Long-term efficacy of uterine artery embolization for adenomyosis and analysis of prognostic factors: a cohort study

Significant differences in baseline characteristics were observed between the two groups (Table  1 ). Patients in the recurrence group were younger and had a longer disease duration ( P  < 0.001). They also had more severe preoperative symptoms, with higher VAS and PBAC scores ( P  < 0.001), and elevated preoperative serum CA-125 levels compared with the non-recurrence group ( P  < 0.001). Regarding imaging parameters, the recurrence group had a higher proportion of diffuse-type adenomyosis ( P  = 0.009), significantly increased JZ thickness ( P  < 0.001), and a greater proportion of lesions with sparse vascularity ( P  < 0.001). No significant differences were observed between the groups in embolization particle size or parity ( P  = 0.877 and P  = 0.057, respectively). Table 1 Comparison of baseline characteristics between non-recurrence and recurrence groups Characteristics Non-recurrence group ( n  = 138) Recurrence group ( n  = 62) t/χ² P Age (years) 43.09 ± 3.11 39.24 ± 3.30 7.953 < 0.001 Obstetric history 2 (2, 2) 2 (1, 2) - 0.057 Disease duration (years) 4.66 ± 1.08 5.97 ± 0.69 -10.338 < 0.001 Preoperative VAS score 7 (6, 8) 8 (7.8, 9) - < 0.001 Preoperative PBAC score 180 (175, 200) 222.5 (210, 240) - < 0.001 Preoperative CA-125 (U/mL) 76.9 (60.1, 101.6) 118.2 (99.9, 135.5) - < 0.001 Adenomyosis type (%)  Diffuse 91 (65.9) 52 (83.9) 6.748 0.009  Focal 47 (34.1) 10 (16.1)  JZ thickness (mm) 11.53 ± 2.24 14.92 ± 1.53 -12.474 < 0.001 Lesion vascularity (%)  Rich/Moderate 116 (84.1) 36 (58.1) 15.847 < 0.001  Sparse 22 (15.9) 26 (41.9) Embolization particle size (%)  Diameter ≥ 500 μm 95 (68.8) 42 (67.7) 0.024 0.877  Diameter < 500 μm 43 (31.2) 20 (32.3) Comparison of baseline characteristics between non-recurrence and recurrence groups At 2-year follow-up (Table  2 ), both groups showed significant improvement in primary symptom scores compared with preoperative values (all P  < 0.001). However, the non-recurrence group experienced more pronounced improvement in pain relief, reduction of menstrual blood loss, and quality of life, with the differences reaching statistical significance ( P  < 0.001). Overall, both groups obtained substantial clinical benefit post-UAE, but the non-recurrence group demonstrated greater improvement in core symptom parameters. Table 2 Comparison of improvement in primary symptom scores at 2 years post-UAE Parameters Groups Preoperative values 2 years postoperatively P value Δ value VAS score Non-recurrence group ( n  = 138) 7 (6, 8) 1 (1, 2)*** < 0.001 5.52 ± 1.54 Recurrence group ( n  = 62) 8 (7.8, 9) 5 (5, 6)*** 3.03 ± 1.53 PBAC score Non-recurrence group ( n  = 138) 180 (170, 200) 62 (55, 70)*** < 0.001 122.51 ± 23.57 Recurrence group ( n  = 62) 222.5 (210, 240) 151 (137.3, 167.8)*** 70.86 ± 30.64 Quality of life Non-recurrence group ( n  = 138) 70.5 (65.8, 75) 26 (19.5, 32)*** < 0.001 44.37 ± 9.91 Recurrence group ( n  = 62) 77.5 (70, 84.3) 49 (40.3, 56.5)*** 29.58 ± 11.80 *** P  < 0.001 vs. preoperative Comparison of improvement in primary symptom scores at 2 years post-UAE *** P  < 0.001 vs. preoperative Dynamic changes in serum CA-125 at 3 months post-UAE (Supplementary Table 1) demonstrated a significant reduction in both groups compared with preoperative values ( P  < 0.001). The non-recurrence group exhibited a greater decrease in CA-125 levels ( P  = 0.044), suggesting that early biomarker reduction may be associated with treatment efficacy and long-term prognosis. Analysis of postoperative complications (Table  3 ) indicated overall favorable safety in both groups, though the incidence differed significantly ( P  = 0.024). The recurrence group exhibited a higher complication rate than the non-recurrence group (38.7% vs. 23.2%). Specific complications, including post-embolization syndrome, inguinal hematoma, and transient ovarian dysfunction, were more frequent in the recurrence group. Overall, patients in the recurrence group faced a higher postoperative complication risk. Table 3 Comparison of UAE-related postoperative complications between groups Type Non-recurrence group ( n  = 138) Recurrence group ( n  = 62) χ² P Post-embolization syndrome 22 (15.9) 16 (25.8) 5.112 0.024 Inguinal hematoma 4 (2.9) 3 (4.8) Transient ovarian dysfunction 6 (4.3) 5 (8.1) Other complications 32 (23.2) 24 (38.7) - - Comparison of UAE-related postoperative complications between groups Kaplan–Meier survival analysis (Fig.  1 ) demonstrated that overall recurrence-free survival after UAE gradually decreased with increasing follow-up time. Early postoperative survival curves remained relatively stable, with a notable downward trend beginning at 6–12 months. Recurrence-free survival was 90% at 12 months (95% CI: 85.9%–94.3%) and declined to approximately 69% at 24 months. These results indicate that UAE provides relatively stable efficacy within the first postoperative year, while the risk of recurrence increases thereafter. Fig. 1 Kaplan–Meier curve of recurrence-free survival after UAE Kaplan–Meier curve of recurrence-free survival after UAE To identify key factors affecting post-UAE recurrence, univariable and multivariable Cox regression analyses were performed. Univariable analysis indicated that age, adenomyosis type, JZ thickness, lesion vascularity, and preoperative CA-125 were associated with recurrence ( P  < 0.10) and were included in the multivariable model. Multivariable analysis (Table  4 ) showed that age (HR = 0.903, 95% CI = 0.838–0.973, P  = 0.008), JZ thickness (HR = 1.333, 95% CI = 1.168–1.522, P  < 0.001), and preoperative CA-125 (HR = 1.029, 95% CI = 1.016–1.042, P  < 0.001) were independent predictors of postoperative recurrence. Table 4 Multivariable Cox regression analysis of independent prognostic factors for recurrence Prognostic factors HR 95% CI P Age 0.903 0.838–0.973 0.008 Adenomyosis type 1.257 0.580–2.727 0.562 JZ thickness 1.333 1.168–1.522 < 0.001 Lesion vascularity 1.274 0.720–2.256 0.406 Preoperative CA-125 1.029 1.016–1.042 < 0.001 Multivariable Cox regression analysis of independent prognostic factors for recurrence

This study was a single-center, retrospective cohort study. The study protocol was approved by the Institutional Ethics Committee of Wuhan Red Cross Hospital (Approval No.: 2024003), and the requirement for informed consent was waived due to the retrospective nature of the study. We retrospectively retrieved and screened all consecutive inpatients who underwent UAE for symptomatic adenomyosis at our institution between March 2019 and March 2022. Pharmacological treatment before and after UAE varied among patients. Medication information was collected and documented for all cases to account for treatment heterogeneity. Based on inclusion and exclusion criteria, a total of 200 patients were ultimately included, with a minimum of 2 years of follow-up data collected for analysis. Inclusion criteria were: (1) Diagnosis of adenomyosis confirmed by clinical symptoms and MRI; (2) Presence of moderate to severe symptoms, defined as either dysmenorrhea with a visual analog scale (VAS) score ≥ 4 or menorrhagia with a pictorial blood loss assessment chart (PBAC) score > 100 (meeting at least one criterion); (3) Age between 18 and 50 years; (4) Availability of complete preoperative baseline and postoperative follow-up data. Exclusion criteria were: (1) Coexisting large uterine fibroids requiring surgical intervention (diameter > 4 cm) or pelvic endometriosis; (2) Explicit fertility plans; (3) Prior endometrial resection or adenomyotic lesion excision; (4) Loss to follow-up or incomplete follow-up data, making it impossible to determine 2-year clinical outcomes. All enrolled patients underwent standardized UAE performed by experienced interventional radiologists. Embolization of the bilateral uterine arteries was completed using polyvinyl alcohol (PVA) particles and/or gelatin sponge particles. No external control group was included in this study. To evaluate long-term efficacy, patients were divided into two groups according to 2-year postoperative clinical outcomes: (1) Non-recurrence group: patients whose symptoms did not recur within 2 years postoperatively and who did not require any additional surgical or interventional treatment; (2) Recurrence group: patients who experienced symptom recurrence or insufficient efficacy within 2 years postoperatively and underwent additional intervention. Clinical and imaging data were retrieved from the hospital electronic medical record system. Baseline variables collected included: age, obstetric history, disease duration, embolization particle size, preoperative VAS score, preoperative PBAC score, and preoperative serum CA-125 level. Imaging parameters were assessed using preoperative MRI and included: adenomyosis type (diffuse or focal), JZ thickness in millimeters, and lesion vascularity (rich, moderate, or sparse). Efficacy and safety outcomes included: procedure date, postoperative complications, timing of any re-intervention, and date of last follow-up. Primary endpoint: 2-year recurrence-free survival after UAE. Secondary endpoints: symptom improvement rate and incidence of postoperative complications within 2 years. All statistical analyses were performed using SPSS software (version 27.0). Continuous variables with normal distribution were expressed as mean ± standard deviation (SD) and compared using the independent-samples t -test. Non-normally distributed data were expressed as median (interquartile range [IQR]) and compared using the Mann-Whitney U test. Categorical variables were presented as counts (percentages) and compared using the chi-square (χ²) test or Fisher’s exact test. Long-term efficacy was evaluated using Kaplan-Meier survival analysis, and 2-year recurrence-free survival rates were calculated. Prognostic factors were analyzed using Cox proportional hazards regression models. Candidate variables with P  < 0.10 in univariable Cox regression were included in the multivariable model. To avoid overfitting and ensure model stability, the multivariable model was constructed considering the number of events per variable (EPV), clinical relevance, and statistical significance. Independent prognostic factors were identified, and results were reported as hazard ratios (HR) with 95% confidence intervals (CI). All tests were two-sided, and P  < 0.05 was considered statistically significant.

UAE provides favorable short-term efficacy and overall safety in the treatment of adenomyosis, significantly relieving pain, reducing menorrhagia, and improving quality of life. However, postoperative recurrence remains a key factor affecting long-term outcomes. We identified age, JZ thickness, and preoperative CA-125 as independent predictors of recurrence. Survival analysis further indicated that the first postoperative year is a critical window for recurrence monitoring, and high-risk patients should undergo closer follow-up and early intervention. These findings emphasize the importance of precise preoperative assessment and individualized postoperative management to optimize long-term UAE efficacy and reduce recurrence risk.

This study systematically evaluated the long-term efficacy of UAE in patients with adenomyosis and investigated factors associated with recurrence. Overall, most patients experienced significant symptom improvement after UAE, particularly in pain relief, reduction of menstrual blood loss, and quality of life, consistent with previous reports on UAE efficacy [ 12 ]. However, a subset of patients experienced recurrence during follow-up, highlighting the heterogeneous nature of adenomyosis, its chronic and relapse-prone course, and the need to identify key clinical predictors for precision therapy. Our results showed that patients in the recurrence group were younger, had longer disease duration, higher symptom burden, a higher proportion of diffuse-type adenomyosis, more pronounced JZ thickening, and elevated preoperative CA-125 levels, all previously reported to correlate with disease severity [ 13 ]. Diffuse-type adenomyosis, with deeper myometrial involvement and complex vascularization, is more prone to residual perfusion or collateral circulation reconstruction after UAE, increasing recurrence risk [ 14 ]. In terms of symptom improvement, the non-recurrence group showed more substantial gains in pain, menstrual blood loss, and quality of life, suggesting that recurrence significantly attenuates the long-term benefits of UAE. This aligns with reports that elevated CA-125 is associated with symptom relapse within 1–3 years postoperatively [ 15 ]. Postoperative complications were generally low in both groups, reaffirming the safety of UAE. However, the recurrence group had a higher complication rate, possibly due to diffuse lesions, abundant collaterals leading to a larger embolization area, and stronger inflammatory response [ 16 ]. Younger age in this group may increase ovarian sensitivity to blood supply changes, elevating the risk of transient ovarian dysfunction [ 17 , 18 ]. Survival analysis further indicated that recurrence-free survival remained stable during the first postoperative year but declined significantly between 12 and 24 months, consistent with prior medium- and long-term follow-up studies [ 19 ]. Therefore, the first postoperative year may represent a critical window for recurrence monitoring, particularly for high-risk patients with extensive lesions, increased JZ thickness, or elevated CA-125, who should undergo closer follow-up and early intervention. In the multivariable Cox regression model, age, JZ thickness, and preoperative CA-125 were confirmed as independent prognostic factors. Age acted as a protective factor (HR < 1), with younger patients exhibiting higher recurrence risk. Previous studies suggest that higher ovarian function and estrogen levels in younger patients may promote proliferation of residual ectopic tissue, facilitating reperfusion and recurrence [ 20 , 21 ]. Adenomyosis in reproductive-age women progresses more rapidly, with higher biological activity in lesions, contributing to less durable symptom relief after UAE in younger patients [ 22 ]. JZ thickness has received increasing attention in adenomyosis pathology, reflecting the degree of myometrial disruption and lesion infiltration. Our study found that thicker JZ significantly increased recurrence risk, consistent with multiple MRI studies [ 23 ]. JZ thickening often indicates deep lesions with high pathological activity and rich vascular supply. Although UAE effectively blocks primary feeding arteries, patients with marked JZ thickening often have extensive collaterals, allowing residual perfusion and continued lesion growth, thus increasing recurrence likelihood [ 24 ]. Deep lesions may also limit postoperative necrosis and scarring, failing to completely inactivate ectopic tissue, explaining why JZ thickness is a strong predictor. Preoperative CA-125 was also an independent predictor. Numerous studies link CA-125 levels with lesion extent, inflammatory activity, and pelvic adhesion severity [ 25 ]. Elevated CA-125 often indicates extensive, metabolically active lesions that may persist after UAE, increasing recurrence risk. In our study, the HR for CA-125, although close to 1, was significantly higher than reference, indicating a stable, linear increase in recurrence risk per unit elevation, consistent with previous cohort studies [ 26 , 27 ]. In addition to baseline CA-125 levels, we also evaluated its early postoperative change at 3 months. A greater decline was observed in the non-recurrence group, indicating that dynamic biochemical response may reflect short-term treatment effectiveness after UAE. As CA-125 is considered a marker of disease burden and inflammatory activity in adenomyosis, its postoperative reduction may suggest adequate lesion devascularization. These findings imply that short-term CA-125 monitoring could provide additional information during postoperative follow-up. Although age, JZ thickness, and CA-125 have been previously reported as prognostic indicators in adenomyosis, this study provides large-cohort validation and risk stratification data in patients undergoing UAE. These findings contribute clinically actionable evidence by identifying high-risk individuals who may benefit from closer follow-up or consolidation therapy, thereby complementing existing literature and supporting precision treatment strategies. These findings have important clinical implications. First, a predictive model combining age, JZ thickness, and CA-125 can guide preoperative risk stratification, helping clinicians select optimal treatment strategies. Rather than positioning UAE as a definitive standalone therapy, our data support its role as a lesion-debulking and rapid symptom-controlling intervention, particularly for patients with severe dysmenorrhea or menorrhagia who require immediate relief or uterine preservation. For high-risk patients, postoperative maintenance therapy with GnRH-a or progestins may prolong symptom relief. In this context, hormonal therapy should be considered a consolidation strategy rather than a substitute for UAE, potentially reducing cumulative hormonal exposure and avoiding prolonged high-dose medical treatment alone. High-risk patients (e.g., JZ > 12–15 mm) may benefit from more frequent follow-up, including MRI, to detect early lesion regrowth. Importantly, this stratified approach may enhance clinical utility and cost-effectiveness by identifying patients most likely to benefit from UAE while avoiding unnecessary interventional procedures in those with a high probability of early recurrence. For carefully selected patients, UAE may delay or reduce the need for hysterectomy and decrease repeated hospitalizations due to uncontrolled symptoms, thereby offering meaningful long-term value beyond short-term symptom control. Second, baseline information can improve patient counseling, enabling clearer explanation of long-term outcomes and enhancing adherence. For example, younger patients with markedly elevated CA-125 should be informed of higher recurrence probability to make informed decisions among UAE, pharmacologic therapy, or surgery. This shared decision-making framework allows individualized comparison between primary hormonal therapy and interventional treatment, taking into account symptom severity, reproductive intentions, tolerance to long-term medication, and economic considerations. Moreover, these prognostic factors can be used for stratification in future clinical trials, helping identify patients suitable for novel interventional techniques (e.g., partial embolization, super-selective embolization, combined ablation) and enhancing comparability of study outcomes. Ovarian reserve markers (e.g., anti-Müllerian hormone) and reproductive outcomes were not collected in this retrospective cohort, as patients with explicit fertility plans were excluded from the study population. Therefore, fertility-related safety and pregnancy outcomes could not be evaluated. This limits conclusions regarding the impact of UAE on ovarian function and subsequent reproductive potential. Emerging evidence has raised concerns regarding potential reproductive risks following UAE, including adverse obstetric outcomes. Although these issues were beyond the scope of the present study, they should be carefully considered in clinical decision-making, particularly in younger patients. Future prospective studies incorporating ovarian reserve assessment and long-term reproductive follow-up are warranted. Additionally, although our analysis suggests potential cost-effectiveness through improved risk stratification, formal health economic evaluation and direct comparison with primary medical therapy were beyond the scope of this study and should be addressed in future research. Future research incorporating health economic evaluation and direct comparison with primary medical therapy will be essential to further clarify the long-term cost-effectiveness and optimal positioning of UAE within the treatment algorithm. Despite providing important clinical evidence, this study has limitations. First, it is a single-center retrospective study with a limited sample size, which may reduce statistical power and contribute to non-significance of some variables. Second, lesion vascularity and adenomyosis type were based on MRI classification, which is subjective; future studies should incorporate MRI radiomics or 3D reconstruction to improve measurement consistency. Third, CA-125 is a non-specific inflammatory marker influenced by menstrual cycle or comorbidities; although confounding factors were minimized, some bias may remain. Finally, follow-up, while spanning multiple years, may be insufficient to fully capture long-term recurrence, and recurrence definitions based on symptoms and imaging may be influenced by patient variability and subjective scoring. Future multicenter, prospective, long-term studies are needed to validate these findings and explore additional molecular or imaging predictors.

Adenomyosis is a common benign uterine disorder in women of reproductive age, characterized pathologically by the ectopic infiltration of endometrial glands and stroma into the myometrium, resulting in uterine enlargement, disruption of myometrial architecture, and chronic inflammatory responses [ 1 ]. Clinically, patients often present with progressive dysmenorrhea, heavy menstrual bleeding, chronic pelvic pain, infertility, and recurrent miscarriage, all of which markedly impair quality of life and reproductive potential [ 2 ]. In recent years, the diagnostic rate of adenomyosis has substantially increased with advances in imaging techniques, particularly magnetic resonance imaging (MRI). Imaging indicators such as junctional zone (JZ) thickness, lesion distribution, and vascularity have been suggested to correlate with disease severity and therapeutic responsiveness [ 3 ]. Ultrasonography is widely regarded as the first-line imaging modality for adenomyosis due to its accessibility and satisfactory diagnostic performance in clinical practice. Typical sonographic features include heterogeneous myometrial echotexture, myometrial cysts, and indistinct endometrial–myometrial junction, which correlate with histopathological findings and aid diagnosis [ 4 ]. Although magnetic resonance imaging provides superior soft-tissue resolution and is useful in complex cases, ultrasonography remains the primary diagnostic tool and facilitates routine assessment and follow-up [ 5 ]. Current treatment options for adenomyosis include pharmacological therapy, surgical intervention, and interventional procedures. Pharmacological treatments, such as gonadotropin-releasing hormone agonists and progestin-based therapies, can temporarily alleviate symptoms; however, symptoms often recur upon drug withdrawal, and long-term use is limited by adverse effects [ 6 ]. Surgical options, including hysterectomy or localized excision of adenomyotic lesions, may not be suitable for patients who wish to preserve the uterus or maintain fertility. In recent years, uterine artery embolization (UAE), a minimally invasive, uterus-sparing interventional procedure, has been increasingly applied in the management of adenomyosis [ 7 ]. Evidence indicates that UAE effectively relieves pain, reduces menstrual blood loss, and improves quality of life, with favorable short- and mid-term outcomes [ 8 ]. Despite its demonstrated symptomatic benefits, the long-term efficacy of UAE remains controversial. Reported recurrence rates vary widely across studies, suggesting that individual patient characteristics and biological heterogeneity of lesions may influence treatment outcomes [ 9 ]. Identifying key prognostic factors that affect long-term efficacy is therefore crucial for optimizing patient selection, guiding personalized management strategies, and predicting recurrence risk. Existing research has attempted to evaluate prognostic indicators from perspectives including imaging features, laboratory markers, and clinical characteristics. For example, some studies have associated larger lesion volume or markedly increased JZ thickness with poorer therapeutic response [ 10 ]. Lesions with rich vascularity may exhibit stronger reperfusion tendencies after embolization, potentially compromising the durability of symptom relief [ 8 ]. Additionally, clinical parameters such as age and carbohydrate antigen 125 (CA-125) levels may reflect underlying disease activity [ 11 ]. However, findings across studies remain inconsistent, and many investigations have been limited by short follow-up periods and small sample sizes, leaving a gap in systematic long-term outcome evaluation. Therefore, this cohort study included patients with adenomyosis treated with UAE and conducted long-term follow-up to systematically assess trends in therapeutic outcomes. Cox proportional hazards regression analysis was employed to identify clinical and imaging factors associated with recurrence. We aim to clarify key prognostic indicators and support further refinement of individualized management strategies for patients undergoing UAE.

Supplementary Material 1. Supplementary Material 1.