were included in this comprehensive review based on the following specific criteria:1) retrospective or prospective observational studies; 2) symptomatic uterine fibroid/leiomyoma patients; 3) high-intensity focused ultrasound ablation was performed; 4) reintervention outcomes were reported in the study; and 5) follow-up time was longer than 12 months. If necessary, potential non-English research articles were translated using translation software or translators.

The exclusion criteria were as follows:1) duplicate studies; 2) HIFU combined with other treatment modalities; 3) articles without original data, such as abstracts, reviews, animal experiments, letters, comments, and consensus articles; and 4) case reports or studies with less than 20 cases.

Study selection

Two individual researchers (DYY and LY) conducted a systematic review of the articles based on predetermined eligibility criteria. Research screening was divided into two stages: first, titles and abstracts were screened, and second, full texts were assessed. Additionally, we reviewed the references of the included studies. Any disagreements were resolved by the senior investigators. Figure 1 shows the literature screening process, and a list of qualified research information is shown in Supplement 2.

Figure 1. Study selection flow chart.

Data extraction

Two investigators (DYY and LY) independently extracted relevant data from the included articles. For each study, the author’s first name, publication year, country, study design, sample size, follow-up time, intraoperative guide method, and reintervention events were recorded. Participant characteristics, including age, fibroid size, year of surgery, signal intensity of fibroids on T2 weighted image (T2WI), and non-perfusion volume ratio (NPVR) were extracted. The reintervention rate was the primary outcome indicator for determining long-term therapeutic efficacy. Based on the length of follow-up, the reintervention rates at 12, 24, 36 and 60-month are summarized.

Assessment of risk of bias

Two investigators, DYY and LY, independently assessed the quality of the studies using the ROBINS-I tool, which is an evaluation tool for non-randomized intervention studies [16]. The bias assessment included seven aspects: confounding bias, participant selection bias, exposure assessment bias, misclassification bias during follow-up, missing data bias, measurement bias of the outcome, and bias in selection of the reported results. Risk assessment was categorized into four tiers: low, moderate, severe, and no information (Supplement 3). Differences were determined using a third evaluator (HM).

Data synthesis and analysis

Meta-analysis of the extracted data was conducted using R statistical language version R 3.5.3. Continuous data are presented as mean ± standard deviation (SD), while enumeration data are expressed as numbers and percentages. The P value was bilateral, and p < 0.05 level was considered significant. I² statistics were used to evaluate the heterogeneity between studies, with 75%, 50%, 25% representing high, moderate, and low heterogeneity, respectively. Random- and fixed- effect models were used to summarize reintervention rate based on relatively I²% >50 and <50, and a 95% confidence interval (95% CI) was reported. To identify subgroup differences and potential heterogeneity between studies, stratification was performed according to the year of surgery (before or after 2010), sample size (<100 or ≥100), intraoperative guide method (MR-guide- or US-guide), and NPVR(<50 or ≥50). The estimated difference between subgroups was considered significant when it was less than 0.05 (p < 0.05). According to Funaki’s study [36], uterine fibroids were classified into three groups based on their signal intensity on T2-weighted imaging (T2WI): hypo-intensity (signal intensity similar to that of the skeletal muscle), iso-intensity (signal intensity between that of the myometrium and the skeletal muscle), and hyper-intensity (signal intensity the same as or greater than that of the myometrium). The relationship between fibroid signal intensity on T2WI and the need for reintervention was further analyzed.

Results

Study screening and study characteristics

Our preliminary search identified 7525 potential references. One hundred and fourteen articles qualified for further evaluation after removing duplicated literature and screening the abstracts of these records. Eighteen full-text studies were included in the meta-analysis based on the inclusion and exclusion criteria. Figure 1 illustrates the process of reviewing the records. Among these 18 studies, there were eight from China [12,17–22], two from the United States [11,23], two from Germany [24,25], two from the Netherlands [26,27], and four from Japan, the United Kingdom, France, and Israel respectively [28–31]. These studies comprised 4 prospective observational studies and 14 retrospective studies. These studies were stratified based on the length of follow-up, eight of which had 12 months follow-up results, five studies with 24 months results, seven studies with 36 months results, and eight studies with 60 months results.

Data source

This study involved a comprehensive analysis of 5216 patients who were diagnosed with symptomatic fibroids and underwent HIFU treatment. Of the 3247 women who had 12-month follow-up results, 90 had reintervention events. Of the 1239 women who had 24-month follow-up results, 73 reintervention events were observed. Of the 1762 women who had 36-month follow-up results, 224 reintervention events were observed. Of the 2535 women who had 60-month follow-up results, 442 reintervention events were observed. Table 1 summarizes the published literature and baseline characteristics.

Table 1. Characteristics of the studies included in this meta-analysis.

NO.	First author, year	Country	Data years	Study design	Sample size	Age, y	Fibroid size/diameter cm,	NPV%	T2WI image (event/n)	Follow-up (m)	Reintervention (event/n)	Monitors
							Volume, mm^3
1	Funaki, 2009	Japan	2004–2008	Prospective cohort	91	40.4 ± 4.6	177.40 ± 164.96	52.7 ± 21.63	Hypo-/iso- 12/80	24	14/91	MR-guide
									Hyper- 2/11
2	Froeling, 2013	Germany	2002–2009	Retrospective	50	36.1 ± 9.85	67.4 (40.5–138.1)	41.2% (25–75%)	None	60.7	1year 15/50	MR-guide
											5year-rei 24/36
3	Ikink, 2014	Netherlands	2010–2013	Retrospective	51	46 ± 4.44	8.5 (6.5–10.7)	<0.5 15/35	None	15(9–26)	24/51	MR-guide
								≥0.5 3/16
4	Gorny, 2014	USA	2005-–2011	Retrospective	138	45.6 ± 5.8	343.3 ± 300.4	45.5 ± 22.7	Hypo-/iso- 23/116	34(12–85)	29/138	MR-guide
									Hyper- 6/21
5	Quinn,2014	England	2003–2010	Retrospective	180	42.2	396.3	44.2	Hypo-/iso- 68/130	None	3year-rei 77/180	MR-guide
									Hyper- 35/44		5year-rei 96/162
6	Mindjuk, 2015	Germany	2010–2012	Retrospective	250	42.1 ± 6.9	5.3 ± 2.1	88.7 ± 14.4%	None	19.4	28/250	MR-guide
7	Thiburce, 2015	France	2008–2012	Retrospective	36	43.5 (41–46)	255			21.4	1Year-rei 1/36	MR-guide
											2year-rei 5/36
8	Zhao,2017	China	2012–2015	Prospective	165	37.1 ± 6.0	6.04 (3.46–10.97)	76.2 ± 20.7	Hypo-/iso- 8/112	12	22/165	US-guide
									Hyper- 12/53
9	Chen,2017	China	2011–2013	Prospective	1357	41.31 ± 5.08		87.2		12	1year-rei 14/1353	US-guide
10	Mohr-Sasson,2018	Israel	2012–2017	Retrospective	68	44 ± 6.67	7 ± 1.85	None	None	36.5(24.2–41.7)	9/68	MR-guide
11	Xu,2018	China, Suining	2010–2014	Retrospective	450	42 ± 5		83 (65–100)		24	1year-rei 6/450	US-guide
											2year-rei 13/450
12	Laughlin-Tommaso, 2019	America	2010–2014	Prospective	43	44.5 ± 5.0	249.2 ± 159.9	None	None	36	3year-Riv 13/43	MR-guide
13	Verpalen, 2020	Netherlands	2010–2017	Retrospective	87	44.6 ± 4.7	9.3 ± 3.1	45.03 ± 27.89	None	63.5 ± 29.0	29/87	MR-guide
14	Li, 2020	China	2012–2014	Retrospective	381	42.4 ± 5.0	5.46 ± 1.52	81.9 (90.36–91.91)	Hypo-/iso- 53/303	70.0 ± 9.0	79/381	US-guide
									Hyper- 26/78
15	Hu,2020	China	2011–2014	Retrospective	412	40 ± 5.6	6.2 ± 1.6	87.1 ± 16.3	2year Hypo-/iso-6/331 Hyper- 7/81	77	1year-rei 3/412	US-guide
											2year-rei 13/412
									5year Hypo-/iso- 25/331 Hyper- 19/81		3year-rei 29/412
									5year-rei 44/412
16	Liu, 2021	China	2011–2016	Retrospective	536	40.59 ± 9.46	5.7 ± 1.94	>74.3	Hypo-/iso- 47/395	69(48–89)	77/536	US-guide
									Hyper- 63/142
17	Liu, 2022	China	2007–2016	Retrospective cohort	730	38.8 ± 6.3	6.4 ± 2	86 ± 32.59	None	84(44–142)	1year-rei 5/730	US-guide
											3year-Riv 30/730
											5year-Riv 50/730
18	Zeng, 2022	China. Nanchong	2016–2017	Retrospective	191	41	<6.58		Hypo-/iso- 28/156	60	5year-rei 43/191	US-guide
									Hyper- 28/77

NPV, non-perfusion volume; Hyper-, Hyper-intensity fibroids on T2WI; Hypo-/iso-, Hypo-/iso-intensity fibroids on T2WI; 1 year-rei, 1-year reintervention; 2 year-rei, 2-year reintervention; 3 year-rei, 3-year reintervention; 5 year-rei, 5-year reintervention; MR-guide, Magnetic resonance guide; US-guide, ultrasound guide.

Risk of bias of the included studies

The included studies were mainly non-randomized controlled trials, resulting in a high confounding bias. Retrospective studies showed a high missing data bias and a lack of follow-up information on misclassification. According to the ROBIN-I tool, nine studies were considered to have a moderate risk of bias and nine studies had a serious risk of bias. The assessment details of the risk of bias are presented in Supplement 3.

Primary outcomes

Reintervention rates were summarized by meta-analysis (Figure 2) and are presented in Table 2. The reintervention rates at 12, 24, 36, and 60-month were 1% (95% CI: 1-1), 7% (95% CI: 4-11), 19% (95% CI: 11-27), and 29% (95% CI: 14-44), respectively. Although the level of heterogeneity between studies was relatively high, sensitivity analysis showed that the results were stable and not affected by the exclusion of studies individually (Supplement 4).

Figure 2. Forest map of overall intervention risk at 12-,24-,36-,60-months. A. Forest map of overall intervention risk at 12-month; B. Forest map of overall intervention risk at 24-month; C. Forest map of overall intervention risk at 36-month; D. Forest map of overall intervention risk at 60-month.

Table 2. Overall reintervention risk.

Time	Rei (95%CI) %	I ^2%	Total sample size
12- month	1 (1–1)	92	3247 [9,12,18,20,22,24,26,30]
24- month	7 (4–11)	86	1239 [12,22,25,28,30]
36- month	19 (11–27)	96	1762 [11,20–21,23,27,29,31]
60- month	29 (14–44)	97	2535 [12,17,20,21,24,27–29]

Rei, reintervention rate.

Sub-analysis different reasons

To further analyze the source of heterogeneity, subgroup analyses were performed to evaluate differences in reintervention rates among different research and demographic characteristics. As shown in Table 3, the results of subgroup analyses revealed that USgHIFU had significantly lower reintervention rates than that of MRgFUS at 12 (0.9% vs. 12%, p = 0.08), 24 (3% vs. 12%, p < 0.0001), 36 (8% vs. 29%, p = 0.004), and 60-month (15% vs. 53%, p = 0.0002) of follow-up. The reintervention rates of patients with NPVR ≥ 50% were significantly lower than those of patients with NPVR < 50% at 12 (1% vs. 38%, p < 0.0001), 36 (5% vs. 36%, p < 0.0001) and 60-month (15% vs. 53%, p = 0.0002), respectively. The data also showed that the different surgical years (before vs. after 2010) led to a significant difference in the reintervention rate of 60-month (61% vs. 16%, p < 0.0001). Sample size had no consistent significant effect on the different lengths of follow-up; specifically, no significant difference at 36- and 60-months, while sample size over 100 had lower reintervention rate than sample size under 100 at 12 (0.9% vs. 12%, p = 0.08) and 24-month (5% vs. 15%, p = 0.007).

Table 3. Subgroup reintervention rate.

	12-month Rei%					24-month Rei%					36-month Rei%					60-month Rei%
Subgroup	(95%CI)	I^2%	No.	Sample size	P-value	(95%CI)	I^2%	No.	Sample size	P-value	(95%CI)	I^2%	No.	Sample size	P-value	(95%CI)	I^2%	No.	Sample size	P-value
the year of surgery																				<0.001
<2010	None					None					None					61 (54–67)	0	2	198
≥2010	None					None					None					16 (14–17)	88	5	1607
sample size					0.08					0.007					0.82					0.11
<100	12 (7–16)	95	3	137		15 (0.8–21)	0	2	127		21 (4–37)	77	2	111		49 (17–82)	92	2	123
>100	0.9 (0.6–1)	83	5	3110		5 (2–9)	87	3	1112		19 (9–28)	97	5	1651		22 (13–31)	98	6	2412
Guide method					0.08					<0.001					0.004					0.0002
MR-guide	12 (7–16)	95	3	137		12 (9–16)	0	3	377		29 (15–42)	90	4	429		53 (33–72)	90	3	285
US-guide	0.9 (0.6–1)	83	5	3110		3 (2–4)	0	2	862		8 (3–12)	88	3	1333		15 (9–20)	93	5	2250
NPVR					<0.001										<0.001					0.0002
<50%	38 (22–55)	69	2	101		None					36 (21–50)	87	2	318		53 (33–72)	90	3	285
>50%	1 (0.3–2)	83	5	3110		None					5 (3–8)	75	2	1142		15 (9–20)	93	5	2250

Rei, reintervention rate; MR-guide, Magnetic resonance guide; US-guide, Ultrasound guide; NPVR, non-perfusion volume ratio; No. Number of studies.

Second outcomes

In addition, the influence of the T2WI signal intensity of uterine fibroids on the reintervention rate was identified. Of the 8 studies reporting the number of patients with different signal intensities and reintervention events, 2200 patients were included, including 1623 patients with hypo-/iso-intensity fibroids and 507 patients with hyper-intensity fibroids on T2WI (no distinction between uniform or mixed high signal intensity). The results demonstrated that patients with hyper-intensity fibroids had a 3.45 times higher risk of reintervention (95% CI: 2.7-4.39; I²: 40%) compared to those with hypo-/iso-intensity fibroids (Figure 3). The sensitivity analysis showed that the results were unaffected by the exclusion of any single study (Supplement 4). The funnel map did not show any publication bias (Supplement 4).

Figure 3. Forest map of different T2WI signal intensity. Hyper-, Hyper-intensity fibroids on T2WI; Hypo-/iso-, Hypo-/iso-intensity fibroids on T2WI.

Discussion

The long-term reintervention rate is a considerable factor in treatment options for patients with uterine fibroids. Based on the length of follow-up after HIFU, we systematically evaluated the reintervention rate and conducted a sub-analysis of the factors affecting the reintervention rate. The reintervention rates at 12, 24, 36, and 60-month after HIFU were 1% (95% CI: 1-1), 7% (95% CI: 4-11), 19% (95% CI: 11-27), and 29% (95% CI: 14-44), respectively. The subgroup analysis results showed that NPVR and fibroids signal intensity on T2WI were important factors that affected the reintervention rate.

Currently, both MRI and ultrasonography are used to guide HIFU treatment. MRI provides exceptional anatomical precision and enables monitoring of temperature levels during HIFU treatment. However, MRI is currently unable to provide real-time anatomical monitoring during HIFU for uterine fibroids. Ultrasound is a cutting-edge technology that enables real-time monitoring and imaging of anatomical structures. Although the temperature changes in the treated area could not be monitored, the alteration in grayscale levels observed in fibroids during HIFU treatment proved to be a dependable indicator for monitoring the effectiveness of HIFU in treating fibroids. In addition, The USgHIFU procedure offers a significant advantage over other treatments in that it eliminates the need for patients to be confined within a small space during HIFU treatment. The patient was positioned in a prone position on the HIFU table, with a nurse and a physician seated nearby to facilitate comfortable communication. Numerous studies have provided compelling evidence to support the effectiveness of USgHIFU as a treatment option for uterine fibroids [3,9,17,30]. A multi-center prospective study demonstrated that patients treated with USgHIFU had comparable long-term therapeutic outcomes and lower postoperative morbidity compared to patients treated with myomectomy or hysterectomy [9]. A meta-analysis revealed that patients treated with USgHIFU had better ablation effects, shorter treatment times, fewer complications, and better long-term effectiveness than those treated with MRgFUS [32]. In that study, NPVR achieved in patients treated with USgHIFU was 81.7%, while it was 58.92% in patients treated with MRgHIFU, and the long-term reintervention rate could drop to 5.2% in patients treated with USgHIFU and 13.4% in patients treated with MRgFUS during 12 months follow-up. A study demonstrated the equivalent safety of MRgFUS and USgHIFU. USgHIFU was superior to MRgFUS in complete ablation rate, which was 43.1% in USgHIFU and 23.3% in MRFUS [33].

Another previous meta-analysis, including seven studies related to MRgFUS, concluded that the reintervention rates of patients with uterine fibroids treated with MRgFUS were 12%(95% CI: 4-20), 14%(95% CI: 7-21), 22%(95% CI: 11-32)and 49%(95% CI: 21-77)at 12, 24, 36, and 60-month, respectively. The reintervention rate at 60-month after MRgFUS increased rapidly [34]. Our meta-analysis showed that the reintervention rate at 60-month after HIFU treatment was 29% (95% CI 14, 44), which was lower than that reported at 49% (95% CI, 21-77) in the previous meta-analysis. This phenomenon may be explained by the fact that this meta-analysis was a comprehensive study that summarized both USgHIFU and MRgFUS. The subgroup analysis also showed that the USgHIFU group had better long-term results than the MRgFUS group.

Several studies have shown that the reintervention rate of myomectomy is approximately 1.1-6% at 12-month, 10% at 24-month, 1.2-9% at 36-month and 12.2-19% at 60-month after myomectomy [34,35]. With regard to UAE, the reintervention rate was 3.6-7% at 12-month, 8% at 24-month, 7.4-14% at 36-month, and 14.4-21% at 60-month, respectively. In comparison, Both MRgFUS and USgHIFU showed that the reintervention rate at 12 and 24-month were similar to those of myomectomy and UAE. The reintervention rates of MRgFUS at 36-month and 60-month were higher than those of myomectomy and UAE. However, the reintervention rate of USgHIFU was consistent with those of myomectomy and UAE. The relatively high reintervention of MRgFUS can be explained by the fact that the non-perfusion volume ratio (NPVR) of fibroids after MRgFUS was lower than that of USgHIFU. Fibroids without ablation may continue to grow, leading to reintervention events [34].

To provide a better treatment strategy and tailor a suitable population for HIFU, we further analyzed the impact of potential risk factors for reintervention. Our results demonstrated that the risk of reintervention for hyper-intensity fibroids was 3.45 times higher than that for hypo-/iso-intensity fibroids. Previous studies reported that hyper-intensity uterine fibroids had higher smooth muscle cell proliferative activity, more vascular index, more tissue deformation, richer fluid content, and less fibrous tissue than that of hypo-intensity fibroids [13,28,36–38]. The reasons for the insufficient ablation effect of hyper-intensity fibroids were that tissue degeneration hindered ultrasound focusing, and rich fluid content and blood flow inhibited the temperature rise. In 2009, Funaki also conducted a cohort study of 91 patients with fibroids receiving MR-guided HIFU, including 80 patients with hypo/iso-intensity fibroids, with a reintervention rate of 14% at 24 months, and 11 patients with hyper-intensity fibroids, with a reintervention rate of 21.6% at 24 months [28]. Quinn et al. also showed that fibroid signal intensity had a significant impact on the reintervention rate, and the reintervention rate of patients with type 3 (hyper-intensity) fibroids reached 79.54% at 60 months after MRgFUS [29].

Currently, a new classification system and multimodal imaging examination based on the heterogeneity of uterine leiomyoma on MRI have been introduced [38,39]. The results indicate that MRI can be used to predict the reintervention rate based on the heterogeneity of uterine fibroids.

Our study indicated that NPVR was associated with long-term effectiveness after HIFU. If NPVR reached higher than 50%, the reintervention rates decreased to 1%, 5%, and 15% at 12, 36, and 60-month after HIFU. The present findings seem to be consistent with those of previous observational studies. In the initial application of MRgFUS for uterine fibroids, NPVR only reached 36.4%, and the reintervention rate was 66.7% at 24 months [23]. Later, when the NPVR increased to 50%, the reintervention rate of fibroids decreased to 15.38% at 24-month [28]. Recently, Liu et al. retrospectively analyzed 536 patients with fibroids who were treated with USgHIFU, and the results also found that NPVR was one of the independent risk factors affecting reintervention (OR 0.944, 95% CI: 0.919–0.969). The overall reintervention rate was 14.37% during an average of 69 months of follow-up [19]. Previous studies have shown that NPVR is related to the characteristics of uterine fibroids [25]. As mentioned in a previous paper, a lower NPVR was found to be associated with fibroids in terms of the interval (NPVR reduced by 8%), subserosal fibroids (NPVR reduced by 3.6%), distance of fibroids to the spine less than 3 cm (NPVR reduced by 3.9%), enhanced fibroids on T1WI (NPVR obviously reduced by 11.3%), and acoustic window distance (NPVR decreased by 1.5% with an increase of every 1 cm).

The data showed a significant difference in the reintervention rate at 60-month of follow-up for HIFU treatment performed before or after 2010. The main reasons included the differences between the restrictive and nonrestrictive applications of MRgFUS in the two stages. Originally, before 2004, MRgFUS was only allowed to ablate 33% of the fibroid volume, less than 100mm³ of a single fibroid, less than 120 min of the operation time, and no repeated surgery. In April 2004, the FDA relaxed the requirement for permission to ablate 50% of fibroid volume, less than 150mm³ of single fibroids, and less than 180 min of operation time for MR-guided HIFU [40]. Finally, nonrestrictive MR-guided HIFU came after 2009. Therefore, the reintervention rates decreased as NPVR increased after 2010.

Recent studies have found a positive correlation between doctor experience and ablation success rates [41,42]. The NPVR in inexperienced centers (starting less than three years) was lower than that in experienced centers (starting > 3 years). Thus, experience may be a problem for centers with a small number of participants in our study.

However, this study has several limitations. First, it primarily relied on observational studies investigating the effects of HIFU treatment on uterine fibroids. This choice was constrained by the limited availability of the literature on the long-term outcomes of this treatment. However, the present study incorporated the most comprehensive meta-analysis that is currently available. Nevertheless, further research utilizing randomized controlled trials (RCTs) about the long-term outcomes of HIFU and myomectomy is necessary.

Additionally, it is important to acknowledge that the prevalence of uterine fibroids can vary across racial groups, potentially influencing the long-term outcomes of reinterventions. Regrettably, owing to limitations in the available data, subgroup analyses based on race were not feasible. Nevertheless, this meta-analysis subgroup has provided valuable insights into the risk factors associated with the rate of reintervention, including the interoperation guidance method employed for HIFU, NPVR, and the MRI signal intensity of fibroids. The findings derived from this analysis can be used to develop individualized treatment strategies for patients diagnosed with uterine fibroids.

Conclusion

This systematic review and meta-analysis comprehensively reviewed the long-term outcomes of HIFU treatment for uterine fibroids. Overall, reintervention rates were acceptable. Subgroup analyses demonstrated that the use of USgHIFU and high NPVR were associated with a decreased long-term reintervention rate. Additionally, hypo-/iso-intensity fibroids showed better long-term effects. However, further research is required to optimize the selection of the patient population and the treatment methods for uterine fibroids.

Acknowledgment

The authors acknowledge all the people who helped us.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

This is a meta-analysis study, and the data that support the findings of this study are available in reference number [9,11,12,17,18,20–31]. The Characteristics of included studies are shown in Supplement 2.

Additional information

Funding

This study was supported by the Opening of the Foundation of State Key Laboratory of Ultrasound in Medicine and Engineering, Chongqing Medical University, in 2023 (No. 2023KFKT006).