Endometrial polypoid adenomyoma outcomes and fertility-sparing treatments efficacy for atypical type: A systematic review and network meta-analysis

Abstract

BACKGROUND

Endometrial polypoid adenomyoma is a kind of estrogen-dependent benign tumor, categorizing into typical [typical polypoid adenomyoma (TPA)] and atypical polypoid adenomyoma (APA). APA primarily occurs in reproductive-aged women and has malignant potential. However, the clinical outcomes between TPA and APA, and efficacy of fertility-sparing treatments for APA are lack of consensus.

AIM

To investigate distinction of clinical outcomes between TPA and APA, and assess the effectiveness of different fertility-sparing treatments for APA.

METHODS

This systematic review included randomized controlled trials and observational studies via searching seven public databases. Conventional meta-analysis was conducted by pairwise comparison of clinical outcomes between APA and TPA. The Bayesian network meta-analysis was performed for evaluating the efficacy of different fertility-sparing therapies in APA. Pooling effect sizes were calculated as odds ratio (OR) with 95% confidence interval (95%CI). Surface under the cumulative ranking curve (SUCRA) values were used to rank efficacy of different fertility-sparing treatments for each outcome in APA.

RESULTS

This study compared 773 patients with TPA to 357 patients with APA. The network meta-analysis examined an additional 400 patients diagnosed with APA only. Results showed that patients with APA had a higher incidence of synchronous endometrial hyperplasia and endometrial cancer than those with TPA, but recurrence rates were similar across both groups. Medroxyprogesterone acetate at 10 mg daily after transcervical resection of polyps (TCRP) was most effective in achieving complete response (vs TCRP: OR = 7.15, 95%CI: 1.47-34.71, SUCRA = 78.3%) and reducing recurrence in APA (vs TCRP: OR = 6.29, 95%CI: 1.28-30.84, SUCRA = 65.3%). High-dose progestin therapy after TCRP showed potential in preventing endometrial cancer (SUCRA = 73.1%) in patients with APA but did not significantly outperform TCRP.

CONCLUSION

APA was linked to a greater malignant risk than TPA. Progestin-based therapy post-TCRP may prevent recurrence, endometrial cancer progression, and achieve complete response in APA patients.

Key Words: Atypical polypoid adenomyoma; Typical polypoid adenomyoma; Fertility-sparing management; Hysteroscopy; Progestin-based treatment

Core Tip: Atypical polypoid adenomyoma (APA) carries a significantly higher risk of concurrent endometrial hyperplasia and cancer than typical polypoid adenomyoma. In this first network meta-analysis comparing fertility-sparing treatments for APA, transcervical resection of polyps (TCRP) followed by medroxyprogesterone acetate (10 mg/day) proved most effective for achieving complete response and preventing recurrence. Additionally, TCRP with subsequent high-dose progestin was most effective for preventing cancer progression. These findings provide critical evidence-based guidance for managing APA in reproductive-aged women.

INTRODUCTION

Endometrial polypoid adenomyoma is a rare, well-circumscribed, and polypoid tumor of mixed epithelial and mesenchymal origin, predominantly composed of smooth muscle[1]. According to its histopathological characteristics, it is classified into typical and atypical subtypes[1,2]. Typical polypoid adenomyoma (TPA) is characterized by architecturally regular, benign endometrial glands dispersed within well-defined bundles of benign smooth muscle resembling the myometrium[3]. In contrast, atypical polypoid adenomyoma (APA) exhibits endometrial glands with varying degrees of architectural complexity and cytological atypia, embedded within a cellular fibromuscular stroma[4]. Therefore, APA has potential to coexist with or progress to endometrial hyperplasia or cancer[5]. Furthermore, despite the well-established histopathological distinction between APA and TPA[6], systematic comparative research on the clinical biological behavior of these two entities remains scarce.

Current therapeutic strategies for polypoid adenomyoma are tailored to the patient's desire for fertility preservation. Management options range from fertility-sparing treatments, including curettage and hysteroscopic polypectomy, to radical definitive surgery[7]. Given the high incidence of APA in reductive-aged women and its substantial risk of malignant progression, fertility-sparing treatment is necessary[8]. Conservative management strategies for APA exhibit considerable heterogeneity, encompassing the use of various types and dosages of progestins (such as oral medroxyprogesterone acetate (MPA) and the variable incorporation of hysteroscopic resection[9]. However, there is a notable absence of direct prospective comparative studies among these different strategies.

The present systematic review and network meta-analysis is designed to comprehensively evaluate and compare the long-term clinical profiles of TPA vs APA of the endometrium. Given the distinct clinical challenges and management dilemmas posed by APA, a pivotal objective of this study is to perform a network meta-analysis.

MATERIALS AND METHODS

Registration

This meta-analysis was reported according to the Systematic Review and Meta-Analysis Protocol and has been registered in the International Prospective Register of Systematic Reviews (PROSPERO), with the registration number CRD420251273319 on March 21, 2026.

Search Strategy and selection criteria

The investigators conducted a comprehensive systematic search across multiple databases, including PubMed, Web of Science, EMBASE, China National Knowledge Infrastructure (CNKI), Chinese BioMedical Literature Database (CBM), Wan Fang Data, and the China Science and Technology Journal Database (VIP). The search utilized the following terms: “Adenomyomatous polyps” OR “polypoid adenomyoma” OR “atypical polypoid adenomyoma” OR “typical polypoid adenomyoma” OR “adenomyoma of the endometrium”.

We included patients who were pathologically diagnosed with TPA and APA, with or without endometrial hyperplasia or endometrial cancer, from randomized controlled trials, cohort studies, case-controlled studies, and cross-sectional studies. In comparative analyses of clinical outcomes between TPA and APA, patients can undergo fertility-sparing treatments or radical surgeries, such as total abdominal hysterectomy with or without bilateral adnexectomy. For the comparison of efficacy among different fertility-preserving therapies for APA, the enrolled patients were diagnosed exclusively with APA. The study included the following conservative treatments for APA: Transcervical resection of polyps (TCRP), transcervical resection of endometrium, polyp resection without hysteroscopy (PWH), hormone therapy alone (including oral high-dose progestin (HDP), MPA, oral progesterone, and levonorgestrel-releasing intrauterine system, combined oral contraceptive), gonadotrophin-releasing hormone agonist, and conservative surgery combined with hormone treatment with or without maintenance therapy. Studies consisting of animal or cell experiments, case reports, scientific protocols, conference abstracts, or correspondence were excluded.

Study outcomes and definitions

The outcome parameters were defined as follows[10]: (1) Complete response: Complete removal of APA or TPA through surgical resection and/or hormone therapy, with no recurrence or progression to endometrial hyperplasia or cancer during the follow-up period; (2) Recurrence: The reappearance of polypoid lesions in the same or a different site from the initially diagnosed uterine area after pathological confirmation; (3) Concurrent endometrial hyperplasia or cancer: The presence of endometrial hyperplasia (with or without atypia) or cancer in relation to APA or TPA, either on the lesions, adjacent to the lesions, or in other endometrial sites not close to the lesions; and (4) Subsequent endometrial hyperplasia or cancer: The emergence of endometrial hyperplasia or cancer that was not present at the time of initial diagnosis of APA or TPA, occurring on or near the lesion, or in other endometrial sites without proximity to the lesion during the follow-up period.

Study selection and data extraction

All studies were retrieved based on the aforementioned search terms and imported into Endnote X9 to eliminate duplicates. Two investigators independently reviewed the titles and abstracts of the articles, initially excluding studies that did not involve patients with polypoid adenomyoma. Any disagreements were resolved through consensus. The selected literature was then meticulously reviewed according to the established inclusion and exclusion criteria.

Data extraction was conducted carefully to ensure accuracy and consistency. We collected key information from each study, including patient demographics, histopathological findings, characteristic of endometrial polypoid adenomyoma (including size and location), clinical symptoms, complications, and clinical outcomes. Two reviewers independently extracted the data, resolving any discrepancies through discussion. The collected data was then organized into a standardized format for analysis.

Quality assessment

Two authors independently reviewed each article included in the study and assessed potential biases using the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies. This 14-item tool evaluates key aspects of internal validity, including information bias, selection bias, confounding bias, and measurement bias. Each item was rated as “yes”, “no,” “cannot determine”, “not applicable”, or “not reported”. For cross-sectional studies, questions 6-7, 10, and 12-13 were marked as not applicable and excluded from scoring process. Any discrepancies between the two reviewers were resolved through discussion and consensus. A quality percentage score for each study was computed by dividing the number of “yes” responses by the total number of applicable questions. Studies were then categorized into quality ratings: Poor (< 60%), adequate/fair (60%-69%), good (70%-79%), and strong (≥ 80%)[11].

Data synthesis

The meta-analysis was performed using STATA software (Version 14.0) and R software (Version 4.3.1). For dichotomous outcomes, odds ratios (ORs) were employed as effect size measures, accompanied by 95% confidence intervals (95%CIs). We extracted and calculated the number of patients and events of interest in each group for studies where data were available, allowing us to derive the study-specific and pooled OR. Statistical heterogeneity, which may arise from clinical or methodological differences, was assessed using the χ² test and the I² test. When high heterogeneity was indicated (P < 0.05 or I² > 50%), a random-effects model was used to compute the pooled effect; otherwise, a fixed-effects model was applied[12,13]. Findings were summarized using forest plots, and publication bias was evaluated with funnel plots and Egger's tests.

In network meta-analysis, Bayesian methods were used to create network diagrams that compared various fertility-preserving treatments for APA by simultaneously analyzing both direct and indirect evidence. If a significant difference was detected (P < 0.05), a global inconsistency test was conducted to identify the source of any inconsistency[14]. In the absence of a significant difference, a random effects consistency model was employed to account for heterogeneity among the treatment effects across studies[14]. Within the Bayesian framework, we estimated the overall treatment rankings by calculating the surface under the cumulative ranking curve (SUCRA), where values range from 0 to 1, with higher SUCRA values indicating superior treatment regimens. Additionally, we presented treatment rankings through forest plots and league tables. Funnel plots and Egger’s test were generated to assess potential small-sample effects or publication bias[15]. Statistical significance was defined as P < 0.05.

RESULTS

Study selection

The PRISMA flowchart was illustrated in Figure 1. We retrieved a total of 32219 articles from several databases, including PubMed, EMBASE, Web of Science, CNKI, Wan Fang Data, CBM, and VIP. Automation tools removed 26865 records due to reasons such as animal or cell experiments, case reports, scientific protocols, conference abstracts, or correspondence. The remaining 5354 studies were imported into EndNote software to check for duplicates, leading to the exclusion of 4263 duplicate records. We meticulously reviewed the titles and abstracts of the 1091 remaining records. Studies were excluded for reasons absence of polypoid adenomyoma samples and ineligible study types. The full text of 71 studies was evaluated, resulting in the removal of studies that involved only TPA patients, APA patients treated exclusively with radical surgery, uncertain interventions, lack of disease outcomes, or incomplete/non-extractable data. Ultimately, 18 articles met the eligibility criteria for inclusion in the study. Among the reviewed articles, seven were identified as originating from China, focusing on the comparison of clinical outcomes between TPA and APA. Of these, two articles were published within the last five years[16,17], while 5 were published more than five years ago[18-22]. Additionally, twelve studies were selected to compare the efficacy of various fertility-sparing treatments in women diagnosed with APA alone. Among these, one study was conducted in the United States[6], two in Japan[7,9], and nine in China. Five of the Chinese studies were published within the past five years[23-27], with four studies published more than five years ago[22,28-30].

Figure 1 The flow diagram of study inclusion. CNKI: China National Knowledge Infrastructure; CBM: Chinese BioMedical Literature Database; VIP: China Science and Technology Journal Database; APA: Atypical polypoid adenomyoma.

Basic characteristics and risk bias of included literature

To compare clinical outcomes between the TPA and APA groups, we included a total of 773 patients with TPA and 357 patients with APA. The characteristics of these patients are detailed in Supplementary Table 1. For the evaluation of different fertility-preserving treatments in APA, we conducted a network meta-analysis that included 400 patients with APA alone. Information on patients receiving conservative treatment for APA was provided in Supplementary Table 2. The risk of bias in the included studies was summarized in Supplementary Table 3. Overall, the quality assessment of the literature revealed that 8 studies were rated as poor, 6 as adequate/fair, and 4 as good.

Comparison of clinical outcomes between TPA and APA

As shown in Figure 2A, there was no significant difference in recurrence rates between patients with APA and those with TPA (OR = 2.23, 95%CI: 0.43 to 11.49, I² = 79.8%). Given the high heterogeneity observed in the analysis of recurrence rates due to small sample sizes in various studies, a sensitivity analysis was performed to further pool the OR (OR = 1.999, 95%CI: 0.48-8.33; Supplementary Table 4). However, the incidence of concurrent both non-atypical and atypical endometrial hyperplasia, was higher in APA patients compared to TPA patients. Specifically, the odds of concurrent non-atypical hyperplasia were higher in APA patients (OR = 2.15, 95%CI: 1.45-3.15, I² = 38.2%; Figure 2B), and the odds of concurrent atypical hyperplasia were significantly higher as well (OR = 73.07, 95%CI: 17.59-304.1, I² = 0.0%; Figure 2C). Additionally, APA patients had a higher proportion of concurrent endometrial cancer (OR = 10.3, 95%CI: 4.26-24.90, I² = 0.0%; Figure 2D). The funnel plots and Egger's test revealed no publication bias for the recurrence rates (P = 0.156), concurrent atypical endometrial hyperplasia (P = 0.156), and concurrent endometrial cancer (P = 0.861; Supplementary Figure 1A-C). However, publication bias was noted in studies evaluating the rates of concurrent non-atypical endometrial hyperplasia (P = 0.029; Supplementary Figure 1D). A robustness analysis using the Metatrim method confirmed the significance of the results for concurrent non-atypical hyperplasia rates (Supplementary Figure 1D).

Figure 2 Comparison of clinical outcomes in patients with atypical polypoid adenomyoma and typical polypoid adenomyoma. A: Recurrence rates; B: Concurrent endometrial non-atypical hyperplasia rates; C: Concurrent endometrial atypical hyperplasia rates; D: Concurrent endometrial cancer. APA: Atypical polypoid adenomyoma; TPA: Typical polypoid adenomyoma; 95%CI: 95% confidence interval.

Comparison of effectiveness for different fertility-sparing treatments on APA

Complete response rates: The network map illustrated 13 different interventions from 9 studies, which were analyzed for their probability of achieving a complete response in patients with APA (Figure 3A). The results indicated that the combination of TCRP and MPA (10 mg/day) had the highest probability of achieving complete response (SUCRA = 78.3%; Figure 3B and Table 1). Compared to using TCRP alone, the TCRP + MPA (10 mg/day) regimen significantly improved complete response rates in APA patients (OR = 7.15, 95%CI: 1.47-34.71; Figure 3C). However, there were no significant differences in complete response rates between TCRP and other conservative treatments (Figure 3C). Notably, the intervention TCRP + MPA (10 mg/day) was more effective than TCRP (OR = 0.14, 95%CI: 0.03-0.68), progesterone therapy (PT; OR = 0.03, 95%CI: 0.00-0.98), polyp PWH combined with subsequent HDP therapy (PWH + HDP; OR = 0.01, 95%CI: 0.00-0.39), and PWH (OR = 0.03, 95%CI: 0.00-0.48) in achieving complete responses in APA patients (Figure 3D).

Figure 3 Network meta-analysis of conservative treatments for complete response rates in patients with atypical polypoid adenomyoma. A: Network structure showing treatment groups, with circle size indicating patient numbers and line thickness representing study comparisons; B: Rankings of interventions based on surface under the cumulative ranking curve analysis; C: Forest plot of achieving complete response for each treatment compared to transcervical resection of polyps; D: Net league diagram comparing treatment effectiveness. TCRP: Transcervical resection of polyps; LNG-IUS: Levonorgestrel-releasing intrauterine system; COC: Combined oral contraceptive; PT: Progesterone therapy; MPA: Medroxyprogesterone acetate; TCRE: Transcervical resection of endometrium; PWH: Polyp resection without hysteroscopy; HDP: High-dose progestin; MT: Hormone maintenance therapy; 95%CI: 95% confidence interval.

Table 1 Rankings of interventions based on surface under the cumulative ranking curve analysis.

Treatment	SUCRA (%)	PrBest (%)	Mean rank
TCRP	40.9	16.8	8.3
TCRP + HDP	49.9	15.8	7
TCRP + MPA (10 mg/day)	78.3	5.7	3.6
TCRP + COC	56.1	10	6.3
TCRP + LNG-IUS	66	14.7	5.1
TCRP + PT	42.3	10.3	7.9
TCRE	54.3	1.1	6.5
TCRE + PT	51.3	4.1	6.9
PWH	38.2	2.6	8.4
PWH + HDP	50.2	1	7
PWH + HDP + MT	49	3.8	7.1
LNG-IUS	40.3	10.1	8.2
PT	43.1	4	7.8

SUCRA: Surface under the cumulative ranking curve; PrBest: Probability of being the best; TCRP: Transcervical resection of polyps; LNG-IUS: Levonorgestrel-releasing intrauterine system; COC: Combined oral contraceptive; PT: Progesterone therapy; MPA: Medroxyprogesterone acetate; TCRE: Transcervical resection of endometrium; PWH: Polyp resection without hysteroscopy; HDP: High-dose progestin; MT: Hormone maintenance therapy.

Recurrence rates: The network analysis included 12 interventions from 10 studies aimed at reducing recurrence in patients with APA (Figure 4A). The results indicated that the combination of TCRP and MPA at 10 mg/day (SUCRA = 65.3%) had the highest probability of effectiveness (Figure 4B and Table 2). When compared to TCRP alone, TCRP + MPA (10 mg/day; OR = 6.29, 95%CI: 1.28-30.84) and TCRP + HDP (OR = 4.49, 95%CI: 1.54-13.06) significantly reduced recurrence rates in APA patients (Figure 4C). However, there was no statistically significant difference in recurrence rates between TCRP and other conservative treatments (Figure 4C). In comparison to using TCRP alone, both TCRP + MPA (10 mg/day; OR = 0.16, 95%CI: 0.03-0.78) and TCRP + HDP (OR = 0.22, 95%CI: 0.08-0.65) were more effective in decreasing APA recurrence (Figure 4D). Additionally, TCRP + MPA (10 mg/day; OR = 0.05, 95%CI: 0.00-0.62) and TCRP + HDP (OR = 0.07, 95%CI: 0.01-0.52) showed greater potential for reducing recurrence compared to using PWH alone (Figure 4D).

Figure 4 Network meta-analysis of conservative treatments for reducing recurrence rates in patients with atypical polypoid adenomyoma. A: Network structure of treatments; B: Rankings of interventions based on their ability to inhibit recurrence as indicated by surface under the cumulative ranking curve analysis; C: Forest plot of decreasing recurrence rates for treatments compared to transcervical resection of polyps; D: Net league diagram comparing the effectiveness of any two treatments. TCRP: Transcervical resection of polyps; LNG-IUS: Levonorgestrel-releasing intrauterine system; COC: Combined oral contraceptive; PT: Progesterone therapy; MPA: Medroxyprogesterone acetate; TCRE: Transcervical resection of endometrium; PWH: Polyp resection without hysteroscopy; HDP: High-dose progestin; MT: Hormone maintenance therapy; GnRH-a: Gonadotrophin-releasing hormone agonist; 95%CI: 95% confidence interval.

Table 2 Rankings of interventions based on their ability to inhibit recurrence as indicated by surface under the cumulative ranking curve analysis.

Treatment	SUCRA (%)	PrBest (%)	Mean rank
TCRP	45.3	2	7
TCRP + HDP	60.3	0	5.4
TCRP + MPA (10 mg/day)	65.3	0.2	4.8
TCRP + COC	47.4	4.8	6.8
TCRP + LNG-IUS	47.6	4.6	6.8
TCRP + GnRH-a	55.3	2.1	5.9
TCRP + PT	50.1	0.6	6.5
TCRE	53.9	22.8	6.1
TCRE + PT	58.2	17.1	5.6
PWH + HDP	43.2	17	7.3
PWH + HDP + MT	59.3	16.2	5.5
PWH	30.4	12.6	8.7

SUCRA: Surface under the cumulative ranking curve; PrBest: Probability of being the best; TCRP: Transcervical resection of polyps; LNG-IUS: Levonorgestrel-releasing intrauterine system; COC: Combined oral contraceptive; PT: Progesterone therapy; MPA: Medroxyprogesterone acetate; TCRE: Transcervical resection of endometrium; PWH: Polyp resection without hysteroscopy; HDP: High-dose progestin; MT: Hormone maintenance therapy; GnRH-a: Gonadotrophin-releasing hormone agonist.

Subsequent endometrial cancer: The network analysis included 13 interventions from 8 studies aimed at inhibiting EC development in patients with APA (Figure 5A). The results indicated that the combination of TCRP and HDP had the highest probability of effectiveness (SUCRA = 73.1%; Figure 5B and Table 3). However, there were no significant differences in reducing subsequent endometrial cancer rates between TCRP and TCRP + HDP (Figure 5C). Reference to TCRP, TCRP may be associated with a lower risk than PT in endometrial cancer progression among APA patients (Figure 5C). The analysis showed that TCRP was more effective than PT in preventing subsequent endometrial cancer (OR = 0.07, 95%CI: 0.01-0.97; Figure 5C). Conversely, the combination of TCRP + HDP demonstrated a greater therapeutic effect compared to PT (OR = 0.03, 95%CI: 0.00-0.45; Figure 5D).

Figure 5 Network meta-analysis of conservative treatments for inhibition of subsequent endometrial cancer in patients with atypical polypoid adenomyoma. A: Network structure of treatments; B: Rankings of interventions based on their ability to suppress endometrial cancer as indicated by surface under the cumulative ranking curve analysis; C: Forest plot of preventing endometrial cancer for treatments compared to transcervical resection of polyps; D: Net league diagram comparing the effectiveness of any two treatments. TCRP: Transcervical resection of polyps; LNG-IUS: Levonorgestrel-releasing intrauterine system; COC: Combined oral contraceptive; PT: Progesterone therapy; TCRE: Transcervical resection of endometrium; HDP: High-dose progestin; MT: Hormone maintenance therapy; PWH: Polyp resection without hysteroscopy; GnRH-a: Gonadotrophin-releasing hormone agonist; 95%CI: 95% confidence interval.

Table 3 Rankings of interventions based on their ability to suppress endometrial cancer as indicated by surface under the cumulative ranking curve analysis.

Treatment	SUCRA (%)	PrBest (%)	Mean rank
TCRP	47.1	15.8	7.7
TCRP + HDP	73.1	11.9	4.2
TCRP + COC	46.8	8	7.4
TCRP + LNG-IUS	47.5	10.9	7.3
TCRP + PT	55.6	8	6.3
TCRP + GnRH-a	38.4	8.7	8.4
TCRE	51	6.2	6.9
TCRE + PT	53.6	5.6	6.6
PWH + HDP	45.9	4.6	7.5
PWH + HDP + MT	57.8	5.9	6.1
PWH	54	10.2	6.5
LNG-IUS	39.8	3.6	8.2
PT	46.9	0.6	7.4

SUCRA: Surface under the cumulative ranking curve; PrBest: Probability of being the best; TCRP: Transcervical resection of polyps; LNG-IUS: Levonorgestrel-releasing intrauterine system; COC: Combined oral contraceptive; PT: Progesterone therapy; TCRE: Transcervical resection of endometrium; HDP: High-dose progestin; MT: Hormone maintenance therapy; PWH: Polyp resection without hysteroscopy; GnRH-a: Gonadotrophin-releasing hormone agonist.

Inconsistency, consistency, and reporting bias: In the network meta-analysis, the global inconsistency test revealed no significant inconsistencies in complete or partial response rates (χ² = 0.01, P = 0.938), recurrence rates (χ² = 0.32, P = 0.957), and subsequent endometrial cancer rates (χ² = 0.80, P = 0.843). As a result, a consistency model was applied to estimate the effect size. Consistent results were observed for TCRP + MPA (10 mg/day) compared to TCRP (P = 0.015) regarding complete response rates, for TCRP + HDP vs TCRP (P = 0.006) and TCRP + MPA (10 mg/day) vs TCRP (P = 0.023) concerning recurrence rates, and for TCRP vs PT (P = 0.047) in subsequent endometrial cancer rates (Supplementary Tables 5-7). The funnel plot and Egger’s test indicated no reporting bias for complete response rates and subsequent endometrial cancer rates (Supplementary Figure 2). Although potential publication bias was noted for recurrence rates (Supplementary Figure 2), robustness analyses using the Metatrim method confirmed that the statistically significant intervention effects persisted after bias correction (Supplementary Figure 3).

Meta-regression analysis: In light of the high proportion of low-quality studies, a meta-regression analysis was performed, incorporating study quality as a covariate. The results indicated no statistically significant difference in the effect of fertility-preserving treatments on complete response rates (β = 0.12, P = 0.537), recurrence rates (β = 0.76, P = 0.832), and subsequent endometrial cancer rates (β = 1.70, P = 0.721).

DISCUSSION

Our systematic analysis revealed that the rates of synchronous endometrial hyperplasia and cancer were notably higher in the APA group than in the TPA, though no significant difference in recurrence rates was observed. For achieving complete response and preventing recurrence in APA, the optimal conservative treatment was identified as MPA (10 mg/day) following TCRP. The most effective intervention for reducing the incidence of endometrial cancer in APA patients was found to be TCRP followed by HDP.

There is currently no consensus on the conservative treatment for APA due to the limited number of cases reported worldwide. Various conservative treatment options for APA have been explored, including progestin-based hormone therapy, with or without maintenance treatment, TCRP, dilation and curettage (D&C), or a combination of hormone therapy with TCRP or D&C[31]. However, no prospective trials have been conducted, and management strategies for APA lack standardization. TCRP has been shown to more accurately locate lesions and reduce the likelihood of residual tissue, resulting in improved pregnancy success rates compared to blind curettage and standard polypectomy[9]. This evidence supported our finding that polypectomy performed without hysteroscopy, with or without adjunctive progestin therapy, yielded poorer therapeutic outcomes in achieving complete response and minimizing recurrence of APA compared to TCRP combined with progestin. However, given the potential for pre-malignant and malignant changes in APA, a four-step hysteroscopic approach, rather than a simple hysteroscopic polypectomy, was recommended for young, nulliparous patients who wish to preserve their fertility[15]. The four-step technique includes: (1) Removal of the APA; (2) Excision of the endometrium adjacent to the APA; (3) Excision of the underlying myometrium; and (4) Multiple random endometrial biopsies[32]. In contrast, a hysterectomy was recommended for perimenopausal and menopausal patients[33].

The effectiveness of hormone therapy in achieving complete response and preventing recurrence of APA remains unclear. Nomura et al[34] found that continuing treatment with MPA in five cases reduced the incidence of recurrences. Chen et al[35] reported that APA patients who wished to conceive and received progestin treatment had no recurrences following hysteroscopic resection of the lesion. Biasioli et al[36] noted that while progestogens were not considered first-line treatments, they could prevent APA recurrence. Consistent with our findings, MPA at a dose of 10 mg/day and HDP following TCRP positively promoted the complete response and recurrence rates of APA. Unlike progesterone alone, adjunct progestin therapy (MPA 10 mg/day) after TCRP yielded more favorable outcomes in achieving complete response. However, a review indicated that recurrence rates of 10%-33% have been reported even with MPA or megestrol acetate therapy[9]. Additionally, one study found that TCRP alone results in significantly higher initial response rates, suggesting that TCRP without hormone therapy may offer a shorter treatment duration for patients seeking earlier conception[31]. A systematic review concluded that the therapeutic effects of MPA, dydrogesterone, and megestrol on recurrence and cure rates in APA patients were generally unsatisfactory[10].

The effectiveness of postoperative adjuvant progestin therapy in preventing endometrial hyperplasia or cancer in patients with APA remains a topic of debate. Tumor heterogeneity is increasingly recognized as key determinants of biological behaviors and therapeutic response. Regrettably, this study was unable to further explore the tumor heterogeneity between TPA and APA due to the limitation of systematic review and network meta-analysis. Single-cell transcriptomics and multi-omics are demanded for exploring tumor heterogeneity and explaining heterogeneity in treatment response between TPA and APA in the future studies[37-39]. A multicenter prospective study indicated that high-dose MPA might be necessary when atypical endometrial hyperplasia was identified during follow-up[40]. Given the expression of estrogen and progesterone receptors in both the stroma and epithelium of APA, elevated progesterone levels are thought to provide a protective effect against the progression to endometrial carcinoma[41,42]. Wang and Guo[43] found that APA patients with both low and high malignant potential experienced a normalization of endometrial tissue after 3-6 months of treatment with HDP, followed by maintenance therapy with MPA (20 mg/day) during the second half of the menstrual cycle for an additional 3 months. This suggested that maintenance therapy may facilitate tumor regression. The use of levonorgestrel-releasing intrauterine system has also been proposed as a maintenance option for APA[44]. Contrary to our findings, HDP as postoperative adjuvant therapy did not show a significant difference in preventing EC development compared to TCRP alone. Similarly, a systematic review indicated no significant difference in cancer progression rates between hysteroscopy alone and hysteroscopy with hormone maintenance[31]. However, our results suggested that TCRP or TCRP combined with subsequent HDP offered greater protection against cancer development than progesterone alone, underscoring the importance of complete resection by TCRP in inhibiting malignant progression in APA. Unfortunately, due to limited literature, we were unable to evaluate the most effective conservative treatments for preventing endometrial hyperplasia or promoting fertility.

Patients with APA were at a heightened risk for developing atypical hyperplasia and cancer, highlighting the necessity for vigilant clinical management and regular follow-up in this patient population[45]. A follow-up period of 5 years was recommended due to the peak recurrence of APA occurring 2 to 5 years after surgical intervention[34]. The follow-up protocol typically included D&C or hysteroscopic biopsy, combined with transvaginal ultrasound every 3 months during the first 2 years, followed by evaluations every 4-6 months for an additional 3 years, and annually thereafter[6]. If two consecutive endometrial biopsies shown normal results, routine hysteroscopy and endometrial biopsy can be discontinued[43].

This study had several limitations. First, we could not compare the development of endometrial hyperplasia or cancer, as well as fertility outcomes between TPA and APA, due to a lack of relevant research. Second, the limited number of studies on conservative treatments for APA may lead to overestimation of outcomes. Furthermore, the high proportion of low-quality studies could diminish the credibility and robustness of our results. Additionally, the incomplete reporting of key covariate information such as body mass index, number of deliveries, or symptoms in our enrolled studies potentially led to confounding bias. Lastly, we were unable to evaluate the effectiveness of fertility-sparing treatments in preventing endometrial hyperplasia and improving fertility in APA patients due to insufficient literature. The ultimate goal of fertility-sparing management is not only disease control but also successful reproductive outcomes. Future prospective studies or well-designed cohort studies should explicitly report post-treatment pregnancy and live birth rates, and consider standardized follow-up protocols that capture both oncological and reproductive endpoints.

CONCLUSION

In summary, patients with APA were at a greater risk for concurrent endometrial hyperplasia and cancer compared to those with TPA. Recurrence rates were similar for both groups. Among the fertility-sparing treatments, TCRP combined with a 10 mg/day course of MPA was the most effective for achieving complete response and preventing recurrence in APA. Additionally, TCRP followed by HDP was effective in preventing cancer progression in APA.

ACKNOWLEDGEMENTS

The authors would like to thank all the participants who contributed to this study.