Study design and patient selection

This was an open-label, single-center, prospective, randomized trial that was approved by the institutional ethics committee of Shanghai Tenth Peoples’ Hospital (No. 21XJS39). The study was registered at www.chictr.org.cn, with the research number ChiCTR2100053582, and conducted following the Declaration of Helsinki. Patients who met the following criteria were enrolled and excluded.

The inclusion criteria were as follows: (1) premenopausal women with symptomatic adenomyosis (≥ 18 years); (2) both US- and magnetic resonance imaging (MRI)-diagnosed adenomyosis; (3) childbirth completed and had no fertility desire; (4) refused other treatments and had no contraindications to PMWA (e.g., severe cardiopulmonary dysfunction, coagulation dysfunction, and mental illness); and (5) willing to participate in the study and sign the informed consent forms. The exclusion criteria were (1) adenomyosis with bilateral wall involvement (n = 14); (2) concomitant submucosal leiomyomas (n = 14); (3) inability to undergo MRI examination due to metal implants in the body or contrast agent allergy (n = 3); (4) pre-ablation MRI revealing IUA or uterine deformities, indicating difficulty in placing a catheter into the uterine cavity (n = 4); (5) endometrial atrophy and thinness (thickness < 3 mm) leading to difficulty in assessing the degree of endometrial impairment (n = 1); or (6) planned to undergo transvaginal MWA treatment (n = 2). Finally, 60 patients were included in this study.

The study process is presented in Fig. 2. From December 2021 to July 2023, ninety-eight eligible patients who wished to undergo PMWA treatment were enrolled prospectively. All patients were informed of the benefits and potential risks of PMWA treatment and intrauterine saline installation, and written informed consent was obtained from each patient. After exclusions of patients who met the exclusion criteria, sixty patients were included and randomly assigned at a 1:1 ratio by a computer-generated randomization list into two groups. The study group received PMWA treatment assisted by intrauterine chilled saline (Fig. 3), while the control group received traditional PMWA treatment without intrauterine chilled saline (Fig. 4). As an open-label design was necessary, the allocation information was available to the clinical staff members, each of whom were blinded to the post-ablation evaluation personnel to avoid subjective factors that might bias the research results.

Fig. 2

Flowchart of the randomized controlled trial

Fig. 3

a–d A case of adenomyosis treated by PMWA assisted by intrauterine chilled saline. Before treatment, US (transverse section) reveals type IV adenomyosis (blue circle) located on the posterior wall (a). The endometrium is thin (3 mm), which leads to a blurred display of uterine cavity lines on B-mode US (white arrows). After injecting 5 mL of saline through a urinary catheter (red arrows), the uterine cavity lines separate from each other and display clearly (blue arrows) on B-mode US (b). Then the deepest part of the lesion (yellow circle) is ablated at first (c), with the uterine cavity displayed clearly (blue arrows). When ablating adenomyosis near the uterine cavity (yellow circle), chilled saline is injected into the uterine cavity slowly (blue arrows) and continuously to maintain a low temperature (d)

Fig. 4

a–d A case of adenomyosis treated by PMWA only. Before treatment, US (sagittal section) reveals type IV adenomyosis (blue circle) located on the posterior wall (a). The endometrium (white arrows) is thin (4 mm), which leads to blurred cavity lines on B-mode US, and the antenna (red arrows) is inserted into the lesion 2 cm from it (b). The ablation starts at the deepest part (yellow circle) (c) when the cavity line still can be displayed at this point (white arrows). When the main part of the lesion is ablated (yellow circle), the endometrial line turns invisible due to gas interference (white arrows), making it impossible to determine whether heat has spread to the uterine cavity (d)

Baseline information collection

The baseline information of all participants, including age, body mass index (BMI), complaints, medication history, and pregnancy and delivery history, among other variables, was collected before treatment. Symptoms were evaluated by an interventional US radiologist who specializes in gynecological interventional treatment. A pictorial blood loss assessment chart (PBAC) recommended by the Chinese expert consensus on the chemical ablation of uterine fibroids [23] was used to quantify the menstrual bleeding volume. Visual analog scale (VAS) scores ranging from 0 (no pain at all) to 10 (severe pain equivalent to childbirth pain) were used to quantify the severity of pain. The symptom severity score (SSS) and health-related quality of life (HRQOL) scale were used to quantify the severity of symptoms and patients’ quality of life [24].

Pre-ablation evaluation

All patients underwent systematic examinations, including laboratory tests, chest X-ray, electrocardiography, pelvic US, and contrast-enhanced MRI, within 2 days before treatment. To rule out pelvic infection and early pregnancy, all patients underwent human chorionic gonadotropin tests and microbial testing of vaginal secretions. Pelvic US examination was performed with a US scanner (LOGIQ E9 system; GE Healthcare, Milwaukee, WI, USA) equipped with a 1–6 MHz convex array probe and 3–12 MHz intracavity probe. In total, 2 mL SonoVueTM (Bracco Suisse SA, Geneva, Switzerland) was injected intravenously for contrast-enhanced ultrasound (CEUS) imaging before ablation. MRI was performed on a 1.5-T MR scanner (Magnetom Verio, Siemens, Germany) with a gradient slew rate of 200 mT/ms and gradient strength of 40 mT/m. Gadopentetate dimeglumine (Magnevist, Bayer Schering, Germany) was used as the contrast agent for contrast-enhanced T1-weighted imaging (CE-T1WI).

Based on pre-ablation T2-weighted imaging (T2WI), adenomyosis was classified into four types according to Kish’s study [25]: type I (intrinsic), internal adenomyosis surrounded by a layer of normal outer myometrium; type II (extrinsic), external adenomyosis with involvement of the outer myometrium only; type III (intramural), intramural isolated adenomyosis without connection to the endometrium or serosa; and type IV (mixed), adenomyosis penetrating the entire myometrium. The baseline volume of the uterine corpus and adenomyosis was measured on T2WI, with the ellipsoid analog volume formula (V = π/6 × height × length × width) adopted in the calculation.

Pre-ablation preparation and intrauterine catheterization

All patients took oral laxatives to clear their intestines one day before treatment. Twenty minutes before ablation, all patients were given conscious sedation and analgesia, namely, slow intravenous injection of midazolam (0.1 mg) to relieve anxiety and fear and intravenous pumping of fentanyl (0.02 mg/kg) to alleviate pain. A portable electrocardiogram monitor was used to monitor the patient’s heart rate, blood pressure, respiration rate, and oxygen saturation throughout the entire procedure. With the patient placed in the lithotomy position, the external genitalia were disinfected. The cervical os was exposed with a speculum inserted into the vagina and then cleaned with povidone-iodine solution. An 8 F urinary catheter was inserted through the cervical os and into the cervical canal very gently with air evacuated. The catheter balloon was then inflated with 2–3 mL of saline, and the speculum was subsequently removed carefully. Finally, a 50 mL syringe filled with chilled sterile saline was attached to the catheter, which was prepared for intrauterine instillation during subsequent ablation therapy.

US-guided PMWA treatment

A monopolar water-cooling MWA applicator (Microwave Ablation system MTI-5A; Nanjing Great Wall Medical Equipment Co. Ltd., Nanjing, China) with a power setting of 60 W was used. Correspondingly, an 18-cm long, 14-G microwave antenna (XR-A2018W; Nanjing Great Wall Medical Equipment Co. Ltd., Nanjing, China) was used. PMWA therapy was performed by two senior interventional US radiologists with more than 5 years of experience with the uterine ablation procedure. The operation process was standardized according to the recommendations of the Chinese expert consensus on MWA of adenomyosis [26], including disinfection and draping, local anesthesia, artificial ascites injection, US-guided biopsy, PMWA, and intra-ablation CEUS.

The patient was placed in the supine position, and the skin of the lower abdomen was disinfected. After administration of local anesthesia, 400–600 mL of warm sterilized saline was instilled into the pelvic cavity around the uterus to protect the surrounding organs. Next, a routine free-hand biopsy was performed under US guidance with an 18-G biopsy core needle (Bard Magnum Biopsy Instrument; Covington, GA, USA) to rule out potential malignancy. The puncture site was selected randomly to avoid large blood vessels. Subsequently, PMWA was performed with an output power of 60 W using the “moving-shot” technique under the guidance of transabdominal US. For patients with adenomyosis located at the posterior wall of a retroverted or retroflexed uterus, Yu’s uteropexy was applied to change the uterine position to avoid damage to the anterior endometrium [11].

To facilitate the precise placement of the microwave antenna, 5 mL of saline was injected into the endometrial cavity to visualize the spatial relationships among the antenna, adenomyosis, and endometrium (Fig. 3a, b). First, the antenna was inserted into the deepest part of the adenomyotic tissue 2 cm away from the serosa, and then microwave radiation was applied. Once the vaporization reaction was induced by heat, a hyperechoic cloud was generated and gradually spread out on B-mode US (Fig. 3c), allowing the ablation zone to be monitored in real time. A safe distance between the edge of the ablation zone and the serosa was maintained at least 5 mm to prevent thermal damage to the surrounding pelvic organs. When the region of adenomyosis near the endometrium (yellow circle) was ablated, chilled saline was slowly and continuously infused into the uterine cavity to maintain a low temperature (Fig. 3d). The amount of saline injected varied depending on the distance of the uterus (aimed to separate the uterine cavity lines by ≥ 5 mm) and patient tolerance. Generally, the total amount of saline instillation was controlled at ≤ 50 mL for each patient. If any pain occurred, the severity of the pain was evaluated and recorded by a dedicated US radiologist using a VAS, as was the pain induced by ablation. Finally, after the hyperechoic cloud covered the entire lesion on B-mode US, the ablation was stopped. After several minutes, intraoperative CEUS was performed to evaluate the instant local treatment response to PMWA, and supplementary ablation was performed if deemed necessary. The ablation was terminated when no enhancement was observed in the target lesion. At the same time, the extent of intra-ablation discomfort was recorded.

Endometrial impairment evaluation

Contrast-enhanced MRI was performed within 2 days after treatment to investigate thermal damage to the endometrium. The assessment was carried out by two senior radiologists, who were blinded to the group allocation information. If a consensus was not reached, the final result was determined after discussion. Based on the enhancement defect on CE-T1WI, the thermal damage to the perfusion of the basal layer of the endometrium on the opposite side of the adenomyotic region was evaluated. In one patient whose fundus and posterior wall were affected, endometrial impairment in the anterior wall was assessed, and in another patient whose left wall and fundus were affected, endometrial impairment in the right wall was assessed. According to Kim’s study [16], endometrial perfusion impairment was divided into four levels: grade 0 (Fig. 5a), normal endometrium; grade 1 (Fig. 5b), mild, pinpoint discontinuity with a size ≤ 3 mm; grade 2 (Fig. 5c), moderate, discontinuity ≤ 1 cm, but > 3 mm; and grade 3 (Fig. 5d), severe, discontinuity > 1 cm. Based on the endometrial signal intensity (SI) displayed on T2WI after ablation, the severity of endometrial dehydration was evaluated and divided into three levels: grade 0 (Fig. 6a), normal high-signal endometrium; grade 1 (Fig. 6b), mild or moderate, with part of the endometrium showing hypointensity; and grade 2 (Fig. 6c), severe, with almost the whole endometrium showing hypointensity.

Fig. 5

a–d Grading the perfusion impairment of the contralateral endometrium. Grade 0, pre-ablation CE-T1WI (a1) reveals uniform enhancement of the basal layer of endometrium, post-ablation CE-T1WI (a2) shows normal endometrium with similar enhancement (white arrows). Grade 1, pre-ablation CE-T1WI (b1) reveals enhanced bilateral endometrium, and post-ablation MRI (b2) shows pin-point enhancement discontinuity measured 2 mm (white arrow) on the endometrium of the anterior wall. Grade 2, pre-ablation CE-T1WI (c1) reveals type IV adenomyosis on the posterior wall, post-ablation CE-T1WI (c3) shows moderate enhancement discontinuity measuring 8 mm (dotted line) on the endometrium and junction zone of the anterior wall. Grade 3, pre-ablation CE-T1WI (d1) reveals adenomyosis on the anterior wall, while post-ablation CE-T1WI (d2) shows severe enhancement discontinuity measuring 15 mm (dotted line) on the endometrium and junction zone of the posterior wall

Fig. 6

a–c Grading of the endometrial signal intensity impairment on T2WI. Grade 0, both pre-ablation (a1) and post-ablation (a2) sagittal fat-suppressed T2WI demonstrate hyper-intense endometrium (arrows). Grade 1, pre-ablation T2WI shows hyper-intense endometrium (b1), but post-ablation T2WI (b2) reveals heterogeneous endometrium with partial low T2 signal intensity (arrow), indicating mild endometrial dehydration and necrosis. Grade 2, almost the whole endometrium (arrows) demonstrates hypo-intensity after treatment (c2), indicating severe endometrial dehydration and necrosis

Statistical analysis

The minimum sample size was calculated as indicated at http://powerandsamplesize.com. The intervention was considered effective if the incidence of endometrial impairment decreased by 0.35 in the study group. A sample of 26 participants per group provided at least 80% power with a 0.05 two-sided significance. Considering that some patients may have been excluded, which resulted in sample dropout (assuming ≤ 15%), the target number of enrolled patients was expanded to 31 per group.

SPSS 25.0 (IBM, Chicago, USA) was used for the statistical analysis. The distribution of the continuous data was tested by the Shapiro‒Wilk normality test. Normally distributed data are expressed as the mean ± SD, and nonnormally distributed data are expressed as the median (P25–P75). A t test was used to compare continuous variables with a normal distribution (age, BMI, and ablation rate), while the Wilcoxon signed-rank test was used to compare continuous data with a skewed distribution (VAS, PBAC, SSS, uterine volume, and adenomyosis volume). The Kruskal‒Wallis test was used to compare the ordered class variables (endometrial impairment grading results). The chi-square test or Fisher’s test was used to compare other categorical variables (uterine position, adenomyosis location, MRI classification, incidence of intraoperative discomforts). Statistical significance was defined as a p value < 0.05.