Evaluation of physician experience in achieving non-perfused volume ratio of high-intensity focused ultrasound ablation for uterine fibroids: a multicentre study.

OBJECTIVE: To evaluate the effect of different levels of physician experience on the high-intensity focused ultrasound (HIFU) ablation of uterine fibroids and to provide a reference for the use of non-perfused volume ratio (NPVR) standards during training.
METHODS: This prospective multicentre study enrolled patients with uterine fibroids. The effect of the physician's level of experience on the outcomes under different NPVR standards and the learning curve of six centres without HIFU experience were analysed. The impact of patient demographic and clinical characteristics were also evaluated.
RESULTS: A total of 1352 patients from 20 centres were included in the study. The median NPVRs were 92.00%, 88.10% and 92.86% in the no experience group, inexperienced group and experienced group, respectively. Posterior wall fibroids, lateral wall fibroids and fundus fibroids were inversely correlated with NPVR, while experienced physicians were positively correlated with NPVR. With NPVR ≥ 70% and NPVR ≥ 80% standards, physicians in the no experience group completed the learning curve on the 11th and 16th procedure, respectively. Physicians under a standard of an NPVR ≥ 90% did not complete the learning curve.
CONCLUSIONS: NPVR ≥ 80% is a standard that is worth using for HIFU treatment of uterine fibroids.

Introduction

Uterine fibroids are benign tumours of the uterus, which are formed by the excessive growth of monoclonal smooth muscle cells of the uterus. Uterine fibroids are asymptomatic in most cases, but sometimes they can cause abnormal uterine bleeding, pelvic pressure and pain, and reproductive dysfunction. There are many therapeutic strategies for uterine fibroids, such as hysterectomy, myomectomy, laparoscopic myomectomy and uterine artery embolization (UAE) therapy.[2,3] However, hysterectomy and myomectomy are associated with a high rate of significant complications and take weeks to recover from; and they can also cause psychological and physical trauma to patients. There is also evidence that prenatal uterine rupture can occur after laparoscopic myomectomy. The rate of major complications associated with uterine artery embolization are estimated to range from 1% to 17%. Every treatment approach should consider the age of the patient, their childbearing plans and the location and size of the leiomyoma.

In recent years, alternative therapies such as high intensity focused ultrasound (HIFU) have been widely used. The advantages of HIFU treatment are that it is completely non-invasive, it has a lower cost associated with patient recovery time and a lower incidence of complications. As a novel and complex therapeutic technique, there have been differences in outcomes between surgeons according to the physician’s experience, with studies showing that physician experience was not always positively correlated with patient outcomes.[8 –10] The non-perfused volume ratio (NPVR) as evaluated by enhanced magnetic resonance imaging (MRI) has been used as an important indicator of the success of HIFU ablation of uterine fibroids.[11,12] The cumulative recurrence rate in 2 years was less than 10% when 70% of NPVR was achieved, which was similar to that of myomectomy.[13,14] Recently, with the increasing clinical application of HIFU technology, the effectiveness of NPVR ≥ 80% and NPVR ≥ 90% have also been reported.[14 –16] Experienced physicians achieved better results than inexperienced physicians when NPVR ≥ 80% was the criteria for technical success. However, there has not been any studies reported on the learning curve of physicians under the three standards of NPVR ≥ 70%, NPVR ≥ 80% and NPVR ≥ 90%. Cumulative summation test for learning curve (LC-CUSUM) analysis can be used to evaluate various clinical techniques.[9,17]

The purpose of this study was to compare NPVR in physicians with different levels of experience from a multicentre study and to verify the significance of different NPVRs with LC-CUSUM analysis for physicians. The aim was to provide a basis for clinical technical training, physician selection, patient selection, NPVR standard formulation and the popularization and application of HIFU technology.

Patients and methods

Patient population

This prospective multicentre study enrolled consecutive patients with uterine fibroids at 20 clinical centres in China between March 2011 and December 2013. Patient data were selected from the Multicentre Research Information System of Therapy of Uterine Fibroids (www.hifuctr.com; Chongqing MicroSea Software Development Co., Ltd., Chongqing, China; Software copyright registration number: 2011SR094656). The inclusion criteria for patients were as follows: (i) premenopausal women with completed planned families and had no recent plan for a further pregnancy; (ii) imaging-confirmed diagnosis of uterine fibroids that had any of the following indications for hysterectomy: (a) enlarged uterus (uterine volume equal to or greater than that at 10 weeks’ gestation); (b) menorrhagia and/or secondary anaemia; (c) pelvic pain, urinary frequency, or constipation; (iii) for patients with multiple fibroids, no more than three fibroids with minimal diameters of 2 cm based on abdominal ultrasound present; (iv) fibroids clearly imaged by abdominal ultrasound; (v) for patients with abdominal surgical scars, the width of image blurring due to acoustic attenuation had to be <10 mm. The exclusion criteria for patients were as follows: (i) patients with uterine adenomyosis; (ii) previous myomectomy; (iii) concurrent pregnancy; (iv) pedunculated subserous or submucosal fibroids; (v) any single fibroid >10 cm maximum diameter; (vi) acute pelvic inflammation or uncontrolled systemic disease; (vii) patients were unable to communicate with physicians or were unwilling to sign informed consent. Patients were provided with written information describing the potential risks and benefits, including the potential impact on fertility and the risk of recurrence of symptoms.

This multicentre study was approved by the Chinese Ethics Committee of Registering Clinical Trials on 26 March 2011 (IRB approval number: ChiECRCT-2011034). Each patient signed informed consent before inclusion. Written informed consent was obtained from each patient before every procedure. The results of the clinical efficacy have been reported. The reporting of this study conforms to the CONSORT statements.

Eligibility criteria for physicians, prior training and quality control

A programme of HIFU treatment training and qualification certification authorized by the Ministry of Health of China were required for physicians. All operators were required to complete ablation treatment for 40–60 patients under guidance and supervision before entering the trial, with no major complications among treated patients, and a mean NPVR of treated uterine fibroids NPVR ≥ 70%.

Before participating in this study, all physicians needed to meet the following inclusion criteria: (i) they had obtained the qualification certificate of practitioners; (ii) they had met the criteria of ‘NPVR should be ≥70% per patient without major complications’; (iii) they had obtained HIFU treatment training programmes and qualification certificates authorized by the Ministry of Health of China through examinations.

HIFU ablation procedure

A single session of HIFU ablation was performed using the Focused Ultrasound Tumor Therapeutic System (Model-JC/JC200; Chongqing Haifu Medical Technology Co., Ltd, Chongqing, China). Equipment parameters used in this study were as follows: the frequency of the transducer was 0.8 MHz, the physical focal area was 1.5 mm × 1.5 mm × 10 mm, and the therapeutic power was 300–400W. A standardized clinical programme for physician training and clinical treatment was used. Patients were placed in the prone position on the HIFU treatment bed and the anterior abdominal wall was fully in contact with the degassed water in the water sac, which contained the ultrasonic transducer. Fentanyl (0.8–1 μg/kg) and midazolam (0.02–0.03 mg/kg) were provided every 30–40 min for conscious sedation. All patients kept awake or in light sleep with their breathing, oxyhaemoglobin saturation, heart rate and blood pressure monitored during the procedure. Using real-time ultrasonography monitoring, the treatment used massive grey-scale changes in the treated area as a measure of treatment effect. The unit time dose was adjusted according to patient tolerance and target grey-scale changes. The sonication was terminated when the grey-scale enhancement area covered the planned treatment area. Dose reference standard: total dose ≤150 kJ, treatment time ≤3 h. Postoperative patients remained prone for observations for 2 h.

MRI evaluation

All patients underwent pelvic MRI scans with a 3.0T MRI system before their operation (GE Medical Systems, Milwaukee, WI, USA). The three-dimensional diameters of the dominant uterine fibroids and the uterus were measured based on T2-weighted images. Enhanced MRI scans was performed again within 1 week after the treatment and used to measure the diameters of the NPV. The fibroid volume and NPV were calculated using the following equation: V = 0.5233 × D1 × D2 × D3, where D1 was longitudinal, D2 was anteroposterior and D3 was transverse. The NPVR was the ratio of the volume of the non-perfusion area in the postablation MRI scans to the volume of the fibroids. The signal intensity of the uterine fibroids was classified into three types according to the pretreatment T2-weighted images: hypointense, isointense and hyperintense.[11,15]

LC-CUSUM analysis

Technical success in terms of the LC-CUSUM analysis was defined as follows. For the three standards NPVR ≥ 70%, NPVR ≥ 80% and NPVR ≥ 90%, the definition of technical success was defined as an NPVR of uterine fibroids of at least 70%, 80% and 90%, respectively, with no major complications. Complications were recorded during a 30-day follow-up period and graded using the guidelines of the Society of Interventional Radiology (SIR): class A and B were considered to be minor and classes C to F were considered to be major complications.

The LC-CUSUM analysis parameters were as follows. For LC-CUSUM, suppose H0: process out of control; H1: process controlled. For standard CUSUM, suppose H0: process controlled; H1: process out of control. Unacceptable failure rate (P0), acceptable failure rate (P1: the level of operation when trainee is available). H represents the limit. CUSUM test performance is usually represented by average run length (ARL), defined as invalid before the signal appears under the hypotheses of null (H0; ARL0) and alternative (H1; ARL1). For standard CUSUM, suppose the P0 corresponding to H0 is 0.1 (failure rate is 10%; process is controlled), while P1 corresponding to H1 is 0.25 (failure rate is 25%; process out of control), select control HC = 2.25 to generate ARL0 and ARL1 values 120 and 24. Parameter settings are as follows: For LC-CUSUM, suppose the P0 corresponding to H0 is 0.2 (failure rate is 20%; process out of control), while P1 corresponding to H1 is 0.1 (failure rate is 10%; the process is under control). According to the computer simulation results of ARL0 and ARL1 under different constraints, the control limit HLC = 1.25 is selected to generate ARL0 and ARL1 values of 48 and 20, respectively. CUSUM analysis results are represented by two-dimensional coordinate diagrams, where the x-axis represents the number of operations and the y-axis represents CUSUM results.

Statistical analyses

All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, NY, USA) and Microsoft Office Excel (Microsoft, Redmond, WA, United States). Continuous data that followed a normal distribution are presented as the mean ± SD. Categorical data are presented as the number. Continuous data without normal distribution are expressed as median (interquartile range [IQR]). Comparison of continuous data between two groups without a normal distribution was undertaken using independent Mann–Whitney U-test. Multigroup comparisons were undertaken using Kruskal–Wallis H-test. Categorical data were compared by using χ2-test. Multivariate analysis was undertaken using binary logistic regression analysis. A P-value <0.05 was considered statistically significant. In logistic regression, the minimum sample size was 15-times the number of independent variables.

Results

A total of 1352 patients were included in this study. One patient was excluded as they were diagnosed with diffuse intravenous leiomyomatosis during follow-up. For the overall patient cohort, the median age was 42.00 years (IQR, 38.00–45.00 years), the median body mass index (BMI) was 22.42 kg/m2 (IQR, 20.70–24.33 kg/m2), the median volume of dominant uterine fibroids was 61.33 cm3 (IQR, 48.15–137.46 cm3), the median uterine volume was 258.56 cm3 (IQR, 186.40–360.30 cm3) and the median NPVR was 91.90% (IQR, 81.41%–100.00%). The median power was 400.00W (IQR,400.00W, 400.00W), the median sonication time was 1041.50 s (IQR, 667.00–1548.50 s) and the median total dose was 415.00 kJ (IQR, 265.34–621.28 kJ).

Among the 20 clinical centres, physicians were assigned to specific groups based on their HIFU experience when they joined the study as follows: five physicians from five centres with ≥3 years of HIFU experience were in the experienced group and completed at least 100 HIFU procedures per year; three physicians from three centres with HIFU experience <3 years were in the inexperienced group and completed no more than 300 HIFU procedures; and 12 physicians from the remaining 12 centres had no HIFU experience and formed the no experience group. A total of 489 patients with fibroids were treated by the experienced group of physicians, 150 patients with fibroids were treated by the inexperienced group and 713 patients with fibroids were treated by the group of physicians with no experience. Demographic and clinical characteristics of the three groups are shown in Table 1. There were significant differences between the three groups in terms of the maximum diameter of fibroids, uterine fibroid volume, uterine volume, number of fibroids, type of fibroids and classification of T2-weighted MRI signal intensity (P < 0.05 for all comparisons).

Table 1.

Demographic and clinical characteristics of the three groups of patients (n = 1352) with uterine fibroids stratified according to the level of experience of the physicians undertaking the high-intensity focused ultrasound ablation for uterine fibroids.

Characteristic	Level of physician experience			H/χ2	Statistical analysisa
	No experience group n = 713	Inexperienced group n = 150	Experienced group n = 489
Age, years	42.00 (38.00–45.00)	43.00 (38.75–46.00)	41.00 (38.50–45.00)	4.349	NS
Body mass index, kg/m2	22.41 (20.70–24.22)	22.95 (20.89–24.41)	22.27 (20.51–24.41)	1.803	NS
Maximum diameter of fibroids, cm	5.80 (5.00–6.90)	5.65 (5.00–7.10)	6.10 (5.10–7.30)	11.945	P = 0.003
Volume of uterine fibroids, cm3	74.88 (46.61–125.41)	75.05 (47.10–146.91)	90.26 (52.38–148.03)	12.553	P = 0.002
Volume of uterus, cm3	250.87 (179.24–351.17)	219.62 (160.62–323.5)	285.72 (204.02–388.52)	32.697	P < 0.001
Position of uterus				3.561	NS
Anteverted	496 (69.57)	112 (74.67)	362 (74.03)
Retroverted	217 (30.43)	38 (25.33)	127 (25.97)
Number of fibroids				9.399	P = 0.009
Solitary	528 (74.05)	127 (84.67)	354 (72.39)
Multiple	185 (25.95)	23 (15.33)	135 (27.61)
Location of fibroids				4.642	NS
Anterior wall	320 (44.88)	76 (50.67)	210 (42.94)
Posterior wall	194 (27.21)	35 (23.33)	148 (30.27)
Lateral wall	155 (21.74)	30 (20.00)	97 (19.84)
Fundus	44 (6.17)	9 (6.00 )	34 (6.95)
Type of fibroids				25.635	P < 0.001
Intramural	551 (77.28)	116 (77.33)	330 (67.48)
Submucosal	48 (6.73)	11 (7.33)	73 (14.93)
Subserosal	114 (15.99)	23 (15.33)	86 (17.59)
T2-weighted signal intensity				24.687	P < 0.001
Hypointense	404 (56.66)	59 (39.33)	220 (44.99)
Isointense	108 (15.15)	35 (23.33)	96 (19.63)
Hyperintense	201 (28.19)	56 (37.33)	173 (35.38)

Data presented as median (interquartile range) or n of patients (%).

aMultigroup comparisons were undertaken using Kruskal–Wallis H-test; categorical data were compared by using χ2-test; NS, no significant between-group difference (P ≥ 0.05).

The median (IQR) NPVR were 92.00% (79.85–100.00%) in the no experience group, 88.10% (80.34–94.01%) in the inexperienced group and 92.86% (83.22–100.00%) in the experienced group (Table 2). Under the standard of an NPVR ≥ 70%, the proportion of patients with an NPVR ≥ 70% in the inexperienced group (140 of 150 patients; 93.33%) was significantly higher than the proportion in the no experience group (618 of 713 patients; 86.68%) and the experienced group (449 of 489 patients; 91.82%) (P = 0.004). Under the standard of an NPVR ≥ 80%, the proportion of patients with an NPVR ≥ 80% in the experienced group (402 of 489 patients; 82.21%) was significantly higher than the proportion in the no experience group (535 of 713 patients; 75.04%) and the inexperienced group (115 of 150 patients; 76.67%) (P = 0.012). Under the standard of an NPVR ≥ 90%, the proportion of patients with an NPVR ≥ 90% in the experienced group (297 of 489 patients; 60.74%) was significantly higher than the proportion in the no experience group (403 of 713 patients; 56.52%)and the inexperienced group (61 of 150 patients; 40.67%) (P < 0.001).

Table 2.

Comparison of the success indicators between the three groups of patients (n = 1352) with uterine fibroids stratified according to the level of experience of the physicians undertaking the high-intensity focused ultrasound ablation for uterine fibroids.

Success indicators	Level of physician experience			H/χ2	Statistical analysisa
	No experience group n = 713	Inexperienced group n = 150	Experienced group n = 489
NPVR, %	92.00 (79.85–100.00)	88.10 (80.34–94.01)	92.86 (83.22–100.00)	21.598	P < 0.001
NPVR ≥ 70%	618 (86.68)	140 (93.33)	449 (91.82)	10.918	P = 0.004
NPVR ≥ 80%	535 (75.04)	115 (76.67)	402 (82.21)	8.773	P = 0.012
NPVR ≥ 90%	403 (56.52)	61 (40.67)	297 (60.74)	18.825	P < 0.001
Major complicationsb	1 (0.14)	1 (0.67)	1 (0.20)	2.110	NS

Data presented as median (interquartile range) or n of patients (%).

aMultigroup comparisons were undertaken using Kruskal–Wallis H-test; categorical data were compared by using χ2-test; NS, no significant between-group difference (P ≥ 0.05).

bMajor complications were defined as Society of Interventional Radiology class C–F complications.

NPVR, non-perfused volume ratio.

Major complications attributable to the intervention occurred in three of 1352 patients (0.22%), one in each group, during the 30-day follow-up period and were classified as SIR class C. There were no significant differences between the three groups. Other complications included the following: 62.28% (842 of 1352 patients) with lower abdominal pain, 16.20% (219 of 1352 patients) with vaginal discharge, 11.24% (152 of 1352 patients) with lumbar and back (sacrum) pain, 10.28% (139 of 1352 patients) with bloody vaginal secretions, 6.51% (88 of 1352 patients) with vaginal bleeding, 2.59% (35 of 1352 patients) with lower limb numbness and pain, 1.55% (21 of 1352 patients) with nausea and vomiting, 0.81% (11 of 1352 patients) with pain and distension of the anus, 0.67% (nine of 1352 patients) with lower limb weakness, 0.81% (11 of 1352 patients) with blurred vision, 0.81% (11 of 1352 patients) with pain and distension of the anus, 0.30% (four of 1352 patients) with inguinal pain and 0.30% (four of 1352 patients) with fever, 0.30% (four of 1352 patients) with urinary retention.

Among the 12 centres in the no experience group, six centres had 60 patients each and a total of 360 patients. A LC-CUSUM analysis of these six centres was performed under the standard of an NPVR ≥ 70% (Figure 1a). The LC-CUSUM provided an alert when it reached the 11th patient of the study group, indicating that the physician was competent and had completed the learning stage. The standard CUSUM was used to monitor whether the physician’s operative skills remained stable after completion of learning. If 60 procedures were completed without alarm, the treatment results were all considered ‘successful’. Under the standard of an NPVR ≥ 80% (Figure 1b), although the LC-CUSUM part of the learning stage in the 5th patient was a ‘failure’, a rapid warning was issued for the 16th procedure indicating that the physician had completed the learning stage. The follow-up standard CUSUM was used to monitor whether the physician’s operative skills remained stable after completion of learning. After completion of learning until the 60th procedure, there were no alarms and the treatment results were all considered ‘successful’. Under the standard of an NPVR ≥ 90% (Figure 1c), the physicians did not complete the learning stage and the treatment result were considered as ‘failure’.

Figure 1.

Cumulative summation test for learning curve (LC-CUSUM) analyses of the procedures undertaken at six centres with a total of 360 patients: (a) LC-CUSUM analysis under the standard of a non-perfused volume ratio (NPVR) ≥ 70%. The LC-CUSUM provided an alert when it reached the 11th procedure of the study group, indicating that the physician was competent and had completed the learning stage. If 60 procedures were completed without alarm, the treatment results were all considered ‘successful’; (b) LC-CUSUM analysis under the standard of an NPVR ≥ 80%. The LC-CUSUM part of the learning stage for the 5th procedure was a ‘failure’ but a rapid warning was issued for the 16th procedure indicating that the physician had completed the learning stage and (c) LC-CUSUM analysis under the standard of an NPVR ≥ 90% showed that the physicians did not complete the learning stage and the treatment results were all considered as ‘failures’.

When the cut-off NPVR was 80%, there were 1052 patients in the NPVR ≥ 80% group and 300 in the NPVR < 80% group. There were no significant associations with age, BMI, maximum diameter of fibroids, volume of uterine fibroids, volume of uterus, position of uterus, number of fibroids, type of fibroids and T2-weighted MRI signal intensity. Multivariate binary logistic regression analysis showed that the location of the fibroids and physician experience were significantly associated with NPVR (P < 0.05) (Table 3). Posterior wall fibroids, lateral wall fibroids and fundus fibroids were inversely correlated with NPVR, while experienced physicians were positively correlated with NPVR (P < 0.05).

Table 3.

Multivariate binary logistic regression analysis to evaluate the correlation between the technical success of physicians and the significant factors from the univariate analysis under the standard of non-perfused volume ratio (NPVR) ≥ 80% in patients (n = 1352) with uterine fibroids undergoing high-intensity focused ultrasound ablation.

Univariate factors	B	SE	Wald	P-value	OR [Exp(B)]	95% CI
						Lower	Upper
Location of fibroids
Anterior wall			16.838	P = 0.001
Posterior wall	–0.577	0.160	13.067	P < 0.001	0.561	0.411	0.768
Lateral wall	–0.505	0.175	8.359	P = 0.004	0.604	0.429	0.85
Fundus	–0.631	0.264	5.711	P = 0.017	0.532	0.317	0.893
Physician experience
No experience group			9.454	P = 0.009
Inexperienced group	0.058	0.213	0.075	NS	1.060	0.698	1.610
Experienced group	0.448	0.148	9.233	P = 0.002	1.566	1.173	2.091

SE, standard error; OR, odds ratio; CI, confidence interval; NS, no significant association (P ≥ 0.05).

Discussion

High-intensity focused ultrasound ablation therapy is a uterine-sparing technique that is reported have the potential to improve the quality of life on patients. As a new technology for the treatment of uterine fibroids, physician training is the key way to increase its clinical application in order to benefit more patients. As NPVR is directly related to long-term remission of clinical symptoms and it has a positive correlation with symptom control, it is considered to be the technical parameter for successful focused ultrasound therapy.[12,21] Early in the application of HIFU technology, NPVR up to 25% can result in the relief of symptoms. However, it was reported that the reintervention rate was 66.7% with a median NPV of 36.4% of all fibroid tissue volume. When the NPVR immediately post-treatment achieved 60%, the probability of undergoing an additional treatment was 10% at a mean 2-years after HIFU treatment of uterine fibroids; results that were comparable to those of myomectomy.[18,19] In early studies, the NPVR was predicted to be >60%, which was used as an evaluation index for the efficacy of HIFU treatment for uterine fibroid training.[22,24] On this basis, the technical requirements of operation training NPVR reached 70%.[22,23] A previous study reported that an HIFU ablation of uterine fibroids that achieved an immediate NPVR of at least 80% was safe. Recently, researchers have set the NPVR at a minimum of 80% as a measure of sufficient ablation in the HIFU ablation therapy of uterine fibroids.[25,26] Even some researchers have shown that an NPVR of 90% can be used as a standard for the HIFU ablation therapy of uterine fibroids.[15,16]

Among the 20 centres included in this current study, the aim was to obtain a higher NPVR on the premise of safety, so as to improve the clinical effect of HIFU ablation of uterine fibroids. Therefore, the performance of different physicians with varying levels of experience using an NPVR ≥ 80% was further studied. Under the standard of an NPVR ≥ 80%, posterior wall fibroids, lateral wall fibroids and fundus fibroids were negatively correlated with NPVR (P < 0.05) and experienced physicians were positively correlated with NPVR (P = 0.002). Patient selection and physician experience have an impact on NPVR ≥ 80%, which provides a basis for physicians to select suitable patients according to their own experience. Another noteworthy fact is that most clinical reports of HIFU treatment of fibroids come from China, while many other countries still have limited clinical experience of this technique. Studying the influence of physician experience on the efficacy of HIFU in the treatment of uterine fibroids can improve the confidence of the centre lacking HIFU treatment and provide a reference for physicians to choose different patients according to their own experience. In this current study, there was no significant difference in the occurrence of major complications (SIR class C) between the three groups. Class A complications do not require any therapy and have no subsequent consequences; and class B complications require nominal therapy and have no subsequent consequences other than overnight admission for observation. Class A and B complications have less impact on patients than class C complications.

Measuring CUSUM is a useful method of providing visual feedback before major quality problems occur. A CUSUM analysis is considered to be the best tool for determining quality control in medicine, especially in the evaluation of surgical technique.[12,13] LC-CUSUM is a variant of the CUSUM method that is dedicated to evaluating the personalized learning cycle of training programmes and it can be easily applied to various clinical techniques. Two previous reports described the first use of the learning curve and CUSUM to analyse HIFU treatment of uterine fibroids.[28,29] On this basis, this current study verified the LC-CUSUM under the three NPVR standards of NPVR ≥ 70%, NPVR ≥ 80% and NPVR ≥ 90%. This current study found that physicians performed well under an NPVR ≥ 70% and NPVR ≥ 80%, completing the learning curve and remaining in control. However, when the standard increased to NPVR ≥ 90%, physicians did not complete the learning curve within 60 procedures, which suggests that physicians without HIFU experience should not use NPVR ≥ 90% as the treatment standard.

This current study was the first to analyse physician experience stratified into three stages (no experience group, inexperienced group and experienced group) and to assess technical standards (NPVR ≥ 70%, NPVR ≥ 80% and NPVR ≥ 90%) using the LC-CUSUM method, so as to provide a reference for the promotion of HIFU ablation therapy of uterine fibroids, standardize training and standardize the setting.

This current study had several limitations. First, 20 clinical centres in China were selected, but it is not clear whether consistent results would be obtained in international centres. There are many differences in medical education models, medical systems and patient understanding of diseases, so whether training programmes would be applicable and whether standards for technical success would be appropriate still need to be further evaluated. Thus, prospective studies with larger sample sizes are needed to investigate the long-term results, including long-term symptom relief, the recurrence rate, as well as the re-intervention rate. The level and nature of the hospitals was different. There may also be differences in the source of patients and the expertise of physicians. Therefore, the influence of HIFU on the technical success needs further study. Possible abdominal scars in burn patients after treatment also need to be studied.

In conclusion, physicians with different levels of experience can successfully undertake HIFU ablation therapy of uterine fibroids and can exceed an NPVR of 80%. Experienced physicians perform best under the criteria of NPVR ≥ 80%. Posterior wall fibroids, lateral wall fibroids and fundus fibroids were negatively correlated with NPVR. Under the standard of an NPVR ≥ 70% and NPVR ≥ 80%, the inexperienced physician group can successfully complete the learning curve on the 11th and 16th procedure, respectively, while physicians did not complete the learning curve under a standard of an NPVR ≥ 90%. NPVR ≥ 80% is a standard that is worth using for HIFU treatment of uterine fibroids. In addition, in the selection of indications, based on the physician’s experience, more patients with uterine fibroids are suitable for HIFU ablation treatment.