Ultrasound-guided percutaneous microwave ablation for adenomyosis with abnormal uterine bleeding: clinical outcome and associated factors

Abstract

Objective

To investigate the factors affecting the efficacy of ultrasound (US)-guided percutaneous microwave ablation (PMWA) for adenomyosis with abnormal uterine bleeding (AUB-A).

Methods

Baseline data of patients with AUB-A who underwent US-guided PMWA treatment between October 2020 and October 2021, including demography characteristics, laboratory and imaging examination results were retrospectively analyzed. 3D reconstruction of magnetic resonance imaging (MRI) was applied to quantitatively assess the local treatment responses, including ratio of non-perfusion volume to adenomyosis volume (NPVr), ablation rate of the endometrial-myometrial junction (EMJ), and surface area (SA) of the ablated part of the EMJ. Patients were followed up at 3, 6, and 12 months after treatment, and divided into two groups: group with complete relief (CR), and group with partial relief (PR) or no relief (NR). Data were compared between them.

Results

Thirty-one patients were analyzed with a mean age of 38.7 ± 6.8 years (range: 24–48): 48.4% (15/31), 63.3% (19/30), and 65.5% (19/29) achieved CR at 3, 6, and 12 months, respectively. In univariate analysis, compared with the PR/NR group, serum CA125 levels were significantly lower in CR group at 3 months, while ablation rates of EMJ and SA of the ablated part of the EMJ were significantly higher at the three time points. Other baseline characteristics and NPVr did not differ between the two groups.

Conclusion

Baseline CA125 and ablation rate of the EMJ and SA of the ablated part of the EMJ are associated with the outcome of AUB-A patients after US-guided PMWA treatment.

KEYSWORDS:

• Adenomyosis
• abnormal uterine bleeding
• ultrasound-guided microwave ablation

Introduction

Abnormal uterine bleeding (AUB) is a burdensome and common gynecological condition with a prevalence of up to 30% in reproductive-aged women [1,2]. It is defined as any bleeding pattern that differs in frequency, duration, and amount from the pattern observed in normal menstrual cycles [3]. The main presentations include heavy menstrual bleeding (HMB), prolonged menses (menstrual length > 7 days), frequent menses (bleeding occurs more often than every 21 days), irregular menses (cycle intervals ≥7 days), and intermenstrual bleeding (IMB). Adenomyosis is a common structural cause of chronic AUB [3,4] according to the PALM-COEIN classification system [5]. It is usually diagnosed by imaging examination and thereafter confirmed by histopathology.

The current management of AUB-adenomyosis (A) mainly comprises medical therapy, surgery, endometrial ablation (EA), and other interventions [3,5–7]. Drug treatment can alleviate symptoms in most patients and is considered the first-line treatment worldwide. However, its effectiveness is usually limited because of side effects, poor adherence, and symptomatic recurrence after drug withdrawal [8]. Hysterectomy is the only radical treatment, but it is not suitable for patients wishing to preserve their uterus [7]. EA can treat HMB by directly damaging the endometrium. However, it cannot be used in women desiring future fertility [7,9]. Additionally, its long-term efficacy is reduced in patients with deep adenomyosis due to the limited penetration of the generated heat [10,11]. Therefore, it is usually not included in the potential treatment options for patients with advanced adenomyosis. Other conservative interventions include uterine artery embolization (UAE), high-intensity focused ultrasound (HIFU), laparoscopy or US-guided radiofrequency ablation, and microwave ablation [12,13]. They have considerably similar technological mechanisms that involve destroying the adenomyotic lesion either directly (HIFU and in-situ thermal ablation) or indirectly (UAE), leading to tissue necrosis and volumetric shrinkage of the uterus and lesion [13,14].

Among several thermal ablation methods, US-guided percutaneous microwave ablation (PMWA) for adenomyosis has gained increasing attention in recent years due to better efficacy, higher efficiency and fewer complications [13,15–17]. Compared to UAE, it has the advantage of less invasiveness, lower incidence of severe complications and no radiation exposure [18]. Its effectiveness and safety have been confirmed in several studies [15,19–22] with a significant percentage of patients showing an improvement in HMB [16,17]. However, no study has investigated the factors influencing the outcomes of patients with AUB-A after PMWA. Therefore, this retrospective study aimed to examine these factors in order to optimize the ablation strategy and ensure better outcomes in patients with AUB-A.

Methods

Participants

From October 2020 to October 2021, 38 adenomyosis patients with AUB who underwent US-guided PMWA and both pre- and post-ablation contrast-enhanced magnetic resonance imaging (CE-MRI) were enrolled retrospectively. This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee of Shanghai Tenth People’s Hospital (No.22K32). All patients were informed of the benefits and potential risks of PMWA and other alternative therapies such as intervention methods (HIFU and UAE) or surgery prior to treatment. Among them, seven had a history of HIFU, two of UAE, and one of trans-cervical RFA. Written informed consent for both the ablation procedure and this retrospective study was obtained from all patients.

Inclusion criteria: (1) age > 18 years with AUB symptoms; (2) uterine adenomyosis confirmed by both US and MRI; (3) patients who wanted to preserve the uterus and received PMWA treatment after giving informed consent; and (4) patients who underwent both pre- and post-ablation MRI performed in our hospital. Exclusion criteria: (1) concomitant submucosal leiomyoma (n = 3); (2) levonorgestrel-releasing intrauterine device (n = 2); (3) lost to follow-up after ablation (n = 1); and (4) long-term continuous oral contraceptive treatment within one year after PMWA, and thus unable to evaluate the clinical response (n = 1). Finally, 31 patients were included in the study.

Baseline pretreatment evaluation

A standard medical history was obtained from all participants before treatment to collect baseline information, including age, body mass index (BMI), main complaints, menstrual history, pregnancy and delivery history, medical history, and previous treatment. AUB symptoms was evaluated by a senior US radiologist who focused on gynecological intervention. Several clinical scoring methods were used to quantitatively evaluate the severity of adenomyosis-related symptoms. A pictorial blood loss assessment chart (PBAC) was used to quantify the menstrual blood loss volume, and HMB was defined as a PBAC score > 100. Visual analog scale (VAS) scores ranging from 0 to 10 were used to quantify the severity of menstrual pain. Zero indicates no pain at all, while 10 indicates intolerable pain, similar to labor pain. Dysmenorrhea was defined as continuous pain with a VAS score of > 4. The symptom severity score (SSS) and health-related quality of life (HRQOL) score were used to quantify the severity of the menstrual disorder and quality of life.

All patients underwent routine pre-ablation examinations to rule out contraindications and other potential causes of AUB, including laboratory tests, chest radiography, electrocardiography, echocardiography, contrast-enhanced MRI, and pelvic US. Both MRI and US were performed 2 days before treatment, and the MRI results were recorded based on Kishi’s study (Figure 1) [22]. A 1.5-T MR scanner with a gradient strength of 40 mT/m and gradient slew rate of 200 mT/ms (Magnetom Verio, Siemens, Germany) was used. T2 weighted images (T2WI) with a resolution of 1056 × 1056 and slice distance of 3–5 mm were obtained. Gadopentetate dimeglumine (Magnevist, Bayer Schering, Germany) was used as an enhanced contrast agent to acquire contrast-enhanced T1-weighted images (CE-T1WI). Pelvic US examination was performed using an US scanner equipped with a 1–6 MHz convex array probe and 3–12 MHz end-fire intracavity probe (LOGIQ E9 system; GE Healthcare, Milwaukee, WI, USA). As the margin of adenomyosis is usually undefined on B-mode US imaging, contrast-enhanced ultrasound imaging (CEUS) was routinely performed before PMWA to better visualize the extent of involvement. A bolus injection of 2 ml Sonovue^TM (Bracco Suisse SA, Geneva, Switzerland) in the unilateral anterior elbow vein, followed by a flush with 5 ml of normal saline, was used.

Figure 1. Schematic drawings of MRI classification of adenomyosis (upper) and the corresponding examples (lower). (A) Type I, intrinsic focal adenomyosis (arrows); (B) type II, extrinsic focal adenomyosis (arrows); (C) type III, intramural focal adenomyosis (arrows); (D) type IVa, diffuse adenomyosis with unilateral myometrium involved (arrows); (E) type IVb, diffuse adenomyosis with bilateral myometrium involved.

US-guided PMWA procedure and follow-up

A monopolar water-cooling MWA applicator (Microwave Ablation system MTI-5A; Nanjing Great Wall Medical Equipment Co. Ltd, Nanjing, China) equipped with a 14-gauge, 18-cm long monopolar MWA antenna (XR-A2018W; Nanjing Great Wall Medical Equipment Co. Ltd, Nanjing, China) with a 1 cm active tip was used to perform PMWA. The output power of the MWA generator was 50–60 w. The PMWA procedure was performed according to the recommendations of the Chinese expert consensus on the clinical application of PMWA for adenomyosis [23]. The main steps included pre-ablation CEUS, anesthesia, disinfection with iodine, US-guided core needle biopsy, PMWA, and intraoperative CEUS. Free-hand biopsies were routinely performed with an 18-gauge automatic biopsy core needle (Bard Magnum Biopsy Instrument; Covington, GA, USA) to rule out potential malignancy. The puncture site was selected in an area with typical US findings, avoiding large blood vessels. Subsequently, PMWA was performed under US guidance. Among eight patients who had large lesions (>129 ml), two were treated with two microwave antennas. The remaining patients were treated with one antenna using a moving-shot technique. Under real-time transabdominal US monitoring, the adenomyotic lesion was ablated from deep to shallow, point by point along the puncture paths. The procedure was stopped when the gasification reaction of hyperechoic cloud covered the entire adenomyotic lesion and reached 5 mm depth from the uterine serosa. Five minutes after ablation, CEUS was performed to quickly evaluate the local treatment response, and supplementary ablation was performed when needed.

Twenty-seven patients returned to our clinic for reevaluation and four patients who failed to return were followed up via phone or the WeChat platform. Based on the results of the pre-ablation evaluation results, the symptom that had the greatest impact on the quality of life was determined, and the outcome after PMWA treatment was evaluated based on the degree of improvement. For example, for patients with HMB as the main manifestation, complete relief (CR) was defined as the complete correction of anemia and PBAC score ≤ 100; partial relief (PR) was defined as an improvement in PBAC score > 20% or hemoglobin (Hb) level > 20 g/L but both of them had not returned to normal levels; and no relief (NR) was defined as no significant change in the PBAC score (< 20%) or Hb level (< 20 g/L) compared to the baseline. For patients with prolonged menses as the main manifestation and a PBAC score ≤ 100, CR was defined as duration of the menstruation ≤ 7 days after treatment, and NR was defined as duration of menstruation > 7 days.

3D Quantitative assessment of uterine anatomy and local treatment responses

The raw data of pre- and post-ablation MRI in DI-COM format were retrospectively retrieved from the imaging data system. 3D MRI ablation planning and evaluation software (Demetics Medical Technology Co. Ltd, Hangzhou, China) was used to perform 3D reconstruction of the uterine corpus, endometrium, adenomyotic lesion, and non-perfused area (Figure 2) as following:

Figure 2. Diagrams of the of 3D reconstruction of MRI. With the adenomyotic lesion (arrows) identified clearly on T2WI (A), the segmentation could be accomplished by marking the target in red and the surrounding structures in blue (D) on 3 ∼ 5 different planes. By activating the reconstruction function, the system could identify the outline of the target automatically (B,C) by analyze the signal intensity and distribution of the marked area. After all the targets were reconstructed, a vivid 3D map reflecting the spacial relationship between the adenomyotic lesion (pink) and endometrium (yellow), uterine corpus (purple), and bladder (red) could be generated (E). The system will automatically identify the numbers of pixels inside and the reconstructed structure, and calculate the volume and surface area measurement values (F).

Administration

To import the raw MRI examination data in DI-COM format in the system and determine the proper sequence for further analysis (Figure 2A).

Segmentation

To select three to five sections at different levels, mark the target area in red and non-target areas in blue, and initiate the ‘segmentation’ process (Figure 2D).

Render and correction

The system renders all pixels with similar signal intensity (SI) inside the drawn region of interest (ROI) and automatically trace the contour of the target (Figure 2B,C). The operator must check the automatic trace of the target in each plane, correct trace errors with trim tools if needed, and then execute the ‘correction’ process untill the target contour tracing is correct in every plane. A vivid 3D map was then generated for direct visualization of the location and size of the adenomyotic lesion and spatial relationship between the lesion and surrounding structures (Figure 2E).

The mechanism of the automatic measurement function (volume and surface area) of this software is to identify the length and width of each pixel spacing and slice thickness (height) of each possessed plane and calculate the volume of a single pixel using the formula: V (single pixel) = Length × Width × Height. Once the target contour was confirmed, the number of pixels in the ROI was automatically recognized. The volume of the target can then be calculated using the formula V (target) = V (single pixel) × number. The surface area automatic calculation function was implemented based on the built-in visualization toolkit (VTK) plugin [24]. The contour extraction algorithm in the surface rendering module was used to process the DICOM and Communications in Medicine data to generate a network model. Then, the stripper module was used to segment the iso-surface of this network into many triangular patches. Finally, the surface area of the three-dimensional structure was obtained by summing the areas of all triangular patches.

After 3D reconstruction of MRI before and after treatment, the baseline uterine anatomy parameters were measured automatically, including the volume of the uterine corpus, endometrium, and adenomyotic lesion (Figure 2B,E). The ratio of nonperfusion volume to the volume of the adenomyotic lesion (NPVr) was assessed, which indicates the local response of the adenomyotic lesion to MWA. Using the surface area measurement tool, the baseline surface area (SA) of the endometrial-myometrial junction (EMJ) was also assessed (Figure 2C,F). The local response of the EMJ was evaluated quantitatively, including the area and proportion of the ablated part of the EMJ. The CE-T1WI sequence was selected for the reconstruction of the ablated part in type II adenomyosis (Figure 3E) and type IV adenomyosis with unilateral myometrium involvement (Figure 4E). T2WI was selected for reconstruction of the ablated EMJ in type IV adenomyosis with bilateral myometrium involvement (Figure 5D). For patients with type II adenomyosis who had EMJ with intact iso-enhancement on post-ablation CE-T1WI (Figure 3E) indicating no damage, no overlap was observed between the endometrium and non-perfused area on the 3D map (Figure 3F). In patients with type IV adenomyosis, the ablated part of the EMJ showed non-enhancement (Figure 4E and Figure 5E), and could be visualized (Figure 4F and Figure 5F) as an overlapping area between the endometrium and non-perfused area in the 3D map (arrows). The quantification results of the area and proportion of the ablated part of the EMJ were obtained using the surface area measurement tool.

Figure 3. 3D reconstruction of MRI before and after ablation in a case of Extrinsic adenomyosis located at the posterior wall (type II). A 37 year-old woman who presented with HMB, had focal adenomyosis located at the posterior wall (type II) (A). 3D reconstruction of pre-ablation MRI showed that the volume of uterine corpus, adenomyotic lesion and endometrium was 209.74 ml, 38.8ml, and 8.76 ml respectively; the baseline ISA of EMJ was 34.1 cm² (B,C). Post-ablation MRI showed obvious edema around the ablation zone (D), without any damage to the ipsilateral EMJ (E). 3D map after treatment showed no overlap between ablation zone and EMJ (arrows) (F). The NPVr reached 98.3%, the ablation rate of EMJ was 0. During the follow-up, this patient had PR at 3 months, NR at 6 and 12 months after treatment.

Figure 4. 3D reconstruction of MRI before and after ablation in a case of diffuse adenomyosis with unilateral myometrium involved (type IV). A 44 year-old woman who presented with HMB, had diffuse adenomyosis with the posterior myometrium involved (type IV) (A). 3D reconstruction of pre-ablation MRI showed that the volume of uterine corpus, adenomyotic lesion and endometrium was 180.2 ml, 44.5 ml, and 7.9 ml, and the baseline ISA of EMJ was 36 cm² (B,C). Post-ablation MRI showed obvious decrease of the SI of the endometrium (D), and no perfusion was observed on the ipsilateral EMJ (arrows) (E). 3D map after treatment showed slight overlap between ablation zone and EMJ (arrows) (F). The NPVr reached 85.7%, and the ablation rate of EMJ was 14.6%. During the follow-up, this patient had PR at 3 months, NR at 6 and 12 months after treatment.

Figure 5. 3D reconstruction of MRI before and after ablation in a case of diffuse adenomyosis with bilateral myometrium involved (type IV). A 44 year-old woman who presented with HMB and frequent menses, had diffuse adenomyosis (arrows) with bilateral myometrium involved (type IV) (A). 3D reconstruction of pre-ablation MRI showed that the volume of uterine corpus, adenomyotic lesion and endometrium was obtained as 187.5 ml, 76.5 ml, and 4 ml respectively, and the baseline ISA of EMJ was 28.9 cm² (B,C). Post-ablation MRI showed obvious decrease in the SI of the endometrium (D), and no perfusion was observed on the ipsilateral EMJ (arrows) (E). 3D reconstruction map after treatment showed that more than half of the endometrium was wrapped by the ablation zone (arrows) (D & F). The NPVr reached 93.6 %, and the ablation rate of EMJ was 84.4 %. During the follow-up, this patient had CR of AUB at 3, 6 and 12 months after treatment.

Statistical analysis

All statistical analyses were performed using SPSS software (version 25, IBM, Chicago, USA). All continuous data were subjected to the Shapiro-Wilk normality test. Data with a normal distribution are expressed as means ± standard deviation, while data with a skewed distribution are expressed as medians (P25-P75). The Wilcoxon signed-rank test was used to compare continuous data with a skewed distribution. Student’s t-test was used to analyze continuous variables with a normal distribution. A two-tailed significance test result was adopted for the comparison analysis, and statistical significance was defined as a p-value < 0.05.

Results

Patient characteristics

The baseline characteristics of all participants are summarized in Table 1. The mean age was 38.7 ± 6.8 years (range: 24–48) and mean BMI was 22.5 ± 3 kg/m² (range: 17.5–28.7 kg/m²); mean Hb level was 112.8 ± 19.3 g/L (range: 72–150 g/L), and the median CA125 level was 88.5 U/ml (range: 18.2–1257 U/ml). After 3D reconstruction of MRI, the distribution of the quantitative parameters was as follows: the median baseline uterine volume was 225.9 ml (range: 86.9–768.8 ml), median adenomyosis volume was 68.4 ml (range: 43.1–129 ml), mean endometrial volume was 7.1 ± 3.2 ml (range: 2.4–14.6 ml) and mean SA of EMJ was 36 ± 11 cm² (range: 19.5–64.5 cm²). Regarding the MRI classification, 26 patients had type IV adenomyosis, five had type II adenomyosis, and none had type I or type III adenomyosis.

Table 1. Baseline characteristics of the study population, means ± SD or median (P25-P75).

Variables	Value
Age	38.7 ± 6.8
BMI (kg/m^2)	22.5 ± 3
Gravidity	2 (1–4)
Parity	1 (1–1)
Menarche age	14 (13–15)
CA125 (U/ml)	88.5 (45.8–150)
Hb (g/L)	112.8 ± 19.3
Uterine volume (ml)	225.9 (151–361.3)
Adenomyosis volume (ml)	66 (44.5–143.4)
Endometrial volume (ml)	7.1 ± 3.2
Inner surface area of EMJ (cm^2)	36 ± 11
MRI classification (I/II/III/IV)	0/5/0/26
Symptoms of AUB
HMB	29/31 (93.5%)
Prolonged menses	10/31 (32.3%)
Frequent menses	2/31 (6.5%)
Irregular menses	1/31 (3.2%)
IMB	0
Symptom scores
SSS	37.8 ± 20.5
HRQOL	52.8 ± 20.4
PBAC	183 (154–277)
VAS	8 (7–9)

Note: BMI: body mass index; Hb: hemoglobin; EMJ: endometrial-myometrial junction; HMB: heavy menstrual bleeding; IMB: intermenstrual bleeding; SSS: symptom severity score; HRQOL: heath related quality of life; PBAC: pictogram blood-loss assessment chart; VAS: visual analog scale.

The mean values of SSS and HRQOL were 37.8 ± 20.5 and 52.8 ± 20.4 respectively. The mean PBAC score was 183 (range: 67–685) and median VAS score was 8 (range: 0–10). The AUB symptoms included HMB (93.5%, 29/31), prolonged menses (32.3%, 10/31), frequent menses (6.5%, 2/31), and irregular menses (3.2%, 1/31). Twenty patients (64.5%, 20/31) presented with HMB only, while two patients (6.5%, 2/31) reported prolonged menses only. The remaining nine patients (29%, 9/31) had two or three positive symptoms, but HMB was the primary complaint in all of them.

Patient outcome

The clinical outcome within one year of PMWA treatment are summarized in Table 2. One patient had PR of AUB at 3 months and chose continuous combined oral contraceptive treatment for symptoms alleviation. Since no menses occurred during the long-term medication treatment, she was excluded from the 6-month and 12-month analyses. Another patient presented with HMB and infertility (type II adenomyosis) and was referred to a local reproductive medical center 6 months after PMWA treatment. Subsequently, she had no menses due to the downregulation treatment. Therefore, she was excluded from the 12-month analysis.

Table 2. Clinical outcome of patients with AUB-a within one year after PMWA treatment.

	3 months (n = 31)	6 months (n = 30)	1 year (n = 29)
Complete relief	48.4% (15/31)	63.3% (19/30)	65.5% (19/29)
Partial relief	45.2% (14/31)	26.7% (8/30)	24.1% (7/29)
No relief	6.5% (2/31)	10% (3/30)	10.3% (3/29)

In the entire cohort, the proportion of patients with different outcome did not differ over time (p = 0.34). The percentage of patients who achieved CR for the primary symptom was 48.4% (15/31) at 3 months, 63.3% (19/30) at 6 months, and 65.5% (19/29) at 12 months after treatment. The percentage of patients achieving NR was 6.5% (2/13) at 3 months, 10% (3/30) at 6 months, and 10.3% (3/29) at 12 months. Among them, two patients showed no relief at any of the three time points. Both patients presented with dysmenorrhea and HMB before treatment had type II adenomyosis on MRI. Since both desired future fertility, PMWA was implemented with preservation of the entire endometrium and EMJ. The NPVr value reached 98.3% and 102% in these two patients, and post-ablation MRI revealed no endometrial impairment.

Univariate analysis of the factors associated with the clinical efficacy

According to the outcomes at 3, 6, and 12 months after PMWA treatment, the patients were divided into two groups: group CR and group PR/NR. The results of the comparative analysis are summarized in Table 3. At 3 months after treatment, patients with CR had significantly lower baseline serum CA125 levels than those with PR/NR, with median values of 79.4 (38–92.5) U/ml and 108 (81.2–231.5) U/ml, respectively (Z = −2.609, p =.009). The ablation rate of EMJ was significantly different between patients with CR and PR/NR, with mean values of 55.3 ± 19.7% and 19.1 ± 15.9%, respectively (p = 0). Similarly, the mean SA of the ablated part of the EMJ for patients with CR and PR/NR were 15.7 (8.3–25.3) cm² and 4.3 (2.1–12.9) cm², respectively (Z = −3.399, p = .001). No significant differences were observed in the NPVr and other baseline characteristics between the two groups. No differences were observed in any parameter between patients with PR and patients with NR.

Table 3. Comparison of the baseline information and local response indicators between patients with CR and PR/NR outcome, means ± SD or median (P25-P75).

	3 months (n = 31)	6 months (n = 30)	1 year (n = 29)
	CR (n = 15) vs. PR/NR (n = 16)	CR (n = 19) vs. PR/NR (n = 11)	CR (n = 19) vs. PR/NR (n = 10)
Baseline characteristics
Age	40.6 ± 4.7 vs. 36.9 ± 8	41(37–44) vs. 37(31–45)	41(37–45) vs. 37(30.8–44.3)
BMI (kg/m^2)	22.9 ± 2.7 vs. 22.1 ± 3.2	22.4 ± 2.7 vs. 22.8 ± 3.7	22.7 ± 3 vs. 22.6 ± 3.1
Hb (g/L)	112.3 ± 22.7 vs. 113.3 ± 16.2	112.8 ± 21.5 vs. 113.9 ± 16.2	113.7 ± 21.3 vs. 112.6 ± 17.5
CA125	79.4 (38–92.5) vs. 108 (81.2–231.5) *	84.5 (42.4–103) vs. 104 (80.2–233)	87.8 (42.4–131) vs. 93.6 (76.4–230.5)
PBAC	185 (155–252) vs. 167.5 (133.5–420.3)	183 (154–252) vs. 169 (132–397)	185 (154–271) vs. 167.5 (116–395.3)
VAS	8 (5–9) vs. 8 (8–9)	8(5–9) vs. 8(8–9)	8 (5–9) vs. 8 (8–9)
SSS	41.3 ± 23 vs. 34.6 ± 18	37.8 ± 22.6 vs. 38.9 ± 18.1	38.8 ± 22.7 vs. 39.1 ± 17.6
HRQOL	49.1 ± 21.8 vs. 56.4 ± 18.9	52.2 ± 24.7 vs. 53.4 ± 11.8	51.5 ± 24.9 vs. 55.4 ± 11.4
Uterine volume (ml)	213.5(138.6–292.8) vs. 167.5(133.5–420.3)	213.5(138.6–292.8) vs. 234.4(171.3–361.3)	213.5 (138.6–361.3) vs. 227.4 (162.3–321)
Lesion volume (ml)	76.5 (55.5–115.1) vs. 49.3 (27.9–189.1)	66.8 (46.1–1115.1) vs. 52.6 (38.8–145.2)	66.8 (46.1–143.5) vs. 49.3 (35.2–132.4)
Endo - V (ml)	6.5 ± 2.8 vs. 7.5 ± 3.5	6.7 ± 2.9 vs. 6.9 ± 3	6.6 ± 2.9 vs. 7.4 ± 2.9
EMJ - ISA (cm^2)	36.4(28.9–44.4) vs. 33.9(25.9–57.5)	36.5 ± 12.1 vs. 33.4 ± 7.9	36.4 ± 12 vs. 34.3 ± 8.1
Local response indicators
NPVr (%)	82.2 ± 15.7 vs. 80.9 ± 23.5	84.7 ± 16.5 vs. 76 ± 25.1	83.8 ± 17.7 vs. 79 ± 25
Post-ablation quantitative assessment of non-perfusion part of the EMJ
Percentage (%)	55.3 ± 19.7 vs. 19.1 ± 15.9 *	47.9 ± 23.3 vs. 15.5 ± 14 *	46.8 ± 24.7 vs. 19 ± 13.4 *
SA (cm^2)	15.7(8.3–25.3) vs. 4.3(2.1–12.9) *	15.5 (8–24.2) vs. 4.3(1.3–6.6) *	14.9 ± 9.3 vs. 6.3 ± 5.2 *

Note: CR: complete relief; PR: partial relief; NR: no relief; HMB: heavy menstrual bleeding; BMI: body mass index; Hb: hemoglobin; PBAC: pictogram blood-loss assessment chart; VAS: visual analogue scale; SSS: symptom severity score; HRQOL: heath related quality of life; Endo – V: endometrial volume; EMJ: endometrial-myometrial junction; SA: surface area; NPVr: non-perfusion volume ratio; *, significant difference, p-value < .05.

Discussion

In recent years, US-guided PMWA has emerged as a new minimally invasive treatment for symptomatic uterine myomas and adenomyosis [15–20,24–27]. However, the clinical efficacy of PMWA in the treatment of AUB-A remains controversial. This study confirmed that PMWA was an effective intervention for AUB-A, in addition to preliminarily exploring the factors associated with the clinical outcome. This is the first highlight of this study.

The percentage of patients who achieved CR increased over time. At 12 months after PMWA treatment, 89.7% (26/29) patients achieved CR or PR, indicating good treatment efficacy compared to that in previous studies [16,17]. According to the application guideline of PMWA in the treatment of adenomyosis in 2016, patients (with or without future fertility desire) with symptomatic adenomyosis diagnosed by MRI (with lesion thickness > 3 cm) are suitable for PMWA treatment [24]. Its application is not limited by lesion size and location; it reportedly has good efficacy and safety in the treatment of adenomyosis located in the posterior wall [22]. Compared to HIFU and transcervical RFA, PMWA has a wider application range, higher efficiency, and less damage on the surrounding organs and skin [13,15]. Compared to UAE, there is no radiation exposure, no impact on ovarian function and less severe complications [18]. Owing to the frequent recurrence of adenomyosis after conservative treatment, each interventional method can serve as a supplementary treatment for failure or recurrence after previous treatment. In this study, PMWA was an effective alternative minimally invasive treatment modality in 10 patients with a previous interventional treatment history. A previous study showed that the ablated lesion continued to shrink for one year after ablation [15], resulting in a gradual improvement in the symptom relief rate over time within one year after ablation [28]. Our study revealed that the number of patients in the CR group increased, and the ablation rate of EMJ and SA of the ablated part of the EMJ in the CR group decreased over time after treatment. This is in consistent with the results of the previous studies.

Our study also revealed that biopsy has a certain false-negative rate in the diagnosis of adenomyosis. In this study, US-guided core needle biopsy was routinely performed for every candidate before treatment. The histopathological results revealed: adenomyosis or adenomyoma (26/31, 83.9%), leiomyomatous hyperplasia (2/31, 6.5%), smooth muscle tissue (1/31, 3.2%), smooth muscle tissue and endometrial tissue (1/31, 3.2%), smooth muscle hyperplasia (1/31, 3.2%). In previous studies, the positive rates of US-guided puncture biopsy and hysteroscopic guided tissue biopsy for adenomyosis diagnosis were reported to be 28% [6] and 55.47% respectively [29]. Although the diagnostic value for adenomyosis is limited by the biopsy site and sample size, it is still necessary before ablation to rule out potential malignancy [17,24].

This study investigated on the associated factors of the clinical outcome of AUB-A patients after PMWA treatment. First, it was revealed that the baseline CA125 level was an influencing factor for the 3-month outcome. A high CA125 level suggested a greater inflammation, large adenomyotic lesions volume, and higher chance of coexisting pelvic adhesion and endometriosis in patients with adenomyosis [30]. Tang et al. found that a higher level of preoperative CA125 (>35 U/mL) level was related to an earlier onset of symptom recurrence after HIFU treatment [31], which was in consist with our findings.

Second, the ablation rate of EMJ and SA of the ablated part of the EMJ were found to influence the outcome of AUB-A after PMWA treatment, instead of NPVr. The reasons why these two parameters were included in this study are as follows. After nearly 10 years of exploration, it has become a common consensus that the most important key to alleviating symptoms effectively and maintaining a low recurrence rate is to destroy the adenomyotic lesion as much as possible [16,17]. Several studies have documented a close relationship between NPVr and mid- and long-term improvement in symptomatic adenomyosis patients after HIFU treatment [32,33]. According to the Chinese expert consensus on the clinical practice of PMWA in the treatment of adenomyosis released in 2016, NPVr above 70% is recommended for effective PMWA treatment [24]. However, a significant difference was observed in the outcomes of AUB after US-guided thermal ablation treatment, even when the NPVr reached an ideal level (79.7%-91.34%) [16]. The symptomatic relief rate reportedly varied from 20.9% to 60.2% for HMB, and from 39.1% to 80.2% for menstrual disorders in different studies [16]. Although the root cause of this large difference is unknown, a recent review discussed the potential reasons [17]. One of the clues is that different standards for the safe distance between the ablation zone and EMJ (ranging from 0 to 10 mm) have been adopted by different researchers [17,34–36]. Therefore, the ablation rate of the EMJ and extent of endometrial necrosis after ablation might vary greatly among different studies, thus contributing to the large heterogeneity of the results.

EA can induce amenorrhea or menstrual bleeding volume reduction by directly destroying the basal layer of the endometrium and its deep junction zone. PMWA and HIFU may also reduce menstrual blood loss volume in adenomyosis patients, but the mechanism is unclear. As both treatments cause endometrial thermal damage and destruction of the EMJ, the effect of the ablation of the EMJ plays on patient’s outcome is unknown. No study has discussed the relationship between the extent of EMJ ablation and outcomes of patients with AUB-A. Li et al. [20] found that a combination of myometrial and endometrial microwave ablation achieved more satisfactory clinical efficacy in the treatment of adenomyosis-induced anemia than myometrial MWA alone. This is indirect evidence that endometrial ablation contributes to the relief of HMB caused by adenomyosis during PMWA treatment. Our study provides direct evidence that ablation of the EMJ or endometrial basal layer is important for a good outcome in patients with AUB-A. In the past decade, post-ablation evaluation of EMJ necrosis has usually been ignored, and endometrial necrosis has only been considered as a post-ablation complication [16]. This was mainly because there was no proper noninvasive assessment tool to accurately evaluate the extent of EMJ or endometrial ablation in the past. Whether endometrial impairment exists and whether the EMJ adjacent to the adenomyotic lesion should be ablated must be thoroughly discussed.

For patients with type IV adenomyosis involving the entire myometrial layer, complete ablation of the adenomyotic lesion leads to direct thermal damage to the EMJ and endometrium covering the ablated zone. Kim et al. reported that 43.6% of patients with submucosal myoma had different extents of endometrial necrosis after HIFU treatment [37]. However, significant differences were observed in the distribution of blood vessels between adenomyosis and leiomyomas. Penetrating blood flow perpendicular to the EMJ in the myometrial lesion is considered a specific US sign of adenomyosis [38]. The blood vessels in adenomyosis belong to the upstream blood supply network of the spiral artery in the adjacent EMJ. In this study, endometrial impairment was observed in all patients (26/26, 100%) with type IV adenomyosis (ablation rate ranging from 13.8% to 86.5%) and three patients (3/5, 60%) with type II adenomyosis (ablation rate ranging from 0% to 25.1%). No obvious damage to the EMJ was observed in the other two patients with type II adenomyosis (Figure 3), but both had poor outcomes.

After EA, complete atrophy, partial adhesions or obliteration of the uterine cavity, fibrosis, and scarring were observed on second-look hysteroscopy and endometrial biopsy [1]. It can increase the risk of uterine inflammation and hemometra, destroy the environment for fertilized egg implantation. At the same time, endometrial regeneration originating from the fundus and cornu uteri was observed in post-treatment MRI in most patients six months after EA [39]. McCausland et al. reported that partial endometrial ablation with unilateral myometrium ablation did not cause intrauterine adhesions [40]. However, it is generally believed that patients with a desire for continued fertility are not suitable for EA because pregnancies after EA are associated with increased rates of morbidly adherent placenta [41]. Spontaneous uterine rupture has been reported in patients undergoing HIFU treatment [42,43]. According to Hai et al. the incidence of intrauterine adhesions after US-guided RFA treatment was 2.3% (2/87), resulting in amenorrhea without other complications [44]. However, Nam reported that after modifying the US-guided RFA treatment strategy to ensure a safe distance of 1 cm from the ablation zone to EMJ, patients who tried to conceive had good outcomes: the pregnancy success rate reached 50%, the live birth rate 86.3% (22/29), and no case of uterine rupture occurred [34]. Therefore, whether the EMJ adjacent to the adenomyotic lesion should be ablated in patients with AUB-A should be decided according to the patient’s future fertility desire. For patients without a desire for fertility, ablation should be performed for good long-term outcomes. For patients with a desire for future fertility, medication or conception should always be recommended as first-line treatment. If the patient strongly desires to receive interventional therapy, the EMJ should be well preserved and damage to the endometrium minimized as much as possible. This is the second highlight of this study.

Concerning the current status and progress of noninvasive quantitative evaluation of uterine structures, the technical bottleneck has been overcome by 3D image reconstruction technology in recent years. It provides a visualization tool to observe the spatial relationship between the adenomyotic lesion and the surrounding structures (Figure 2D), and also facilitates the precise volumetric measurement of the irregularly-shaped target. Kim et al. used 3D-MRI to reconstruct uterine myomas, the endometrium, and blood vessels as a helpful preoperative adjunct to uterine myomectomy [45]. Zhang et al. used 3D reconstruction of MRI as the standard method to measure the volume of the submandibular gland in patients with sialorrhea [46]. The interobserver agreement of the volumetric measurement of uterine masses and NPV by 3D reconstruction of CE-MRI was confirmed to be better than that of 3D-CEUS and 2D-CEUS [47,48]. In this study, 3D reconstruction of CE-MRI was used to calculate the NPVr and ablation rate of the EMJ after PMWA treatment. This initial experience can inspire further studies in this field and can be considered the third highlight of this study.

This study had several limitations. First, the sample size of this study was small, and further studies with larger sample sizes are required. The BMI of the study population was relatively low compared to that reported from other parts of the world. A multicenter study is needed to address external validity in the future. Second, 29 patients presented with HMB, and two patients presented with only prolonged menses in the study population. The outcome was based on the patient’s primary complaint. Owing to the limited sample size, we were unable to identify the factors influencing the outcome corresponding to each symptom. Further studies are required to solve this problem. Third, only type IV and type II adenomyosis were observed in this population. Whether MRI classification of adenomyosis has an impact on the clinical outcome of patients with AUB-A after PMWA treatment remains unknown. Finally, we focused on the clinical outcomes of AUB-A patients in this study; therefore, other symptoms of adenomyosis, such as dysmenorrhea, infertility, and pelvic compression, were not included. Pelvic pain and infertility are two major issues associated with adenomyosis, that will be the focus of future studies.

Conclusions

In summary, baseline CA125, ablation rate of the EMJ and SA of the ablated part of the EMJ are associated with the clinical efficacy of PMWA in adenomyosis patients with AUB. Therefore, they may act as predictors of outcomes after PMWA treatment for AUB-A in future clinical practice.

Acknowledgments

The authors thank Taylor & Francis for the English language editing and the reviewers for their helpful comments.

Disclosure statement

No conflict of interest was reported.

Additional information

Funding

This work was supported by the Science and Technology Commission of Shanghai Municipality under Grant 21Y11910800 and 19DZ2251100, Shanghai Municipal Health Commission under Grant 2019LJ21 and SHSLCZDZK 03502, and the Scientific Research and Development Fund of Zhongshan Hospital of Fudan University under Grant 2022ZSQD07.