Advertisement

Reproducibility of Endometrial Pathologic Findings Obtained on Hysteroscopy, Transvaginal Sonography, and Gel Infusion Sonography in Women With Postmenopausal Bleeding

      Abstract

      Study Objective

      To evaluate and compare interobserver variation in endometrial pattern recognition with hysteroscopy (HY) and transvaginal sonography (TVS) and gel infusion sonography (GIS) with regard to the diagnosis of endometrial pathology.

      Design

      Prospective study (Canadian Task Force II-1).

      Setting

      University clinic.

      Patients

      One hundred twenty-two consecutive women with postmenopausal bleeding and an endometrium thickness ≥ 5 mm.

      Intervention

      Two observers using HY and 2 others using TVS and GIS evaluated the endometrial pattern in recorded video clips. Interobserver agreement regarding findings obtained with TVS, GIS, and HY for a diagnosis of cancer, hyperplasia, polyps, and no endometrial pathology was expressed by κ coefficients and compared.

      Measurement and Main Results

      Interobserver agreement (κ) was as follows: identification of normal endometrium: HY (.74), TVS (.68), and GIS (.48); diagnosis of cancer: HY (.56), TVS (.59), and GIS (.34); classification in all categories of endometrial pathology: HY (.70), TVS (.47), and GIS (.41) (p < .05 HY vs GIS). The presence of additional endometrial polyps decreased agreement on HY in patients with hyperplasia or cancer. Observer agreement was poor regarding the diagnosis of hyperplasia by all techniques.

      Conclusion

      Observer agreement regarding both HY and TVS was reliable for the diagnosis of a normal endometrium but poor with HY, TVS, and especially GIS for a diagnosis of cancer. In patients with hyperplasia or cancer, agreement between observers was especially low in the presence of additional polyps when HY was used. These findings call attention to the need for systematic methods to improve reliability in endometrial pattern recognition.

      Keywords

       Discuss

      You can discuss this article with its authors and with other AAGL members at http://www.AAGL.org/jmig-22-5-JMIG-D-15-00242.
      Endometrial carcinoma is the most common gynecologic malignancy, and postmenopausal bleeding (PMB) is the cardinal symptom. The survival of patients with endometrial cancer is good when it is diagnosed at an early stage [
      • Lajer H.
      • Elnegaard S.
      • Christensen R.D.
      • et al.
      Survival after stage IA endometrial cancer: can follow-up be altered? A prospective nationwide Danish survey.
      ]. Therefore, early, accurate, and timely diagnosis in women with PMB is important. Transvaginal sonography (TVS) is the first-line diagnostic technique with the lowest cost [
      • Clark T.J.
      • Barton P.M.
      • Coomarasamy A.
      • Gupta J.K.
      • Khan K.S.
      Investigating postmenopausal bleeding for endometrial cancer: cost-effectiveness of initial diagnostic strategies.
      ]. In a large multicenter study [
      • Karlsson B.
      • Granberg S.
      • Wikland M.
      • et al.
      Transvaginal ultrasonography of the endometrium in women with postmenopausal bleeding—a Nordic multicenter study.
      ] and a meta-analysis [
      • Smith-Bindman R.
      • Kerlikowske K.
      • Feldstein V.A.
      • et al.
      Endovaginal ultrasound to exclude endometrial cancer and other endometrial abnormalities.
      ], an endometrial thickness less than or equal to 4 to 5 mm effectively excluded malignancy. Furthermore, according to the American College of Obstetrics and Gynecology, in women with PMB, when an endometrial thickness of less than or equal to 4 mm is identified by ultrasound, endometrial sampling is not required [
      American College of Obstetricians and Gynecologists
      ACOG committee opinion no. 426: the role of transvaginal ultrasonography in the evaluation of postmenopausal bleeding.
      ]. Conversely, an endometrial thickness at least 4 to 5 mm may indicate malignancy [
      • Timmermans A.
      • Opmeer B.C.
      • Khan K.S.
      • et al.
      Endometrial thickness measurement for detecting endometrial cancer in women with postmenopausal bleeding: a systematic review and meta-analysis.
      ], and second-line invasive tools are recommended. There is still no consensus or evidence regarding the most optimal second-line diagnostic technique [
      • van H.N.
      • Breijer M.C.
      • Khan K.S.
      • et al.
      Diagnostic evaluation of the endometrium in postmenopausal bleeding: an evidence-based approach.
      ], which include saline infusion sonography (SIS), gel infusion sonography (GIS), hysteroscopy (HY) [
      • Dreisler E.
      • Poulsen L.G.
      • Antonsen S.L.
      • et al.
      EMAS clinical guide: Assessment of the endometrium in peri and postmenopausal women.
      ], and endometrial samples (ES) [
      • Dreisler E.
      • Poulsen L.G.
      • Antonsen S.L.
      • et al.
      EMAS clinical guide: Assessment of the endometrium in peri and postmenopausal women.
      ,
      • Munro M.G.
      Investigation of women with postmenopausal uterine bleeding: clinical practice recommendations.
      ].
      ES is the most simple second-line technique used either in all women with increased endometrial thickness [
      • Munro M.G.
      Investigation of women with postmenopausal uterine bleeding: clinical practice recommendations.
      ] or in the small group of women without localized pathology on GIS or SIS [
      • Dreisler E.
      • Poulsen L.G.
      • Antonsen S.L.
      • et al.
      EMAS clinical guide: Assessment of the endometrium in peri and postmenopausal women.
      ]. In the presence of focal changes of the endometrium, ES is less efficient [
      • Epstein E.
      • Ramirez A.
      • Skoog L.
      • Valentin L.
      Dilatation and curettage fails to detect most focal lesions in the uterine cavity in women with postmenopausal bleeding.
      ,
      • Angioni S.
      • Loddo A.
      • Milano F.
      • et al.
      Detection of benign intracavitary lesions in postmenopausal women with abnormal uterine bleeding: a prospective comparative study on outpatient hysteroscopy and blind biopsy.
      ]; SIS or GIS [
      • Werbrouck E.
      • Veldman J.
      • Luts J.
      • et al.
      Detection of endometrial pathology using saline infusion sonography versus gel instillation sonography: a prospective cohort study.
      ] may be added to identify those women who have focal changes. In these women HY is often performed [
      • Epstein E.
      • Skoog L.
      • Isberg P.E.
      • et al.
      An algorithm including results of gray-scale and power Doppler ultrasound examination to predict endometrial malignancy in women with postmenopausal bleeding.
      ,
      • Goldstein R.B.
      • Bree R.L.
      • Benson C.B.
      • et al.
      Evaluation of the woman with postmenopausal bleeding: Society of Radiologists in Ultrasound-Sponsored Consensus Conference statement.
      ,
      American Association of Gynecologic Laparoscopists
      AAGL practice report: practice guidelines for the diagnosis and management of endometrial polyps.
      ].
      In women with PMB the main shortcoming of ES is a high failure rate (inadequate sample or inability to perform the biopsy). In meta-analysis the reported range of sampling failure was 0 to 58% with pipelle endometrial sampling [
      • Dijkhuizen F.P.
      • Mol B.W.
      • Brolmann H.A.
      • Heintz A.P.
      The accuracy of endometrial sampling in the diagnosis of patients with endometrial carcinoma and hyperplasia: a meta-analysis.
      ]. Two studies confirmed a high failure rate (51% and 64%, of which 30% and 48% were insufficient samples) [
      • Visser N.C.
      • Breijer M.C.
      • Herman M.C.
      • et al.
      Factors attributing to the failure of endometrial sampling in women with postmenopausal bleeding.
      ,
      • Williams A.R.
      • Brechin S.
      • Porter A.J.
      • Warner P.
      • Critchley H.O.
      Factors affecting adequacy of Pipelle and Tao Brush endometrial sampling.
      ]. Malignancy cannot be excluded in women with insufficient ES [
      • van Doorn H.C.
      • Opmeer B.C.
      • Burger C.W.
      • et al.
      Inadequate office endometrial sample requires further evaluation in women with postmenopausal bleeding and abnormal ultrasound results.
      ,
      • Farrell T.
      • Jones N.
      • Owen P.
      • Baird A.
      The significance of an “insufficient” Pipelle sample in the investigation of post-menopausal bleeding.
      ], and HY is often added. Insufficient ES is costly and painful for patients, and the wait time can cause anxiety. Hysteroscopic samples have a higher diagnostic accuracy compared with ES [
      • Bakour S.H.
      • Jones S.E.
      • O'Donovan P.
      Ambulatory hysteroscopy: evidence-based guide to diagnosis and therapy.
      ,
      • Ceci O.
      • Bettocchi S.
      • Pellegrino A.
      • et al.
      Comparison of hysteroscopic and hysterectomy findings for assessing the diagnostic accuracy of office hysteroscopy.
      ], and HY can be performed as “see-and-treat” procedures with small diameter (mini) hysteroscopes without anesthesia [
      • Di Spiezio S.A.
      • Bettocchi S.
      • Spinelli M.
      • et al.
      Review of new office-based hysteroscopic procedures 2003–2009.
      ,
      • Bettocchi S.
      • Ceci O.
      • Nappi L.
      • et al.
      Operative office hysteroscopy without anesthesia: analysis of 4863 cases performed with mechanical instruments.
      ,
      • Bettocchi S.
      • Ceci O.
      • Di V.R.
      • et al.
      Advanced operative office hysteroscopy without anaesthesia: analysis of 501 cases treated with a 5 Fr. bipolar electrode.
      ].
      See-and-treat HY is therefore being increasingly used [
      • Lotfallah H.
      • Farag K.
      • Hassan I.
      • Watson R.
      One-stop hysteroscopy clinic for postmenopausal bleeding.
      ]. Eye-directed biopsies are performed with the use of small forceps, and this technique has a high accuracy in the hands of experienced hysteroscopists [
      • Loizzi V.
      • Bettocchi S.
      • Vimercati A.
      • et al.
      Hysteroscopic evaluation of menopausal women with endometrial thickness of 4 mm or more.
      ]. However, the accuracy of this technique depends on the recognition of suspect endometrial pathology. Small-sized pathology can be removed, and see-and-treat HY may be feasible in half of the patients with PMB [
      • Gambadauro P.
      • Martinez-Maestre M.A.
      • Torrejon R.
      When is see-and-treat hysteroscopic polypectomy successful?.
      ], whereas a narrow cervix and the total removal of larger polyps may indicate another operative HY [
      • Cicinelli E.
      Hysteroscopy without anesthesia: review of recent literature.
      ]. Moreover, the safety of HY has to be considered in patients with malignancy. Perforation during HY of a soft, fragile uterus with a malignancy poses the risk of spread, and HY should only be performed by experienced operators.
      Evaluation of the endometrial pattern at first-line TVS or GIS could be used to avoid the use of HY in women with suspected malignancy. These women could be offered ES and fast-track reference to oncologic centers, whereas women with a benign pattern based on cervical findings and the measured size and type of the endometrial pathology could be saved for the high numbers of insufficient samples or HYs and, when needed, undergo the most optimal type of HY, thereby increasing the feasibility and safety of see-and-treat HY [
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ]. Evaluation of the endometrial pattern on TVS or GIS had a high diagnostic accuracy for the diagnosis of endometrial cancer [
      • Epstein E.
      • Skoog L.
      • Isberg P.E.
      • et al.
      An algorithm including results of gray-scale and power Doppler ultrasound examination to predict endometrial malignancy in women with postmenopausal bleeding.
      ,
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ,
      • Epstein E.
      • Valentin L.
      Gray-scale ultrasound morphology in the presence or absence of intrauterine fluid and vascularity as assessed by color Doppler for discrimination between benign and malignant endometrium in women with postmenopausal bleeding.
      ,
      • Opolskiene G.
      • Sladkevicius P.
      • Valentin L.
      Prediction of endometrial malignancy in women with postmenopausal bleeding and sonographic endometrial thickness >/= 4.5 mm.
      ,
      • Dueholm M.
      • Møller C.
      • Rydbjerg S.
      • Hansen E.
      An ultrasound algorithm for the identification of endometrial cancer.
      ], which was slightly lower but almost comparable with the accuracy found with HY [
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ].
      However, pattern evaluations on TVS, GIS, or HY are in general practice performed by different observers, and the reliability depends on the reproducibility between observers. Reproducibility between observers can be evaluated on the basis of interobserver agreement.
      In women with PMB, observer agreement at TVS and SIS for endometrial thickness measurements [
      • Warming L.
      • Ravn P.
      • Skouby S.
      • Christiansen C.
      Measurement precision and normal range of endometrial thickness in a postmenopausal population by transvaginal ultrasound.
      ,
      • Epstein E.
      • Valentin L.
      Intraobserver and interobserver reproducibility of ultrasound measurements of endometrial thickness in postmenopausal women.
      ,
      • Wolman I.
      • Amster R.
      • Hartoov J.
      • et al.
      Reproducibility of transvaginal ultrasonographic measurements of endometrial thickness in patients with postmenopausal bleeding.
      ,
      • Karlsson B.
      • Granberg S.
      • Ridell B.
      • Wikland M.
      Endometrial thickness as measured by transvaginal sonography: interobserver variation.
      ] and identification of focal lesions [
      • Opolskiene G.
      • Sladkevicius P.
      • Valentin L.
      Ultrasound assessment of endometrial morphology and vascularity to predict endometrial malignancy in women with postmenopausal bleeding and sonographic endometrial thickness >or= 4.5 mm.
      ,
      • Opolskiene G.
      • Sladkevicius P.
      • Valentin L.
      Two- and three-dimensional saline contrast sonohysterography: interobserver agreement, agreement with hysteroscopy and diagnosis of endometrial malignancy.
      ] are described, whereas observer agreement with HY has only been described in a few premenopausal women [
      • Dueholm M.
      • Lundorf E.
      • Sorensen J.S.
      • et al.
      Reproducibility of evaluation of the uterus by transvaginal sonography, hysterosonographic examination, hysteroscopy and magnetic resonance imaging.
      ,
      • Kasius J.C.
      • Broekmans F.J.
      • Veersema S.
      • et al.
      Observer agreement in the evaluation of the uterine cavity by hysteroscopy prior to in vitro fertilization.
      ]. HY has a higher image resolution than TVS, and a lower observer variation on HY could be expected. No study has described or compared the reproducibility of visual pattern evaluation on HY and TVS or GIS for the diagnosis of endometrial pathology. We accessed and compared the reproducibility of HY, TVS, and GIS in differentiating endometrial pathology and evaluated cases in which reproducibility was poor in women with malignant or premalignant endometrial pathology to suggest improvements.

      Methods

      HY, TVS, and GIS video recordings were obtained in a prospective trial of the diagnostic accuracy of ultrasound and HY for the diagnosis of endometrial cancer [
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ,
      • Dueholm M.
      • Møller C.
      • Rydbjerg S.
      • Hansen E.
      An ultrasound algorithm for the identification of endometrial cancer.
      ,
      • Dueholm M.
      • Holm J.W.
      • Rydbjerg S.
      • Hansen E.S.
      • Ortoft G.
      Two-dimensional and 3-dimensional transvaginal ultrasound with addition of power Doppler angiography and gel infusion sonography for diagnosis of endometrial malignancy.
      ]. In the present study these videos were evaluated by different observers to measure interobserver variation.

       Observers

      Three months to 1 year after completion of the investigation, all video clips were evaluated in personal computers by the observers blinded to prior evaluations and in random order without knowledge of the patient's identity and pathology. Hysteroscopic clips were evaluated by 2 gynecologists from another university hospital (PS and AJ). One had special training in hysteroscopic staging (observer 1) in patients with endometrial cancer, whereas the other (observer 2) was a general specialist in HY and had more than 10 years of special experience in HY.
      Two other observers (MD and GO) evaluated ultrasound videos. Observer A had extensive experience in the staging of endometrial cancer and GIS, whereas observer B had more than 10 years of general specialist experience in the evaluation of endometrial pathology by TVS.

       Patients

      The study group comprised consecutive women with PMB and an endometrial thickness of at least 5 mm either referred directly or referred from other hospitals. Two experienced gynecologists (MD and another hysteroscopist) performed TVS, GIS, and HY in 164 of 174 women referred to Aarhus University Hospital, Denmark, from October 2010 to February 2012 and evaluated endometrial patterns on TVS, GIS, and HY. Systematic videos of endometrium obtained with TVS, GIS, and HY were recorded at the end of each examination. Video clips were not performed or correctly stored for both TVS with Doppler GIS and HY in 42 patients, leaving 122 patients with video clips for observer variation of endometrial patterns by all 3 techniques.
      Hysteroscopy (HY) was performed using a Storz resectoscope (26F 12-degree oblique optic; Storz, N. Westport, CT). Hysteroscopic video clips (2–3 clips of 2–4 minutes) were obtained according to a standardized form. Every recording systematically visualized the anterior wall, the fundus, the ostia, sidewalls, and posterior wall and covered the whole uterine cavity. Evaluation of video clips was performed according to defined standard criteria based on morphologic evaluation of the endometrium [
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ,
      • Garuti G.
      • Mirra M.
      • Luerti M.
      Hysteroscopic view in atypical endometrial hyperplasias: a correlation with pathologic findings on hysterectomy specimens.
      ] and recorded on a standard form. The endometrial pattern was evaluated systematically by each observer with regard to localized or diffuse lesions, surface structure, necrosis, and vascular pattern. Diagnoses of hyperplasia, malignancy, benign polyps, myomas, proliferative endometrium, atrophy, and other specified abnormalities were recorded and categorized into endometrial diagnoses of malignancy, hyperplasia, benign polyps, and normal endometrium. The diagnosis was made by subjective morphologic evaluation of the endometrium based on prior studies [
      • Garuti G.
      • Mirra M.
      • Luerti M.
      Hysteroscopic view in atypical endometrial hyperplasias: a correlation with pathologic findings on hysterectomy specimens.
      ,
      • Dotto J.E.
      • Lema B.
      • Dotto Jr., J.E.
      • Hamou J.
      Classification of microhysteroscopic images and their correlation with histologic diagnoses.
      ,
      • Lasmar R.B.
      • Barrozo P.R.
      • de Oliveira M.A.
      • Coutinho E.S.
      • Dias R.
      Validation of hysteroscopic view in cases of endometrial hyperplasia and cancer in patients with abnormal uterine bleeding.
      ] described in detail elsewhere [
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ,

      Dueholm M, Hjorth IM, Secher P, Jørgensen A, Ørtoft G. Structured hysteroscopic evaluation of endometrium in women with postmenopausal bleeding. J Minim Invasive Gynecol. 2015 Jun 30. [Epub ahead of print].

      ].

       TVS and GIS

      TVS and GIS were performed with a Voluson E8 expert (GE Healthcare, Milwaukee, WI) equipped with a multifrequency (5–12 MHz) endovaginal probe according to a predetermined scanning protocol as previously described [
      • Dueholm M.
      • Møller C.
      • Rydbjerg S.
      • Hansen E.
      An ultrasound algorithm for the identification of endometrial cancer.
      ]. GIS was performed with a small flexible sterile catheter (infant-feeding tube Unometric no. 227581; Abena A/S, Aabenraa, Denmark) mounted with a 10-mL syringe with Instillagel (E. Tjellesen A/S, Lynge, Denmark).
      At each examination (TVS, TVS with Doppler, and GIS), 3 two-dimensional clips were obtained, including 2 sweeps in the sagittal plane and 1 in the horizontal plane. Clips of 5 to 10 seconds (retrospective) were obtained. All video clips from each women were evaluated in the order TVS, TVS with Doppler, and GIS for each woman using accepted criteria [
      • Leone F.P.
      • Timmerman D.
      • Bourne T.
      • et al.
      Terms, definitions and measurements to describe the sonographic features of the endometrium and intrauterine lesions: a consensus opinion from the International Endometrial Tumor Analysis (IETA) group.
      ] on a standard form [
      • Dueholm M.
      • Møller C.
      • Rydbjerg S.
      • Hansen E.
      An ultrasound algorithm for the identification of endometrial cancer.
      ]. Patient age was noted on the form. The quality of the images (score 1–5) was evaluated, and a subjective diagnosis was first made after evaluation of video clips obtained on TVS with Doppler and then video clips by GIS. Subjective diagnoses of malignancy, benign polyps, hyperplasia, normal endometrium, myomas, and other abnormalities were interpreted as previously described [
      • Dueholm M.
      • Holm J.W.
      • Rydbjerg S.
      • Hansen E.S.
      • Ortoft G.
      Two-dimensional and 3-dimensional transvaginal ultrasound with addition of power Doppler angiography and gel infusion sonography for diagnosis of endometrial malignancy.
      ].

       Reference Standard at Pathology

      The reference standard of all examinations was the microscopic pathology of the uterus if hysterectomy was performed or microscopic evaluation of resectoscopic biopsy and additional curettage specimens [
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ,
      • Dueholm M.
      • Møller C.
      • Rydbjerg S.
      • Hansen E.
      An ultrasound algorithm for the identification of endometrial cancer.
      ]. All localized focal changes were removed during HY, and curettage was performed. One of 2 oncopathologists evaluated all specimens.

       Review of Video Clips From Women With Endometrial Cancer

      After final evaluation, all videos from women with endometrial cancer and those in which there was a discrepancy between 1 of the 2 observers on either HY, TVS, or GIS were reviewed by the main author (MD, secondary review). To study discrepancies, findings were compared with prior evaluation during live HY, TVS, and GIS.

       Statistics and Analysis

      Data were analyzed using STATA (Statistic Data Analysis, STATA Corp, College Station, TX). Interobserver agreement for video analysis of TVS, GIS, and HY was expressed in a κ coefficient. A κ value of .20 represents slight agreement, a value between .21 and .40 fair agreement, a value between .41 and .60 moderate agreement, a value between .61 and .80 substantial agreement, and a value between .81 1.00 almost perfect agreement. Weighted κ analysis was used for calculation of observer variation of a total diagnosis in 4 categories: (1) no endometrial abnormality, (2) polyps, (3) endometrial hyperplasia, (4) endometrial cancer. The weights were given by 1−[(i−j)/(κ−1)]2, where i and j index the rows of columns of the ratings by the 2 raters and κ is the maximum number of possible ratings. Confidence limits for the weighted κ were calculated by 1000 bootstraps replications (STATA “kapci” command).
      Chi-square tests and McNemar's test were used for comparison of disagreements. In variables with 4 categories, disagreement of 0 to 1 category and 2 to 3 categories were calculated. A p < .05 was considered statistically significant. A sample size of 122 women was used, based on an expected disagreement between observers in 25% of women by TVS; we wanted to detect a difference in disagreement of 15% between TVS and HY (Type I error, 5%; Type II error, 80%).

      Results

      Patient characteristics are given in Table 1. Additional benign endometrial polyps were diagnosed on pathologic examination in 36% of patients with hyperplasia or endometrial cancer. The agreement between 2 observers was moderate for a localized lesion using GIS and HY (percent of agreement and κ [95% confidence interval]: GIS, 74% and .49 [.34–.63]; HY, 73% and .46 [.31–.62]).
      Table 1Endometrial pathology in resectoscopic samples obtained at operative hysteroscopy or on pathologic examination of tissue removed at hysterectomy
      N (%)No. of patients with additional
      MyomasPolyps
      Atrophic or proliferative endometrium19
      One patient had endometritis.
      (13.1)
      5
      Endometrial polyps37 (30.3)1
      Hyperplasia low grade (simple or complex)12 (9.8)26
      Hyperplasia with atypia5 (4.1)14
      Cancer49 (40.2)114
      Total1221024
      Patients for observer evaluation (N = 122).
      One patient had endometritis.
      The agreement between observers using TVS, GIS, and HY is shown in Table 2. There was moderate agreement between the 2 observers (A and B) using TVS and between observer 1 and observer 2 using HY for the diagnosis of cancer. However, for a diagnosis of no endometrial pathology, substantial agreement was attained between the 2 observers who used TVS and HY. No agreement was obtained by any technique for a diagnosis of hyperplasia, and the agreement between 2 observers for a diagnosis of endometrial polyps was fair to moderate by all 3 techniques.
      Table 2Interobserver agreement between 2 observers for the diagnosis of endometrial cancer (yes/no), endometrial hyperplasia (yes/no), endometrial polyp (yes/no), and no endometrial abnormality (yes/no) and for classification of a total diagnosis in all in 4 categories (cancer, hyperplasia, polyps, no endometrial abnormalities)
      Observer A/observer BObserver A/observer BObserver 1/observer 2
      TVS with DopplerGISHY
      κ (95% confidence interval)nA/nBPercent agreementκ (95% confidence interval)nA/nBPercent agreementκ (95% confidence interval)n1/n2Percent agreement
      Endometrial cancer.59 (.44–.74)45/4281.2.34 (.18–.51)48/3270.5.56 (.41–.71)46/2781.2
      Endometrial hyperplasia.07 (−.15 to .30)11/588.5.07 (−.14 to .28)10/1582.8.05 (−.13 to .23)26/1573.0
      Endometrial polyp.37 (.21–.53)46/5868.9.43 (.27–.59)38/5373.0.42 (.27–.58)27/4874.6
      No endometrial pathology.68 (.52–.85)29/1790.2.48 (.29–.68)26/2283.6.74 (.60–.88)23/3291.0
      Total all 4 categories
      Disagreements of more than 1 category (McNemar's test): GIS vs HY (p < .05); TVS vs HY (not significant).
      .47 (.30–.63)85.2.41 (.23–.56)82.8.70 (.59–.79)90.9
      Total in 4 categories with clips of high quality only.73 (.56–.90)67
      Total numbers of high-quality video clips obtained with TVS and GIS.
      .49 (.27–.72)67
      Total numbers of high-quality video clips obtained with TVS and GIS.
      .56 (.38–.73)101
      Total numbers of high-quality video clips obtained with HY.
      GIS = gel infusion sonography; HY = hysteroscopy; n1 = number of cases diagnosed by observer 1; n2 = number of cases diagnosed by observer 2; nA = number of cases diagnosed by observer A; nB = number of cases diagnosed by observer B; TVS = transvaginal ultrasound.
      N = 122. Observer A and observer B evaluated videos obtained with TVS with Doppler and GIS. Videos of the same patients obtained at HY were evaluated by observers 1 and 2. Weighted κ was calculated for all 4 categories: malignancy, hyperplasia, polyps, and no endometrial abnormality. A κ value of 0 to .20 represents slight agreement, between .21 and .40 fair agreement, between .41 and .60 moderate agreement, between .61 and .80 substantial agreement, and between .81 and 1.00 almost perfect agreement.
      Disagreements of more than 1 category (McNemar's test): GIS vs HY (p < .05); TVS vs HY (not significant).
      Total numbers of high-quality video clips obtained with TVS and GIS.
      Total numbers of high-quality video clips obtained with HY.
      Regarding the classification of diagnoses in all categories, substantial agreement was achieved between the 2 observers using HY, whereas moderate agreement was obtained by those using TVS and GIS. Significantly lower agreement was found with the use of GIS compared with HY in the total classification of all 4 diagnoses. The agreement for a diagnosis of cancer improved in 67 women with high-quality video clips obtained during both TVS and GIS but was unchanged in 101 women with high-quality video clips obtained with HY.
      Table 3 presents the individual agreement by observer (1, 2, A, and B) with final pathologic diagnosis classified into the following categories: cancer, hyperplasia, benign endometrial polyps, and normal endometrium. There was substantial agreement regarding final pathologic findings for observers 1 and 2 with use of HY and for the most experienced observers with TVS and GIS. The total number of disagreements in patients with cancer (all observers and all techniques) was higher in patients with both polyps and cancer compared with patients with cancer and no polyps.
      Table 3Agreement between each observer in 4 categories of endometrial pathology (no endometrial pathology, polyps, hyperplasia, and cancer) with pathologic findings in samples obtained at operative hysteroscopy or hysterectomy
      Microscopic diagnosis at hysterectomy or operative hysteroscopy
      κ
      Weighted κ analysis.
      (95% confidence interval)
      Percent agreement
      TVS observer A.71 (.60–.81)91.4
      TVS observer B.40 (.22–.55)84.0
      GIS observer A.67 (.53–.77)89.4
      GIS observer B.40 (.24–.54)83.9
      HY observer 1.73 (.62–.83)90.8
      HY observer 2.68 (.57–.77)89.7
      GIS = gel infusion sonography; HY = hysteroscopy; TVS = transvaginal sonography.
      N = 122. Observers A and B used TVS and GIS, whereas observers 1 and 2 used HY. Total numbers of disagreements in 49 of 122 patients with cancer regarding a diagnosis of malignancy was in total (sum of TVS by observer A, TVS by observer B, GIS observer A, GIS observer B HY observer 1, HY observer 2) higher in patients with additional polyps (Wilcoxon rank-sum test p = .03) compared with patients without additional polyps. There was no difference in total numbers of disagreement in patients without cancer and with or without polyps. More disagreements were seen between observer B at TVS and GIS compared with TVS and GIS observer A and hysteroscopy observer 1 and 2 (p < .05).
      Weighted κ analysis.
      In Table 4, the agreement between the 2 most experienced investigators (observer 1 and observer A) were substantial with regard to ultrasound and HY, whereas the agreements between the less experienced investigators (observer 2 and observer B) were only moderate. Table 5 displays all cases with a final pathologic diagnosis of malignancy not seen by 1 of the 4 observers. Twelve of the 24 malignant cases missed by both observers during video evaluation of HY, TVS, or GIS were also missed during live evaluation using the same technique. Additional polyps were present in several cases of missed malignancy by both observers. The quality of the videos was low in 36% of discrepancies with GIS, 37% with TVS, and 20% with HY.
      Table 4Agreement between the most experienced investigators with TVS (A) and GIS (A) with most experienced investigator with HY (1) and agreement between observers with standard experience at ultrasound B and hysteroscopy 2 is seen
      Interobserver agreement
      All four categories of endometrial pathology
      All 4 categories of endometrial pathology: no endometrial pathology, polyps, hyperplasia, and cancer.
      Cancer diagnosis
      κ
      Weighted κ analysis.
      (95% confidence interval)
      Percent agreementκ (95% confidence interval)Percent agreement
      Observer 1 (most experienced) at HY
       Observer A GIS.61 (.48–.74)87.9.65 (.52–.79)83.6
       Observer A TVS.67 (.56–.77)90.5.63 (.49–.77)82.8
      Observer 2 (standard experience) at HY
       Observer B GIS.47 (.31–.62)86.3.44 (.26–.63)79.5
       Observer B TVS.43 (.27–.59)84.4.43 (.26–.59)76.2
      GIS = gel infusion sonography; HY = hysteroscopy; TVS = transvaginal sonography.
      N = 122. Disagreements regarding more than 1 category between observer 2 at HY and observer B on GIS or TVS vs observer 1 on HY and observer A at TVS (p < .05, McNemar's test).
      All 4 categories of endometrial pathology: no endometrial pathology, polyps, hyperplasia, and cancer.
      Weighted κ analysis.
      Table 5Total number of disagreements with final diagnosis of endometrial cancer by all observers
      Diagnosis givenTotal numbers missedDiagnosis given by live evaluation (n)Poorer video quality (n)Remarks after pathologic and video revision
      CancerHyperplasiaPolyps
      Cancers missed by 1 observer
       HY: polyp (6), hyperplasia (12)18
      Numbers missed by less experienced observer 2 or B: HY 16 of 18; TVS 9 of 10; GIS 18 of 20. In cases of missed malignancy by both observers additional polyps were present in 4 of 7 cases with HY videos; 4 of 9 with TVS and 5 of 8 with GIS.
      1846 Ca and Po, 2 Ca in Po, 2 polypoid cancers
       TVS: polyp (9), normal (1)10
      Numbers missed by less experienced observer 2 or B: HY 16 of 18; TVS 9 of 10; GIS 18 of 20. In cases of missed malignancy by both observers additional polyps were present in 4 of 7 cases with HY videos; 4 of 9 with TVS and 5 of 8 with GIS.
      8233 additional Po
       GIS: polyp (12), hyperplasia (7), normal (1)20
      Numbers missed by less experienced observer 2 or B: HY 16 of 18; TVS 9 of 10; GIS 18 of 20. In cases of missed malignancy by both observers additional polyps were present in 4 of 7 cases with HY videos; 4 of 9 with TVS and 5 of 8 with GIS.
      161385 additional Po
      Cancers missed by both observers with HY7
       Hyperplasia (at least 1 observer) (a1, a2, b, c)421112 small Ca identified (a1, a2), 1 Ca in Po without clear sign (b), 1 Ca behind Po not seen in video (c)
       Polyps both observers (d, e)2111 Ca in Po in an area (d), Ca and Po (e)- good videos
       Normal both observers (f)10(f) Not seen in video displaying the whole uterine cavity, removed by prior sampling?
      Cancers missed by both observers with TVS9
       Hyperplasia 1 observer (f, h, i, j)4221(f) Discrete changes, removed by prior sampling?
       Polyp both observers (a2, b, d, e, g)5323a2, b, d, e described above. Additional polyps (b, d, e, g)
      Cancers missed by both observers with GIS8
       Hyperplasia 1 observer (c, k, l, m)42111Additional polyps (c, k)
       Polyps both observers (a2, b, d, g)42111Additional polyps (b, d, g)
      Ca = cancer; GIS = gel infusion sonography; HY = hysteroscopy; Po = polyps; TVS = transvaginal sonography.
      Disagreements by 1 observer and both observers in evaluation of video clips HY, TVS, and GIS are compared with diagnosis during live evaluation. Video quality, additional polyps, and remarks during secondary revision of videos are given. For cases missed by both observers, individual patients are numbered a1, a2, b, c, d, e, f, g, h, i, j, k, l, m.
      Numbers missed by less experienced observer 2 or B: HY 16 of 18; TVS 9 of 10; GIS 18 of 20. In cases of missed malignancy by both observers additional polyps were present in 4 of 7 cases with HY videos; 4 of 9 with TVS and 5 of 8 with GIS.
      In 24 of 66 women (31%) with cancers or hyperplasia, additional benign polyps were seen on microscopic examination. In these, 24 of 66 women with either cancer/hyperplasia and benign polyps, there were 15 disagreements (63%) in the diagnosis of hyperplasia or cancer on hysteroscopic video evaluation compared with only 13 disagreements (31%) in 42 of these 66 patients without additional polyps (p = .02). However, there was no difference in the number of disagreements in these women with the use of TVS.

      Discussion

      There was no difference in the fair to moderate level of diagnostic discrepancy between observers obtained with the use of TVS, HY and GIS for a specific diagnosis of malignancy or polyps, and there was no clear advantage with regard to a lower observer variation by any techniques in differentiating between malignancy, hyperplasia, or endometrial polyps. However, observer variation for overall classification of pathology was lower on GIS than HY. However, a substantial agreement with use of both TVS and HY for the diagnosis of a normal uterine cavity indicated a general high reliability of both techniques in ruling out any pathology but a slightly general lower reliability of GIS.
      The present findings of moderate observer agreement with HY and ultrasound for the evaluation of the endometrial cavity are in agreement with studies of premenopausal patients [
      • Dueholm M.
      • Lundorf E.
      • Sorensen J.S.
      • et al.
      Reproducibility of evaluation of the uterus by transvaginal sonography, hysterosonographic examination, hysteroscopy and magnetic resonance imaging.
      ,
      • Kasius J.C.
      • Broekmans F.J.
      • Veersema S.
      • et al.
      Observer agreement in the evaluation of the uterine cavity by hysteroscopy prior to in vitro fertilization.
      ,
      • Van den Bosch T.
      • Valentin L.
      • Van S.D.
      • et al.
      The detection of intracavitary uterine pathology using off-line analysis of 3-dimensional ultrasound volumes: interobserver agreement and diagnostic accuracy.
      ]. A higher observer disagreement with GIS but not between the experienced observers using GIS and HY indicated that especially the evaluation of GIS is critically dependent on observer experience. This dependence may be essential in the case of physicians without special experience with GIS, also seen in other studies [
      • Dueholm M.
      • Lundorf E.
      • Sorensen J.S.
      • et al.
      Reproducibility of evaluation of the uterus by transvaginal sonography, hysterosonographic examination, hysteroscopy and magnetic resonance imaging.
      ,
      • Beemsterboer S.N.
      • Thurkow A.L.
      • Verstraeten R.
      • Brolmann H.A.
      Reproducibility of saline contrast sonohysterography for the detection of intracavitary abnormalities in women with abnormal uterine bleeding.
      ].
      A pattern diagnosis of endometrial hyperplasia was not reproducible by use of either ultrasound or HY. TVS and GIS have varying efficiencies with regard to making a diagnosis of endometrial hyperplasia [
      • Bingol B.
      • Gunenc M.Z.
      • Gedikbasi A.
      • et al.
      Comparison of diagnostic accuracy of saline infusion sonohysterography, transvaginal sonography and hysteroscopy in postmenopausal bleeding.
      ,
      • Farquhar C.
      • Ekeroma A.
      • Furness S.
      • Arroll B.
      A systematic review of transvaginal ultrasonography, sonohysterography and hysteroscopy for the investigation of abnormal uterine bleeding in premenopausal women.
      ] because this diagnosis relies on a nonspecific thickening of the endometrium. The difficulties involved in using HY to differentiate between hyperplasia and other abnormalities have previously been described [
      • de Wit A.C.
      • Vleugels M.P.
      • de Kruif J.H.
      Diagnostic hysteroscopy: a valuable diagnostic tool in the diagnosis of structural intra-cavital pathology and endometrial hyperplasia or carcinoma? Six years of experience with non-clinical diagnostic hysteroscopy.
      ,
      • Liberis V.
      • Tsikouras P.
      • Christos Z.
      • et al.
      The contribution of hysteroscopy to the detection malignancy in symptomatic postmenopausal women.
      ,
      • Clark T.J.
      • Voit D.
      • Gupta J.K.
      • et al.
      Accuracy of hysteroscopy in the diagnosis of endometrial cancer and hyperplasia: a systematic quantitative review.
      ] and are even noted by experienced hysteroscopists [
      • Lasmar R.B.
      • Barrozo P.R.
      • de Oliveira M.A.
      • Coutinho E.S.
      • Dias R.
      Validation of hysteroscopic view in cases of endometrial hyperplasia and cancer in patients with abnormal uterine bleeding.
      ,
      • Garuti G.
      • Cellani F.
      • Garzia D.
      • Colonnelli M.
      • Luerti M.
      Accuracy of hysteroscopic diagnosis of endometrial hyperplasia: a retrospective study of 323 patients.
      ].
      There was an increased observer bias toward the diagnosis of malignancy or hyperplasia using HY in the large numbers of women with concomitant endometrial polyps. Concurrent hyperplasia in polyps and nonpolypoid endometrium is common [
      • Kelly P.
      • Dobbs S.P.
      • McCluggage W.G.
      Endometrial hyperplasia involving endometrial polyps: report of a series and discussion of the significance in an endometrial biopsy specimen.
      ], and the presence of both benign polyps and hyperplasia has been described in 25% to 34% of postmenopausal patients with endometrial polyps [
      • Rahimi S.
      • Marani C.
      • Renzi C.
      • et al.
      Endometrial polyps and the risk of atypical hyperplasia on biopsies of unremarkable endometrium: a study on 694 patients with benign endometrial polyps.
      ,
      • Savelli L.
      • De I.P.
      • Santini D.
      • et al.
      Histopathologic features and risk factors for benignity, hyperplasia, and cancer in endometrial polyps.
      ]. Hyperplasia and cancer are often present at the same time [
      • Trimble C.L.
      • Kauderer J.
      • Zaino R.
      • et al.
      Concurrent endometrial carcinoma in women with a biopsy diagnosis of atypical endometrial hyperplasia: a Gynecologic Oncology Group study.
      ], and although not described before, the prevalence of concomitant endometrial polyps was at the same high level in patients with endometrial cancer as it was in patients with hyperplasia.
      The rationale for the removal of symptomatic endometrial polyps has been questioned because of their low risk of malignancy, most often below 5% [
      • Bakour S.H.
      • Khan K.S.
      • Gupta J.K.
      The risk of premalignant and malignant pathology in endometrial polyps.
      ,
      • Baiocchi G.
      • Manci N.
      • Pazzaglia M.
      • et al.
      Malignancy in endometrial polyps: a 12-year experience.
      ,
      • Costa-Paiva L.
      • Godoy Jr., C.E.
      • Antunes Jr., A.
      • et al.
      Risk of malignancy in endometrial polyps in premenopausal and postmenopausal women according to clinicopathologic characteristics.
      ,
      • Lee S.C.
      • Kaunitz A.M.
      • Sanchez-Ramos L.
      • Rhatigan R.M.
      The oncogenic potential of endometrial polyps: a systematic review and meta-analysis.
      ]. However, the reported low risk of malignancy was mostly based on studies on removed endometrial polyps in selected pathologic reports, where the adjacent endometrium was normal. The high numbers of women with concomitant hyperplasia or cancer in adjacent endometrium were not included.
      The present study supports currently recommended complete removal of localized pathology by HY in women with PMB [
      American Association of Gynecologic Laparoscopists
      AAGL practice report: practice guidelines for the diagnosis and management of endometrial polyps.
      ,
      • Lieng M.
      • Istre O.
      • Qvigstad E.
      Treatment of endometrial polyps: a systematic review.
      ]. Malignancy was overlooked with HY not only when located in a part of a polyp but also when it was behind and beside benign endometrial polyps. ES was not an attractive alternative, because ES is often insufficient also when performed together with HY [
      • Williams A.R.
      • Brechin S.
      • Porter A.J.
      • Warner P.
      • Critchley H.O.
      Factors affecting adequacy of Pipelle and Tao Brush endometrial sampling.
      ], and ES had a lower sensitivity for the diagnosis of endometrial cancer than that obtained by pattern evaluation during GIS or HY without histology [
      • Dueholm M.
      • Marinovskij E.
      • Hansen E.S.
      • Moller C.
      • Ortoft G.
      Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
      ]. Two meta-analyses have established a high diagnostic accuracy of ES for diagnosis of endometrial cancer [
      • Dijkhuizen F.P.
      • Mol B.W.
      • Brolmann H.A.
      • Heintz A.P.
      The accuracy of endometrial sampling in the diagnosis of patients with endometrial carcinoma and hyperplasia: a meta-analysis.
      ,
      • Clark T.J.
      • Mann C.H.
      • Shah N.
      • et al.
      Accuracy of outpatient endometrial biopsy in the diagnosis of endometrial cancer: a systematic quantitative review.
      ]. However, these meta-analyses included only a few studies of high standard and were not performed in women with PMB and increased endometrial thickness. In patients with known endometrial cancer, only 83% of these cancers were detected, and no cancers were discovered occupying less than 5% of the endometrium cavity or when it was localized in a polyp [
      • Guido R.S.
      • Kanbour-Shakir A.
      • Rulin M.C.
      • Christopherson W.A.
      Pipelle endometrial sampling. Sensitivity in the detection of endometrial cancer.
      ]. This is consistent with the finding that a Pipelle sample had on average only 4% (range, 0–12%) of the endometrial lining [
      • Rodriguez G.C.
      • Yaqub N.
      • King M.E.
      A comparison of the Pipelle device and the Vabra aspirator as measured by endometrial denudation in hysterectomy specimens: the Pipelle device samples significantly less of the endometrial surface than the Vabra aspirator.
      ] and a lower accuracy of ES in the presence of focal pathology [
      • Epstein E.
      • Ramirez A.
      • Skoog L.
      • Valentin L.
      Dilatation and curettage fails to detect most focal lesions in the uterine cavity in women with postmenopausal bleeding.
      ,
      • Angioni S.
      • Loddo A.
      • Milano F.
      • et al.
      Detection of benign intracavitary lesions in postmenopausal women with abnormal uterine bleeding: a prospective comparative study on outpatient hysteroscopy and blind biopsy.
      ]. Moreover, ES has an insufficient low accuracy for diagnosis of hyperplasia [
      • Dijkhuizen F.P.
      • Mol B.W.
      • Brolmann H.A.
      • Heintz A.P.
      The accuracy of endometrial sampling in the diagnosis of patients with endometrial carcinoma and hyperplasia: a meta-analysis.
      ,
      • Clark T.J.
      • Mann C.H.
      • Shah N.
      • et al.
      Accuracy of outpatient endometrial biopsy in the diagnosis of endometrial hyperplasia.
      ] compared with hysteroscopic samples [
      Committee opinion no. 631: endometrial intraepithelial neoplasia.
      ].
      Thus, HY may be added in the following situations: when ES is insufficient [
      • Dreisler E.
      • Poulsen L.G.
      • Antonsen S.L.
      • et al.
      EMAS clinical guide: Assessment of the endometrium in peri and postmenopausal women.
      ] and to remove focal pathology like benign endometrial polyps [
      • Goldstein R.B.
      • Bree R.L.
      • Benson C.B.
      • et al.
      Evaluation of the woman with postmenopausal bleeding: Society of Radiologists in Ultrasound-Sponsored Consensus Conference statement.
      ,
      American Association of Gynecologic Laparoscopists
      AAGL practice report: practice guidelines for the diagnosis and management of endometrial polyps.
      ] or hyperplasia [
      Committee opinion no. 631: endometrial intraepithelial neoplasia.
      ]. In women with PMB and an increased endometrial thickness, the routine use of ES and addition of HY in all these situations may be a waste of time and resources. In women with PMB and endometrium at least 4.5 mm, this setup may imply that only 12% of women with benign pathology are investigated with ES without HY [
      • Opolskiene G.
      • Sladkevicius P.
      • Valentin L.
      Prediction of endometrial malignancy in women with postmenopausal bleeding and sonographic endometrial thickness >/= 4.5 mm.
      ]. Therefore, see-and-treat HY has been introduced [

      Institute of Obstetricians and Gynaecologists, Royal College of Physicians of Ireland and Directorate of Clinical Strategy and Programmes, Health Service Executive. Investigation of postmenopausal bleeding. Available at: http://www.hse.ie/eng/about/Who/clinical/natclinprog/obsandgynaeprogramme/postmeno.pdf. Accessed April 1, 2015.

      ]. A general preference for see-and-treat HY as an efficient minimally invasive first choice in all women with PMB and thickened endometrium can be questioned because there is often the need for a second operative HY to remove residual larger localized pathology after see-and-treat HY.
      Neither in systematic reviews nor in international guidelines is consensus found regarding the best sequence of diagnostic procedures for women with PMB and endometrium thickness at least 4 to 5 mm [
      • van H.N.
      • Breijer M.C.
      • Khan K.S.
      • et al.
      Diagnostic evaluation of the endometrium in postmenopausal bleeding: an evidence-based approach.
      ]. Nevertheless, a sufficient standardized pattern analysis during a first-line TVS and GIS may help in selection of the right type of strategy. SIS or GIS is used to identify benign pattern and the type of HYs. See-and treat HY or operative HY may be offered based on the size of pathology to be removed. Women with a specifically defined benign pattern may be diagnosed with ES without HY, even when ES is insufficient. Women with a pattern that is suspect for cancer may be investigated with ES, and women with cancer may thereby avoid see-and-treat HY, which could be safer in general practice.
      Interpretation of simple image techniques (TVS, GIS, and HY) during a general gynecologic examination has the benefit of immediate decision regarding treatment but the disadvantage of observer dependence. ES is supposed to compensate for observer dependence, but ES is not a strong test. Therefore, the selection of a strategy in the individual patient with a decision of localized pathology to be removed and handling of failed samples and hyperplasia is highly observer dependent. Observer dependence is not decreased by see-and-treat HY, and operative HY with removal of tissue may be performed.
      The high observer variation between individual gynecologists may hinder or delay implementation of new improvements in image techniques. Image conferences are used to improve and implement interpretation of new diagnostic advantages and reduce observer variation. High observer variation may intensify primary referral to magnetic resonance imaging centers, with the advantage of a higher accuracy that may not be due to the image technique itself but rather to lower observer variation because of the consensus image conferences by experienced image analysis at these centers.
      Strategies to improve observer variation may be important to preserve the future use of simple-image techniques in the hands of general clinicians. Implementation of online image conferences by telemedicine could be a future solution. Challenging findings with ultrasound and HY may be evaluated online by highly experienced physicians, which may even be situated in distant centers. Obtained 3-dimensional volumes or video clips may have lower image quality but may be used without online facilities [
      • Dueholm M.
      • Holm J.W.
      • Rydbjerg S.
      • Hansen E.S.
      • Ortoft G.
      Two-dimensional and 3-dimensional transvaginal ultrasound with addition of power Doppler angiography and gel infusion sonography for diagnosis of endometrial malignancy.
      ]. This approach seems faster, has lower cost, and is more convenient than sending the patient to another hospital, conducting invasive operative investigations, or using expensive image techniques like magnetic resonance imaging. This may facilitate pattern evaluation during first-line TVS with and without GIS and may be used to decide on the need for an operative HY, with residual pathology after see-and-treat HY. At present, a structured evaluation of the endometrial pattern by a scoring system may improve reproducibility. A structured evaluation by scoring systems improved κ for a malignant diagnosis using GIS from .34 to .63 [
      • Dueholm M.
      • Møller C.
      • Rydbjerg S.
      • Hansen E.
      An ultrasound algorithm for the identification of endometrial cancer.
      ]. Moreover, concomitant structured evaluation of endometrial pattern with TVS, GIS, and HY may improve diagnosis. Areas with abnormal vascularity by TVS may be evaluated during HY to target microforceps biopsies to these areas.
      As a strength, we used videos clips instead of live evaluations to eliminate observer bias induced by not being blinded to a patient's characteristics and prior image findings, and we compared videos by different techniques in the same consecutive patients. The quality of the video clips, either the use of video clips or lower image quality in general, seemed to contribute to the discrepancies. Video clips were obtained by highly experienced observers, which may actually have increased observer reproducibility, because skill in visualization is a main factor in visual diagnosis in endoscopy and ultrasound.
      Lower observer variation would be expected in unselected women with PMB, in whom discrimination between the normal and abnormal cavity is the main issue. The present study evaluated observer variation with regard to the ability to differentiate between malignant and other pathology, and this differentiation is most important in a population with a high prevalence of malignancy. The high numbers of benign polyps in patients with hyperplasia or cancer should be evaluated in future studies.
      In conclusion, a high level of diagnostic discrepancy was found between observers for a specific diagnosis of endometrial pathology with use of TVS and HY, which increased with the addition of GIS. The discrepancy obtained with HY was increased when both benign polyps and cancer or hyperplasia were present. Concomitant evaluation with TVS, GIS, and HY and the introduction of systematic systems and telemedicine for endometrial pattern recognition may improve reproducibility.

      Acknowledgments

      The authors gratefully acknowledge the help of Elisabeth Melin and Charlotte Møller in obtaining the videos and the staff at the day surgery unit. The study was approved by The Ethics Committee of the Central Denmark Region 9th of August 2010 (M-20100139). Because hysteroscopy and transvaginal sonography was a standard procedure in the department, the committee found that no informed consent for this procedure was needed.

      References

        • Lajer H.
        • Elnegaard S.
        • Christensen R.D.
        • et al.
        Survival after stage IA endometrial cancer: can follow-up be altered? A prospective nationwide Danish survey.
        Acta Obstet Gynaecol Scand. 2012; 91: 976-982
        • Clark T.J.
        • Barton P.M.
        • Coomarasamy A.
        • Gupta J.K.
        • Khan K.S.
        Investigating postmenopausal bleeding for endometrial cancer: cost-effectiveness of initial diagnostic strategies.
        Br J Obstet Gynaecol. 2006; 113: 502-510
        • Karlsson B.
        • Granberg S.
        • Wikland M.
        • et al.
        Transvaginal ultrasonography of the endometrium in women with postmenopausal bleeding—a Nordic multicenter study.
        Am J Obstet Gynecol. 1995; 172: 1488-1494
        • Smith-Bindman R.
        • Kerlikowske K.
        • Feldstein V.A.
        • et al.
        Endovaginal ultrasound to exclude endometrial cancer and other endometrial abnormalities.
        JAMA. 1998; 280: 1510-1517
        • American College of Obstetricians and Gynecologists
        ACOG committee opinion no. 426: the role of transvaginal ultrasonography in the evaluation of postmenopausal bleeding.
        Obstet Gynecol. 2009; 113: 462-464
        • Timmermans A.
        • Opmeer B.C.
        • Khan K.S.
        • et al.
        Endometrial thickness measurement for detecting endometrial cancer in women with postmenopausal bleeding: a systematic review and meta-analysis.
        Obstet Gynecol. 2010; 116: 160-167
        • van H.N.
        • Breijer M.C.
        • Khan K.S.
        • et al.
        Diagnostic evaluation of the endometrium in postmenopausal bleeding: an evidence-based approach.
        Maturitas. 2011; 68: 155-164
        • Dreisler E.
        • Poulsen L.G.
        • Antonsen S.L.
        • et al.
        EMAS clinical guide: Assessment of the endometrium in peri and postmenopausal women.
        Maturitas. 2013; 75: 181-190
        • Munro M.G.
        Investigation of women with postmenopausal uterine bleeding: clinical practice recommendations.
        Perm J. 2014; 18: 55-70
        • Epstein E.
        • Ramirez A.
        • Skoog L.
        • Valentin L.
        Dilatation and curettage fails to detect most focal lesions in the uterine cavity in women with postmenopausal bleeding.
        Acta Obstet Gynecol Scand. 2001; 80: 1131-1136
        • Angioni S.
        • Loddo A.
        • Milano F.
        • et al.
        Detection of benign intracavitary lesions in postmenopausal women with abnormal uterine bleeding: a prospective comparative study on outpatient hysteroscopy and blind biopsy.
        J Minim Invasive Gynecol. 2008; 15: 87-91
        • Werbrouck E.
        • Veldman J.
        • Luts J.
        • et al.
        Detection of endometrial pathology using saline infusion sonography versus gel instillation sonography: a prospective cohort study.
        Fertil Steril. 2011; 95: 285-288
        • Epstein E.
        • Skoog L.
        • Isberg P.E.
        • et al.
        An algorithm including results of gray-scale and power Doppler ultrasound examination to predict endometrial malignancy in women with postmenopausal bleeding.
        Ultrasound Obstet Gynecol. 2002; 20: 370-376
        • Goldstein R.B.
        • Bree R.L.
        • Benson C.B.
        • et al.
        Evaluation of the woman with postmenopausal bleeding: Society of Radiologists in Ultrasound-Sponsored Consensus Conference statement.
        J Ultrasound Med. 2001; 20: 1025-1036
        • American Association of Gynecologic Laparoscopists
        AAGL practice report: practice guidelines for the diagnosis and management of endometrial polyps.
        J Minim Invasive Gynecol. 2012; 19: 3-10
        • Dijkhuizen F.P.
        • Mol B.W.
        • Brolmann H.A.
        • Heintz A.P.
        The accuracy of endometrial sampling in the diagnosis of patients with endometrial carcinoma and hyperplasia: a meta-analysis.
        Cancer. 2000; 89: 1765-1772
        • Visser N.C.
        • Breijer M.C.
        • Herman M.C.
        • et al.
        Factors attributing to the failure of endometrial sampling in women with postmenopausal bleeding.
        Acta Obstet Gynecol Scand. 2013; 92: 1216-1222
        • Williams A.R.
        • Brechin S.
        • Porter A.J.
        • Warner P.
        • Critchley H.O.
        Factors affecting adequacy of Pipelle and Tao Brush endometrial sampling.
        Br J Obstet Gynecol. 2008; 115: 1028-1036
        • van Doorn H.C.
        • Opmeer B.C.
        • Burger C.W.
        • et al.
        Inadequate office endometrial sample requires further evaluation in women with postmenopausal bleeding and abnormal ultrasound results.
        Int J Gynaecol Obstet. 2007; 99: 100-104
        • Farrell T.
        • Jones N.
        • Owen P.
        • Baird A.
        The significance of an “insufficient” Pipelle sample in the investigation of post-menopausal bleeding.
        Acta Obstet Gynaecol Scand. 1999; 78: 810-812
        • Bakour S.H.
        • Jones S.E.
        • O'Donovan P.
        Ambulatory hysteroscopy: evidence-based guide to diagnosis and therapy.
        Best Pract Res Clin Obstet Gynaecol. 2006; 20: 953-975
        • Ceci O.
        • Bettocchi S.
        • Pellegrino A.
        • et al.
        Comparison of hysteroscopic and hysterectomy findings for assessing the diagnostic accuracy of office hysteroscopy.
        Fertil Steril. 2002; 78: 628-631
        • Di Spiezio S.A.
        • Bettocchi S.
        • Spinelli M.
        • et al.
        Review of new office-based hysteroscopic procedures 2003–2009.
        J Minim Invasive Gynecol. 2010; 17: 436-448
        • Bettocchi S.
        • Ceci O.
        • Nappi L.
        • et al.
        Operative office hysteroscopy without anesthesia: analysis of 4863 cases performed with mechanical instruments.
        J Am Assoc Gynecol Laparosc. 2004; 11: 59-61
        • Bettocchi S.
        • Ceci O.
        • Di V.R.
        • et al.
        Advanced operative office hysteroscopy without anaesthesia: analysis of 501 cases treated with a 5 Fr. bipolar electrode.
        Hum Reprod. 2002; 17: 2435-2438
        • Lotfallah H.
        • Farag K.
        • Hassan I.
        • Watson R.
        One-stop hysteroscopy clinic for postmenopausal bleeding.
        J Reprod Med. 2005; 50: 101-107
        • Loizzi V.
        • Bettocchi S.
        • Vimercati A.
        • et al.
        Hysteroscopic evaluation of menopausal women with endometrial thickness of 4 mm or more.
        J Am Assoc Gynecol Laparosc. 2000; 7: 191-195
        • Gambadauro P.
        • Martinez-Maestre M.A.
        • Torrejon R.
        When is see-and-treat hysteroscopic polypectomy successful?.
        Eur J Obstet Gynaecol Reprod Biol. 2014; 178: 70-73
        • Cicinelli E.
        Hysteroscopy without anesthesia: review of recent literature.
        J Minim Invasive Gynecol. 2010; 17: 703-708
        • Dueholm M.
        • Marinovskij E.
        • Hansen E.S.
        • Moller C.
        • Ortoft G.
        Diagnostic methods for fast-track identification of endometrial cancer in women with postmenopausal bleeding and endometrial thickness greater than 5 mm.
        Menopause. 2015; 22: 616-626
        • Epstein E.
        • Valentin L.
        Gray-scale ultrasound morphology in the presence or absence of intrauterine fluid and vascularity as assessed by color Doppler for discrimination between benign and malignant endometrium in women with postmenopausal bleeding.
        Ultrasound Obstet Gynecol. 2006; 28: 89-95
        • Opolskiene G.
        • Sladkevicius P.
        • Valentin L.
        Prediction of endometrial malignancy in women with postmenopausal bleeding and sonographic endometrial thickness >/= 4.5 mm.
        Ultrasound Obstet Gynecol. 2011; 37: 232-240
        • Dueholm M.
        • Møller C.
        • Rydbjerg S.
        • Hansen E.
        An ultrasound algorithm for the identification of endometrial cancer.
        Ultrasound Obstet Gynecol. 2014; 43: 557-568
        • Warming L.
        • Ravn P.
        • Skouby S.
        • Christiansen C.
        Measurement precision and normal range of endometrial thickness in a postmenopausal population by transvaginal ultrasound.
        Ultrasound Obstet Gynecol. 2002; 20: 492-495
        • Epstein E.
        • Valentin L.
        Intraobserver and interobserver reproducibility of ultrasound measurements of endometrial thickness in postmenopausal women.
        Ultrasound Obstet Gynecol. 2002; 20: 486-491
        • Wolman I.
        • Amster R.
        • Hartoov J.
        • et al.
        Reproducibility of transvaginal ultrasonographic measurements of endometrial thickness in patients with postmenopausal bleeding.
        Gynecol Obstet Invest. 1998; 46: 191-194
        • Karlsson B.
        • Granberg S.
        • Ridell B.
        • Wikland M.
        Endometrial thickness as measured by transvaginal sonography: interobserver variation.
        Ultrasound Obstet Gynecol. 1994; 4: 320-325
        • Opolskiene G.
        • Sladkevicius P.
        • Valentin L.
        Ultrasound assessment of endometrial morphology and vascularity to predict endometrial malignancy in women with postmenopausal bleeding and sonographic endometrial thickness >or= 4.5 mm.
        Ultrasound Obstet Gynecol. 2007; 30: 332-340
        • Opolskiene G.
        • Sladkevicius P.
        • Valentin L.
        Two- and three-dimensional saline contrast sonohysterography: interobserver agreement, agreement with hysteroscopy and diagnosis of endometrial malignancy.
        Ultrasound Obstet Gynecol. 2009; 33: 574-582
        • Dueholm M.
        • Lundorf E.
        • Sorensen J.S.
        • et al.
        Reproducibility of evaluation of the uterus by transvaginal sonography, hysterosonographic examination, hysteroscopy and magnetic resonance imaging.
        Hum Reprod. 2002; 17: 195-200
        • Kasius J.C.
        • Broekmans F.J.
        • Veersema S.
        • et al.
        Observer agreement in the evaluation of the uterine cavity by hysteroscopy prior to in vitro fertilization.
        Hum Reprod. 2011; 26: 801-807
        • Dueholm M.
        • Holm J.W.
        • Rydbjerg S.
        • Hansen E.S.
        • Ortoft G.
        Two-dimensional and 3-dimensional transvaginal ultrasound with addition of power Doppler angiography and gel infusion sonography for diagnosis of endometrial malignancy.
        Ultrasound Obstet Gynecol. 2015; 45: 734-743
        • Garuti G.
        • Mirra M.
        • Luerti M.
        Hysteroscopic view in atypical endometrial hyperplasias: a correlation with pathologic findings on hysterectomy specimens.
        J Minim Invasive Gynecol. 2006; 13: 325-330
        • Dotto J.E.
        • Lema B.
        • Dotto Jr., J.E.
        • Hamou J.
        Classification of microhysteroscopic images and their correlation with histologic diagnoses.
        J Am Assoc Gynecol Laparosc. 2003; 10: 233-246
        • Lasmar R.B.
        • Barrozo P.R.
        • de Oliveira M.A.
        • Coutinho E.S.
        • Dias R.
        Validation of hysteroscopic view in cases of endometrial hyperplasia and cancer in patients with abnormal uterine bleeding.
        J Minim Invasive Gynecol. 2006; 13: 409-412
      1. Dueholm M, Hjorth IM, Secher P, Jørgensen A, Ørtoft G. Structured hysteroscopic evaluation of endometrium in women with postmenopausal bleeding. J Minim Invasive Gynecol. 2015 Jun 30. [Epub ahead of print].

        • Leone F.P.
        • Timmerman D.
        • Bourne T.
        • et al.
        Terms, definitions and measurements to describe the sonographic features of the endometrium and intrauterine lesions: a consensus opinion from the International Endometrial Tumor Analysis (IETA) group.
        Ultrasound Obstet Gynecol. 2010; 35: 103-112
        • Van den Bosch T.
        • Valentin L.
        • Van S.D.
        • et al.
        The detection of intracavitary uterine pathology using off-line analysis of 3-dimensional ultrasound volumes: interobserver agreement and diagnostic accuracy.
        Ultrasound Obstet Gynecol. 2012; 40: 459-463
        • Beemsterboer S.N.
        • Thurkow A.L.
        • Verstraeten R.
        • Brolmann H.A.
        Reproducibility of saline contrast sonohysterography for the detection of intracavitary abnormalities in women with abnormal uterine bleeding.
        Ultrasound Obstet Gynecol. 2008; 31: 445-449
        • Bingol B.
        • Gunenc M.Z.
        • Gedikbasi A.
        • et al.
        Comparison of diagnostic accuracy of saline infusion sonohysterography, transvaginal sonography and hysteroscopy in postmenopausal bleeding.
        Arch Gynecol Obstet. 2011; 284: 111-117
        • Farquhar C.
        • Ekeroma A.
        • Furness S.
        • Arroll B.
        A systematic review of transvaginal ultrasonography, sonohysterography and hysteroscopy for the investigation of abnormal uterine bleeding in premenopausal women.
        Acta Obstet Gynecol Scand. 2003; 82: 493-504
        • de Wit A.C.
        • Vleugels M.P.
        • de Kruif J.H.
        Diagnostic hysteroscopy: a valuable diagnostic tool in the diagnosis of structural intra-cavital pathology and endometrial hyperplasia or carcinoma? Six years of experience with non-clinical diagnostic hysteroscopy.
        Eur J Obstet Gynecol Reprod Biol. 2003; 110: 79-82
        • Liberis V.
        • Tsikouras P.
        • Christos Z.
        • et al.
        The contribution of hysteroscopy to the detection malignancy in symptomatic postmenopausal women.
        Minim Invasive Ther Allied Technol. 2010; 19: 83-93
        • Clark T.J.
        • Voit D.
        • Gupta J.K.
        • et al.
        Accuracy of hysteroscopy in the diagnosis of endometrial cancer and hyperplasia: a systematic quantitative review.
        JAMA. 2002; 288: 1610-1621
        • Garuti G.
        • Cellani F.
        • Garzia D.
        • Colonnelli M.
        • Luerti M.
        Accuracy of hysteroscopic diagnosis of endometrial hyperplasia: a retrospective study of 323 patients.
        J Minim Invasive Gynecol. 2005; 12: 247-253
        • Kelly P.
        • Dobbs S.P.
        • McCluggage W.G.
        Endometrial hyperplasia involving endometrial polyps: report of a series and discussion of the significance in an endometrial biopsy specimen.
        Br J Obstet Gynaecol. 2007; 114: 944-950
        • Rahimi S.
        • Marani C.
        • Renzi C.
        • et al.
        Endometrial polyps and the risk of atypical hyperplasia on biopsies of unremarkable endometrium: a study on 694 patients with benign endometrial polyps.
        Int J Gynecol Pathol. 2009; 28: 522-528
        • Savelli L.
        • De I.P.
        • Santini D.
        • et al.
        Histopathologic features and risk factors for benignity, hyperplasia, and cancer in endometrial polyps.
        Am J Obstet Gynecol. 2003; 188: 927-931
        • Trimble C.L.
        • Kauderer J.
        • Zaino R.
        • et al.
        Concurrent endometrial carcinoma in women with a biopsy diagnosis of atypical endometrial hyperplasia: a Gynecologic Oncology Group study.
        Cancer. 2006; 106: 812-819
        • Bakour S.H.
        • Khan K.S.
        • Gupta J.K.
        The risk of premalignant and malignant pathology in endometrial polyps.
        Acta Obstet Gynecol Scand. 2000; 79: 317-320
        • Baiocchi G.
        • Manci N.
        • Pazzaglia M.
        • et al.
        Malignancy in endometrial polyps: a 12-year experience.
        Am J Obstet Gynecol. 2009; 201: 462-464
        • Costa-Paiva L.
        • Godoy Jr., C.E.
        • Antunes Jr., A.
        • et al.
        Risk of malignancy in endometrial polyps in premenopausal and postmenopausal women according to clinicopathologic characteristics.
        Menopause. 2011; 18: 1278-1282
        • Lee S.C.
        • Kaunitz A.M.
        • Sanchez-Ramos L.
        • Rhatigan R.M.
        The oncogenic potential of endometrial polyps: a systematic review and meta-analysis.
        Obstet Gynecol. 2010; 116: 1197-1205
        • Lieng M.
        • Istre O.
        • Qvigstad E.
        Treatment of endometrial polyps: a systematic review.
        Acta Obstet Gynecol Scand. 2010; 89: 992-1002
        • Clark T.J.
        • Mann C.H.
        • Shah N.
        • et al.
        Accuracy of outpatient endometrial biopsy in the diagnosis of endometrial cancer: a systematic quantitative review.
        Br J Obstet Gynaecol. 2002; 109: 313-321
        • Guido R.S.
        • Kanbour-Shakir A.
        • Rulin M.C.
        • Christopherson W.A.
        Pipelle endometrial sampling. Sensitivity in the detection of endometrial cancer.
        J Reprod Med. 1995; 40: 553-555
        • Rodriguez G.C.
        • Yaqub N.
        • King M.E.
        A comparison of the Pipelle device and the Vabra aspirator as measured by endometrial denudation in hysterectomy specimens: the Pipelle device samples significantly less of the endometrial surface than the Vabra aspirator.
        Am J Obstet Gynecol. 1993; 168: 55-59
        • Clark T.J.
        • Mann C.H.
        • Shah N.
        • et al.
        Accuracy of outpatient endometrial biopsy in the diagnosis of endometrial hyperplasia.
        Acta Obstet Gynaecol Scand. 2001; 80: 784-793
      2. Committee opinion no. 631: endometrial intraepithelial neoplasia.
        Obstet Gynecol. 2015; 125: 1272-1278
      3. Institute of Obstetricians and Gynaecologists, Royal College of Physicians of Ireland and Directorate of Clinical Strategy and Programmes, Health Service Executive. Investigation of postmenopausal bleeding. Available at: http://www.hse.ie/eng/about/Who/clinical/natclinprog/obsandgynaeprogramme/postmeno.pdf. Accessed April 1, 2015.