Asymptomatic endometrial thickening in postmenopausal women: predictor of malignant pathology?

Abstract It is not standardised what is the endometrial thickness that discriminates between normal and potentially malignant. The objective of this study was to determine the endometrial thickness cut-off point from which the risk of endometrial cancer (EC) increases in asymptomatic postmenopausal women; and to evaluate the risk factors linked to malignant endometrial pathology as well as other associated ultrasound findings. This was a retrospective observational study that included hysteroscopies performed at the Hospital Materno-Infantil on 267 asymptomatic menopausal women with an increase in endometrial thickness (AET) >5 mm, from 2015 to 2019. The results shows that the prevalence of malignant pathology in asymptomatic postmenopausal women with a casual finding of endometrial thickening was 3.7%. This percentage was 16.3% when the cut-off point of AET was established at 10 mm. There was a significant association for the diagnosis of malignant pathology with this cut-off point. There is a significant association between the 10 mm endometrial thickness cut-off point from which the risk of EC increases in asymptomatic postmenopausal women. Impact statement What is already known on this subject? Several studies have established the cut-off point for asymptomatic endometrial thickening (AET) for atypical endometrial hyperplasia and endometrial cancer at 10 mm. Although no cut-off point has optimal accuracy for the diagnosis of malignant endometrial pathology, it has been found that with a cut-off value of AET >10 mm no cases are missed. Likewise, a cut-off point of AET > 11 mm may provide a balance between cancer detection and histopathological workup extension. What do the results of this study add? A significant association was found at the cut-off point of AET > 10 mm, which suggests that screening postmenopausal women at this thickness is acceptable and unlikely to miss cases of endometrial hyperplasia and endometrial cancer. What are the implications of these findings for clinical practice and/or further research? After analysing our results we can conclude, like other published studies, that by establishing a cut-off point of 10 mm we obtain a good discrimination between benign and malignant pathology, which would allow us to diagnose 100% of malignant pathology. Above this cut-off point, the risk of endometrial cancer increases, and it would therefore be advisable to extend the study. A multicentre study is needed to confirm the cut-off point at which the risk of endometrial cancer increases in postmenopausal women with asymptomatic endometrial thickening.


Introduction
Transvaginal ultrasound (USTV) is a tool that complements the gynecological examination and allows an optimal approach to the uterus, as well as an excellent assessment of the endometrium.Several authors (Ciatto et al. 1995, Fleischer et al. 2001, Gerber et al. 2001) have assessed the usefulness of transvaginal ultrasound in endometrial cancer (EC) screening in asymptomatic women.The Canadian Cancer Society, in 2010 and the American Cancer Society in 2008, do not recommend screening for EC by USTV as there is no evidence that it reduces mortality.
Asymptomatic Endometrial Thickening (AET) is defined as an endometrium greater than 5 mm, without evidence of uterine bleeding, in a woman after menopause (Gambacciani et al. 2004 andWolfman et al. 2010).There is currently no consensus on the significance of this incidental finding, which implies variations in clinical practice (Giannella et al. 2014).Likewise, it is considered that the endometrial thickness limit of 5 mm used in menopausal women with uterine bleeding would not be applicable to asymptomatic women (Goldstein et al. 2001).
It has been described that the probability of malignant neoplasia and atypical endometrial hyperplasia in this population is estimated to be around 0.007% and 4.9%, as discussed by several authors (Goldstein et al. 2001, Gambacciani et al. 2004, Aedo et al. 2008a, 2008b, Giannella et al. 2014), with which, Goldstein (2010), states that the study of AET should be defined according to the conditions of each woman.In 2009, the American College of Obstetricians and Gynaecologists (ACOG) stated that there is no evidence to recommend routine screening in women with AET.
Other studies show that the likelihood of finding EC is proportional to greater endometrial thickness.Smith-Bindman et al. (1998), estimated the risk of EC with a threshold value of 11 mm, obtaining a prevalence of EC in AET > 11 mm of 6.7% versus 0.002% in AET < 11 mm.Likewise, Gerber, in 2001, using 10 mm as a cut-off point for AET, observed a prevalence of 13% of EC, which contrasts with positive predictive values of less than 5% when cut-off points less than or equal to 6 mm are used by other authors (Ciatto et al. 1995, Fleischer et al. 2001, Gambacciani et al. 2004, Tsuda et al. 2005, Smith et al. 2008).
In another study by Ghoubara et al. in 2018, the prevalence of endometrial atypical hyperplasia and cancer was 4.9% versus 24.7% for benign endometrial polyps in asymptomatic postmenopausal women with AET > 4 mm.
According to Schmidt et al. 2009, andWolfman et al. 2010, the finding of AET also allows the presence of benign pathology such as polyps and endometrial hyperplasia to be suspected.The estimated prevalence of endometrial polyps (EP) in women with postmenopausal bleeding varies between 13% and 50%, as discussed by several authors (Tjarks and Van Voorhis 2000, Fern� andez-Parra et al. 2006, Aedo et al. 2008a, 2008b, Ferrazzi et al. 2009, Baiocchi et al. 2009, Gregoriou et al. 2009).Polyps in menopausal women are a risk factor for developing EC, especially if the size exceeds 15 mm, as 0.5%-4.8% of polyps have findings of malignancy (Bakour et al. 2000, Savelli et al. 2003, Antunes et al. 2007, Dom� ıngues et al. 2009).According to S� ocrates Aedo et al. 2013, women with asymptomatic EP findings should be evaluated according to size, age and other risk factors to define their management.
In addition to endometrial thickness, existing risk factors for endometrial cancer should be taken into account.
Tamoxifen (TMX) is a selective oestrogen receptor modulator (SERM) with antagonistic action on breast tissue and agonist action on the endometrium (Machado et al. 2005, Abdulkareem and Zurmi 2012, Cuzick et al. 2013).TMX use is a risk factor for benign endometrial pathology such as endometrial polyps and EC, with a risk of 2.3‰ women, as discussed by several authors (Berliere et al. 2000, Fishman et al. 2006, Schmidt 2006).Screening with USTV for endometrial pathology in asymptomatic users with TMX treatment is currently not recommended by ACOG (2006).
Taking into account the above, it is recommended to extend the histological study in patients with endometrial thickening >5 mm associated with other ultrasound findings such as increased vascularisation, endometrial heterogenicity, intracavitary fluid or endometrial thickening greater than 10-11 mm (Opolskiene et al. 2007).

Materials and methods
The main objective of this study is to determine the endometrial thickness cut-off point above which the risk of endometrial cancer increases in asymptomatic postmenopausal women.
As a secondary objective, we evaluated the risk factors associated with malignant endometrial pathology as well as other associated ultrasound findings.
For this purpose, a retrospective observational study was performed from January 2015 to December 2019 including hysteroscopies performed in the Service of Gynaecology at the Complejo Hospitalario-Universitario Insular Materno-Infantil (CHUIMI) to asymptomatic menopausal women.The indication of the hysteroscopy was an increase in endometrial thickness in the ultrasound study (>5 mm) in a routinary exam.These women were referred by the gynaecologist to the hysteroscopist to take an endometrial sample with hysteroscopy.
In the sampling process, the inclusion criteria were: postmenopausal women, with increased endometrial thickness in ultrasound study, who had not presented postmenopausal bleeding.
Patients who had presented postmenopausal bleeding were excluded.
Endometrial thickness in ultrasound study was subdivided into three categories: 5-10 mm; 11-15 mm and >15 mm and categorised into two groups for the main analysis: 5-9 mm and >10 mm, with the aim of establishing statistical association between endometrial thickness and histopathological findings of malignancy.
Other ultrasound findings such as increased vascularisation, heterogeneous endometrium or intracavitary fluid were also considered.
The procedure was done with a flexible hysteroscopy.The endometrial sample was taken with a biopsy forceps and sent to histopathologic study.The histopathologist was not blind to the endometrial thickness measurement neither other hysteroscopic findings.
Group 1 included patients with a histological diagnosis of benignity (including endometrial polyps) and group 2 included patients with a diagnosis of atypical endometrial hyperplasia and endometrial cancer, including malignant polyps.
The epidemiological characteristics of the patients were studied: age; years since menopause, parity, body mass index (BMI); personal history (arterial hypertension; Diabetes Mellitus, breast cancer, Lynch syndrome) and drug intake (antihypertensive drugs, oral antidiabetic drugs, Tamoxifen).The variables are detailed in Appendix 1 (Supplementary Material).
Information was obtained by reviewing computerised medical records and the Hysteroscopy database of the CHUIMI Gynaecology Service.An anonymized database was developed in Microsoft Excel spreadsheet (Microsoft Corporation 2010, Redmond, WA).Statistical analyses were carried out using IBM# SPSS# Statistics for Microsoft Windows Software Version 20 (International Business Machines (IBM) Corporation, Armonk, NY).
A descriptive analysis of the main qualitative variables studied was performed expressing them as absolute (n) and relative (%) frequencies.
A receiver operator characteristic (ROC) curve was used to identify the ET cut-off point for pathological endometrium.From the ROC curve, we calculated the area under the curve (AUC) with 95% confidence interval (95% CI), p value, sensitivity and specificity for the relevant cut-off point.
Fisher's exact test was used to find the statistical association between the ET cut-off point and the outcome.A value of p < 0.05 was considered significant.

Results
This study included 267 asymptomatic postmenopausal patients who underwent hysteroscopy following the finding of increased endometrial thickness.The mean age was 63.93 ± 8.89 years (minimum 33 years, maximum 92 years).
The prevalence of malignant findings in the histopathology study was 10/267 (3.7%) and benign findings of 257/ 267 (96.3%).The descriptive analysis of the variables studied and the risk factors as well as the ultrasound, hysteroscopic studies and histopathologic findings are described in Table 1.The characteristics of the women diagnosed with malignant pathology are shown in Table 2.
The malignant pathology is described as follows: 2 cases of hyperplasia with atypia, 1 case typified as neoplasia and the rest, 7 cases, included in the group of endometrial polyps (4 classified as complex hyperplasia with atypia and 3 as adenocarcinoma in polyp size >15 mm).
The results obtained in relation to malignant pathology in asymptomatic postmenopausal women with AET > 10 mm are summarised in Table 3.
The results showed a significant association (p ¼ 0.008) in diagnosing malignant pathology when establishing the cutoff point at AET > 10 mm, which allowed us to diagnose 100% of postmenopausal women with malignant pathology in the absence of bleeding.

Discussion
The results obtained show a prevalence of malignant pathology of 3.7% and of benign pathology of 96.3%.A significant association was obtained at the cut-off point of AET > 10 mm, which allows us to affirm that studying postmenopausal women from this thickness is acceptable and it is unlikely to miss cases of endometrial hyperplasia and endometrial cancer.
Our results are similar to those published in other studies.A more recent study (Ghoubara et al. 2018) has obtained a prevalence of atypical endometrial hyperplasia and endometrial cancer of 4.9% and 24.7% of endometrial polyps.The ROC curve identified the cut-off level of AET for atypical endometrial hyperplasia and endometrial cancer at 10 mm.This study included 81 asymptomatic postmenopausal women with AET > 4 mm.Schmidt et al. (2009) obtained a prevalence of atypical endometrial hyperplasia and cancer of 4.9%, with higher prevalence of endometrial polyps (73.3%), while Giannella et al. (2014) obtained a prevalence of 2.1% and 34%, respectively.The other three series published in the last decade (Famuyide et al. 2014, Saatli et al. 2014, Yasa et al. 2016) are retrospective in nature, showing wide variations in the prevalence of atypical hyperplasia and endometrial cancer (1.3-13.2%)as well as polyps (27.2-59.1%).
Ninety percent of women with endometrial cancer present with postmenopausal bleeding (Bachmann et al. 2003).However, some cancers may be detected before the development of symptoms, according to the NICE guideline in 2016.Giannella et al. (2014) studied the diagnostic accuracy of various endometrial thickness cut-off points by comparing histologic values and hysteroscopic findings in asymptomatic postmenopausal women with AET > 4 mm.They found that no single AET cut-off point had optimal accuracy for the diagnosis of endometrial malignant pathology, but that with an AET cut-off value >10 mm no cases were missed.With this cut-off point, the rate of atypical hyperplasia and cancer was 9.4%.When using the cut-off point of AET >4 mm, 97% of the hysteroscopies performed showed benign intrauterine pathology.
The Smith-Bindman group in 2014 concluded that patients with endometrial cancer are likely to be missed, but a cut-off point of AET > 11 mm may provide a balance between cancer detection and extension of the histopathologic study upon the accidental finding of increased endometrial thickness in asymptomatic postmenopausal women.
After analysing our results we can conclude, like the study published by Giannella et al. (2014) and that of Ghoubara et al. (2018), that by establishing a cut-off point of 10 mm we obtain a good discrimination between benign and malignant pathology, which would allow us to diagnose 100% of malignant pathology.In addition, we included in our analysis characteristics and risk factors for endometrial hyperplasia and cancer.In our setting, the rate of obesity and overweight is one of the highest in the country, and in our study we observed an association between overweight and/or obesity and malignant endometrial pathology, although it is not significant.Among the women diagnosed with malignant pathology, 90% had a BMI > 25.The only patient with BMI < 25 was an 85-year-old woman with a previous diagnosis of gastric cancer, whose histopathologic result showed metastasis of ring cell cancer.
However, we cannot assume that the association of BMI with malignant pathology is significant since 83.2% (198/267) of the patients had a BMI > 25.
The limitations of this study are those inherent to a retrospective observational study, since the data obtained could be biased by recording errors, as was the case with some patients in whom it was not possible to collect information on exposure to certain risk factors that were not recorded in the clinical histories.It is also a small-scale study limited by the small sample.On the other hand, the main disadvantage of endometrial thickness assessment by transvaginal ultrasound is the subjectivity of the observer in the interpretation of the scan.
We can conclude that it is safe to establish 10 mm as the cut-off point from which the risk of endometrial cancer increases, and therefore it would be advisable to extend the study.Likewise, there is an association, although not significant, between BMI > 25, as a risk factor in the development of endometrial malignant pathology.
A well-designed multicenter study would be necessary to establish the optimal cut-off point for endometrial thickness above which it would be justified to initiate histopathologic study in asymptomatic menopausal women with an incidental finding of thickened endometrium.

Table 2 .
Characteristics of asymptomatic postmenopausal women with anatomical pathology finding of malignancy (n ¼ 10).

Table 3 .
Results of the study using an endometrial thickness of 10 mm as a cut-off point in asymptomatic postmenopausal women.Comparing benign and malignant pathology in the sample.The p-value was obtained using Fisher's test.It is considered significant when it is <0.05.AET: Endometrial thickening.