Factors influencing USgHIFU ablation for adenomyosis with NPVR ≥ 50%

Abstract

Objective

To investigate the influencing factors of ultrasound-guided HIFU (USgHIFU) ablation for adenomyosis with a non-perfused volume ratio (NPVR)≥50%.

Methods

A total of 299 patients with adenomyosis who underwent USgHIFU ablation were enrolled. Quantitative signal intensity (SI) analysis was performed on T2WI and dynamic enhancement type. The energy efficiency factor (EEF) was defined as the ultrasound energy delivered for ablating 1 mm³ of tissue. NPVR ≥ 50% was used as the criterion for technical success. Adverse effects and complications were recorded. Logistic regression analyses of variables were conducted to identify the factors affecting NPVR ≥ 50%.

Results

The median NPVR was 53.5% (34.7%). There were 159 cases in the NPVR ≥ 50% group and 140 cases in the NPVR < 50% group. The EEF in NPVR < 50.0% group was significantly higher than that in NPVR ≥ 50% group (p < 0.05). The incidence of intraoperative adverse effects and postoperative adverse events in the NPVR < 50% group were higher than those in the NPVR ≥ 50% group (p < 0.05 for both). Logistic regression analysis showed that abdominal wall thickness, SI difference on T2WI between adenomyosis and rectus abdominis, and enhancement type on T1WI were protective factors for NPVR ≥ 50% (p < 0.05), while the history of childbirth was an independent risk factor (p < 0.001).

Conclusions

Compared with NPVR < 50%, NPVR ≥ 50% did not increase the intraprocedural and postprocedural adverse reactions. The possibility of NPVR ≥ 50% was higher in patients with thinner abdominal walls, showed slight enhancement of adenomyosis on T1WI, with a history of childbirth, or in whom the SI difference on T2WI between adenomyosis and rectus abdominis was more minor.

Keywords:

• High-intensity focused ultrasound (HIFU)
• adenomyosis
• non-perfused
• volume ratio (NPVR)
• safety
• magnetic resonance imaging (MRI)

1. Introduction

Adenomyosis is one of the common benign diseases of the uterus in women of reproductive age. The reported prevalence in women worldwide ranges from 5% to 70%[1]. Symptoms of adenomyosis often include abnormal uterine bleeding, chronic pelvic pain, dysmenorrhea, and infertility [2]. Hysterectomy is a radical treatment. However, the blood supply to the ovaries is affected after a hysterectomy, leading to perimenopausal symptoms [3]; most women do not want their uterus removed. Other treatment modalities include medications, minimally invasive procedures, and surgical procedures. Medications include oral contraceptives, danazol, progestins, levonorgestrel-releasing intrauterine device (LNG-IUD), gonadotropin-releasing hormone agonists (GnRH-a), etc. The most promising medical therapy was LNG-IUD. A study showed that LNG-IUD has a 68.8% success rate in decreasing the symptoms of dysmenorrhea [4]. Although medical treatments can relieve symptoms, the side effects are significant, and the symptoms still recur even after medical therapies are terminated [5]. Minimally invasive procedures include uterine artery embolization (UAE), and studies have shown that 75.7% of patients improved symptoms after UAE [6]. Possible side effects associated with UAE include postembolization syndrome, ovarian dysfunction, endometrial atrophy, and sepsis [7]. Surgical treatment includes excisional or cytoreductive surgery, endometrial ablation, and adenomyomectomy have proven effective in more than 50% of patients [8]. Surgical treatment is limited due to ill-defined endometrial-myometrial boundaries and may have complications such as uterine rupture [5,9,10].

Ultrasound-guided high-intensity focused ultrasound (USgHIFU) is a noninvasive therapy. The ultrasonic waves transmitted transcutaneously focus on the target area in vivo, which increases the tissue temperature above 60 °C and causes thermal ablation of the targeted tissue without damaging the surrounding tissue. Several studies have demonstrated the safety and efficacy of HIFU in treating adenomyosis [11,12]. The non-perfused volume ratio (NPVR) is the ratio of the non-perfused volume (NPV) of the diseased area measured on contrast-enhanced magnetic resonance images (MRI) to the total adenomyosis volume. Several studies have shown that the NPVR is highly correlated with the degree and duration of symptom relief [11,13,14]. There is a consensus that a higher NPVR is beneficial for a longer-term improvement in clinical symptoms [11,15–17]. With the continuous development of HIFU technology, it has become possible to achieve a higher NPVR. Several studies have shown that the average or median NPVR of USgHIFU ablation of adenomyosis has reached 50% to 70% [13,14,18]. Regarding MRI-guided HIFU treatment of adenomyosis, the NPVR has been reported to reach 80% or even 90% [19,20].

Adenomyosis is characterized by the invasion of endometrium and stroma into the myometrium [21]. The agreement is that a more than 12 mm in width junctional zone is strongly associated with the disease [5]. Zhou et al. found that satisfactory clinical outcomes can be achieved when the NPVR reaches 50% or above [13]. In a recent study, the average NPVR remained at about 50%, and the degree of dysmenorrhea and menstrual volume of the patients was significantly improved during the 18-month follow-up [22]. Therefore, in this study, NPVR ≥50% was used as the goal of technical success. The aim of this study was two-fold: (1) to investigate the safety of HIFU ablation for adenomyosis, and (2) to explore the factors affecting the technical success and the relative importance of its contribution to NPVR combined with multidimensional clinical data. These results will provide a more accurate NPVR prediction and the factors influencing NPVR could optimize patient screening for HIFU ablation.

2. Materials and methods

This retrospective study was approved by the ethics committee at our institution (IRB-2021006). The Chinese Clinical Trial Registry provided full approval for the study protocol, recruitment materials, and consent form (Registration No. ChiCTR2100050655).

2.1. Patients

A total of 299 patients from January 2018 to December 2020 in our hospital were enrolled in this retrospective study. All patients signed an informed consent before being enrolled in the study. The diagnosis of adenomyosis was suspected by clinical evaluation and ultrasonography, which was confirmed by MR imaging. Ultrasonography or MRI criteria for diagnosing adenomyosis included an enlarged uterus, asymmetric myometrial thickness, heterogeneous myometrial echotexture, poorly defined endomyometrial junction, and endomyometrial junctional zone measurement >12 mm on MRI [23].

The inclusion criteria were as follows: (1) Age ≥18 years old; (2) Patients had menorrhagia and/or dysmenorrhea; (3) Localized or diffuse thickening of the uterine wall, with a thickness of ≥30 mm as shown by MR imaging; (4) The MR scanning data before and after HIFU were complete, and the same parameters were used; (5) For patients with abdominal surgical scars, the width of image blurring due to acoustic attenuation had to be <10 mm through B-mode ultrasound detection.

The exclusion criteria were as follows: (1) Patients with endometriosis; (2) Patients with uterine fibroids; (3) Patients treated with sex hormone analogs, such as GnRHa or antagonistic hormone drugs within 3 months before HIFU treatment; (4) Patients with the levonorgestrel intrauterine device removed for less than 3 months; (5) Inconsistent MRI scanning parameters.

2.2. Magnetic resonance imaging evaluation

All patients underwent pre-HIFU and 1-day post-HIFU MRI examinations with a 3.0 T MRI system (GE Medical System, Milwaukee, WI, USA). T1-weighted fast spin-echo (FSE) images [repetition time/echo time (TR/TE), 270/2.1; slice thickness, 5 mm; slice interval, 1 mm], T2-weighted FSE images (TR/TE, 3600/105; slice thickness, 5 mm; slice interval, 1 mm), and dynamic contrast-enhanced T1-weighted liver acquisition with acceleration volume acquisition (LAVA) images (TR/TE, 4.2/2.0; slice thickness, 2.5 mm; slice interval, 0.5 mm) were obtained. An intravenous mass injection of 15–20 ml contrast agent gadodiamide (0.5 mmol/ml, Omniscan) was administered at 2 ml/s, and dynamic contrast-enhanced MRI was observed until 120 s after injection.

The digital imaging and communications in medicine (DICOM) format of the MRI information was imported into the MicroSea-HIFU treatment 3D image assistance system (Chongqing MicroSea Software Development Co., Ltd., Chongqing, China). The MR images were analyzed by a radiologist who had completed 5 years of specialization in abdominal MR imaging (Reader 1) and they were validated by another radiologist (Reader 2) who had 15 years of experience in abdominal MR imaging; in case of disagreement, Reader 2 retraced the image and this was considered as the final decision.

2.2.1. Measurement of data

The thickness of the abdominal wall, the distance from the anterior surface of the lesion to the skin, and the distance from the posterior surface of the lesion to the sacrococcygeal surface were analyzed and measured based on the MR T2WI sagittal images (Figure S1). According to the classification proposed by Kishi et al. adenomyosis was divided into the following four subtypes: type I (intrinsic), type II (extrinsic), type III (intramural), and type IV (indeterminate)[24]. The adenomyosis volume was measured slice-by-slice on preoperative T2WI and the NPV was measured slice-by-slice on the contrast-enhanced T1WI acquired immediately after the treatment. The NPVR (%) was defined as NPV divided by the percentage of adenomyosis volume. The energy efficiency factor (EEF) was defined as the ultrasound energy delivered for ablating 1 mm³ of tissue; and the equation was EEF = 0.7 × P × t/V(J/mm³), where P was the power (W), t was the sonication time (s), and V was the NPV [25].

2.2.2. Quantitative measurements on T2WI

All quantitative measurements of signal intensity (SI) values and standard deviation (SD) values were performed using the previously mentioned software. The largest three slices of adenomyosis on sagittal T2WI were selected to outline the region of interest (ROI) for automatically achieving the SI value and SD value of each slice [26]. The average SI value and SD value from the three slices were recorded, and the SD value indicated the signal homogeneity of adenomyosis [27]. SI values of the rectus abdominis muscle, myometrium, and endometrium were measured by the same method (Figure S2).

2.2.3. Classification based on dynamic contrast-enhanced MRI

The largest three slices of adenomyosis on dynamic contrast-enhanced MRI within 60 s after the injection of gadolinium were selected. The average SI and average SD values of adenomyosis and myometrium were recorded. By comparing the SI value of the adenomyosis lesion (SI_L) with the SI value of the myometrium (SI_M) and combining it with the SD value of the adenomyosis lesion (SD_L), adenomyosis was classified into the following three types: (1) slight enhancement: the SI value of the adenomyosis lesion was lower than that of the myometrium (SI_L < SI_M); (2) irregular enhancement: the SI value of the adenomyosis was equal to or higher than that of the myometrium and the SD value of the adenomyosis lesion was ≥150 (SI_L ≥ SI_M and SD_L ≥150); (3) regular enhancement: the SI value of the adenomyosis lesion was equal to or higher than that of the myometrium, and the SD value of the adenomyosis lesion was < 150 (SI_L ≥ SI_M and SD_L < 150) (Table S1) (Figure S3).

2.3. HIFU ablation procedure

The treatment was performed using the Focused Ultrasound Tumor Therapeutic System (Model-JC200, Chongqing Haifu Medical Technology Co. Ltd., China). The operating frequency of the US transducer was adjustable in the range of 0.5–1.5 MHz, and power was adjustable in the range of 0–400 W. All patients underwent bowel preparation and acoustic pathway skin preparation, degreasing, and degassing before HIFU treatment. Patients were placed in the prone position with the anterior abdominal wall in contact with the degassed water. An ultrasound imaging device (Esaote, MyLab 70, Italy) was used for real-time guidance during the procedure. The monitoring ultrasonography probe was integrated into the therapeutic transducer’s center, and the ultrasonic frequency was 3.5 MHz. Fentanyl (0.8–1 g/kg) and midazolam hydrochloride (0.02–0.03 mg/kg) were administered to maintain conscious sedation while reducing patient discomfort and movement. Patients were kept awake or in shallow sleep throughout the procedure. The treated adenomyosis lesions were at least 5 mm from the endometrium and the serosa layer. If the patient had burning pain over the abdominal skin, or lumbosacral, hip, or lower limb pain, sonication was stopped and the focal area was replaced to continue treatment. The ultrasonic dose and focal position were adjusted according to the patient’s tolerance. The gray-scale changes in the target area were displayed by monitoring ultrasound imaging, using massive grey-scale changes in the treated area as a measure of the ablation effect. The numerical rating scale (NRS) was used to assess the pain severity (0–10 points, 0: no pain, 10: maximum pain), and NRS scores were measured immediately after treatment. All patients were observed for 2 h before discharge or return to the ward. Complications were defined based on the classification standard by the Society of Interventional Radiology (SIR). In the SIR classification, classes C to F were regarded as major [28].

2.4. Statistical analysis

SPSS version 26.0 (IBM, Armonk, NY, USA) was used for statistical analysis. Normally distributed data were reported as the mean ± standard deviation (SD), and non-normally distributed data were reported as medians and interquartile range. Categorical data were expressed as numerals and percentages (%). The Chi-square test and Mann–Whitney U-test were used for univariate analysis. The Spearman correlation analysis was used for analyzing the correlation between two parameters. Binary logistic regression analysis was used for multivariate analysis. The cutoff values, area under the curve (AUC), sensitivity, and specificity for the prediction of achieving NPVR ≥ 50% were determined based on receiver operating characteristic (ROC) curve analysis. The optimal sensitivity and specificity cutoff point for the prediction model was established by maximizing the Youden index (sensitivity + specificity − 1). p < 0.05 was considered statistically significant.

3. Results

3.1. Patients

A total of 299 patients with an average age of 38.0 ± 5.8 (18.0–51.0) years old were enrolled. The median uterine volume was 236.8 (132.6) cm³, and the median adenomyosis volume was 73.6 (68.6) cm³. Among them, 220 patients (73.5%) had a history of childbirth, 238 patients (79.5%) had a history of abortion or miscarriage, and 112 patients (37.4%) had scars on the lower abdominal wall. Of the 112 patients with prior abdominal surgical scars, 86 had cesarean sections and 26 experienced other abdominal surgeries that did not involve the uterus.

3.2. Ultrasound ablation results

The average acoustic power was 388.4 ± 27.7 W (143.0-400.0 W). The median sonication time was 714.0 (778.0) seconds, the median treatment time was 71.0 (63.0) minutes, the median dose was 275.6 (315.1) kJ, the median EEF was 5.8 (8.5) J/mm³, and the median NPVR was 53.5% (34.7%). Technical success was achieved in 159 patients (53.2%) with NPVR ≥50% (Figure 1). The median NPVR in the slight enhancement, irregular enhancement, and regular enhancement types was 61.3% (30.0%), 48.1% (31.7%), and 43.9% (33.7%), respectively. The median NPVR for the slight enhancement type was higher than that for the irregular and regular enhancement types, and the difference was statistically significant (p = 0.006; p = 0.001). There was no statistically significant difference between the irregular and regular enhancement types (P > 0.05).

Figure 1. Sagittal T2-weighted images of adenomyosis and enhanced MR images with different non-perfused volume ratios (NPVR) after HIFU treatment. In a 38-year-old patient, the volume of the adenomyotic lesion was 61.2 cm³ (A1) and the non-perfused volume (NPV) was 45.3 cm³ (A2). The NPVR was 74.0%. In a 32-year-old patient, the volume of the adenomyotic lesion was 57.4 cm³ (B1) and the non-perfused volume (NPV) was 20.6 cm³(B2). The NPVR was 35.9%.

3.3. Safety of patients

All patients tolerated the HIFU procedure well; the median pain score was 3.0 (1.0) points during treatment. The median pain score was 3.0 (1.0) points in the NPVR <50% group and 2.0 (2.0) points in the NPVR ≥50% group, respectively, and there was a significant difference between the two groups (p < 0.05). During the procedure, adverse effects were reported in 82.6% (247/299) of the cases. They included lower limb radiation pain (24.1%), sacrococcygeal pain (48.8%), skin-burning pain (30.8%), treatment area pain (68.6%), groin pain (7.7%), and hip discomfort (6.4%). The incidences of sacrococcygeal pain and treatment area pain were 58.6% and 74.3% in the NPVR < 50% group and 40.3% and 63.5% in the NPVR ≥50% group, respectively, and there were significant differences between the two groups (p < 0.05). No significant difference was observed between the two groups in terms of additional adverse effects (p > 0.05). Through the correlation analysis, EEF was associated with the occurrence of intraprocedural sacrococcygeal pain (r = 0.215, p < 0.001) and treatment area pain (r = 0.148, p = 0.010). Distance from the dorsal side of adenomyosis to the sacrum was associated with intraprocedural sacrococcygeal pain (r= −0.150, p = 0.010) (Table 1).

Table 1. Comparison of intraprocedural adverse effects between NPVR ≥50% group and NPVR < 50% group.

Variable	%, n	NPVR <50% (n = 140)	NPVR ≥50% (n = 159)	p Value
Lower limb radiation pain	24.1 (72)	23.6 (33)	24.5 (39)	0.847
Sacrococcygeal pain^a	48.8 (146)	58.6 (82)	40.3 (64)	0.002
Skin burning pain	30.8 (92)	34.3 (48)	27.7 (44)	0.216
Treatment area pain^a	68.6 (205)	74.3 (104)	63.5 (101)	0.045
Groin pain	7.7 (23)	10.7 (15)	5.0 (8)	0.066
Discomfort in hip folds	6.4 (19)	5.7 (8)	6.9 (11)	0.670
Straining feeling in anus	1.7 (5)	2.1 (3)	1.3 (2)	0.552
Urethral pain	0.3 (1)	0.7 (1)	0	0.286
Vomiting	0.3 (1)	0	0.6 (1)	0.347

^aStatistically significant, p < 0.05.

No major complications occurred in patients after treatment. The main complication reported by 73.2% of patients was lower abdominal pain. Other adverse events included sacral tail or hip pain (19.4%), vaginal discharge (18.4%), numbness and pain in the lower limb (2.0%), fever (0.7%), nausea or vomiting (0.7%), and pain and distension of the anus (0.7%). In addition, one patient with an old surgical scar on the abdominal wall developed an orange peel appearance (2 mm × 3 mm) of the skin within the ultrasonic pathway. Skin lesions healed with dressing change. Another patient had dysuria and urinated successfully on the night of the HIFU treatment. A total of 65 patients (21.7%) experienced nominal therapy determined by their SIR grade B complications. All patients recovered within 7 days after treatment. The incidence of sacral tail or hip pain was 25.0% in the NPVR <50% group and 14.5% in the NPVR ≥50% group, and there was a significant difference between the two groups (p < 0.05) (Table 2).

Table 2. Comparison of postprocedural complications between NPVR ≥ 50% group and NPVR < 50% group.

Variable	%, n	NPVR <50% (n = 140)		NPVR ≥50% (n = 159)
		SIR grade A	SIR grade B	SIR grade A	SIR grade B
Fever	0.7 (2)	0	0.7 (1)	0	0.6 (1)
Lower abdominal pain	73.2 (219)	56.4 (79)	18.6 (26)	57.2 (90)	14.5 (24)
Sacral tail or hip pain^a	19.4 (58)	19.3 (27)	5.7 (8)	14.5 (23)	0
Numbness and pain in the lower limb	2.0 (6)	0.7 (1)	1.4 (2)	1.3 (2)	0.6 (1)
Vaginal discharge	18.4 (55)	17.1 (24)	0	19.5 (31)	0
Skin burn	0.3 (1)	0	0.7 (1)	0	0
Pain and distension of anus	0.7 (2)	0.7 (1)	0	0.6 (1)	0
Nausea or vomiting	0.7 (2)	0.7 (1)	0	0.6 (1)	0
Dysuria	0.3 (1)	0	0	0	0.6 (1)

^aStatistically significant, p < 0.05.

3.4. Comparison of the characteristics and ultrasound ablation results between the NPVR ≥50% group and NPVR <50% group

A total of 159 patients were classified in the NPVR ≥50.0% group, while 140 patients were classified in the NPVR <50.0% group. The characteristic data of patients were grouped into the following three dimensions of the data set: general information of patients, MRI evaluation, and laboratory test. Univariate analysis showed that history of childbirth, abdominal wall thickness, SI value of adenomyosis on T2WI, SI difference on T2WI between adenomyosis and rectus abdominis muscle, enhancement type on T1WI, leukocyte count and hemoglobin content were significant factors between the two groups (p < 0.05). The treatment time and sonication time in the NPVR <50.0% group were significantly longer than those in the NPVR ≥ 50% group (p < 0.05). The EEF in the NPVR <50.0% group was significantly higher than that in the NPVR ≥50% group (p < 0.05) (Table 3).

Table 3. Comparison of characteristics and ultrasound ablation results between two groups.

Variable	NPVR < 50% (n = 140)	NPVR ≥ 50% (n = 159)	p Value
General patient data
Age (years)	37.2 (9.0)	39.0 (6.0)	0.050
BMI (kg/m^2)	22.6 (3.0)	22.1 (3.7)	0.086
History of childbirth (yes/no) (n)^d	91/49	129/30	0.002
History of abortion or miscarriage (yes/no) (n)	106/34	132/27	0.118
Scar in the lower abdominal wall (yes/no) (n)	52/88	60/99	0.916
MR imaging
Volume of uterus (cm^3)	230.8 (118.7)	245.2 (141.0)	0.826
Volume of adenomyosis (cm^3)	75.1 (67.8)	70.8 (73.6)	0.391
Abdominal wall thickness (cm)^d	2.5 (1.4)	2.3 (1.1)	0.030
Adenomyosis ventral side to skin (cm)	5.4 (3.6)	4.9 (4.0)	0.629
Adenomyosis dorsal side to sacrum (cm)	1.4 (1.5)	1.3 (1.7)	0.315
Location of uterus (A/M/R)^a (n)	67/22/51	81/19/59	0.631
Location of adenomyosis (A/P/L/F)^b (n)	32/99/17/32	43/95/19/38	0.632
Type of adenomyosis (I/II/III/IV)^c (n)	47/31/11/51	45/33/19/62	0.550
SI value of adenomyosis on T2WI^d	259.8 (195.0)	231.7 (133.3)	0.011
SD value of adenomyosis on T2WI	53.9 (46.8)	46.7 (35.5)	0.107
SI difference between adenomyosis and myometrium	49.5 (96.3)	53.0 (83.3)	0.661
SI difference between adenomyosis and rectus abdominis muscle^d	114.3 (136.5)	95.7 (103.6)	0.023
SI difference between adenomyosis and endometrium	272.0 (191.8)	245.3 (183.7)	0.318
Enhancement type on T1WI (slight/ irregular/regular/) (n)^d	31/56/53	72/51/36/	<0.001
Laboratory test data
Leukocyte count (10^9/L)^d	5.8 (2.0)	5.4 (1.5)	0.024
Red blood cell count (10^12/L)	4.4 (0.5)	4.4 (0.5)	0.683
Hemoglobins (g/L)^d	123.0 (23.0)	118.1 (24.0)	0.046
Platelet count (10^9/L)	255.6 (90.0)	255.0 (86.0)	0.433
Ultrasound ablation results
Power (W)	400.0 (17.5)	400.0 (2.0)	0.162
Treatment time (min)^d	85.0 (69.8)	64.0 (51.0)	<0.001
Sonication time (s)	870.0 (899.0)	666.0 (702.0)	0.037
Dose (kJ)	322.5 (378.6)	244.8 (295.0)	0.113
EEF (J/mm^3)^d	9.4 (12.8)	3.6 (4.8)	<0.001

^aLocation of uterus: A, anteverted; M, mid position; R, retroverted.

^bLocation of adenomyosis: A, Anterior Wall; P, Posterior Wall; L, Lateral Wall; F, fundus of uterus.

^cType of adenomyosis: I, intrinsic adenomyosis; II, extrinsic adenomyosis; III, intramural adenomyosis; IV, indeterminate adenomyosis.

^dStatistically significant, p < 0.05.

Data are median value; interquartile range in brackets.

3.5. Evaluation of the factors affecting NPVR ≥50%

In Table 4, NPVR ≥50% was set as a dependent variable and no childbirth history was as the control group designated as the categorical variable. The multivariate logistic regression analysis showed that abdominal wall thickness, SI difference on T2WI between adenomyosis and rectus abdominis, and enhancement type on T1WI were all protective factors for NPVR ≥50% (p < 0.05), while the history of childbirth was an independent risk factor (p < 0.001). Patients with a history of childbirth were more likely to achieve NPVR ≥50% than those without a history of childbirth. The possibility of NPVR ≥50% was higher in patients with thinner abdominal walls, showed slight enhancement of adenomyosis on T1WI, or in whom the SI difference on T2WI between adenomyosis and rectus abdominis was more minor. Compared to the absolute value of standardized coefficients, the most important factor affecting NPVR ≥50% was the enhancement type on T1WI, followed by a history of childbirth, SI difference on T2WI between adenomyosis and rectus abdominis, and abdominal wall thickness (Figure 2).

Figure 2. Comparison of the degree among the factors influencing the non-perfused volume ratio (NPVR) ≥50%. AWT: Abdominal wall thickness; T1WI: Enhancement type on T1WI; T2WI: SI difference on T2WI between adenomyosis and rectus abdominis.

Table 4. The binary logistic regression analysis of variance.

Variable	β	S.E.	Wald	OR (95% CI)	p Value
(Intercept)	3.352	1.028	10.635	28.554	0.001
History of childbirth	1.054	0.300	12.314	2.868 (1.592, 5.167)	<0.001
Abdominal wall thickness	−0.368	0.163	5.117	0.692 (0.503, 0.952)	0.024
Enhancement type on T1WI	−0.731	0.163	20.145	0.482 (0.350, 0.663)	<0.001
SI difference on T2WI between adenomyosis and rectus abdominis	−0.004	0.001	7.184	0.996 (0.993, 0.999)	0.007

SE: standard error; OR: odd ratio; CI: confidence interval.

3.6. ROC curves of the prediction model

Logistic regression analysis was used to model the prediction of an NPVR ≥50%: $\mathit{Logit}\mathrm{P}1=3.352+1.054\mathrm{*}{\mathrm{X}}_1-0.368\mathrm{*}{\mathrm{X}}_2-0.731\mathrm{*}{\mathrm{X}}_3-0.004\mathrm{*}{\mathrm{X}}_4$ [Dependent variable: NPVR ≥ 50% equal to 1, NPVR < 50% equal to 0; Covariates: History of childbirth = X₁ (with history of childbirth = 1, without a history of childbirth =0), Abdominal wall thickness = X₂, enhancement type on T1WI = X₃ (slight enhancement = 1, irregular enhancement = 2, regular enhancement = 3), SI difference on T2WI between adenomyosis and rectus abdominis muscle = X₄, P1: incidence of NPVR ≥ 50%]. In the ROC curve analysis performed in this study, the AUC of this prediction model was 0.720 (95% CI, 0.662–0.778). A cutoff value of 0.387 yielded a sensitivity of 0.730 and a specificity of 0.657 (Figure 3).

Figure 3. The receiver operating characteristic (ROC) curves of the prediction model in predicting the non-perfused volume ratio (NPVR) ≥ 50%.

4. Discussion

The safety and efficacy of USgHIFU in treating uterine fibroids and adenomyosis have been confirmed in many studies [29–33]. In HIFU ablation of uterine fibroids, the average or median NPVR is up to 80%; sometimes, it is even more than 90% [34–36]. Adenomyosis does not have the characteristic pseudocapsule of uterine fibroids and the energy cannot be diffused homogeneously to form coagulative necrosis with regular boundary [37]. The normal uterine wall thickness in postpuberty women is 10–15mm [38]. To preserve sufficient uterine wall thickness and prevent unnecessary thermal damage to adjacent tissues, protecting the patient’s endometrium and serosa layer from ablation for at least 5 mm is necessary. This ablation protocol can preserve the uterine wall thickness of 10–15 mm after the coagulative necrotic tissue is absorbed. The four subtypes proposed by Kishi et al. based on the distribution and location of adenomyosis on MRI could guide the selection of treatment modalities [24]. Intrinsic adenomyosis is closer to the endometrium; hence, there are residual lesions on the endometrium side; extrinsic adenomyosis is closer to the serosa. Thus, there are residual lesions on the serosa side; intramural adenomyosis is easier to treat than other types; for indeterminate adenomyosis that invades the entire layer of the uterus, a safe distance should be maintained from the endometrium and serosa layer [22,37]. Therefore, the ablation protocol for adenomyosis involves volume reduction rather than complete conformal ablation of lesions. The ablation goals and results of adenomyosis cannot be the same as those of uterine fibroids. This study selected an ablation protocol to protect the endometrium and seromuscular layer, and 159 cases (53.2%) obtained NPVR ≥50% of adenomyosis.

In this study, all patients completed HIFU treatment safely and effectively. Conscious sedation analgesia was used for intraoperative response control; however, the feeling of pain or discomfort for patients with adenomyosis is relatively obvious. The possible reason for this occurrence is that patients with adenomyosis have excessive uterine prostaglandin secretion, particularly that of PGF2a and PGF2, thus resulting in increased uterine tone and high-amplitude contractions [39,40]. Patients with NPVR <50% reported a higher incidence of pain in the sacrococcygeal region and treatment area during the procedure. Our analysis showed that EEF and distance from the dorsal side of adenomyosis to the sacrum were associated with sacrococcygeal pain. The EEF of the NPVR＜50% group was significantly higher than that of the NPVR ≥ 50% group. A previous study has demonstrated a positive linear relationship between the EEF and the distance from the ventral side of the adenomyosis to the skin, enhancement type on T1WI, SI on T2WI, and abdominal wall thickness [41]. In this study, there were statistical differences in the abdominal wall thickness, the SI value of adenomyosis on T2WI, the SI difference between adenomyosis and rectus abdominis, and the enhancement type on T1WI between the two groups. The quantification of SI values validates the results of previous grading comparisons. Fan et al. found the distance from the fibroid’s ventral side to the skin, enhancement type on T1WI, and SI on T2WI all had a line relationship with EEF [42]. Fibroids manifest as hyperintense on T2WI are mainly composed of cellular components. Cellular fibroids are copiously supplied by vessels, so the energy could not be easily deposited and ablation is difficult [43]. Adenomyosis tissue is quite different from uterine fibroids, and the hyperintense on T2WI may be due to the presence of ectopic endometrial tissue, endometrial cyst, or hemorrhage on histology. Repeated bleeding of the lesion during menstruation may lead to changes in the tissue structure, so the acoustic environment of the lesion tissue is more complex, resulting in a decrease in the sensitivity of the tissue to ultrasound and difficulty in energy deposition at the focal point. Therefore, we speculated that more energy inputting and poorer local energy deposition induce a higher incidence of pain caused by posterior field energy.

There were no major adverse events related to the treatment in the two groups after HIFU ablation. The incidence of sacral tail or hip pain in the NPVR <50% group was higher than that in the NPVR ≥50% group, and the difference was statistically significant. This suggests that NPVR ≥ 50% did not increase the intraoperative and postoperative adverse reactions compared with NPVR <50%. Although 112 patients had scarring in the lower abdomen, these patients were also treated safely and efficiently. Xiong et al. demonstrated that patients with prior abdominal surgical scars have a higher risk of developing skin burns than those without abdominal scars [44]. In this study, only one patient experienced skin burns after treatment and had a prior abdominal surgical scar. After the active dressing change, the patient’s skin lesions healed. Therefore, patients with adenomyosis and abdominal scar width <10 mm can be safely treated with USgHIFU.

Gong et al. used linear regression to identify the following factors influencing the NPVR of HIFU ablation of adenomyosis: the enhancement type on T1WI, SI on T2WI, the volume of adenomyosis, location of the uterus and adenomyosis, number of hyperintense points, abdominal wall thickness, and distance from the skin to the ventral side of the adenomyotic lesion [41]. Keserci et al. used MRI features to predict the effect of HIFU ablation on adenomyosis. It was found that, in addition to the thickness of the subcutaneous fat layer and adenomyosis volume, the T2 SI ratio of adenomyosis to the myometrium and the Ktrans ratio of adenomyosis to the myometrium could be used to predict whether the NPVR could reach 90%[19]. In this study, MRI signal values were quantified. The multivariate logistic regression study found that among the 23 variables in three dimensions, the four most important independent factors influencing NPVR in order of importance were enhanced type on T1WI, history of childbirth, SI difference on T2WI between adenomyosis and rectus abdominis, and abdominal wall thickness. The ROC curve analysis confirmed that the logistic regression prediction model could predict the possibility of NPVR ≥50%.

Dynamic enhanced MRI is performed using a T1-weighted gradient echo sequence after injection of 0.5 mmol/ml gadolinium. The arterial vessels within adenomyosis are visible within 10–30 s after gadolinium injection, followed by the myometrial vessels, and the arterial vessels disappear after 30–60 s. Therefore, the enhanced MRI can indicate the perfusion and permeability of the adenomyotic lesion within 60 s after gadolinium injection [45,46]. Keserci et al. introduced a classification method based on MRI T1 perfusion-based time–SI curves for the adenomyosis tissue compared with those for the myometrium [20]. They found that the mean NPVR was significantly higher in group A (time–SI curve of adenomyosis lower than that of the myometrium) than in group B (time–SI curve of adenomyosis equal to or higher than that of the myometrium). Based on the classification of Keserci et al. we proposed a quantitative classification method using MR values that could further categorize the patients with adenomyosis SI equal to or higher than that of the myometrium into the irregular enhancement type and the regular enhancement type. The median NPVR for the slight enhancement type was significantly higher than that for the irregular and regular enhancement types. The above results were consistent with the reports by Keserci et al. Zhou et al. found that patients with NPVR ≥50.0% had better treatment outcomes in terms of reduction of symptoms [13]. Therefore, it can be speculated that adenomyosis with irregular and regular enhancement may be associated with a lower treatment efficiency compared with adenomyosis with slight enhancement. In contrast, adenomyosis with slight enhancement is easily ablated by HIFU as it has reduced blood supply. Therefore, adenomyosis with rich blood flow has a great influence on the results of HIFU ablation.

In this study, patients with a history of childbirth were more likely to achieve technical success than those without a history of childbirth. Transient uterine ischemia that occurs after childbirth to avoid excessive blood loss after placental separation could directly affect the survival of adenomyosis. Uterine involution after childbirth involves the mechanisms of apoptosis and proliferation, with the ultimate effect of remodeling the existing tissues [47,48]. Uterine remodeling in patients with a history of childbirth may lead to increased fibrosis, which is conducive to ultrasonic energy deposition.

Previous studies suggested a relationship between the SI of adenomyosis on T2WI and the therapeutic efficacy [19,49]. In this study, multivariate logistic regression found that the NPVR was negatively correlated with the SI difference between adenomyosis and rectus abdominis; the greater the difference, the lower the NPVR. Compared with the rectus abdominis, the higher the SI of the adenomyosis lesion, the worse the ablation efficiency. Pathological research indicates that high-intensity adenomyosis consists of more focal cystic dilatation of the fluid-filled glands [50]. These tissue structures can affect energy deposition, resulting in reduced ablation efficiency.

In this study, the abdominal wall thickness was also one of the factors affecting the NPVR. Subcutaneous fat, rectus abdominis, and sheath are the main structures constituting the abdominal wall of the acoustic pathway; these findings are consistent with those of previous studies [19,41]. Fat tissue can absorb and deposit ultrasonic energy, resulting in energy deficiency in the target tissue. The conversion of ultrasonic energy into thermal energy is the highest efficiency at the fat/muscle interface. Studies have shown that increased abdominal wall thickness and BMI generate more thermal energy in the ultrasonic pathway and are more prone to thermal injury [51]. In this study, there was no statistically significant difference in BMI between the two groups. It may be that most patients displayed abdominal obesity; the thicker the subcutaneous fat, the greater the abdominal wall thickness.

Liu et al. found that the patient group with an age of 40 years had a higher chance of clinical success rate than the patient group with an age <40 years [14]. The reason for this occurrence may be attributed to the higher level of estrogen in younger women, which can lead to a stronger invasion of the lesions. In this study, although the median age of the NPVR ≥50% group was higher than that of the NPVR <50% group, the difference was not statistically significant (p = 0.05). One possible reason for this phenomenon was the relatively small sample size.

The advantage of this study is the quantitative acquisition of MR values to determine the SI of tissues on T2WI and the type of dynamic contrast-enhanced MRI. This method is more accurate and objective than a visual judgment based on experience. The limitation of this study is that the data were obtained from a single center; thus, multi-center and large sample size studies are required to support the conclusion. In addition, the thickness of the abdominal wall, the distance from the anterior surface of the lesion to the skin, and the distance from the posterior surface of the lesion to the sacrococcygeal surface were measured based on the preoperative MR T2WI sagittal images, and these measurements may change on the day of treatment. This is another limitation of this study.

5. Conclusion

Compared with NPVR <50%, NPVR ≥50% did not increase the intraoperative and postoperative adverse reactions. Enhancement type on T1WI, history of childbirth, SI difference on T2WI between adenomyosis and rectus abdominis, and abdominal wall thickness can be used as factors affecting the technical success.

Disclosure statement

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

Data availability statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to their containing information that could compromise the privacy of research participants.

Additional information

Funding

This study was supported by the Natural Science Foundation of Chongqing [grant number cstc2021jcyj-msxmX0514] and the Young Project of Science and Technology Research Program of Chongqing Education Commission of China [No. KJQN202200478].