Transvaginal ultrasound and diagnostic hysteroscopy as a predictor of endometrial polyps: risk factors for premalignancy and malignancy

Endometrium segmentation on transvaginal ultrasound image using key-point discriminator

PURPOSE

Transvaginal ultrasound imaging provides useful information for diagnosing endometrial pathologies and reproductive health. Endometrium segmentation in transvaginal ultrasound (TVUS) images is very challenging due to ambiguous boundaries and heterogeneous textures. In this study, we developed a new segmentation framework which provides robust segmentation against ambiguous boundaries and heterogeneous textures of TVUS images.

METHODS

To achieve endometrium segmentation from TVUS images, we propose a new segmentation framework with a discriminator guided by four key points of the endometrium (namely, the endometrium cavity tip, the internal os of the cervix, and the two thickest points between the two basal layers on the anterior and posterior uterine walls). The key points of the endometrium are defined as meaningful points that are related to the characteristics of the endometrial morphology, namely the length and thickness of the endometrium. In the proposed segmentation framework, the key-point discriminator distinguishes a predicted segmentation map from a ground-truth segmentation map according to the key-point maps. Meanwhile, the endometrium segmentation network predicts accurate segmentation results that the key-point discriminator cannot discriminate. In this adversarial way, the key-point information containing endometrial morphology characteristics is effectively incorporated in the segmentation network. The segmentation network can accurately find the segmentation boundary while the key-point discriminator learns the shape distribution of the endometrium. Moreover, the endometrium segmentation can be robust to the heterogeneous texture of the endometrium. We conducted an experiment on a TVUS dataset that contained 3,372 sagittal TVUS images and the corresponding key points. The dataset was collected by three hospitals (Ewha Woman's University School of Medicine, Asan Medical Center, and Yonsei University College of Medicine) with the approval of the three hospitals' Institutional Review Board. For verification, fivefold cross-validation was performed.

RESULT

The proposed key-point discriminator improved the performance of the endometrium segmentation, achieving 82.67 % for the Dice coefficient and 70.46% for the Jaccard coefficient. In comparison, on the TVUS images UNet, showed 58.69 % for the Dice coefficient and 41.59 % for the Jaccard coefficient. The qualitative performance of the endometrium segmentation was also improved over the conventional deep learning segmentation networks. Our experimental results indicated robust segmentation by the proposed method on TVUS images with heterogeneous texture and unclear boundary. In addition, the effect of the key-point discriminator was verified by an ablation study.

CONCLUSION

We proposed a key-point discriminator to train a segmentation network for robust segmentation of the endometrium with TVUS images. By utilizing the key-point information, the proposed method showed more reliable and accurate segmentation performance and outperformed the conventional segmentation networks both in qualitative and quantitative comparisons.

Is there a survival advantage in diagnosing endometrial cancer in asymptomatic postmenopausal patients? An Israeli Gynecology Oncology Group study

BACKGROUND

Incidental ultrasonographic findings in asymptomatic postmenopausal women, such as thickened endometrium or polyps, often lead to invasive procedures and to the occasional diagnosis of endometrial cancer. Data supporting a survival advantage of endometrial cancer diagnosed prior to the onset of postmenopausal bleeding are lacking.

OBJECTIVE

To compare the survival of asymptomatic and bleeding postmenopausal patients diagnosed with endometrial cancer.

STUDY DESIGN

This was an Israeli Gynecology Oncology Group retrospective multicenter study of 1607 postmenopausal patients with endometrial cancer: 233 asymptomatic patients and 1374 presenting with postmenopausal bleeding. Clinical, pathological, and survival measures were compared.

RESULTS

There was no significant difference between the asymptomatic and the postmenopausal bleeding groups in the proportion of patients in stage II-IV (23.5% vs 23.8%; P = .9) or in high-grade histology (41.0% vs 38.4%; P = .12). Among patients with stage-I tumors, asymptomatic patients had a greater proportion than postmenopausal bleeding patients of stage IA (82.1% vs 66.2%; P < .01) and a smaller proportion received adjuvant postoperative radiotherapy (30.5% vs 40.6%; P = .02). There was no difference between asymptomatic and postmenopausal bleeding patients in the 5-year recurrence-free survival (79.1% vs 79.4%; P = .85), disease-specific survival (83.2% vs 82.2%; P = .57), or overall survival (79.7% vs 76.8%; P = .37).

CONCLUSION

Endometrial cancer diagnosed in asymptomatic postmenopausal women is not associated with higher survival rates. Operative hysteroscopy/curettage procedures in asymptomatic patients with ultrasonographically diagnosed endometrial polyps or thick endometrium are rarely indicated. It is reasonable to reserve these procedures for patients whose ultrasonographic findings demonstrate significant change over time.

Sonographic Evaluation for Endometrial Polyps: The Interrupted Mucosa Sign

OBJECTIVES

To evaluate the interrupted mucosa sign for identification of endometrial polyps, using pathologic confirmation as the reference standard, compared to other accepted sonographic findings.

METHODS

We reviewed 195 patients referred for pelvic sonographic evaluations for suspected endometrial polyps in this retrospective Institutional Review Board-approved study. Of these, 82 had tissue sampling of the endometrium and constituted the final study group. Patient data, including age, menopausal status, last menstrual period, and final pathologic diagnosis, were recorded. Sonograms were reviewed by 2 blinded board-certified radiologists for endometrial features, including thickness, echogenicity, vascularity, presence of a mass, and the interrupted mucosa sign. Descriptive statistics and multivariate logistic regression analysis were performed.

RESULTS

The mean age of the patients was 44.99 (SD, 9.88) years, 79.1% of whom were premenopausal. Pathologic diagnosis confirmed polyps in 58 (70.73%). A single feeding vessel was visualized in 36 patients with polyps (62.07%), whereas the interrupted mucosa sign was visualized in 34 (58.62%). The presence of a feeding vessel, the interrupted mucosa sign, or both detected 48 (82.76%) of the polyps. In the multivariate analysis, only the interrupted mucosa sign was a statistically significant predictor of pathologic diagnosis of a polyp (P= .035), with an odds ratio of 3.83 (95% confidence interval, 1.10-13.29). Other sonographic findings were not independent predictors of a polyp: mass (P = .35), single feeding vessel (P = .31), endometrial thickness (P = .88), and endometrial echogenicity (P = .45). The sensitivity, specificity, and positive predictive value of the interrupted mucosa sign were 59%, 75%, and 85%, respectively.

CONCLUSIONS

The interrupted mucosa sign is a promising sonographic sign for identification of endometrial polyps, with greater predictive power than previously described signs. It has the potential to improve the diagnostic performance of sonography, especially when used in combination with other described signs.

Thickness of preperitoneal fat as a predictor of malignancy in overweight and obese women with endometrial polyps

The aim of the present study was to determine the association between the thickness of preperitoneal fat (PFT), utilized as an indicator of visceral fat deposition, and the risk of premalignant and malignant changes of endometrial polyps (EPs) in overweight and obese women. Overweight and obese women who had undergone diagnostic outpatient hysteroscopy and subsequent endometrial polypectomy between January 2010 and May 2013 were analyzed. Each patient underwent a transabdominal ultrasound for the purpose of measuring the thickness of preperitoneal fat. A total of 146 overweight or obese women were analyzed. The overall incidence of malignant or premalignant lesions was 8.2%. Notably, 5 patients (3.4%) received a histopathological diagnosis of complex hyperplasia with atypia, while polyps harboring carcinoma were diagnosed in 7 cases (4.8%). A significantly increased PFT was observed in women exhibiting preneoplastic and neoplastic lesions, compared with women with benign EPs (mean ± SD, 23.2±3.7 vs. 15.9±8.3; P<0.01). Patient age of >60 years was significantly associated with malignant progression of EPs, while body mass index, menopausal status, arterial hypertension, diabetes, abnormal uterine bleeding, hormone replacement therapy and tamoxifen treatment demonstrated no significant association with the development of preneoplastic and neoplastic lesions of the endometrium. In a multivariate analysis, only PFT maintained a significant correlation with the diagnosis of preneoplastic and neoplastic lesions on EPs (odds ratio, 1.14; 95% confidence interval, 1.04-1.26). Ultrasound evaluation of PFT in overweight and obese women may be useful for the identification of a particularly high-risk subgroup of women. Therefore, regardless of symptoms or additional clinical variables, these particularly high-risk women require appropriate counseling and prompt surgical removal of EPs.

Malignant and premalignant changes in the endometrium of women with an ultrasound diagnosis of endometrial polyp

The association between polyps and endometrial cancer is under debate. The main objectives of this study were to study the frequency and the characteristics of malignant and premalignant endometrial changes in women with endometrial polyp at ultrasound. The study population consisted of 1,390 consecutive patients that were referred to office hysteroscopy because of the ultrasonographic diagnosis of endometrial polyps. A total of 16 cases of endometrial neoplasia were diagnosed (1.15%). The frequencies of atypia and cancer in our population were 0.14% and 1.01%, respectively. All patients, except one, were postmenopausal (93.8%). All had undergone the initial ultrasonographic assessment because of symptoms (bleeding in the 93.8%). The neoplasia was not confined to the polyp in 75% of the cases. Nine cases had a lower risk disease (56.25%; atypical hyperplasia or endometrial cancer stage IA-G1,2), while seven had a higher risk cancer (43.75%; ≥ stage IA-G3). Patients with a higher risk disease were found to be significantly younger, and their polyps were smaller, albeit non-significantly. In spite of the common practice to refer all women with an ultrasound diagnosis of polyp to hysteroscopy, our data show how the prevalence of endometrial neoplasia in these patients is low (1.15%). Moreover, the malignancy is not confined to a polyp in most of the cases.

Malignant Potential of Endometrial Polyps: A retrospective study

To estimate the potential risk of malignancy in endometrial polyps and to investigate the possible clinical risk factors. Clinical, sonographic, hysteroscopic and pathological data were reviewed of 337 women with pathologically confirmed endometrial polyps diagnosed in Qatar between 2000 and 2010. The polyps were benign in 95.6% of cases, pre-malignant in 2.3%, and malignant in 2.1 % of cases. Advanced age, menopausal status and vaginal bleeding were strongly associated with atypical hyperplasia and cancer. There appeared to be no association between the risk of malignancy and other clinical risk factors.

Conclusions: There is a small risk of malignancy in endometrial polyps, mainly in post-menopausal patients of advanced age with symptomatic vaginal bleeding. In these circumstances, endometrial polyps should be managed aggressively as potential cancer and removalfor histology is mandatory.

Risk and predictors of malignancy in women with endometrial polyps

BACKGROUND

Endometrial polyps commonly affect premenopausal and postmenopausal women and carry a small risk of cancer. Consensus guidelines to direct the management of women with endometrial polyps are lacking. We examined the risk of malignancy in symptomatic and asymptomatic women with endometrial polyps.

METHODS

Institutional databases were analyzed to identify women with pathologically confirmed endometrial polyps diagnosed from 2002 to 2007. Demographic, clinical, and pathologic outcomes were reviewed. The most significant pathologic diagnosis was recorded for each subject. Endometrial hyperplasia and cancer were characterized as arising in the polyp or the adjacent endometrium. Factors associated with atypical hyperplasia and cancer were analyzed.

RESULTS

A total of 1011 women with endometrial polyps were identified. On pathology review, 964 (95.4%) polyps were reported as benign, 13 (1.3%) as hyperplasia without atypia, 5 (0.5%) as hyperplasia with atypia, and 13 (1.3%) as endometrial cancer. The only clinical or demographic factor associated with atypical hyperplasia and cancer was menopausal status (P = .02). Among premenopausal women the risk of cancer or atypical hyperplasia was 0.9% in patients without bleeding and 1.0% in women with bleeding. In postmenopausal women cancer or atypical hyperplasia was found in 1.9% of patients without bleeding and in 3.8% of women with bleeding.

CONCLUSIONS

The risk of endometrial cancer in women with endometrial polyps is 1.3%, while cancers confined to a polyp were found in only 0.3%. The risk is greatest in postmenopausal women with vaginal bleeding.

Polycystic ovary syndrome and increased polyp numbers as risk factors for malignant transformation of endometrial polyps in premenopausal women

OBJECTIVE

To determine the pre-malignant and malignant potential of endometrial polyps and to assess whether different clinical parameters are associated with malignancy in the polyps of premenopausal women.

METHODS

The clinical records of operative office hysteroscopic and resectoscopic procedures for endometrial polyps in 417 premenopausal women who attended Baskent University were examined over a retrospective period of 30 months. Only premenopausal patients were included in the study.

RESULTS

In 97.8% of women, histology showed benign endometrial pathology. In 2.2% of women, pre-malignant or malignant conditions were found in the polyp. Polycystic ovary syndrome (PCOS) and the presence of 2 or more polyps were associated with significant pre-malignant or malignant changes.

CONCLUSION

The presence of irregular vaginal bleeding was not a predictor of malignancy in the polyp. Premenopausal women with PCOS and those with 2 or more polyps had an increased prevalence of polyp malignancy. These groups of patients, whether symptomatic or not, should be evaluated by hysteroscopic resection of the polyps.

Are endometrial polyps true cancer precursors?

OBJECTIVE

The purpose of this study was to assess whether endometrial polyps (EMPs) represent cancer precursors.

STUDY DESIGN

Age standardized incidence ratios (SIRs) of histologically verified endometrial cancers (EmCas) were estimated in women with EMPs and in women with uterine leiomyomata, which is a condition that is unrelated to endometrial carcinogenesis. SIRs were calculated as the ratio of observed to expected EmCas based on age-specific incidence rates for female Montreal residents during the same period.

RESULTS

Of 1467 women with EMPs, 125 (8.5%) had EmCa. Of 1138 patients with uterine leiomyomata, 133 (11.7%) had EmCa. The SIRs of EmCa for women with EMPs (odds ratio, 8.0; 95% confidence interval, 6.6-9.5) were significantly lower than that in women with leiomyomata (odds ratio, 19.1; 95% confidence interval, 16.0-22.6). Abnormal uterine bleeding was the main reason for evaluating patients with EMP with or without associated EmCa.

CONCLUSION

The findings of higher EmCa incidence are consistent with enhanced detection opportunity rather than with the endometrial cancer precursor potential of EMPs.

How often are endometrial polyps malignant in asymptomatic postmenopausal women? A multicenter study

OBJECTIVE

The objective of the study was to evaluate the prevalence of cancer and premalignant lesions in polyps on atrophic endometrium in asymptomatic postmenopausal women to compare these findings with a similar cohort of patients with abnormal uterine bleeding.

STUDY DESIGN

One thousand one hundred fifty-two asymptomatic and 770 consecutive postmenopausal women with abnormal uterine bleeding were included in a retrospective multicenter study. Recruited patients underwent hysteroscopic polypectomy based on a sonohysterographic or hysteroscopic diagnosis. The pathologic report was the main outcome measure.

RESULTS

One single case of stage 1 grade 1 endometrial carcinoma on a polyp with a mean diameter of 40 mm (0.1%) was observed in asymptomatic women. This prevalence was 10 times lower than in symptomatic patients (P < .0001). The prevalence of atypical hyperplastic polyps was 1.2% in asymptomatic women (2.2% in symptomatic patients; P < .005). At multivariate analysis, polyps' diameter was the only variable significantly associated to an abnormal histology (cancer, polypoid cancer, and atypical hyperplasia) in asymptomatic women (odds ratio for polyps with mean diameter > 18 mm, 6.9; confidence interval, 2.2-21.4).

CONCLUSION

Follow-up and/or treatment of endometrial polyps incidentally diagnosed in asymptomatic postmenopausal patients could be safely restricted to few selected cases based on polyp diameter.

[Prevalence and characteristics of endometrial polyps in patients with abnormal uterine bleeding]

BACKGROUND/AIM

The prevalence of endometrial polyps (EPs) in the general female population is about 24%. Abnormal uterine bleeding is frequently the presenting symptom of EPs. The aim of this study was to determine the prevalence and characteristics of EPs in patients with abnormal uterine bleeding.

METHODS

The prevalence and characteristics of EPs were investigated in 961 patients with abnormal uterine bleeding who underwent dilatation and curettage between January and December 2006. Regarding histopathological features of EPs (presence of atypical hyperplasia or endometrial carcinoma), patients were divided into two groups: group A--patients who had EPs and EPs with hyperplasia without atypia (n = 204) and group B--patients who had EPs with atypical hyperplasia and EPs with carcinoma (n = 7).

RESULTS

In 211 (21.94%) patients EPs were found with abnormal uterine bleeding. Histopathologically, there were 175 (82.94%) EPs, 29 (13.74%) EPs with hyperplasia without atypia, 5 (2.37%) EPs with atypical hyperplasia, and 2 (0.95%) EPs with endometrial carcinoma. Contrary to the patients with EPs and EPs with hyperplasia without atypia (group A), patients who had EPs with atypical hyperplasia and EPs with carcinoma (group B) were older (p < 0.05), and more commonly postmenopausal (p < 0.05) and with hypertension (p < 0.05), all of statistical significance.

CONCLUSION

The prevalence of endometrial polyps in patients with abnormal uterine bleeding according to our data was 21.95%. Atypical hyperplasia and endometrial carcinoma were rarely confined to a polyp. Older age, postmenopausal period and hypertension may increase the risk of premalignant and malignant changes in endometrial polyps.

Endometrial directed biopsy during sonohysterography using the NiGo device: prospective study in women with abnormal uterine bleeding

OBJECTIVE

To determine the accuracy of sonohysterography (SHG), combined with sonography-guided biopsies that are performed by using the NiGo device, in the diagnosis of endometrial pathologies in women with abnormal uterine bleeding (AUB).

DESIGN

Prospective study.

SETTING

Outpatient clinic.

PATIENT(S)

One hundred sixteen women with AUB (55 postmenopausal and 61 premenopausal).

INTERVENTION(S)

Patients underwent transvaginal sonography and SHG; when an endometrial condition needing a biopsy was diagnosed, the biopsy was performed both by using the NiGo device (ASL 1, Imperia, Italy) and by using Pipelle.

MAIN OUTCOME MEASURE(S)

Feasibility of SHG and endometrial biopsy by NiGo.

RESULT(S)

Sonohysterography was successfully performed by using the NiGo device in 87.4% of the cases (92.6% of premenopausal women and 81.5% of postmenopausal women). Seventy-eight (86.7%) women were judged to require endometrial biopsy; the NiGo device allowed obtaining an adequate biopsy for histological evaluation in 74 cases (94.9%). There was no statistically significant difference in efficacy between the NiGo and Pipelle in providing a specimen that was adequate for histological evaluation. Although no significant difference was observed in the histological diagnosis performed on the specimens obtained by the two techniques, NiGo allowed diagnosis of more focal lesions (simple hyperplasia, endometrial polyps, and myomas).

CONCLUSION(S)

The NiGo device allows an adequate biopsy for histological evaluation to be obtained during SHG in an outpatient setting.

Clinical utility of liquid-based cytology for the characterization and management of endometrial polyps in postmenopausal age

The proper management of endometrial polyps still represents a clinical ongoing challenge, especially when they are asymptomatic and occasionally discovered. The aim of this study was to evaluate liquid-based endometrial cytology to manage endometrial polyps in postmenopausal age by its ability to exclude hidden premalignant and malignant changes within polyps. Three hundred fifty-nine consecutive postmenopausal patients who underwent hysteroscopic diagnosis of endometrial polyp over a 3-year period and who were scheduled for surgical removal within the three subsequent months were retrospectively evaluated. Histologic results after resection during operative hysteroscopy or during hysterectomy were compared with liquid-based cytology and endometrial biopsy obtained at the time of diagnostic hysteroscopy. Eight of 359 patients (2.2%) had malignant or premalignant polyps interpreted as benign finding at hysteroscopy. Unsatisfactory samples were higher for endometrial biopsy compared to liquid-based cytology in the whole series and in the subgroup of low-risk asymptomatic patients (P< 0.001). Endometrial biopsy and liquid-based cytology revealed a sensitivity of 62% and 87.5%, respectively and a 100% specificity. Considering the subgroup of low-risk asymptomatic patients, liquid-based cytology disclosed all the five pathologic lesions with a 100% sensitivity and specificity. In conclusion, liquid-based cytology proved to be a useful tool to establish the nature of endometrial polyps in postmenopausal patients. Complete removal of the lesion should be offered to all symptomatic patients and those with established risk factors for endometrial cancer. Conversely, a wait and see attitude should be considered in case of asymptomatic low-risk polyps with typical appearance on hysteroscopy and negative liquid-based cytology.

Endometrial polyps in pre- and postmenopausal women: Factors associated with malignancy

OBJECTIVE

To evaluate the prevalence of premalignant and malignant polyps and their association with menopausal status, hormone therapy and clinical characteristics in perimenopausal and postmenopausal women.

METHODS

A surgical database was used to select pre- and postmenopausal women >or=40 years of age, submitted to hysteroscopic resection of endometrial polyps. The medical records of 475 women were reviewed and clinical characteristics and histological diagnosis of resected polyps were assessed.

RESULTS

The majority of women had benign endometrial lesions, 78.53% of which were endometrial polyps and 13.47% polyps with simple or complex endometrial hyperplasia without atypia. Polyps with endometrial hyperplasia with atypia comprised 1.05% of cases, while 2.74% were carcinomatous polyps. Analysis using prevalence ratios showed that premalignant and malignant lesions were associated with age and postmenopausal bleeding. Women >60 years of age had a prevalence ratio 3.28 times greater (95%CI: 1.19-9.07) of premalignant or malignant polyps. When only postmenopausal women were evaluated for the effect of age, those over 60 years of age had a prevalence 5.31 times greater (95%CI: 1.22-23.09), while those with postmenopausal bleeding had an age-adjusted prevalence ratio of 3.71 (95%CI: 1.21-11.34) compared to asymptomatic women. No significant association was found between arterial hypertension, diabetes mellitus, obesity, use of hormone therapy or tamoxifen and premalignancy or malignancy.

CONCLUSIONS

There was a low prevalence of premalignant and malignant lesions in endometrial polyps. Older women and those with postmenopausal bleeding had a greater prevalence of malignancy and in these cases hysteroscopic resection of endometrial polyps is mandatory.

Progesterone receptor modulators and the endometrium: changes and consequences

Progesterone receptor modulators (PRMs) have been used for contraceptive research, as well as for treatment of fibroids, endometriosis and heavy or irregular menstrual bleeding. Long-term treatment with these compounds results in changes to the endometrium resulting in potential confusion in trying to characterize endometrial biopsies. A meeting was held to discuss the properties of PRMs, the effects of perturbed hormonal control of the endometrium and the need for further understanding of the biology of progesterone receptor action to facilitate the development of new PRMs. A panel of pathologists was convened to evaluate endometrial changes associated with a minimum of three months of chronic treatment with PRMs. Four different agents were used in the treatment regimens but the pathologists were blinded to treatment regimen or agent. The panel agreed that the endometrial biopsies did not fit into a classification of either proliferative or secretory endometrium but exhibited an unusual architecture that could be characterized as glandular dilatation. There was little evidence of mitosis, consistent with a proposed anti-proliferative effect of PRMs. The panel concluded that the biopsies did not reveal evidence of safety concern and that pathologists and investigators familiar with endometrial effects of chronic PRM exposure should consider working with pharmaceutical companies and regulatory agencies to develop standard descriptions of PRM-associated endometrial changes as well as the types of histologic changes that would signal a need for intervention.

A new technique to obtain endometrial directed biopsy during sonohysterography: the NiGo device

The primary goal of the clinical evaluation of abnormal uterine bleeding is to establish a specific diagnosis in the most efficient and least invasive manner possible. Hysteroscopy (HS) allows physicians to obtain directed biopsy in addition to direct visualization of the uterine cavity and lesions, but often requires ambulatory surgery centers and some anesthesia, or can potentially result in significant patient discomfort. Sonohysterography (SHG) is less invasive than HS but does not allow a histologic sample. A new technique, called the Nicoletti-Gorlero (NiGo) device, was developed and evaluated to obtain histologic results during SHG. This study is a prospective, observational evaluation between the NiGo device technique and standard HS to obtain endometrial samples. The NiGo device was used to obtain an endometrial sample during SHG in 18 women. In a subsequent session, HS was performed to verify the results. From January 2005 through November 2005 both procedures (SHG + biopsy and HS) were performed on 18 patients. The SHG procedures using the NiGo forcep were performed successfully on 15 out of 18 women, and the endometrial sample was obtained in 14 out of 15 patients. In one patient, the endometrial biopsy obtained provided too little tissue to accomplish histologic evaluation. All 13 pathologic results obtained with the NiGo device were identical with those subsequently obtained with HS. The office-based HS procedure was not successful in two women; in these women, an HS procedure performed in the operating room was necessary. The NiGo device technique allows the physician to obtain sonographic-guided biopsies of the entire endometrium during SHG. The technique is less invasive compared with HS. In our small series, there were no complications during the procedure.