Next Article in Journal
Radiomics Analysis of QUS Spectral Parametric Images for Predicting the Risk of Breast Cancer Recurrence
Previous Article in Journal
Digital Pathology with AI for Cervical Biopsies: Diagnostic Accuracy at the CIN2+ Threshold
Previous Article in Special Issue
Evaluation of Different Risk Factors for Metastatic Sentinel Lymph Nodes in Endometrial Cancer
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Cancers
Volume 17
Issue 23
10.3390/cancers17233809
cancers-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

When Should We Biopsy? A Risk Factor-Based Predictive Model for EIN and Endometrial Cancer

by
Shina Jang
* and
Sung Ook Hwang
Department of Obstetrics and Gynecology, Inha University Hospital, Inha University College of Medicine, Incheon 22212, Republic of Korea
*
Author to whom correspondence should be addressed.
Cancers 2025, 17(23), 3809; https://doi.org/10.3390/cancers17233809
Submission received: 28 September 2025 / Revised: 2 November 2025 / Accepted: 26 November 2025 / Published: 27 November 2025
(This article belongs to the Special Issue Clinical Research Advances in Endometrial Carcinoma)
Browse Figures
Versions Notes

Simple Summary

Endometrial cancer is becoming increasingly common worldwide, including among younger women. Because abnormal uterine bleeding often has benign causes, timely diagnosis of precancerous or malignant lesions can be challenging. This study analyzed over one thousand women who underwent hysteroscopy to identify clinical features predictive of endometrial intraepithelial neoplasia or cancer. Six independent risk factors were found: postmenopausal, obesity, abnormal uterine bleeding, polycystic ovary syndrome, multiple polyps, and increased endometrial thickness. The developed six-factor model showed good predictive performance and may help clinicians make timely, risk-based decisions for endometrial sampling. A body mass index of 30 kg/m2 or higher was confirmed as a meaningful threshold for Asian women, consistent with the World Health Organization definition of obesity.

Abstract

Background: The incidence of endometrial cancer (EC) is rising globally across all age groups. Endometrial intraepithelial neoplasia (EIN) is a premalignant lesion that may progress to EC if untreated. A clinical model is needed to efficiently identify women requiring prompt evaluation while avoiding unnecessary invasive procedures. Obesity is a major risk factor, but whether Asian women require a lower body mass index (BMI) cutoff than the World Health Organization (WHO) definition remains debated. This study aimed to develop a multivariable risk prediction model to guide biopsy decisions and determine an appropriate BMI cutoff for predicting EIN/EC risk among Asian women. Methods: This study retrospectively reviewed 1192 women aged ≥18 years who underwent hysteroscopy between 2010 and 2023 at a tertiary hospital. Candidate predictors included patient age, parity, BMI, postmenopausal status, symptom of abnormal uterine bleeding (AUB), diabetes mellitus, hypertension, polycystic ovary syndrome (PCOS), use of oral contraceptives, intrauterine devices, or menopausal hormone therapy, tamoxifen treatment, presence of multiple polyps, and endometrial thickness (EMT) measured by transvaginal ultrasonography. Multivariable logistic regression with stepwise selection identified independent predictors, and model stability and calibration were assessed using 1000 bootstrap resamples. Results: EIN/EC was diagnosed in 55 patients (4.6%). Six independent predictors were identified: postmenopausal status (adjusted odds ratio [aOR] 5.93, 95% CI 2.92–12.04), AUB (aOR 4.07, 1.51–10.97), multiple polyps (aOR 2.49, 1.33–4.66), PCOS (aOR 2.37, 1.08–5.22), BMI (aOR 1.13 per kg/m2; 1.84 per +5 kg/m2), and EMT (aOR 1.07 per mm, 1.02–1.11). When using categorical cutoffs, Obese II (BMI ≥ 30 kg/m2) and markedly increased EMT (≥20 mm) remained significant. Predicted probabilities ranged from 0.3% with no risk factors to 90.9% with all six risk factors present. The final model demonstrated good discrimination (AUC 0.79, 95% CI 0.73–0.86) and excellent calibration on bootstrap validation (mean absolute error 0.005). Conclusions: This six-factor clinical model stratifies individual EIN/EC risk using readily available variables and may guide timely, risk-based biopsy decisions by identifying high-risk patients while minimizing unnecessary procedures in low-risk cases. BMI ≥ 30 kg/m2 (WHO obesity threshold) was confirmed as a meaningful cutoff, but external validation is warranted to confirm its generalizability and clinical applicability.

1. Introduction

The global burden of endometrial cancer (EC) has been rising over the past decade, with considerable geographic variation in incidence and mortality rate [1,2]. Notably, population-based data indicate increasing EC incidence among younger women, underscoring the need for risk-based evaluation even when presenting symptoms are nonspecific [3,4]. Endometrial intraepithelial neoplasia (EIN), also referred to as atypical hyperplasia (AH), is recognized as a precursor lesion for EC. EIN is reported to co-occur with EC in approximately one-third of cases and carries an estimated 30% risk of progression to EC if left untreated [5,6,7].
Obesity is recognized as a major preventable determinant in the development of EC. In Asian populations, lower body mass index (BMI) thresholds than the standard World Health Organization (WHO) definition are often used to define metabolic risk, but it remains unclear whether a lower BMI cutoff should also be applied to define elevated EIN/EC risk. Several Asian studies have suggested that the BMI threshold associated with increased EC risk varies across populations [8,9,10]. Only a limited number of investigations have directly examined EIN or identified a consistent BMI threshold that remains independently predictive after adjusting for other factors. Defining a clinically applicable, population-appropriate BMI cutoff for EIN/EC risk therefore remains an open question [5,8,9].
Postmenopausal bleeding (PMB) is widely recognized as a key warning sign of underlying endometrial disease and should be investigated without delay. Beyond PMB, several patient level features are cited as risk factors for EC, including nulliparity, polycystic ovary syndrome (PCOS), and increased endometrial thickness (EMT) on transvaginal ultrasonography [11,12,13,14]. However, the magnitude of risk associated with these factors and optimal clinical thresholds vary widely across studies and populations, including those proposed for EMT. Abnormal uterine bleeding (AUB) in premenopausal and perimenopausal patients is frequently due to benign causes, which may lead to diagnostic delays if risk is underestimated in this younger group.
Accordingly, this study developed a multivariable risk prediction model to generate scenario-based probabilities of EIN/EC and to support decisions regarding the need for endometrial biopsy. The study also determined a clinically applicable, population-appropriate BMI cutoff for discriminating EIN/EC risk in an Asian cohort by analyzing BMI as both a continuous and categorical variable.

2. Materials and Methods

This investigation was designed as a retrospective observational review at Inha University Hospital (Incheon, Republic of Korea). Consecutive women aged 18 years or older who underwent hysteroscopic endometrial sampling between 2010 and 2023 were included. Institutional Review Board approval was obtained (IUH-IRB 2025-09-023).
Data on demographic and clinical characteristics were collected through review of electronic medical records. These included patient age; reproductive history (parity); BMI; postmenopausal status; AUB; comorbid conditions such as diabetes mellitus and hypertension; diagnosis of PCOS; ongoing use of oral contraceptives (COCs), intrauterine devices (IUDs), or menopausal hormone therapy (MHT); tamoxifen treatment in women with breast cancer; presence of multiple polyps; and EMT measured by ultrasonography. BMI was calculated from height and weight measured at outpatient visits. The categorization of BMI followed the Asia-Pacific guidelines jointly introduced by the WHO Western Pacific Regional Office (WPRO), the International Obesity Task Force, and the International Association for the Study of Obesity, while also considering recommendations from the WHO Expert Consultation addressing Asian BMI cutoff values [15,16]. Menopause was determined as the absence of menstruation for at least 12 consecutive months in women aged 50 years or older, and in those younger than 50, it was assessed by clinical symptoms and laboratory evidence. AUB and PCOS were identified based on patients’ initial medical records, diagnostic information, self-reported questionnaires, and, for PCOS, ultrasonographic findings. AUB was defined according to the FIGO Menstrual Disorders Group (2009), as non-pregnancy-related uterine bleeding that deviates from normal cyclic patterns in frequency, regularity, duration, or volume, including heavy menstrual bleeding, intermenstrual bleeding or spotting, and irregular cycles [17]. PCOS was defined according to the revised 2003 Rotterdam criteria, requiring at least two of the following: oligo/anovulation, clinical or biochemical hyperandrogenism, or polycystic ovarian morphology [18]. Active tamoxifen therapy was noted if the patient was receiving tamoxifen at the time of evaluation. The presence of two or more polyps visualized during hysteroscopic evaluation, confirmed through operative notes or surgical images, was considered multiple endometrial polyps. EMT was measured via transvaginal (or, when necessary, transrectal) ultrasonography in the mid-sagittal plane as the maximum double-layer thickness (mm). Focal intracavitary lesions (e.g., endometrial polyps, submucosal leiomyomas) were recorded by number and location [19]. Ultrasonographic examinations were performed either by attending physicians or by trained sonographers under physician supervision. All of the above clinical assessments and ultrasonographic examinations were conducted according to standardized protocols.
Indications for hysteroscopic endometrial biopsy included PMB; AUB refractory to medical therapy for at least 1–6 months in premenopausal women; abnormal ultrasound findings such as a suspected endometrial polyp, submucosal myoma, cystic change, or thickened endometrium; and tamoxifen use accompanied by AUB or abnormal endometrial findings on ultrasonography. Exclusion criteria were pregnancy-related bleeding, intraoperative biopsies performed during surgery for ectopic pregnancy or retained placenta, inadequate specimens or absence of endometrial tissue, and follow-up biopsies in patients previously diagnosed with EIN or EC undergoing conservative management. After applying these criteria, 1192 patients remained eligible for analysis.
Hysteroscopic procedures were performed under general anesthesia. Preoperatively, oral misoprostol was administered for cervical softening. Cervical dilation was achieved using Hegar dilators, and the endometrial cavity was evaluated under direct visualization. Most procedures were performed using a gravity-flow system with an infusion height maintained at 1–1.5 m. The average operating time ranged from 10 to 30 min. Depending on the intrauterine findings, therapeutic or diagnostic steps—including removal of polyps, excision of submucosal myomas, or targeted sampling of the endometrium—were undertaken using either a 5 mm rigid hysteroscope (scissors and forceps) or a 10 mm resectoscope fitted with a monopolar loop electrode (Karl Storz, Tuttlingen, Germany).
All endometrial specimens obtained during hysteroscopy were evaluated by board-certified gynecologic pathologists at the institutional pathology department. Histopathologic diagnoses were made according to the WHO classification. Biopsy specimens insufficient for diagnosis (n = 7) were excluded from the analysis. The main study endpoint was the identification of clinically relevant endometrial pathology, determined by histological confirmation of EIN/EC. Cases with benign endometrial histology served as the reference group.
Sample Size Adequacy: A post hoc assessment of sample size adequacy was conducted for the multivariable model following the framework of Riley et al. for prediction models [20,21]. Assuming a Cox–Snell R2 of 0.10, the pmsampsize calculation in R indicated a minimum required sample of 510 patients with 24 outcome events, corresponding to approximately 3.9 events per predictor (for six predictors). This analyzed cohort (N = 1192 with 55 events) exceeded these requirements, yielding about 9.2 events per predictor. This satisfied the recommended shrinkage criterion (≥0.90), and the difference between the model’s apparent and optimism-adjusted R2 was ≤0.05, indicating a low risk of overfitting.
Statistical Analysis: Baseline characteristics were compared between benign and pathology (EIN/EC) groups using the Wilcoxon rank-sum test (Mann–Whitney U) for continuous variables, chi-square test or Fisher’s exact test for categorical variables, and the Cochran–Armitage trend test for ordinal variables. Any variable demonstrating a univariable association with a p-value < 0.10 was included in a multivariable logistic regression model. Adjusted odds ratios (aORs) with 95% CIs were calculated to describe independent predictors. Scenario-based predicted probabilities of EIN/EC were then estimated from the final model to illustrate absolute risk under different clinical scenarios. Internal validation of the final model was performed using 1000 bootstrap resamples to assess model stability and calibration. All statistical procedures were implemented in R (version 4.5.0), with significance defined at p < 0.05 (two-sided).

3. Results

3.1. Cohort and Baseline Characteristics

Among 1192 women who underwent hysteroscopic endometrial sampling, 55 (4.6%) comprised the pathology group (EIN/EC) and 1137 (95.4%) the benign group (Table 1). Compared with the benign group, the pathology group had a higher BMI (median 28.8 vs. 23.2 kg/m2, p < 0.001) and a greater prevalence of Obese II (BMI ≥ 30 kg/m2; 43.6% vs. 9.8%, p < 0.001). The pathology group was also more likely to be postmenopausal (32.7% vs. 16.4%, p = 0.003), to present with AUB (90.9% vs. 72.0%, p = 0.003), and to have PCOS (25.5% vs. 9.6%, p < 0.001). Metabolic comorbidities were more frequent (diabetes 14.5% vs. 4.3%, p = 0.003; hypertension 29.1% vs. 11.7%, p < 0.001). Multiple endometrial polyps were more common (43.6% vs. 22.9%, p = 0.001). The median EMT was greater in the pathology group (13.0 mm vs. 10.4 mm, p = 0.002), and a markedly thick endometrium (EMT ≥ 20 mm) was also more frequent (23.6% vs. 6.6%, p < 0.001).

3.2. Histopathology

The histopathological diagnoses for all 1192 patients are summarized in Table 2. Benign endometrial findings accounted for 95.4% of cases (1137 out of 1192), whereas endometrial pathology (EIN/EC) was identified in 4.6% (55/1192). The most common benign diagnosis was endometrial polyp, which was present in 682 cases (57.2% of the total cohort). Other benign findings included disordered proliferative or secretory endometrium (8.1%), submucosal leiomyoma (7.0%), proliferative (7.0%), secretory (6.4%), atrophy (2.3%), endometrial hyperplasia without atypia (1.8%), and a variety of less frequent benign lesions.
Among the 55 cases with premalignant or malignant pathology, AH/EIN was diagnosed in 24 patients (2.0% of the total cohort). Endometrioid endometrial carcinoma was found in 27 patients (grades 1, 2, and 3 in 15, 7, and 5 cases, respectively, totaling 2.3%), and less common histologic types included serous carcinoma (3 cases, 0.2%) and clear cell carcinoma (1 case, 0.1%).

3.3. Multivariable Modeling

In the multivariable logistic regression model treating BMI and EMT as continuous variables, six predictors remained independently associated with the likelihood of EIN/EC (Figure 1, Table 3). These independent risk factors were: postmenopausal status (aOR 5.93, 95% CI 2.92–12.04), AUB (aOR 4.07, 95% CI 1.51–10.97), multiple endometrial polyps (aOR 2.49, 95% CI 1.33–4.66), PCOS (aOR 2.37, 95% CI 1.08–5.22), BMI (aOR 1.13 per 1 kg/m2 increase, 95% CI 1.08–1.19; aOR 1.84 per +5 kg/m2), and EMT (aOR 1.07 per 1 mm increase, 95% CI 1.02–1.11).
In addition, an alternative model specification was evaluated in which BMI was modeled using WHO Asia-Pacific categories and EMT was analyzed as categorical variables. The categorical model produced results consistent with the continuous-variable analysis (Table 3). Only the Obese II category (≥30 kg/m2) showed a significant association with EIN/EC (aOR 5.17, 95% CI 2.43–11.01), whereas Obese I, Overweight, and Underweight were not significant. For EMT, values ≥20 mm were associated with higher risk compared to <15 mm (aOR 2.74, 95% CI 1.23–6.06), while 15–19.9 mm was not different from the reference. In the final analysis, the multivariable logistic regression model demonstrated good discrimination, with an AUC of 0.79 (95% CI 0.73–0.86). The Hosmer–Lemeshow test confirmed adequate model fit (χ2 = 2.25, df = 4, p = 0.69). Internal validation with 1000 bootstrap resamples showed excellent calibration and model stability, yielding a mean absolute error of 0.005 and a mean squared error of 0.00008, indicating a minimal risk of overfitting (Figure 2).

3.4. Predicted Probabilities

Scenario-based predicted probabilities for EIN/EC were generated from the final multivariable model (Figure 3). These estimates demonstrated a wide separation in absolute risk across different clinical profiles. For example, the baseline scenario of a premenopausal woman with no AUB, no PCOS, no multiple polyps, BMI < 30 kg/m2, and EMT < 20 mm had an estimated risk of approximately 0.3%. In contrast, a premenopausal woman with AUB (but BMI < 30 and EMT < 20 mm) had an estimated 1.4–1.5% risk. For a postmenopausal woman with AUB, BMI < 30, and EMT < 20 mm, the risk was around 7.5% (95% CI approximately 4.0–13.4%), which increased to 17.7% if the same patient had an EMT ≥ 20 mm. In a premenopausal obese (Obese II) woman with AUB and EMT < 20 mm (and no PCOS or multiple polyps), the predicted risk was about 6.7%, rising to 16.1% if EMT was ≥20 mm. Finally, a high-risk constellation—defined by postmenopausal status, AUB, PCOS, multiple polyps, BMI ≥ 30 kg/m2, and EMT ≥ 20 mm—corresponded to an approximate 90.9% predicted probability of EIN/EC.
Overall, the largest increases in risk were associated with postmenopausal status, the presence of AUB, obesity (BMI ≥ 30 kg/m2), EMT ≥ 20 mm, and the presence of multiple polyps or PCOS. These trends were consistent with the dose–response relationships observed when BMI and EMT were treated as continuous variables in the model. A complete table of scenario-based predicted probabilities is provided in the Supplementary Materials (Table S1) for reference.

4. Discussion

In this study, postmenopausal status, AUB, PCOS, multiple endometrial polyps, higher BMI, and increased EMT were identified as independent predictors of EIN/EC in women undergoing hysteroscopic evaluation. Both BMI and EMT showed clear dose–response associations with risk when analyzed as continuous variables or stratified into categories, with risk particularly elevated at BMI ≥ 30 kg/m2 and EMT ≥ 20 mm. These findings reinforce that obesity, endometrial thickening, and other clinical factors contribute significantly to the risk of premalignant or malignant endometrial pathology, and they should be considered when assessing patients for possible endometrial sampling.
Obesity and EC risk. Obesity and the risk of EC have been extensively studied. Large-scale cohorts and meta-analyses consistently indicate that every 5 kg/m2 increment in BMI confers a 1.5–1.8-fold elevation in risk [22,23]. Furthermore, The World Cancer Research Fund has identified body fatness as a definitive causal factor for EC [24]. In the present analysis, each 1 kg/m2 rise in BMI was linked with an aOR of 1.13 (95% CI 1.08–1.19), which corresponds to an approximately 1.8-fold higher odds of EIN/EC per 5 kg/m2. This is consistent with prior evidence and confirms that adiposity contributes independently to endometrial carcinogenic risk.
Endometrial thickness. EMT likewise showed a dose–response relationship with pathology risk. In this study, each 1 mm increase in EMT was associated with an aOR of 1.07 (95% CI 1.02–1.11) for EIN/EC. When categorized, an EMT ≥ 20 mm conferred a higher risk of EIN/EC (approximately 2.7-fold compared to <15 mm), whereas intermediate thickness (15–19.9 mm) did not significantly differ from the <15 mm reference. In the context of PMB, an EMT ≤ 4 mm is widely accepted as a safe threshold below which significant pathology is unlikely [25,26]. In premenopausal or perimenopausal women, however, proposed EMT cutoffs (often between 8–14 mm) have shown inconsistent performance for cancer prediction [27]. The present findings indicate that markedly increased EMT (≥20 mm) can serve as a red flag for endometrial neoplasia in a referred cohort, whereas moderate thickening appears to confer little additional risk once other variables are taken into account.
Other factors. Within the model, AUB and postmenopausal status proved to be the most influential factors associated with EIN/EC risk. This aligns with clinical guidance that any bleeding in a postmenopausal woman warrants prompt evaluation and with data showing higher pretest probabilities of malignancy in symptomatic populations [25]. Notably, in this cohort AUB remained significantly associated with EIN/EC despite adjustment for BMI, EMT, and polyp status. This indicates that abnormal bleeding is best regarded as a direct clinical manifestation of underlying pathology rather than an indirect reflection of other factors.
PCOS remained independently associated with EIN/EC, in line with recent meta-analyses reporting a higher risk of EC among women with PCOS [11,28]. The underlying mechanisms are thought to involve chronic anovulation with prolonged unopposed estrogen exposure, insulin resistance with compensatory hyperinsulinemia, low-grade inflammation, and features of progesterone resistance [28,29]. These results highlight the need to consider PCOS in risk-stratified diagnostic approaches for women presenting with AUB.
Multiple endometrial polyps were also an independent predictor of EIN/EC. Although most polyps are benign, previous studies have reported that the presence of multiple polyps is associated with a higher risk of atypical hyperplasia or carcinoma compared with cases involving a single polyp [30,31]. This association may reflect an increased proliferative potential associated with greater lesion volume or an inflammatory background that promotes endometrial neoplastic transformation [31].
Several variables showed associations in univariable analysis but did not remain significant predictors in the multivariable model. Nulliparity, for example, lost significance after adjustment. This finding may be explained by overlapping mechanisms with PCOS and AUB, such as anovulatory cycles and prolonged estrogen exposure, together with the limited number of events, which reduced the ability to detect small independent effects. Similarly, diabetes mellitus and hypertension were associated with EIN/EC in unadjusted comparisons but were no longer significant after controlling for BMI, consistent with literature suggesting that the apparent links between these metabolic conditions and EC risk are largely confounded or mediated by adiposity and related metabolic disturbances [32,33].
Recent studies have developed predictive models for endometrial malignancy using various approaches. The BLUSH model combined clinical, ultrasonographic, and hysteroscopic features with high diagnostic accuracy, while the SIR-En model integrated systemic inflammatory indices with EMT and showed moderate performance [34,35]. This study adds to this growing body of work by presenting a model based on routinely available clinical factors, which demonstrated good discrimination and satisfactory calibration after internal validation.
Strengths and limitations. A key strength of this study is the relatively large sample size for a hysteroscopy-based analysis, together with histopathological confirmation of all outcomes. Risk factors were assessed using both continuous and categorical specifications, and BMI categories were defined according to region-specific criteria, which allowed identification of a potential threshold relevant to risk. The multivariable model was carefully evaluated to minimize overfitting, and it demonstrated favorable performance (AUC = 0.79; Hosmer–Lemeshow χ2 = 2.25, p = 0.69). Furthermore, internal bootstrap validation confirmed excellent calibration and model stability, supporting the reliability of these findings despite the absence of external validation.
Several limitations should be noted. The study was retrospective and performed at a single tertiary referral hospital, potentially limiting generalizability and introducing selection bias, since only women referred for hysteroscopy were included. The number of EIN/EC events was modest (n = 55). Although this exceeded the calculated requirement from the sample size analysis, it nonetheless constrained model complexity and reduced the precision of some estimates. External validation was not available, leaving uncertainty about the model’s performance in other populations. In addition, EMT in premenopausal women was measured without accounting for cycle phase, which may have introduced variability. These limitations indicate that caution is warranted when interpreting the absolute risk estimates, and they highlight the importance of future studies to validate and extend these findings.

5. Conclusions

In this cohort, BMI ≥ 30 kg/m2 and EMT ≥ 20 mm, along with postmenopausal status, AUB, PCOS, and multiple endometrial polyps, were independently associated with EIN/EC. A predictive model incorporating these six factors produced estimated probabilities ranging from <1% to nearly 91%, demonstrating wide variation in absolute risk clinical profiles. These findings suggest that common clinical variables can be integrated for individualized risk estimation and may aid in the risk stratification of women undergoing hysteroscopic evaluation. External validation of this model in diverse populations or clinical settings is warranted to confirm its generalizability and applicability.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cancers17233809/s1, Table S1: Scenario-based predicted probabilities of EIN/EC according to clinical risk factors.

Author Contributions

Conceptualization, S.J.; methodology, S.J.; formal analysis, S.J.; investigation, S.J.; data curation, S.J.; writing—original draft preparation, S.J.; writing—review and editing, S.O.H.; visualization, S.J.; supervision, S.O.H.; project administration, S.J. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by an Inha University Research Grant (funder: Inha University, funding number: 74980).

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Inha University Hospital (protocol code IUH-IRB 2025-09-023, date of approval: 29 September 2025).

Informed Consent Statement

Patient consent was waived by the Institutional Review Board due to the retrospective nature of the study and the use of anonymized data.

Data Availability Statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Abbreviations

The following abbreviations are used in this manuscript:
AHAtypical Hyperplasia
AUBAbnormal Uterine Bleeding
BMIBody Mass Index
ECEndometrial Cancer
EINEndometrial Intraepithelial Neoplasia
EMTEndometrial Thickness
MHTMenopausal Hormone Therapy
PCOSPolycystic Ovary Syndrome
PMBPostmenopausal Bleeding
WPROWorld Health Organization Western Pacific Regional Office

References

  1. Makker, V.; MacKay, H.; Ray-Coquard, I.; Levine, D.A.; Westin, S.N.; Aoki, D.; Oaknin, A.; Ledermann, J.A. Endometrial Cancer. Nat. Rev. Dis. Primers 2021, 7, 88. [Google Scholar] [CrossRef]
  2. Bray, F.; Laversanne, M.; Sung, H.; Ferlay, J.; Siegel, R.L.; Soerjomataram, I.; Jemal, A. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA A Cancer J. Clin. 2024, 74, 229–263. [Google Scholar] [CrossRef]
  3. Liu, L.; Habeshian, T.S.; Zhang, J.; Peeri, N.C.; Du, M.; De Vivo, I.; Setiawan, V.W. Differential trends in rising endometrial cancer incidence by age, race, and ethnicity. JNCI Cancer Spectr. 2023, 7, pkad001. [Google Scholar] [CrossRef] [PubMed]
  4. Wijayabahu, A.T.; Shiels, M.S.; Arend, R.C.; Clarke, M.A. Uterine cancer incidence trends and 5-year relative survival by race/ethnicity and histology among women under 50 years. Am. J. Obstet. Gynecol. 2024, 231, 526.e1–526.e22. [Google Scholar] [CrossRef] [PubMed]
  5. Nees, L.K.; Heublein, S.; Steinmacher, S.; Juhasz-Böss, I.; Brucker, S.; Tempfer, C.B.; Wallwiener, M. Endometrial Hyperplasia as a Risk Factor of Endometrial Cancer. Arch. Gynecol. Obstet. 2022, 306, 407–421. [Google Scholar] [CrossRef] [PubMed]
  6. Giannella, L.; Grelloni, C.; Bernardi, M.; Cicoli, C.; Lavezzo, F.; Sartini, G.; Natalini, L.; Bordini, M.; Petrini, M.; Petrucci, J.; et al. Atypical Endometrial Hyperplasia and Concurrent Cancer: A Comprehensive Overview on a Challenging Clinical Condition. Cancers 2024, 16, 914. [Google Scholar] [CrossRef]
  7. Doherty, M.T.; Sanni, O.B.; Coleman, H.G.; Cardwell, C.R.; McCluggage, W.G.; Quinn, D.; Wylie, J.; McMenamin, Ú.C. Concurrent and future risk of endometrial cancer in women with endometrial hyperplasia: A systematic review and meta-analysis. PLoS ONE 2020, 15, e0232231. [Google Scholar] [CrossRef]
  8. Miyata, H.; Shirai, K.; Muraki, I.; Iso, H.; Tamakoshi, A. Associations of Body Mass Index, Weight Change, Physical Activity, and Sedentary Behavior with Endometrial Cancer Risk Among Japanese Women: The Japan Collaborative Cohort Study. J. Epidemiol. 2021, 31, 621–627. [Google Scholar] [CrossRef]
  9. Kawachi, A.; Shimazu, T.; Budhathoki, S.; Sawada, N.; Yamaji, T.; Iwasaki, M.; Inoue, M.; Tsugane, S. Association of BMI and height with the risk of endometrial cancer, overall and by histological subtype: A population-based prospective cohort study in Japan. Eur. J. Cancer Prev. 2019, 28, 196–202. [Google Scholar] [CrossRef]
  10. Liu, F.; Cheung, E.C.; Lao, T.T. Obesity increases endometrial cancer risk in Chinese women with postmenopausal bleeding. Menopause 2021, 28, 1093–1098. [Google Scholar] [CrossRef]
  11. Johnson, J.-E.; Daley, D.; Tarta, C.; Stanciu, P.I. Risk of endometrial cancer in patients with polycystic ovarian syndrome: A meta-analysis. Oncol. Lett. 2023, 25, 168. [Google Scholar] [CrossRef] [PubMed]
  12. Jaya-Bodestyne, S.L.; Goh, M.S.; Gwan, M.C.H.; Chonkar, S.P.; Merchant, K.; Mathur, M. To do or not to do?—Endometrial biopsy in younger women with abnormal uterine bleeding. Eur. J. Obstet. Gynecol. Reprod. Biol. X 2025, 25, 100368. [Google Scholar] [CrossRef] [PubMed]
  13. Katagiri, R.; Iwasaki, M.; Abe, S.K.; Islam, R.; Rahman, S.; Saito, E.; Merritt, M.A.; Choi, J.-Y.; Shin, A.; Sawada, N.; et al. Reproductive Factors and Endometrial Cancer Risk Among Women. JAMA Netw. Open 2023, 6, e2332296. [Google Scholar] [CrossRef] [PubMed]
  14. Wang, L.; Quan, S.; Bai, E.; Yang, X. Analysis of clinical data of different endometrial pathological types in perimenopausal women with abnormal uterine bleeding. Front. Oncol. 2024, 14, 1370681. [Google Scholar] [CrossRef]
  15. World Health Organization; International Association for the Study of Obesity; International Obesity Task Force. The Asia-Pacific Perspective: Redefining Obesity and Its Treatment; Health Communications Australia: Sydney, Australia, 2000. [Google Scholar]
  16. Tan, K. Appropriate Body-Mass Index for Asian Populations and Its Implications for Policy and Intervention Strategies. Lancet 2004, 363, 157–163. [Google Scholar] [CrossRef]
  17. Whitaker, L.; Critchley, H.O. Abnormal Uterine Bleeding. Best Pract. Res. Clin. Obstet. Gynaecol. 2016, 34, 54–65. [Google Scholar] [CrossRef]
  18. ESHRE/ASRM Rotterdam Consensus Workshop Group. Revised 2003 Consensus on Diagnostic Criteria and Long-Term Health Risks Related to Polycystic Ovary Syndrome (PCOS). Fertil. Steril. 2004, 81, 19–25. [Google Scholar] [CrossRef]
  19. Liu, W.; Bai, W. Association of endometrial thickness with lesions in postmenopausal asymptomatic women: Risk factors and diagnostic thresholds. BMC Women’s Health 2025, 25, 105. [Google Scholar] [CrossRef]
  20. Riley, R.D.; Snell, K.I.; Ensor, J.; Burke, D.L.; Harrell, F.E., Jr.; Moons, K.G.M.; Collins, G.S. Minimum Sample Size for Developing a Multivariable Prediction Model: Part II—Binary and Time-to-Event Outcomes. Stat. Med. 2019, 38, 1276–1296. [Google Scholar] [CrossRef]
  21. Riley, R.D.; Ensor, J.; I E Snell, K.; E Harrell, F.; Martin, G.P.; Reitsma, J.B.; Moons, K.G.M.; Collins, G.; van Smeden, M. Calculating the sample size required for developing a clinical prediction model. BMJ 2020, 368, m441. [Google Scholar] [CrossRef]
  22. Hazelwood, E.; Sanderson, E.; Tan, V.Y.; Ruth, K.S.; Frayling, T.M.; Dimou, N.; Gunter, M.J.; Dossus, L.; Newton, C.; Ryan, N.; et al. Identifying molecular mediators of the relationship between body mass index and endometrial cancer risk: A Mendelian randomization analysis. BMC Med. 2022, 20, 125. [Google Scholar] [CrossRef]
  23. Renehan, A.G.; Tyson, M.; Egger, M.; Heller, R.F.; Zwahlen, M. Body-mass index and incidence of cancer: A systematic review and meta-analysis of prospective observational studies. Lancet 2008, 371, 569–578. [Google Scholar] [CrossRef]
  24. World Cancer Research Fund/American Institute for Cancer Research. Diet, Nutrition, Physical Activity and Endometrial Cancer; World Cancer Research Fund International: London, UK, 2018. [Google Scholar]
  25. American College of Obstetricians and Gynecologists. ACOG Committee Opinion No. 734: The Role of Transvaginal Ultraso-nography in Evaluating the Endometrium of Women with Postmenopausal Bleeding. Obstet. Gynecol. 2018, 131, 945–946. [Google Scholar] [CrossRef]
  26. Chee, R.K.; Koshy, R.M.; Haidey, J.; Murad, M.H.; Low, G.; Wilson, M.P. Re-evaluating Endometrial Thickness in Symptomatic Postmenopausal Patients for Excluding Cancer: Systematic Review and Meta-Analysis. J. Am. Coll. Radiol. 2024, 22, 425–435. [Google Scholar] [CrossRef]
  27. Kaur, H.; Qadri, S.; Nevill, A.M.; Ewies, A.A. The optimal endometrial thickness threshold for prediction of endometrial cancer in postmenopausal women without bleeding remains uncertain–Systematic review and meta-analysis. J. Gynecol. Obstet. Hum. Reprod. 2024, 53, 102831. [Google Scholar] [CrossRef]
  28. Amiri, M.; Bidhendi-Yarandi, R.; Fallahzadeh, A.; Marzban, Z.; Tehrani, F.R. Risk of endometrial, ovarian, and breast cancers in women with polycystic ovary syndrome: A systematic review and meta-analysis. Int. J. Reprod. Biomed. 2022, 20, 893. [Google Scholar] [CrossRef]
  29. Emanuel, R.H.; Roberts, J.; Docherty, P.D.; Lunt, H.; Campbell, R.E.; Möller, K. A review of the hormones involved in the endocrine dysfunctions of polycystic ovary syndrome and their interactions. Front. Endocrinol. 2022, 13, 1017468. [Google Scholar] [CrossRef]
  30. Jang, S.; Hwang, S.O. The risk factors for premalignant and malignant endometrial polyps in premenopausal and postmenopausal women and trends over the past decade: A retrospective study in a single center, South Korea. Eur. J. Obstet. Gynecol. Reprod. Biol. 2024, 295, 118–123. [Google Scholar] [CrossRef] [PubMed]
  31. McGowan, M.A.; Davies, J.M.; Addley, S.; Honeyman, L.J.; Kolhe, S.N.; Phillips, A.J. Does the presence of single compared to multiple endometrial polyps alter the risk of cancer in post-menopausal women? Eur. J. Obstet. Gynecol. Reprod. Biol. 2022, 279, 118–121. [Google Scholar] [CrossRef] [PubMed]
  32. Friberg, E.; Orsini, N.; Mantzoros, C.S.; Wolk, A. Diabetes mellitus and risk of endometrial cancer: A meta-analysis. Diabetologia 2007, 50, 1365–1374. [Google Scholar] [CrossRef] [PubMed]
  33. Arthur, R.S.; Kabat, G.C.; Kim, M.Y.; Wild, R.A.; Shadyab, A.H.; Wactawski-Wende, J.; Ho, G.Y.F.; Reeves, K.W.; Kuller, L.H.; Luo, J.; et al. Metabolic syndrome and risk of endometrial cancer in postmenopausal women: A prospective study. Cancer Causes Control 2019, 30, 355–363. [Google Scholar] [CrossRef]
  34. Vitale, S.G.; Angioni, S.; D’Alterio, M.N.; Ronsini, C.; Saponara, S.; De Franciscis, P.; Riemma, G. Risk of endometrial malignancy in women treated for breast cancer: The BLUSH prediction model—Evidence from a comprehensive multicentric retrospective cohort study. Climacteric 2024, 27, 482–488. [Google Scholar] [CrossRef]
  35. Ronsini, C.; Iavarone, I.; Vastarella, M.G.; Della Corte, L.; Andreoli, G.; Bifulco, G.; Cobellis, L.; De Franciscis, P. SIR-EN-New Biomarker for Identifying Patients at Risk of Endometrial Carcinoma in Abnormal Uterine Bleeding at Menopause. Cancers 2024, 16, 3567. [Google Scholar] [CrossRef]
Cancers 17 03809 g001
Figure 1. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) for predictors of EIN/EC from the final multivariable logistic regression model, with BMI (per 1 kg/m2) and endometrial thickness (EMT; per 1 mm) entered as continuous predictors. Significance codes: *** p < 0.001; ** p < 0.01; * p < 0.05 (two-sided Wald tests). Abbreviations: MP, postmenopausal; AUB, abnormal uterine bleeding; PCOS, polycystic ovary syndrome; BMI, body mass index.
Figure 1. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) for predictors of EIN/EC from the final multivariable logistic regression model, with BMI (per 1 kg/m2) and endometrial thickness (EMT; per 1 mm) entered as continuous predictors. Significance codes: *** p < 0.001; ** p < 0.01; * p < 0.05 (two-sided Wald tests). Abbreviations: MP, postmenopausal; AUB, abnormal uterine bleeding; PCOS, polycystic ovary syndrome; BMI, body mass index.
Cancers 17 03809 g001
Cancers 17 03809 g002
Figure 2. Bootstrap-corrected calibration plot for the final logistic regression model predicting endometrial intraepithelial neoplasia or endometrial cancer (EIN/EC). The dashed line represents perfect calibration (ideal line), the dotted line shows the apparent performance of the model, and the solid line indicates the bias-corrected performance obtained through 1000 bootstrap resamples. The close alignment between the bias-corrected and ideal lines demonstrates excellent agreement between predicted and observed probabilities.
Figure 2. Bootstrap-corrected calibration plot for the final logistic regression model predicting endometrial intraepithelial neoplasia or endometrial cancer (EIN/EC). The dashed line represents perfect calibration (ideal line), the dotted line shows the apparent performance of the model, and the solid line indicates the bias-corrected performance obtained through 1000 bootstrap resamples. The close alignment between the bias-corrected and ideal lines demonstrates excellent agreement between predicted and observed probabilities.
Cancers 17 03809 g002
Cancers 17 03809 g003
Figure 3. Predicted probability of EIN/EC across clinical scenarios. Bars show model-predicted risk with error bars representing 95% CIs for selected patient profiles, ordered by increasing risk. Gray bars indicate scenarios without Obese II; red bars indicate scenarios with Obese II (BMI ≥ 30 kg/m2). Abbreviations: MP, postmenopausal; AUB, abnormal uterine bleeding; PCOS, polycystic ovary syndrome; EMT, endometrial thickness.
Figure 3. Predicted probability of EIN/EC across clinical scenarios. Bars show model-predicted risk with error bars representing 95% CIs for selected patient profiles, ordered by increasing risk. Gray bars indicate scenarios without Obese II; red bars indicate scenarios with Obese II (BMI ≥ 30 kg/m2). Abbreviations: MP, postmenopausal; AUB, abnormal uterine bleeding; PCOS, polycystic ovary syndrome; EMT, endometrial thickness.
Cancers 17 03809 g003
Table 1. Baseline demographic and clinical characteristics by outcome.
Table 1. Baseline demographic and clinical characteristics by outcome.
GroupBenign
(N = 1137)		EIN/EC
(N = 55)		Total
(N = 1192)		p
Age43.0 [36.0;49.0]44.0 [34.5;61.0]43.0 [36.0;49.0]0.377
Postmenopausal status 0.003
no951 (83.6%)37 (67.3%)988 (82.9%)
yes186 (16.4%)18 (32.7%)204 (17.1%)
BMI (kg/m2)23.2 [21.1;26.2]28.8 [22.8;33.0]23.3 [21.2;26.6]<0.001
BMI_WPRO 1 <0.001
Under weight53 (4.7%)2 (3.6%)55 (4.6%)
Normal491 (43.2%)13 (23.6%)504 (42.3%)
Overweight200 (17.6%)4 (7.3%)204 (17.1%)
Obese I282 (24.8%)12 (21.8%)294 (24.7%)
Obese II111 (9.8%)24 (43.6%)135 (11.3%)
Nulliparous 0.015
no824 (72.5%)31 (56.4%)855 (71.7%)
yes313 (27.5%)24 (43.6%)337 (28.3%)
AUB 0.003
absent318 (28.0%)5 (9.1%)323 (27.1%)
present819 (72.0%)50 (90.9%)869 (72.9%)
PCOS <0.001
no1028 (90.4%)41 (74.5%)1069 (89.7%)
yes109 (9.6%)14 (25.5%)123 (10.3%)
Diabetes 0.003
absent1088 (95.7%)47 (85.5%)1135 (95.2%)
present49 (4.3%)8 (14.5%)57 (4.8%)
Hypertension <0.001
absent1004 (88.3%)39 (70.9%)1043 (87.5%)
present133 (11.7%)16 (29.1%)149 (12.5%)
COCs/IUD/MHT use 0.213
no1028 (90.4%)53 (96.4%)1081 (90.7%)
yes109 (9.6%)2 (3.6%)111 (9.3%)
Multiple polyps 0.001
absent877 (77.1%)31 (56.4%)908 (76.2%)
present260 (22.9%)24 (43.6%)284 (23.8%)
Tamoxifen use 1.000
no1077 (94.7%)53 (96.4%)1130 (94.8%)
yes60 (5.3%)2 (3.6%)62 (5.2%)
EMT (mm)10.4 [7.1;14.0]13.0 [8.4;19.2]10.5 [7.2;14.2]0.002
EMT (category) <0.001
<15 mm898 (79.0%)33 (60.0%)931 (78.1%)
15 mm–19.9 mm164 (14.4%)9 (16.4%)173 (14.5%)
≥20 mm75 (6.6%)13 (23.6%)88 (7.4%)
Abbreviations: BMI, body mass index; AUB, abnormal uterine bleeding; PCOS, polycystic ovary syndrome; EIN, endometrial intraepithelial neoplasia; EC, endometrial carcinoma; EMT, endometrial thickness; COCs, combined oral contraceptives; IUD, intrauterine device (levonorgestrel-releasing intrauterine system, LNG-IUS); MHT, menopausal hormone therapy; WPRO, World Health Organization Western Pacific Regional Office. 1 BMI category (WPRO): categorized according to the WHO Western Pacific Regional Office (WPRO) classification: Underweight (<18.5 kg/m2), Normal (18.5–22.9 kg/m2), Overweight (23.0–24.9 kg/m2), Obese I (25.0–29.9 kg/m2), and Obese II (≥30 kg/m2).
Table 2. Histopathological diagnoses in the study population (N = 1192).
Table 2. Histopathological diagnoses in the study population (N = 1192).
Benign group1137 (95.4%)
Endometrial polyp682 (57.2%)
Disordered proliferative/secretory phase96 (8.1%)
Submucosal leiomyoma84 (7.0%)
Proliferative phase84 (7.0%)
Secretory phase76 (6.4%)
Atrophy27 (2.3%)
Endometrial hyperplasia without atypia (EH)21 (1.8%)
Inactive17 (1.4%)
Acute and chronic endometritis16 (1.3%)
Glandular and stromal breakdown12 (1.0%)
Adenomyomatous polyp7 (0.6%)
Others *15 (1.3%)
Premalignant/malignant group55 (4.6%)
Atypical hyperplasia/endometrial intraepithelial neoplasia (AH/EIN)24 (2.0%)
Endometrioid adenocarcinoma grade 1 (EC)15 (1.3%)
Endometrioid adenocarcinoma grade 2 (EC)7 (0.6%)
Endometrioid adenocarcinoma grade 3 (EC)5 (0.4%)
Serous carcinoma3 (0.2%)
Clear cell carcinoma1 (0.1%)
Counts (percentages) are shown for the entire cohort. * Benign “Others” include benign endometrial tissue (n = 8), decidualized stroma (n = 2), progestin effect (n = 2), granulomatous inflammation with necrosis (n = 2), and epithelioid mesenchymal tumor (n = 1).
Table 3. Multivariable logistic regression results (final model), comparing continuous vs. categorical specifications of BMI and EMT.
Table 3. Multivariable logistic regression results (final model), comparing continuous vs. categorical specifications of BMI and EMT.
VariableContinuous ModelCategorical Model
aOR (95% CI)paOR (95% CI)p
Postmenopausal status5.93 (2.92–12.04)<0.0015.96 (2.87–12.37)<0.001
BMI (per 1 kg/m2)1.13 (1.08–1.19)<0.001
BMI_WPRO: Obese II vs. Normal 5.17 (2.43–11.01)<0.001
BMI_WPRO: Under vs. Normal 1.50 (0.32–7.07)0.610
BMI_WPRO: Obese I vs. Normal 1.17 (0.50–2.70)0.720
BMI_WPRO: Over vs. Normal 0.64 (0.20–2.07)0.461
AUB4.07 (1.51–10.97)0.0054.11 (1.51–11.2)0.006
Multiple polyps2.49 (1.33–4.66)0.0052.97 (1.59–5.56)<0.001
PCOS2.37 (1.08–5.22)0.0322.91 (1.36–6.22)0.006
EMT (mm)1.07 (1.02–1.11)0.004
EMT ≥ 20 mm vs. EMT < 15 mm 2.74 (1.23–6.06)0.013
15 mm–19.9 mm vs. EMT < 15 mm 1.11 (0.49–2.49)0.804
Abbreviations: aOR, adjusted odds ratio; BMI, body mass index; WPRO, World Health Organization Western Pacific Regional Office; AUB, abnormal uterine bleeding; PCOS, polycystic ovary syndrome; EMT, endometrial thickness.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Jang, S.; Hwang, S.O. When Should We Biopsy? A Risk Factor-Based Predictive Model for EIN and Endometrial Cancer. Cancers 2025, 17, 3809. https://doi.org/10.3390/cancers17233809

AMA Style

Jang S, Hwang SO. When Should We Biopsy? A Risk Factor-Based Predictive Model for EIN and Endometrial Cancer. Cancers. 2025; 17(23):3809. https://doi.org/10.3390/cancers17233809

Chicago/Turabian Style

Jang, Shina, and Sung Ook Hwang. 2025. "When Should We Biopsy? A Risk Factor-Based Predictive Model for EIN and Endometrial Cancer" Cancers 17, no. 23: 3809. https://doi.org/10.3390/cancers17233809

APA Style

Jang, S., & Hwang, S. O. (2025). When Should We Biopsy? A Risk Factor-Based Predictive Model for EIN and Endometrial Cancer. Cancers, 17(23), 3809. https://doi.org/10.3390/cancers17233809

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop